<%BANNER%>

DNA Aptamers for Bioanalytical Applications

Permanent Link: http://ufdc.ufl.edu/UFE0041958/00001

Material Information

Title: DNA Aptamers for Bioanalytical Applications
Physical Description: 1 online resource (170 p.)
Language: english
Creator: Martin, Jennifer
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2010

Subjects

Subjects / Keywords: aptamer, capillary, cell, chromatography, dna, selex
Chemistry -- Dissertations, Academic -- UF
Genre: Chemistry thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Aptamers are oligonucleotides selected by the SELEX (Systematic Evolution of Ligands by EXponential enrichment) method that bind to a target with high affinity and specificity based on the three dimensional conformation they adopt. Selected for targets ranging from small molecules to proteins or whole cells, aptamers may similarly assume diverse roles such as drug delivery vehicles, decoys, inhibitory therapeutics, and affinity separation probes. This work investigates aptamers as inhibitory drugs and affinity ligands in two bioanalytical applications. In the first set of studies, an aptamer was selected to the peptide anticoagulant bivalirudin to serve as an antidote to the drug in instances of severe patient bleeding. The drug was immobilized on a monolithic column and binding sequences were eluted by salt gradient. The elution profile was compared to that of a blank column (no drug), and fractions with a chromatographic difference between drug-immobilized and blank counterparts were analyzed via real-time PCR (polymerase chain reaction) and used for further selection. Sequences were identified which demonstrated low micromolar dissociation constants through fluorescence anisotropy after only two rounds of selection. One aptamer displayed a dose-dependent reduction of the clotting time in buffer, with a 20 ?M aptamer concentration achieving a nearly complete antidote effect. A DNA microarray was used to truncate two of the aptamers to decrease the cost of aptamer synthesis for medical applications. The second group of experiments tested a simple aptamer-based square capillary cell affinity chromatography system for selective capture of target cancer cells from a flowing suspension. Aptamers were immobilized on the inner surface of the capillary through biotin-avidin chemistry, the extent of which was controlled by adjusting the aptamer concentration. The device captured target leukemia cells in higher amounts than nontarget cells under a variety of conditions, and the capture efficiency was optimized to retain > 90% of cells. In addition, the system was used to capture two colon cancer cell lines by their respective aptamers at high efficiency. The capillary system could also detect stained cancer cells spiked in blood by imaging the length of the tubing and counting the fluorescent captured cells.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Jennifer Martin.
Thesis: Thesis (Ph.D.)--University of Florida, 2010.
Local: Adviser: Tan, Weihong.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2011-02-28

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2010
System ID: UFE0041958:00001

Permanent Link: http://ufdc.ufl.edu/UFE0041958/00001

Material Information

Title: DNA Aptamers for Bioanalytical Applications
Physical Description: 1 online resource (170 p.)
Language: english
Creator: Martin, Jennifer
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2010

Subjects

Subjects / Keywords: aptamer, capillary, cell, chromatography, dna, selex
Chemistry -- Dissertations, Academic -- UF
Genre: Chemistry thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Aptamers are oligonucleotides selected by the SELEX (Systematic Evolution of Ligands by EXponential enrichment) method that bind to a target with high affinity and specificity based on the three dimensional conformation they adopt. Selected for targets ranging from small molecules to proteins or whole cells, aptamers may similarly assume diverse roles such as drug delivery vehicles, decoys, inhibitory therapeutics, and affinity separation probes. This work investigates aptamers as inhibitory drugs and affinity ligands in two bioanalytical applications. In the first set of studies, an aptamer was selected to the peptide anticoagulant bivalirudin to serve as an antidote to the drug in instances of severe patient bleeding. The drug was immobilized on a monolithic column and binding sequences were eluted by salt gradient. The elution profile was compared to that of a blank column (no drug), and fractions with a chromatographic difference between drug-immobilized and blank counterparts were analyzed via real-time PCR (polymerase chain reaction) and used for further selection. Sequences were identified which demonstrated low micromolar dissociation constants through fluorescence anisotropy after only two rounds of selection. One aptamer displayed a dose-dependent reduction of the clotting time in buffer, with a 20 ?M aptamer concentration achieving a nearly complete antidote effect. A DNA microarray was used to truncate two of the aptamers to decrease the cost of aptamer synthesis for medical applications. The second group of experiments tested a simple aptamer-based square capillary cell affinity chromatography system for selective capture of target cancer cells from a flowing suspension. Aptamers were immobilized on the inner surface of the capillary through biotin-avidin chemistry, the extent of which was controlled by adjusting the aptamer concentration. The device captured target leukemia cells in higher amounts than nontarget cells under a variety of conditions, and the capture efficiency was optimized to retain > 90% of cells. In addition, the system was used to capture two colon cancer cell lines by their respective aptamers at high efficiency. The capillary system could also detect stained cancer cells spiked in blood by imaging the length of the tubing and counting the fluorescent captured cells.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Jennifer Martin.
Thesis: Thesis (Ph.D.)--University of Florida, 2010.
Local: Adviser: Tan, Weihong.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2011-02-28

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2010
System ID: UFE0041958:00001


This item has the following downloads:


Full Text





DNA APTAMERS FOR BIOANALYTICAL APPLICATIONS


By

JENNIFER ANNE MARTIN














A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA

2010

































2010 Jennifer Anne Martin
































To my family









ACKNOWLEDGMENTS

I thank my advisor Dr. Weihong Tan for his support and encouragement

throughout my graduate studies at the University of Florida. I also thank the many past

and present members of Tan Group for their aid in various aspects of the research

process. In particular, I appreciate the patience and support from Dr. Kwame Sefah and

Dr. Youngmi Kim. Dr. Parag Parekh has been invaluable in collaborating on the SELEX

project, while Dr. Joseph Phillips was irreplaceable in his assistance in the cell capture

studies. In addition, I am deeply grateful for the many discussions and friendships

cultivated during my tenure in the group, including Dimitri Van Simaeys, Elizabeth

Jimenez, Dalia Lopez Colon, Meghan O'Donoghue, Dr. Karen Martinez, Dr. Josh Smith,

Dr. Colin Medley, Dr. M.-Carmen Estevez, and Dr. Prabodhika Mallikarachy. Angela

Bojarski, Dr. Kathryn Williams, and Lori Clark have been of exceptional aid as well.

I wish to recognize my committee members, Dr. Nicolo Omenetto, Dr. Nicole

Horenstein, Dr. David Powell, Dr. Charles Cao, and Dr. Donn Dennis, who have guided

my development with constructive dialogue. Dr. Donn Dennis, along with Dr. Tim

Morey, Dr. Mark Rice, and Dr. Nikolas Gravenstein has been instrumental in the

development of the SELEX project through collaboration with the medical school.

Above all, I have relied on the support of my family to achieve all successes

throughout my life. My parents, Jim and Deborah Martin have encouraged me in all

pursuits, and my brothers, James and Justin have been right there beside them.

Without the inspiration from my family, my new Gainesville friends such as Megan and

Dan, fellow group members, and my crews from college and home, I would not be in the

position I am in today, and for that I am eternally grateful.









TABLE OF CONTENTS

page

ACKNOW LEDGMENTS ........................................ ............... 4

LIST O F TA BLES ......... ................ ..................... ...... ............... 9

LIST O F FIG URES........................................... ............... 10

LIST O F A BBR EV IAT IO N S ......... .................... ......... .................................. 13

A B S T R A C T .............. ..... ............ ................. .................................................. 1 6

CHAPTER

1 INTRODUCTION .................. ...... ......... ......... 18

Molecular Recognition ................ ......... ................. 18
Antibodies ...................... ................................ 19
Aptam ers ................................................................. ... ................. 20
C om prison to antibodies .................................................................... 21
SELEX procedure .......................................................... 22
Analytical applications of aptamers ............ .............. ....... .............. 23
Biomedical/therapeutic applications of aptamers ................. .................. 25
Introduction to Specific Dissertation Projects......................................................... 27
Selection of an Aptamer Antidote to an Anticoagulant ............... ............... 27
Heparin, protamine, and the importance of thrombin.............................. 28
Direct throm bin inhibitors and bivalirudin ............................. ... ............... 29
Square Capillaries as Simple Microfluidic Channels for Bioanalysis ............... 33
M icrofluidic devices .......................... .. .. ......... .._.. .............. 33
Antibodies and aptamers as stationary phase ligands............................. 37
Theory of cell capture ........... ................ ....................... 38
Overview of the Dissertation .............................. ..............._. ....................... 43

2 METHODOLOGY AND DEVELOPMENT OF BEAD-BASED APTAMER
SELECTION ...................... ............. ....................... 44

Introduction ....................... ...... ... ...... ......... ................ 44
Anticoagulation with Bivalirudin ........... ... ....... .. ............. ............. 44
Aptamer Development................................... ............... 45
Materials and Methods................................ ............... 46
Experim ental Design ...................... ............ .... .. .. ............................. 46
Design of PCR Primers and Library ................................................. 48
Optimization of PCR Conditions ...... ...... .......... .............. ....... .. 49
S election P procedure ....................... .. ........ ........ .. ............... .......... 50
PCR Amplification Procedure ......... .............. ....... ...... ...................... 51
PCR amplification of entire first pool .......................... ............... 51









Cycle optimization for preparative PCR .................. ............... 52
Preparative PCR ......................... ..... ..... ...... ..... ..... 53
Gel electrophoresis monitoring of products....................... ............... 53
Preparation of ssDNA.... ..................... ............ ........... 54
D esa lting of ssD N A ............................ ............................. ............. 55
Second and Succeeding Rounds of Selection............... ............ ............ 55
Flow Cytometry Monitoring of Selection Progression................................ 57
Flow cytom etry for DNA blocking step .................................................... 59
Flow cytometry for TCEP cleavage.............. ........................................... 59
R e su lts ..................................... ...................................................... 59
E nrichm ent of A ptam er P oo l.................................................... ... .. ............... 59
Selection M odification #1 .................................... ...................... ... ..... 61
Selection Modification #2......................................... .. ............... 63
Preparation of New Library and Prim ers................................. .................... 66
S e le c tio n .................. ......................................................... 6 7
Selection Modification #3......................................... .. ............... 68
C conclusions ................................................................ .. ..... ......... 69

3 MONOLITHIC SELECTION FOR BIVALIRUDIN .............................................. 71

Intro d uctio n ......... ........................................................................ .. 7 1
Materials and Methods................................................... 72
C olum n Im m obilization ...................... ....... ......... .. ............................ 72
Method Development ............................................... 73
P re lim inary S e lectio n .................................................................... ........ 78
Selection Conditions............................ ............... 79
Monitoring by Real-time PCR ............._. ................ ................... 80
Second and Subsequent Selection Rounds ...................... ................ 81
AlphaScreen Analysis of Selected Pools................... .... ... .............. 82
Sequencing of Selected Pool .................. ......................... 83
Sequence Alignment ....................... ......... ............... 85
Binding Studies ......................... .......... ......... 86
Results ........... .... .................. ............ ................... 87
S e election R esults................................................... ........... 87
AlphaScreen for Pool Selection ............... ............................................. 92
Sequence Alignment .......... .... ......... ............... 93
Binding Studies ......................... .......... ......... 94
C o nclusio ns ............ ........... .......... ... ....... .................................. 95

4 CHARACTERIZATION OF APTAMER AFFINITY AND APTAMER ANTIDOTE
T E S T IN G ................................................................................................... 9 8

Introduction .......................... ............... 98
Materials and Methods......................................... 100
B uffe r.................... ....... .. ........................................................... .. 100
FA Dissociation Constant Measurements............................. ........... 100
Clotting Experiments in Plasma................ .......... ............................. 103


6









Bivalirudin dose response curve ............ ... ........ .... ........ ...... 104
C oagulation testing ........................................ .. .. .................. 105
Clotting Experiments in Buffer............................... ............... 106
Optimizing fibrinogen concentration............. ....... .... ............... 106
Bivalirudin dose response curve ............ ... ........ .... ........ ...... 106
Coagulation testing ............ .... ...... ... ............. .......... 107
Truncation via DNA Microarray ...... ..................... ............... 108
Design.................................. ......... .......... 108
Microarray studies................................................ 109
Results and Discussion....................................... 110
Kd Characterization by FA ...................... ..................... ....... ........ 110
Clotting Studies in Plasma ............. .............. ............ ............. 112
Clotting Studies in Buffer...................................... ................ 113
Microarray Truncation ........................... ................ .. .................. 116
C conclusions .............. ......... ............ ............ ........... ........... 117

5 CANCER CELL CAPTURE USING APTAMER-IMMOBILIZED SQUARE
CAPILLARY CHANNELS: PROOF-OF-PRINCIPLE......................... ................ 119

Introduction ............... ...... .. ....... ...... ..... ................ ........... 119
M materials and M methods .......... .......... ......... ................ ......... ...... 120
Cell Culture and Buffers ........ ................................... ............... 120
Device C construction ........... ............................ ........... ... ............ 121
DNA Synthesis ......................... .......... ......... 122
Device Characterization ............................ ............... .. ............... 122
Controlling Degree of Sgc8 Immobilization ...... ........ .................... 123
Cell Capture Assays ........................................ ................. 123
Flow Rate Experiments ...................................... 124
Stained Cell Imaging Using Fluorescence Microscopy............................... 124
Cell Elution Efficiency ................ ... ....... .. ............. 124
Square and Circular Capillary Cell Capture Comparison .............................. 125
Microscopy and Image Analysis .......... ............ ........... ... ........ .... 125
Square and Circular Capillary Imaging ........ ....... ..... .......................... 126
Results and D discussion ........................ ....... ......... .. .. ........................... 126
S gc8 Im m mobilization ......... ................ ............ ................. .. .......... 126
Cell Volumetric Flow Rate Trends ...... ............... .. ......... ............... 127
Fluorescence Microscopy Imaging of Stained Cancer Cells........................ 128
Cell Capture Performance ......... ......................... ............... 130
Square and Circular Capillary Cell Capture Comparison .............................. 131
Imaging Properties of Square and Round Capillaries................ ............... 132
Conclusions .............. ......... .. ............... .................. .......... 133

6 OPTIMIZATION AND APPLICATIONS OF SQUARE CAPILLARY DEVICE........ 135

Introduction ............... ...... .. ....... ...... ..... ................ ........... 135
M materials and M ethods............................................. .............. ............... 136
Cell Culture, Buffers, Aptamer Synthesis, and Device Construction .............. 136









Flow Rate Optimization ................. ....... .... ......... .. .. ..... .......... 136
Optimization of Cell Washing Flow Rate ......... ... ...... ................. 137
Detecting CEM Cells in Blood ......... ............................... ............... 137
Cell staining .............. .... ... .............. ... ....... ........ 137
Removal of red blood cells................... ...................... ...... .. 138
C e ll ca ptu re .............................................. ........... ...... 13 9
C olon C ancer C ell C culture ................................. .......................... .............. 139
Colon Cancer Aptamer Sequences ......... .......... .. .......... ............ 139
Cell Capture of Colon Cells .......................... ....... .......................... 140
Testing efficiency of aptamers (without linker) for DLD-1 capture........... 140
DLD-1 cell capture using aptamers with linker....................................... 140
DLD-1 cell capture at low concentration using aptamer KDED2a-3......... 141
HCT 116 cell capture by aptamer KCHA10.......................................... 141
HCT 116 cell capture in non-enzymatic buffer by aptamer KCHA10 ....... 141
Microscopy and Image Analysis ...... ...................... .............. 141
Results and Discussion.......................... ......................... 142
O ptim ization of Flow Rate .......................................................................... 142
Optimization of Buffer Wash Flow Rate........ ....... ..................... 143
Detection of CEM Cells in Blood ........................................ ......... ............... 144
Colon Cancer Aptamers (without Linker) for DLD-1 Capture....................... 146
Colon Cancer Aptamers with poly(T)10 Linker for DLD-1 Capture.................. 147
HCT 116 Cell Capture by Aptamer KCHA10 in Non-enzymatic Buffer........... 148
C o n c lus io n s ............. ......... .. .. ......... .. .. ......... ................................ 15 0

7 SUMMARY AND FUTURE WORK ..................... ..................... .............. 154

S u m m a ry ............. ......... .. .............. .. .................................................. 1 5 4
Future Work .......................................... ........ ............... 155
Aptam er Antidote Project ................. ... ................... ...... ......... 155
C e ll A ffin ity P roje ct ............. ......... .. .................... ..................... ........... 156

LIST O F REFERENCES .. ................................. ........................................... 158

BIOGRAPHICAL SKETCH .................................................. 170









LIST OF TABLES

Table page

1-1 Aptamers in the clinical pipeline ......................... ..... ................... 25

1-2 Diffusion constants and times for samples of different diameters.................... 39

2-1 PCR preparation of entire pool ........................... ......................... 51

2-2 PCR cycle optimization ............................................ ................. 52

2-3 Preparative PCR.................. ................ ............... 53

2 -4 C o nd itio ns for initial l S E LE X ....................................................... .. .................. 56

2-5 Conditions for selection modification #1 ................................. .................. 61

2-6 Conditions for selection modification #2 with new library................................. 67

2-7 Conditions for selection modification #3 ................................. ........ ......... 68

3-1 Method development conditions #1 ...... ...................... .............74

3-2 Method development conditions #2 ...... ...................... ............. 74

3-3 Method development conditions #3...... ...................... ............. 78

3-4 PC R Preparation ........................................................... .............. 83

3-5 Fusion Primer PCR..................... ............................... 84

3-6 Probe Sequences ......... ......... ... ......... ................ ... ............. 94

4-1 M icroarray Truncation.............................. ............... 117

5-1 Physical properties of capillary ................................................ 122

5-2 Performance of capillary system............................... ............... 131









LIST OF FIGURES


Figure page

1-1 Secondary structure of aptamer sgc8........................................... 21

1-2 Structure of LNA ......................................................... ....................... 22

1-3 The S ELEX procedure .......................................... ................ .............. 24

1-4 Aptamer Ch-9.3t and cDNA strand for use as anticoagulant/antidote pair for
fa c to r IX ...................................................................................................... 2 6

1-5 Thrombin structure (left), heparin binding (center), and steric constraints
associated with heparin binding fibrin-bound thrombin prior to antithrombin
(A T ) (rig ht). ..................................................................... ............. 2 8

1-6 Bivalirudin peptide sequence and binding/cleavage by thrombin .................... 30

1-7 Comparison of square and circular capillary geometries............... ........... 36

1-8 Antibody immobilization to stationary phase ................................................. 37

1-9 Processes occurring in cell adhesion. ............... ........ ..................... 41

2-1 Mechanisms of aptamer/peptide interactions for restoring anticoagulant
activity .............................. .................. ................................ 45

2-2 PCR amplification of initial library....... .................................. 52

2-3 Schematic of flow cytometry instrumentation.................... ................... 58

2-4 Flow cytometry data of initial selection.. ............................................................ 60

2-5 Flow cytometry data of selection modification #1. ....................... ............... 62

2-6 PCR of unlabeled DNA library ........................ ......................... ....... ........ 63

2-7 Streptavidin blocking with excess random library. ....................... 64

2-8 TCEP concentration optim ization. .............. ................. ................................ 64

2-9 TCEP interference with streptavidin/biotin binding. ................. .................. 66

2-10 PCR amplification program for second initial library. ................ ................ 66

2-11 Flow cytometry of selection modification #2. ................ ............. ............... 67

2-12 Flow cytometry data of selection modification #3. .............. .... ...... ........... 69









3-1 Chromatograms from method development.. .............. .......... .................... 76

3-2 Confirmation of LPC method. .......... .... ..... ... ............... ....... ........ 77

3-3 Chromatogram under SELEX-like conditions. .............. ......... ....... ............... 78

3-4 Overlay of proof-of-concept studies of selections on drug and blank disks........ 79

3-5 Amplification plots of IL-4 plasmid cDNA .................. ....................... 81

3-6 Schematic of AlphaScreen assay ....................................... ......................... 82

3-7 Optimization of pool volume for fusion primer PCR........................... ........ 84

3-8 Agarose gel electrophoresis analysis of fusion primer-amplified, purified pool... 85

3-9 Sample alignment of JPB2 using MAFFT............. ................................... 86

3-10 Chromatograms of selection rounds ........ .... .......... .. .......... .. ............ 88

3-11 qPCR results for each round of selection.. ................ ........ ....................... .. 91

3-12 AlphaScreen assay of binding of different pools with 100 nM or 100 pM drug... 92

3-13 Flow cytometry binding studies of drug and aptamer candidates .................... 95

4-1 Secondary structure and modifications of the anti-VEGF aptamer pegaptanib
(M acugen). .............. ......... ..... .......... .............................. 99

4-2 M odes of coagulation. .................. ............ ............. ....... ............... 104

4-3 Truncation of JPB5 ....................................... ............... 109

4-4 Kd plot of thrombin/15-mer thrombin aptamer using FA ............................. 110

4-5 Kd curves of aptamers and controls........... .... ........ ....... ... ............ 111

4-6 Dose response curve of bivalirudin in human plasma. .................................. 112

4-7 Effect of JPB5 and TV03 control sequence on clotting time in plasma............. 113

4-8 Optimization of conditions for buffer experiments...... ......... .... ...... ... ...... 114

4-9 Effect of JPB5 and TV03 control on clotting time................ ........................... 115

4-10 Image of DNA microarray target binding. ........ ... ..................................... 116

5-1 Schem atic of setup and im m obilization. .................................. .................. .. 121

5-2 Fluorescence intensity increase as aptamer concentration is increased......... 127


11









5-3 Confocal images of FAM-labeled sgc8 immobilized inside a capillary at
various aptam er concentrations ............... .............................................. 127

5-4 Confocal images of cells captured in square capillaries at different volumetric
flow rates. ............ ................... ........ ..... ........... .......... 129

5-5 Simultaneous fluorescence microscopy imaging of cells in a square capillary
stained with two different dyes ................................. 130

5-6 Imaging comparison of square and round capillaries .............. ............... 132

6-1 Comparison of volumetric flow rate to capture efficiency of cells ................ 143

6-2 Comparison of capture efficiency at two different cell washing flow rates........ 144

6-3 Capture of CEM cells spiked in blood....................... ............... 145

6-4 Comparison of capture efficiency of colon cancer cell line DLD-1 with
aptamers KDED2a-3 and KCHA10 ........... ............................. ....... ............ 146

6-5 Flow cytometry comparison of aptamers binding to DLD-1 with and without
poly(T)10 linker. .................................................................... ... ......... 147

6-6 Capture efficiency of DLD-1 aptamers with 3'-poly(T)o1 linker.......................... 148

6-7 Flow cytometric comparison of aptamer binding with HCT 116 cells diluted in
BB2 or with non-enzymatic buffer (NEB). ...................... ..... ..... ............ 149

6-8 Overall capture efficiency of colon cancer cell lines DLD-1 and HCT 116 in
buffer. ............. ........ .......................................................... 150









LIST OF ABBREVIATIONS


AT antithrombin

ATP adenosine triphosphate

BB binding buffer

BB2 binding buffer #2

BSA bovine serum albumin

CAC cell affinity chromatography

CCD charge-coupled device

CD cluster of differentiation

cDNA complementary DNA

CIM convective interaction media

CTC circulating tumor cell

Cy5 cyanine derivative 5

DFP diisopropyl fluorophosphate

DHFR dihydrofolate reductase protein

DMSO dimethyl sulfoxide

DNA deoxyribonucleic acid

DNAse deoxyribonuclease

dNTP deoxyribonucleotide triphosphate

dsDNA double-stranded DNA

DTI direct thrombin inhibitor

EB elution buffer

EDTA ethylenediaminetetraacetic acid

ELISA enzyme-linked immunosorbent assay

EpCAM epithelial cell adhesion molecule









FA

FACS

FAM

FBS

FDA

FITC

HIT

HPLC

ICBR

IDT

IgG

ITC

IU

LB

LMWH

LPC

MACS

MALDI/TOF

NEB

NHS

PB

PBS

PCR

PE

PEG


fluorescence anisotropy

fluorescence activated cell sorting

5/6-carboxyfluorescein

fetal bovine serum

Food and Drug Administration

fluorescein isothiocyanate

heparin-induced thrombocytopenia

high performance liquid chromatography

Interdisciplinary Center for Biotechnology Research

Integrated DNA Technologies

immunoglobulin G

isothermal calorimetry

international unit

red blood cell lysis buffer

low molecular weight heparin

low-pressure chromatography

magnetic activated cell sorting

matrix-assisted laser desorption ionization/time-of-flight spectrometry

non-enzymatic buffer

N-hydroxysuccinimide

physiological buffer

phosphate buffered saline

polymerase chain reaction

R-phycoerythrin

polyethylene glycol









PF4 platelet factor 4

PTFE polytetrafluoroethylene

qPCR quantitative PCR (real-time PCR)

RNA ribonucleic acid

SELEX systematic evolution of ligands by exponential enrichment

S-Hir53-64 sulfated exosite 1 binding region of hirudin

ssDNA single-stranded DNA

TBE tris-borate-EDTA

TCEP tris(2-carboxyethyl)phosphine

TFA trifluoroacetic acid

TMR tetramethylrhodamine

tris tris(hydroxymethyl)aminomethane

tRNA transfer RNA

UV ultraviolet

WB washing buffer

WB2 washing buffer #2

ARn baseline subtracted fluorescence









Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy

DNA APTAMERS FOR BIOANALYTICAL APPLICATIONS

By

Jennifer Anne Martin

August 2010

Chair: Weihong Tan
Major: Chemistry

Aptamers are oligonucleotides selected by the SELEX (Systematic Evolution of

Ligands by EXponential enrichment) method that bind to a target with high affinity and

specificity based on the three dimensional conformation they adopt. Selected for

targets ranging from small molecules to proteins or whole cells, aptamers may similarly

assume diverse roles such as drug delivery vehicles, decoys, inhibitory therapeutics,

and affinity separation probes. This work investigates aptamers as inhibitory drugs and

affinity ligands in two bioanalytical applications.

In the first set of studies, an aptamer was selected to the peptide anticoagulant

bivalirudin to serve as an antidote to the drug in instances of severe patient bleeding.

The drug was immobilized on a monolithic column and binding sequences were eluted

by salt gradient. The elution profile was compared to that of a blank column (no drug),

and fractions with a chromatographic difference between drug-immobilized and blank

counterparts were analyzed via real-time PCR (polymerase chain reaction) and used for

further selection. Sequences were identified which demonstrated low micromolar

dissociation constants through fluorescence anisotropy after only two rounds of

selection. One aptamer displayed a dose-dependent reduction of the clotting time in









buffer, with a 20 pM aptamer concentration achieving a nearly complete antidote effect.

A DNA microarray was used to truncate two of the aptamers to decrease the cost of

aptamer synthesis for medical applications.

The second group of experiments tested a simple aptamer-based square

capillary cell affinity chromatography system for selective capture of target cancer cells

from a flowing suspension. Aptamers were immobilized on the inner surface of the

capillary through biotin-avidin chemistry, the extent of which was controlled by adjusting

the aptamer concentration. The device captured target leukemia cells in higher

amounts than nontarget cells under a variety of conditions, and the capture efficiency

was optimized to retain >90% of cells. In addition, the system was used to capture two

colon cancer cell lines by their respective aptamers at high efficiency. The capillary

system could also detect stained cancer cells spiked in blood by imaging the length of

the tubing and counting the fluorescent captured cells.









CHAPTER 1
INTRODUCTION

Molecular Recognition

Molecular recognition between two biomolecules has long been recognized as a

cornerstone of specificity and control of complex biological processes.1 A number of

factors mediate this development, including the shape of the two binding surfaces,

hydrogen bonding, ionic, dipole, and solvent interactions, and van der Waals forces.1

These binding events may occur between proteins and small molecules to induce

signaling processes,2 amongst proteins and nucleic acids as in DNA/histone

interactions, and in catalysis of various substrates by enzymes.3 Enzymes catalyzing

reactions in sugars (protein/carbohydrate) are well-known interactions, for example the

a-amylase enzyme family recognizing starch.4 RNA/small molecule interactions are

also of interest, such as those displayed by riboswitches.5 Additional interactions of

significance include protein/protein binding, for example dynein or kinesin interacting

with microtubule proteins.6

Researchers have taken advantage of molecular recognition principles to form the

foundation of current target detection methods and modern diagnosis/treatment of

disease. At the molecular level, biosensors have been designed for detection and study

of different properties of small molecule/protein and protein/protein interactions.79

Molecular recognition has also been utilized in designing novel therapeutic drugs.

Examples include targeting of dihydrofolate reductase protein (DHFR) by the

chemotherapy drug Methotrexatelo and the antibiotic Trimethoprim.11 Also, the

aminoglycoside paromomycin was found to bind to specific RNA in E. coli to exhibit

antibiotic effects.12









Despite the term "molecular recognition" the phrase also refers to specific cellular

recognition events resulting from the affinity between a ligand and cell surface marker.

Analytically, ligands immobilized on a surface have been used to detect and/or purify

cellular targets.13 Furthermore, biomarker expression has been manipulated in disease

detection, or targeted for pharmacological response.14

Molecular recognition principles have been extremely effective for both analytical

and pharmaceutical applications regardless of the target. In particular, antibodies

raised for targets either expressed on cell surfaces or found in the blood have made

tremendous progress as detection reagents and therapeutics for various diseases.15

Antibodies

Antibodies are well-known molecular recognition agents deserving special

attention due to rapid development in the last 25 years.15 Antibodies can be

immobilized as a stationary phase for selective detection or capture of target

molecules.16, 17 The proteins are also used to both preconcentrate cellular samples and

stain cancer cells for immunocytochemical detection.18 Additionally, vaccines utilizing

antibody/antigen response have resulted in protection against influenza and the

eradication of smallpox, a disease which was responsible for hundreds of millions of

deaths in the twentieth century.19 More recently, the utility of antibodies used in disease

treatment has been investigated with a high degree of success. Since the first CD3-

specific antibody was approved for the treatment of acute transplant rejection in 1986,20

more than 30 antibodies have been approved for use in various indications.15 Of the

$15 billion in revenue, 78% derives from monoclonal antibodies targeting cancer and

autoimmune diseases.20 Some FDA-approved antibodies include tumor-necrosis factor

specific antibodies infliximab and adalimumab, the human epidermal growth factor









receptor 2 antibody trastuzumab, the vascular endothelial growth factor targeting

antibody bevacizumab, and the CD20-specific rituximab antibody. Phase III trials are

underway on 26 more antibody drug candidates, 9 of which are for diseases designated

"orphan diseases," or rare diseases afflicting only a small number of patients.15' 21

Despite the tremendous achievements of antibodies, several challenges remain.

Antibody generation requires a biological system, limiting the conditions for optimization

to physiological conditions, possibly reducing their utility for in vitro applications such as

detection platforms.22' 23 Due to the biological requirement, antibodies cannot be raised

against biological toxins which would damage the host system.24 Antibodies also suffer

from significant batch-to-batch variability in function. Furthermore, it is difficult to modify

antibodies with labels to facilitate capture or detection while still retaining function, and

the proteins are susceptible to degradation upon temperature changes and prolonged

storage. Another concern is the time and expense associated with the antibody

screening process.25 In addition, a limited number of available antibodies,26 difficulties

with cross-species reactivity (i.e. applying a murine-generated antibody to detect the

human version of the target), and generation of an immune response upon

administration24 also present serious challenges for antibody applications.23 Molecules

designed to address these limitations would be highly attractive candidates for various

target/ligand binding functions.

Aptamers

Aptamers are DNA or RNA oligonucleotides that bind specifically to a target based

on their three dimensional conformation. This structure is the result of a combination of

hydrogen bonding, electrostatic and van der Waals interactions, Watson-Crick base

pairing, and stacking of aromatic rings.22 27 Aptamers usually adopt stem/loop









secondary structures, as seen in Figure 1-1. The loop regions are primarily responsible

for target interactions, while the remainder of the sequences aid in stabilizing the

structure.













*A
-o0 OC
SOG
OT



Figure 1-1. Secondary structure of aptamer sgc8.28

Aptamers were first described in 1990 by two separate groups targeting either an

organic dye29 or a DNA polymerase.30 Since the inception of the technology, aptamers

have been selected for a vast range of targets including metal ions,31' 32 small organic

molecules,29, 33-38 nucleotides and derivatives,39-41 cofactors,42-44 nucleic acids,45' 46

amino acids,47-49 carbohydrates,50' 51 antibiotics,52-54 peptides,55, 56 proteins,57-61 whole

cells,28, 62-67 viruses and virus infected cells,68' 69 and bacteria.70,71

Comparison to antibodies

When compared with antibodies, aptamers demonstrate similar binding affinities

(pM-nM), yet have the following advantages: 1) They are chemically synthesized, and

can therefore be selected for virtually any target, including biological toxins; 2) Chemical

synthesis eliminates batch-to-batch variability and reduces cost of production; 3)









Experimental conditions can be manipulated to select aptamers with specific properties;

4) Aptamers can be easily modified with labels to facilitate target purification or

detection by fluorescent dyes; 5) The sequences are relatively stable to degradation

from high temperatures or long storage times; 6) No evidence of immunogenicity has

been observed; 7) Animals are not required for aptamer selection.22-24'72

One major drawback of aptamers is degradation by nucleases when in vivo

applications are desired. Since most polymerases are unable to incorporate modified

bases in the PCR, several groups have proposed different post-SELEX modifications to

minimize the effects. For RNA, typically attacked by RNAse present in nearly all

biological samples,26 the 2'-OH groups can be substituted with 2'-F, -NH2, or -OCH3

groups, which have proven successful at inhibiting degradation.73-75 Substitution of

pyramidines at the 5-position with I, Br, CI, NH3, or N3 has also been shown to improve

stability.26 Additionally, 3'- and/or 5'-end capping (ex. polyethylene glycol) is able to

protect sequences against exonuclease activity.22 Incorporating locked nucleic acids

(LNA) consisting of a methylene bridge from the 02' to C4' of the sugar into the aptamer

structure is also useful in stabilizing the structure of oligonucleotides (Figure 1-2).76

S- Base






Figure 1-2. Structure of LNA.76

SELEX procedure

Aptamers are selected using a method known as SELEX (Systematic Evolution of

Ligands by EXponential enrichment). The SELEX method is depicted in Figure 1-3, and









certain aspects are described in more detail in future chapters. The SELEX procedure

begins with 1013-1015 unique sequences from a chemically synthesized, randomized

oligonucleotide library competing for binding to the target. Nonbinding sequences are

partitioned from binding oligonucleotides, and binders are eluted from the target. This

partitioning step is the main determinant of the efficiency of the selection. The binding

sequences are PCR-amplified and converted to single-stranded DNA (ssDNA) for the

next round of selection. After the first round, researchers may institute a counter-

selection step, in which sequences binding to a control (such as support matrix) are

removed from solution. Sequences that do not bind to the control are retained for future

SELEX rounds. The process is repeated in cyclical fashion until the final pool is

enriched for sequences binding to the target. The number of cycles varies depending

on parameters such as selection stringency and the efficiency of the partitioning

method; however, a typical selection requires an average of 12 cycles and a timeline of

2-3 months.24' 77 Following enrichment, the oligonucleotide pool is sequenced, and the

oligonucleotides are grouped according to intermolecular homology using programs

such as ClustalX or MAFFT. Several sequences from each family are synthesized to

test for target binding, and successful candidates are characterized in terms of binding

affinity, target specificity, etc.

Analytical applications of aptamers

Antibodies are frequently utilized as immunoaffinity column ligands, but their

linkage to the column is nontrivial, resulting in couplings that are not uniform. This

reduces the binding capacity and in some cases leads to leaching of the separation

phase from the column.78 In contrast, aptamers are easily chemically modified for

linkage to a column, and are smaller in size than antibodies, producing a higher ligand









density for chromatography stationary phases. Aptamers have been employed as

affinity ligands for small molecules and proteins,79-81 and to discriminate between chiral

molecules in a separation due to their exquisite specificity properties.82 83 Aptamers

have also been used as biosensors for specific targets due to their high ligand density

and ease of chemical modification. The probes have been chemically modified to bind

to a surface for acoustic sensing,84 optical sensing,85-87 cantilever-based biosensors

enabling label-free detection,88 and fluorescent signal sensing.89 90



I2C












Channel are capable of selectively capturing target cells from buffer or blood.91-94



discussed later in this chapter.
Figure 1-3. The SELEX procedure.28

At the cellular level, aptamers have found use as ligands for separation,

enrichment, and/or detection. For example, aptamers immobilized in a microfluidic

channel are capable of selectively capturing target cells from buffer or blood.91'94 Cell

separations have important implications in areas such as CD4+ T lymphocyte counting,

blood cell isolation, bacterial identification, rare cell capture, and cancer cell detection.95

Cell separation plays a central role in the work completed for this dissertation, as

discussed later in this chapter.









Biomedical/therapeutic applications of aptamers

The unique properties of aptamers present exciting opportunities for biomedical

diagnostic and therapeutic applications. Aptamers have been selected to serve as

histological detection reagents for glioblastoma96 or in proximity-dependent assays

(ELISA),97 and as throughput validation ligands for drug discovery.98 The probes have

also been developed for in vivo imaging99' 100 and have demonstrated promising results

for reducing side-effects by targeted drug delivery.101-103

Table 1-1. Aptamers in the clinical pipeline104
Aptamer, Target Company Disease Stage of Development
FDA-approved AMD 2004;
Pegaptanib Eyetech AMD, Phase II for diabetic
(Macugen), Pharmaceuticals/ diabetic retinopathy: improved vision,
VEGF-165 Pfizer retinopathy decreased edema
REG-1
(RB006/RB007), Regado
Factor IXa Biosciences PCI, CABG Phase II trial completed
AS1411,
Nucleolin Antisoma RCC, AML Phase II underway
NU172, Thrombin Nuvelo/Archemix PCI, CABG Phase II planned
ARC1779, von TMA, TTP,
Willebrand factor Archemix CEA Phase II underway
Abbreviations: AMD, age-related macular degeneration; PCI, percutaneous coronary
intervention; CABG, coronary artery bypass graft; RCC, renal cell carcinoma; AML,
acute myeloid leukemia; TMA, thrombotic microangiopathies; TTP,
thrombocytopenic purpura; CEA, carotid endarterectomy.

The FDA approval of the aptamer drug Macugen targeting vascular endothelial

growth factor (VEGF) in 2004 was a breakthrough for aptamer pharmaceuticals,

renewing interest in aptamer based therapeutics.105 Therapeutic aptamers can be

divided into several categories104 including decoy-like aptamers targeting proteins that

naturally bind nucleic acids,106,107 regulatable aptamers which effectively serve as their

own antidote,108 multivalent aptamers to improve aptamer binding,109' 110 and inhibitory

aptamers that target proteins, transcription factors, cell-surface markers, etc. in order to










treat human disease.111' 112 Aptamer drug candidates currently in the clinical pipeline

are summarized in Table 1-1. Several of these aptamers are progressing to Phase II

studies, and many of the targets are coagulation factors.

Inhibitory aptamers have shown particular promise in the role of anticoagulants,

targeting various points of the coagulation cascade. Aptamers to thrombin,113 factor

IX,114 factor VII,115 factor X,116 protein C,117 and von Willebrand factor118 have been

shown to successfully modulate thrombus formation, with cDNA antidotes able to

restore normal activity. In particular, Rusconi and coworkers developed the factor IX

aptamer (Table 1-1) that is now in the clinical pipeline. The aptamer was modified with

a 5'-cholesterol and a 3'-inverted deoxythymidine, and consists of pyramidines modified

at the 2-position with fluorine (Figure 1-4).108 Upon addition of 2'-O-methylated cDNA

corresponding to the contiguous stem region at the 5'-end of the aptamer, the

anticoagulant effect was neutralized and clotting activity commenced in animal models.

Aptamer/cDNA development is relatively cost-prohibitive, so studies have been carried

out to find different methods to reverse anticoagulant binding.119

Ch-9.3t Antidote 5-2C Complex
A U
A G UAAUcCU
U C 5 c 5'c- G G
G-U g g-C C
C-G c c-G C
-C g g-C U
: g\ g-C c
u\ u-A C
a a-U C

S gu- C Ui dT 3'
--U u u-A
G--C c c-G
G--C c c-G
G-C c c-G
G--C c c-G
U--A a a-U
5'Ch-A-UidT 3' 3 3'u-A-Ch 5'
Active Inactive

Figure 1-4. Aptamer Ch-9.3t and cDNA strand for use as anticoagulant/antidote pair for
factor IX.108









Recent work has investigated the use of light to photoregulate aptamer antidote

activity by introducing a "caged" structure into an aptamer,120 or an azobenzene moiety

to reverse anticoagulation.121 Clearly the use of light to inhibit internal bleeding is not

feasible, since the light would not penetrate through the skin, but the design may have

topical applications. Researchers have used aptamers and cationic porphyrins to serve

as an anticoagulant/antidote pair, however this method will only work for aptamers

known to form a G-quartet.122 A polymeric antidote has shown a promising universal

antidote response for various anticoagulant aptamers, but the mechanism of action is

still unclear and the research is still in the preliminary phase.119 Therefore, developing a

DNA aptamer to a currently available pharmaceutical agent will be another major theme

for this dissertation work.

Introduction to Specific Dissertation Projects

Selection of an Aptamer Antidote to an Anticoagulant

Anticoagulants are one of the drug classes with the highest instances of adverse

drug reactions and medication errors.123 These actions directly correlate to an

increased occurrence of complications such as severe bleeding that increase patient

morbidity and mortality.119 Additionally, blood transfusions are required for 5-10% of

patients with severe bleeding, at an estimated cost of $8,000-$12,000 per incident.108

An ideal anticoagulant would exhibit minimal adverse drug reactions, and would have a

safe and effective antidote readily available to reverse severe patient bleeding.

Heparin and protamine are the only anticoagulant/antidote pair commonly used in

clinics, but both drugs have considerable risk associated with their use. Heparin

demonstrates an unpredictable anticoagulant response, and some patients cannot be

administered heparin in any circumstance due to severe physiological responses. A









variety of synthetic anticoagulant drugs were developed to improve upon the challenges

created by heparin use. Bivalirudin is one such synthetic peptide anticoagulant that has

several advantages over heparin, including a more predictable drug response, yet does

not currently have an antidote available. Therefore, the objective of this work was to

provide an antidote to bivalirudin to introduce a safe and reliable anticoagulant/antidote

pair.

Heparin, protamine, and the importance of thrombin

Thrombin is a common target for anticoagulant drugs due to its position at the

juncture of the coagulation cascade and the variety of roles it plays in the clotting

process.124 Most importantly, thrombin activation of soluble fibrinogen to insoluble fibrin

and fibrin crosslinking by thrombin-activation of factor XIII is a critical step of the clotting

process.



Thrombin Soluble thrombin Fibrin-bound thrombin

Active
(catalytic)
site

To Fibrinrin

Apolar X
site Heparin-binding X Heparin
site (exosite 2)
Fibrinogen-binding
site (exosite 1)
Heparin
Figure 1-5. Thrombin structure (left), heparin binding (center), and steric constraints
associated with heparin binding fibrin-bound thrombin prior to antithrombin
(AT) (right).125

The binding sites of thrombin consist of two positively charged exosites and the

active site where catalytic activity takes place (Figure 1-5).125-128 Heparin indirectly









targets the protein thrombin by binding to both antithrombin (AT; a naturally occurring

thrombin inhibitor) and exosite 2 of thrombin.129 One problem with heparin is that it

cannot inhibit fibrin bound thrombin, possibly due to steric constraints. If heparin docks

to exosite 2 without previously binding antithrombin, it can form a bond with thrombin-

bound fibrin, actually strengthening the clot (Figure 1-5).126 Heparin also binds to

certain plasma proteins in the blood, resulting in an unpredictable anticoagulant

response and increased patient monitoring. Heparin is neutralized by platelet factor 4

(PF4) a product of activated platelets.130 A major challenge to heparin use is that

heparin completed with PF4 or other plasma proteins can stimulate a response called

heparin-induced thrombocytopenia (HIT), which can cause severe reactions in some

patients. The use of low molecular weight heparin (LMWH) alleviates several, but not

all of these concerns. The antidote to heparin, protamine, also has serious side effects

associated with it's administration, including increased pulmonary artery pressure,

decreased systolic and diastolic blood pressure, impaired myocardial oxygen

consumption, and reduced cardiac output, heart rate, and systemic vascular

resistence.119 Therefore, there is an obvious need for alternative anticoagulant/antidote

pairs with a safer therapeutic profile.

Direct thrombin inhibitors and bivalirudin

In order to improve the safety profile of anticoagulants, synthetic thrombin

inhibitors which directly interact with thrombin, termed direct thrombin inhibitors (DTI),

were designed. One such drug, bivalirudin, is a promising alternative to heparin

anticoagulation. In comparison to heparin, bivalirudin generates a more predictable

anticoagulant response because it does not bind to other plasma proteins. It also binds

both fibrin-bound and free thrombin, is not inactivated in the presence of PF4, and does









not induce HIT.126, 131 In addition, bivalirudin can be chemically synthesized using solid

state chemistry, while heparin is obtained from animals.

Bivalirudin is a 20 amino acid peptide (Figure 1-6) used in the treatment of

coronary angioplasty in unstable angina patients. The drug is a derivative of the leach

anticoagulant hirudin that demonstrates essentially irreversible binding to thrombin at

(inhibition constant) Ki= 231 fM.132 Femtomolar inhibition constants are generally

considered to represent irreversible binding. The engineering of hirudin to bivalirudin

has produced a reversible anticoagulant with Ki= 2 nM.125 The peptide has a molecular

weight of 2180 Daltons, and forms a 1:1 complex with thrombin that is slowly cleaved at

the Arg3-Pro4 bond with a half-life of 25 minutes, but currently no antidote exists for

rapid reversal in the instance of severe bleeding.133

Proteolytic cleavage

+H 3N -OOP O. (N: 0P0 0000 coo-

Binds to active site iGfl')4 Binds to -iirrir:-, n-bincling
-'giojn of thrombin region of thrombin
(exosite 1)


Hepanr.-r.rng ~sre^ + H3N %F,

+H3N Thrombin Thrombin
:b.- Fl:nricgailn-tirirg Mle
.'o-


C-terminal dodecapeptide

Figure 1-6. Bivalirudin peptide sequence and binding/cleavage by thrombin.125

The N-terminus of bivalirudin binds to the active site of thrombin, and the C-

terminus docks to exosite 1 (Figure 1-6). Initially, bivalirudin competes with fibrinogen

for access to the exosite 1.132 Binding to thrombin can be considered to occur in 4









steps: 1) The C-terminus binds to exosite 1 with a Kd of 0.75 pM; 2) A conformational

change occurs in thrombin; 3) Arg3 of bivalirudin interacts with the active site; 4)

Another thrombin conformational change occurs, with a rate constant of 30 s-1.134-136

Steps 3 and 4 result in a stable inhibitor-thrombin complex. After cleavage of the

Arg3/Pro4 bond, the N-terminus is released from the active site, and the C-terminal

fragment interaction at exosite 1 becomes a low-affinity weakly-competitive inhibitor of

fibrinogen.131 Fibrinogen easily displaces the fragment from the binding site and

resumes conversion to fibrin. The mechanism of action of the peptide is important to

consider for future work on selecting an aptamer for the peptide.

In initial studies, the drug marketed today as bivalirudin (Angiomax) was referred

to as hirulog-1, short for hirudin analog 1. Hirulog-1 was designed based on linking the

exosite 1 and active site (with slight modifications to make the binding reversible)

binding domains of hirudin with a (Gly)4 linker.137 Binding tests comparing hirulog-1 with

(Gly)6 and (Gly)8 linkers produced very similar Ki values (2.1-3.0 nM) of which the (Gly)4

was the lowest. Interestingly, when the D-Phel of hirulog-1 was replaced by L-Phel,

the inhibition constant was increased from 2.1 nM to 156.0 nM, demonstrating the

specificity of D-Phe for thrombin binding. These studies also showed that the active

site-binding region and the exosite 1 binding domain separately lacked the ability to

significantly inhibit thrombin, as both inhibition constants were >2 pM. Instead, the

combination of these domains results in the anticoagulant properties of the drug.

Furthermore, the group studied which terminus of hirulog-1 occupies the binding site by

testing the ability of hirulog-1 to block a modification of Ser195 by [14C] DFP (diisopropyl

fluorophosphate). The results show that a 3- and 30-fold excess of drug to thrombin









blocked the modification, while much higher concentrations of S-Hir53-64 (sulfated

exosite 1 binding region) failed to block the same modification. This highlights the fact

that the N-terminus of hirulog-1 occupies the active site of thrombin. Further studies in

the same literature proved the specificity of the interaction of hirulog-1 with thrombin by

showing that it does not affect human plasmin or bovine trypsin activity.

Additional studies on hirulog-1 probed the binding of the drug to various forms of

thrombin.136 Nanomolar Ki's were reported for human a- and -thrombin, but no

inhibition was observed for y-thrombin or trypsin with 7.85 pM hirulog-1 concentrations.

The binding of the drug to -thrombin can actually be considered to be a positive result,

since of all of the proteolytic cleavage products of prothrombin (a-, 3-, y-, E-, and 5-

thrombin), only a- and -thrombin retain the ability to significantly interact with

fibrinogen. However, the other forms are still capable of cleaving small synthetic

molecules and proteins, which include factor XIII, AT, and prothrombin.136' 138

Therefore, hirulog-1 is a specific inhibitor of the two reported fibrinogen-catalyzing forms

of human thrombin.

Final studies relevant to the scope of this project report on the crystal structure of

hirulog-1 binding to thrombin.139 Only 8 of the 20 amino acid residues were observed in

the electron density map, corresponding to the first 3 residues of the N-terminus, and

the Asp1 1-lle15 of the C-terminus. The disorder of Pro4-Gly10 is believed to be a result

of the cleavage of the Arg3-Pro4 bond. The residues adjacent to the C-terminus and

not involved in the electron density map, Glu17-Tyr20, coil into one turn of a 310 helix.

Additionally, the side-chains of D-Phel-Arg3 play a major role in stabilization inside the









active site.128,140 This structural information will form the basis for the decisions made

regarding the methodology described in Chapters 2-3.

Square Capillaries as Simple Microfluidic Channels for Bioanalysis

Cell separations are important in a variety of applications for both basic scientists

and clinicians. Standard microfluidic devices have been successful in cell capture, yet

key drawbacks in design and fabrication may promote the use of capillary-based

systems. This work explores the utility of aptamers as stationary phase ligands in

simple capillary affinity chromatography systems for the selective capture of target

cancer cells.

Microfluidic devices

Microfluidic technology has emerged as a major research area in the last 10 years,

as evidenced by publication of nearly 10,000 papers on the topic.141 The blanket term

"microfluidics" refers to technology ranging from "dipsticks" for simple measurements

(for example pH testing), to more complex lateral flow tests and "lab-on-a-chip" devices

with intricate microvalves and pumps integrated into the device. Some attractive

features of microfluidic platforms encouraging the growth in research are: 1) Portability;

2) High sensitivity; 3) Low cost per test; 4) Short data acquisition time; 5) Less

laboratory space required. Additional advantages incurred by microfluidic

minuturization include lower sample/reagent consumption, consistent and controlled

laminar flow properties, and the possibility to multiplex measurements.

Thus far, microfluidic systems have been administered in a wide variety of

applications. Analytical devices for detecting pregnancyl2 and drug use143 are

particularly well known to the general population. Detection of biowarfare agents has

become increasingly more important as well, with devices capable of recognizing









bacterial144 and molecular targets such as ricin, Shiga toxin I, and Staphylococcal

enterotoxin B.145 Microfluidic systems have also been applied to detection of

contaminants in food and water.146' 147

Biochemical applications are also of major interest for microfluidic devices,

observed in research in microfluidic bioreactors,148 radiopharmaceutical synthesis,149

and high throughput screening for drug discovery.150 Particularly interesting are the

possibilities of microfluidic systems for cell capture and detection, a process known as

cell affinity chromatography (CAC). Cell separations have thus far been utilized for

applications such as CD4+ T lymphocyte counting, blood cell isolation, bacterial

identification, rare cell capture, and cancer cell detection.95 The development of devices

capable of selectively capturing cancer cells may enable advancement in cancer

diagnosis and monitoring treatment progress, which significantly enhances the

outcomes for cancer patients.92-94, 151,152

Either physical or chemical methods may be used to sort cell subpopulations.

Physical methods such as dielectrophoresis and sedimentation have previously been

successful, but these methods rely on significant differences between the physical

properties of the target cells and the matrix cells.95 Chemical methods were therefore

designed to separate cells based on their unique surface chemistry. Most commonly,

fluorescence activated cell sorting (FACS) is utilized for cell separation.92'95 While the

method tends to reproducibly generate highly pure subpopulations, FACS can be

expensive to operate, requires highly trained personnel, and along with a similar

technique termed MACS (magnetic activated cell sorting) necessitates preprocessing

the cells with dyes or fluorescently-labeled antibodies, dramatically increasing the









sorting time of the sample. Similarly, in circulating tumor cell (CTC) analysis the

prevailing method is the commercially-available CellSearch technology, which also

employs immunological assays to differentiate tumor cells from normal cells, in

conjunction with traditional cell immunophenotyping and morphologic analysis.153

Alternatively, microarrays do not require a preprocessing step and are able to

simultaneously detect multiple targets, but tend to be low throughput, and cells cannot

be eluted without mixing the different subpopulations.91 In contrast, CAC is a relatively

simple process that can be multiplexed to perform the same function as an array;

however, CAC allows for the elution of cells, enabling culture and further study of the

population. Due to these advantages, CAC was chosen as the method of cell

separation used in this work.

Despite the many successes of microfluidic technology, many drawbacks are

associated with their use which may promote the use of simple capillary devices for

certain CAC applications. Most importantly, capillary systems remove the difficulties

associated with the design and fabrication of microfluidic devices, which typically require

the use of expensive clean rooms, not easily accessible for many laboratories.154

Microfluidic channels may also be more complicated to interface with a benchtop fluidic

system for sample flow rate control than a capillary. In contrast to microfluidics,

capillaries often have simple, commercially available connection options, and are

available in a variety of internal and external sizes and geometries, depending on the

desired application. Capillaries are commercially manufactured to have consistent

properties throughout, thus eliminating concerns about batch-to-batch variability, a

frequent problem when fabricating microfluidic devices. Also, capillaries can easily









undergo surface modification using standard chemistries to facilitate the immobilization

of cell capture ligands. The most prominent disadvantage of capillaries lies in their low

throughput, which results in challenges for detecting targets present in low

concentration. However, capilliaries can be arranged either in parallel or in series to

increase throughput or to facilitate the capture of multiple targets, with the possibility of

simultaneous detection of several types of cancer.91 Also, a preconcentration step to

enrich the sample for target cells may alleviate this concern. The commercialized

CellSearch technology platform for CTC detection preconcentrates the sample for

epithelial cells using the EpCAM antibody.

Sample Beam a Scattered Beam --------.----->
Clliialed Bam Fcud Bmam Collimared 6eam





::: SILICA SAMPLED VOLUME


Figure 1-7. Comparison of square and circular capillary geometries.155

Specifically, square capillaries have properties with distinct benefits when

compared to circular tubes. They have a higher surface area and volume than circular

capillaries of the same dimensions,156 leading to more capture probe immobilized on the

walls and a higher sample throughput. Additionally, the sensitivity of path length-

dependent detection methods, such as absorption and fluorescence, can be enhanced

by an order of magnitude over circular capillaries. Most importantly, flat-walled

capillaries (square or rectangular) have less optical distortion and scatter than the

curved walls of circular tubes.156'157 Figure 1-7 shows this effect, as collimated light









passes directly through the square capillary, but light in the circular capillary is scattered

at the curved walls of each interface.155 This is especially significant when direct

observation is used for detection, as may be advisable in the case of potentially

extremely low numbers of cancer cells or CTCs present in the blood.

Antibodies and aptamers as stationary phase ligands





(b)




Stationary i ,
y) Phase





Figure 1-8. Antibody immobilization to stationary phase. In a) no antigen recognition
sites are available for binding. In b) and c) either 1 or 2 sites are oriented
correctly (as designated with arrows) for antigen binding.158

An ideal stationary phase should possess the following characteristics: 1)

Chemical stability; 2) Low nonspecific binding; 3) Mechanical stability for favorable flow

rates; 4) Sufficient surface area for binding.158 As previously discussed, antibodies are

not robust in terms of stability, as changes in temperature and buffer conditions can

reduce binding. Antibodies may also bind an epitope of a protein found on many cell

surfaces, depending on the conditions used for their generation (for example, EpCAM

antibody will bind to cancerous and normal epithelial cells). However, aptamers can be

selected to bind to one specific type of cancer, or even a specific cell line within the









disease. Aptamer synthesis is more straightforward than antibody generation, and an

affinity label such as biotin facilitates immobilization to a (strept)avidin-coated surface.

Linkers such as polyT are often incorporated to provide distance from the solid support

and flexibility for aptamer binding.159 Antibodies can be labeled by various chemistries

to promote immobilization, but this requires an extra purification step and often it is

uncertain exactly where on the protein the tags end up. This can cause the protein to

be immobilized in a way that renders their antigen binding sites unavailable, as

demonstrated in Figure 1-8.158 Also, it may interfere with antigen binding, or reduce the

stability of the protein, as was the case in immunoglobulin G (IgG) biotinylation.160

Therefore, aptamers are immobilized on capillary walls as CAC stationary phases in this

work.

Theory of cell capture

The theory dictating flow in square capillaries is similar for square and circular

capillaries, although the equations may vary, especially when area is used as a

parameter. Fluid flow is characterized by either exhibiting laminar or turbulent

properties; laminar flow is desired because it is an ordered flow of fluid throughout the

tubing, while turbulent flow is chaotic. The Reynolds number (Re) is calculated to

determine whether the flow is laminar or turbulent (equation 1-1). Reynolds numbers

below ~2000 dictate that laminar flow will occur within the system.161

pvd
Re= pvd [1-1]


In equation 1-1, p describes the density of the fluid (kg/m3), v is the fluid velocity (m/s), d

is the diameter of the tube (m), and p is the viscosity of the fluid (Pa*s). For fluid flow in

a square or rectangular (noncircular) tube, the diameter is actually defined as hydraulic









diameter (Dh), calculated via equation 1-2, where L and W represent length and width of

the cross-sectional area of the capillary in meters.

2LW
Dh =(L [1-2]
(L + W)

After determining that the system will operate by laminar flow guidelines, it is

important to take into consideration the diffusion times of the target so proper incubation

times can be utilized. Stokes' Law (equation 1-3) describes the force exerted on

spherical objects (such as cells) at low Reynolds number through a fluid.161

FfY = 6-yrv [1-3]

Ffr represents the frictional force (N), r is the cell radius (m), and v is the fluid velocity

(m/s). When Stokes' Law holds (as in conditions of low Re), the Einstein relation can be

used to calculate the diffusion constant of the target (equation 1-4).

kT
D= [1-4]


T is the temperature (K) and k depicts Boltzman's constant (J/K*mol). The diffusion

constant is then used to calculate the diffusion time (t) by equation 1-5, where x is the

average distance (m).

2
t= [1-5]
2D

Table 1-2. Diffusion constants and times for samples of different diameters
Sample r (m) D (m2/s) t (s)
ATP 1.00 x 10-9 2.45 x 10-10 2.9
Thrombin 3.00 x 10-9 8.17 x 10-11 8.6
Nanoparticle 5.00 x 10-9 4.90 x 10-11 14.3
Cell 5.00 x 10-6 4.90 x 10-14 14341.6










The diffusion times of targets of several different sizes are listed in Table 1-2 for

a capillary with an inner diameter of 75 pm. The value of x is assumed to be

approximately half of the diameter, 37.5 pm. Note that for a cell, the diffusion time is

close to 4 hours.

However, the effects of gravity on a particle as large as a cell must be considered.

When Stokes' Law applies, the settling velocity (Vs) of a cell can be calculated by

equation 1-6.


2( )gr [1-6]
9,u

p represents the density of the particles (pp) or fluid (pf) (kg/m3), and g is the

gravitational acceleration constant (m/s2). Using a dynamic viscosity of 9.0 x 10-4 Pa*s,

Vs is calculated as 5.51 x 10-5 m/s. This value is then used to calculate the settling time

for a cell (ts) as 0.68 sec for a 75 pm diameter capillary by equation 1-7.

ts = [1-7]
Vs

Therefore, cells/affinity ligand incubation times should be set with consideration of Vs,

since the system is dominated by sedimentation forces, not diffusion forces.

Figure 1-9 represents several crucial parameters determining whether a cell will

adhere to a surface.162 When the cell closely approaches (100-300 A) the ligand-

immobilized surface, substantial pressure builds up on the region of the cell closest to

the binding surface, and that area of the cell flattens. This area becomes the contact

area between the cell and surface. The cell is assumed to possess receptors on the

surface which bind to the capture ligand. Upon binding, the inherent diffusional mobility

of the receptors in the cell membrane allow for additional contacts to be made. Once









bound, shear forces of the flowing liquid act on the cell, with the strongest forces acting

on the cell apex.163


Figure 1-9. Processes occurring in cell adhesion.162

The cell will bind to the surface if the adhesion forces are greater than the shear

forces (equation 1-3) acting to break the ligand/cell bonds. The adhesion force (FA) is

given by equation 1-8,151 and the critical force (fc) required to rupture the ligand/cell

bond is given by equation 1-9.164

FA fcAcCs [1-8]

Ac is the cell contact area with the ligand-bound surface (~1 pm2 for cell radius of 5

pm),165 and Cs is the number of receptors on the ligand-bound surface within the cell

contact area.


kTac
o--
ro


[1-9]


ro represents the separation distance between receptors at the minimum breaking force,

and ac= KCs, where K is the equilibrium constant between the cell receptor and ligand.









When substituting KCs in for the ac term and Kd= 1/K for the K term, Kd is inversely

proportional to fc, meaning a lower Kd (high affinity) will require a stronger force to break

the ligand/cell bond.164

Another similar way to predict cell adhesion is by comparing the bonds formed

during a collision (B*) to the number of bonds required to resist the forces acting against

the bonds (B,). When B*>Bc, cell adhesion to the surface will occur.165 Collisions of

cells with the surface will commence during a specified amount of time called the

collision duration (To), in which a is the radius of the contact area (equation 1-10). This

is applied to equation 1-11 to determine B*, in which case B is the density of bonds

formed. Be is determined by the Stokes force and the critical force required to break the

ligand/receptor bond (equation 1-12).95

"c=a [1-10]
v

B" = TAcB [1-11]

BcFf- [1-12]
fc

Following cell capture, the cells can be eluted from the capillary and quantified by

off-line measurements using tools such as a hemocytometer, or retained for additional

studies. Traditional CAC elution methods include addition of a competitive agent or

implementing high mechanical shear rates, however, bubble-induced detachment has

recently shown attractive attributes.91 The air/liquid interface exerts shear forces on the

captured cells, breaking the cell/ligand bonds, and the cells can be eluted directly

without further dilution.166 The bubble-induced detachment method also demonstrates a









higher elution efficiency than previous elution methods, and higher cell viability for

applications where the eluted cells are cultured.91

Overview of the Dissertation

The data presented in this dissertation will demonstrate the utility of DNA

aptamers for two different bioanalytical applications. Chapters 2-4 will detail the work

performed to generate an aptamer which can serve as an antidote to the anticoagulant

bivalirudin. Chapter 2 features some in depth general discussion of several SELEX

steps with a bead-based selection strategy, while Chapters 3-4 describe an aptamer

selection and characterization using a monolithic column. In Chapter 5, the focus

switches to proof-of-concept studies for the aptamer-based capture of leukemia cells in

a square capillary. Chapter 6 illustrates the improvements in the system operation upon

optimization, particularly highlighting an increase in capture efficiency, leading to

detection of leukemia cells in blood samples. The capture of two different colon cancer

cell lines by the specific aptamer for each is also described with this system. The

concluding chapter summarizes the global significance of this work, and outlines

prospective directions of the research.









CHAPTER 2
METHODOLOGY AND DEVELOPMENT OF BEAD-BASED APTAMER SELECTION

Introduction

Anticoagulation with Bivalirudin

Bivalirudin is an anticoagulant used as an alternative to heparin in patients with

unstable angina undergoing percutaneous coronary intervention (PCI).167 The drug is a

bivalent direct thrombin inhibitor, binding to both exosite 1 and the active site of

thrombin with an inhibition constant of 2 nM. The 20 amino acid peptide has a

molecular weight of 2180 Da, and is cleaved at the Arg3-Pro4 bond with a half life of

approximately 25 min. Among the most important advantages of using bivalirudin in

place of heparin (described in detail in Chapter 1) are that it inhibits soluble and clot-

bound thrombin and it will not instigate HIT. Approximately 600,000 (5%) patients out of

an annual total of 12 million receiving heparin develop HIT and can no longer continue

heparin administration.168

Despite the many advantages, one reason why the use of bivalirudin has not

become more widespread is the cost of the treatment. Administration of the drug could

reach upwards of $2000, while a 24 hour dose of heparin is only $3.50.131' 132,169

However, these estimates do not account for possible cost savings due to decreased

bleeding and lower need for monitoring from bivalirudin, which have been reported as

~$400 lower with bivalirudin than other inhibitors.134

Perhaps the most significant explanation of why bivalirudin use is not more

prevalent is because an antidote to bivalirudin is not currently available. This critical

issue is the inspiration for much of the work performed in this dissertation. For this

project, an aptamer antidote will be selected for bivalirudin which can restore









coagulation activity in the instance of severe patient bleeding. The aptamer is expected

to function according to the schematic depicted in Figure 2-1.128 The selected aptamer

may compete with bound thrombin for peptide binding, releasing the drug from the

complex and restoring coagulation, and/or inhibit free peptide from interacting with

thrombin. High-affinity aptamers (similar affinity to the drug/thrombin complex) will likely

display both functions, while lower affinity aptamers may still be able to operate

according to the second mechanism.

Apta mer

Thrombin ,,





Aptamer interacts with thrombin- Aptamer binds free peptide,
bound peptide, releasing it from inhibiting binding to thrombin
complex





Figure 2-1. Mechanisms of aptamer/peptide interactions for restoring coagulation. The
aptamer may compete with thrombin for peptide binding, releasing the drug
from the complex (left) and/or bind to the free peptide, rendering it unavailable
to interact with thrombin (right).128

Aptamer Development

Aptamers are generated via an in vitro process known as SELEX, depicted in

Figure 1-3. The critical step determining the efficiency of the selection is the partitioning

of the binding sequences from the nonbinding sequences. A variety of methods have

proven effective for this purpose, including nitrocellulose membranes, affinity capture,

column immobilization, cross-linking, gel electrophoresis, immunoprecipitation,

centrifugation, and capillary electrophoresis.77 However, many of these methods rely on









a significant difference between the bound and unbound forms of aptamer/target, which

is problematic with small molecules such as peptides because the difference between

the bound and free DNA is frequently insufficient for separation. For example,

nitrocellulose membranes employ a molecular weight cutoff in which the disparity

between the free aptamer and bound complex is not large enough for the nonbinding

sequences to be partitioned by the membrane. Similarly, gel electrophoresis and

capillary electrophoresis separate samples based on a change in electrophoretic

mobility between the bound and unbound sequences. This difference may not be

enough for separation with small molecule targets depending on the properties of the

system.

This work focuses on using affinity tags to capture binding sequences with

micrometer-sized beads, followed by centrifugal separation from nonbinding sequences.

Specifically, the work presented in this chapter outlines the SELEX steps and uses a

streptavidin-coated bead-based partitioning method to separate DNA/biotinylated

complexes from nonbinding sequences to select an aptamer for bivalirudin.

The SELEX process is divided into several steps which are generally applicable to

any selection, although the exact methods used may change. These subcategories

include: 1) Experimental design; 2) Design of PCR primers and library; 3) Optimization

of PCR conditions; 4) SELEX procedure; 5) PCR procedures; 6) Monitoring selection

progress; 7) Sequencing and post-selection characterization studies (discussed later).

Materials and Methods

Experimental Design

Partitioning of the binding sequences from nonbinding sequences was achieved

by immobilization of the DNA/biotinylated target to streptavidin-coated micrometer-sized









beads, then centrifuging the beads and discarding the supernatant liquid. The

streptavidin/biotin method was chosen for further study because the interaction exhibits

a femtomolar (~10-15 M) dissociation constant, enabling an efficient capture of biotin-

tagged species.170

The peptide was biotinylated at the N-terminus by an overnight incubation with

Sulfo-NHS-SS-biotin (Pierce) in PBS at a 1:0.8 drug-to-biotin ratio. The limiting amount

of biotin was to assure the drug would not be doubly biotinylated, possibly altering the

properties. The N-terminus was utilized for capture (and immobilization in Chapter 3)

due to the drug structure and mechanism of binding discussed in Chapter 1. Since the

drug binds to exosite 1 via the C-terminus prior to active site binding, aptamer inhibition

of C-terminus binding will most likely be effective in reducing the anticoagulant effect.

Also, most of the tertiary structure of bivalirudin which presents a probable area for

aptamer binding is observed at the C-terminus. Therefore, this region should remain

free for aptamer binding.

Next, the biotinylated drug was purified on an analytical scale using reverse-phase

HPLC (1.0 mL/min; C18 column; 4.6 mm ID; 25 cm length; a gradient beginning with

water/TFA (100/0.1), and continuing to acetonitrile/TFA (100/0.1), 10-80% over 32 min,

monitoring at 214 nm). The samples corresponding to the major peaks were analyzed

by MALDI/TOF (a-cyano-4-hydroxycinnamic acid matrix) in order to determine which

fraction contained the correct mass-to-charge ratio (m/z ~2569). A negative ninhydrin

test, (a test for a free a-amino group) indicated that the N-terminus was biotinylated

instead of amino acid side-chains. Finally, the process was repeated on the preparative

scale to generate enough product for the ensuing research. All peptide purifications









and mass spectrometry were performed at the Interdisciplinary Center for Biotechnology

Research (ICBR) at the University of Florida (Gainesville, FL)

Design of PCR Primers and Library

The design of a functional set of primers and DNA library is a crucial step in

determining the efficiency of the PCR, in order to optimize the yield and avoid

mispriming. One consideration in primer design is nucleotide length, which is typically

around 20 nucleotides. This length is long enough to specifically recognize the priming

region of the template, yet short enough to bind at annealing temperatures from 52-

60C. This temperature is optimal because it is sufficiently lower than the polymerase

extension temperature of 72C, so extension will not begin until the primers have

annealed. Also, the GC content must be carefully controlled in order to maintain a

desirable melting temperature (Tm), which is approximately 5C of the annealing

temperature used. The primers are designed to have similar Tm (within 1 C), and avoid

stretches of repeated sequences or base runs (such as AAAAAA or CGCGCG) to

encourage the proper primer annealing positions. Finally, the primers should not

demonstrate significant secondary structure formation (hairpin) near the Tm, or form

either homo- or heterodimers. These interactions will lower the yield of the PCR by

decreasing the amount of primers available for annealing to the template strand. The

primers and library were designed using IDT OligoAnalyzer software.

The forward (sense) primer was labeled with 5'-FAM (5/6-carboxyfluorescein) for

monitoring the progress of the selection by flow cytometry. The reverse (anti-sense)

primer was 5'-biotinylated for capture by streptavidin coated beads used for preparing

ssDNA for subsequent selection rounds after PCR amplification.









The library template is constructed based on the sequences of the primers. The

library consists of two PCR primer regions flanking a random region of ~30-50

sequences. This random region generally contains sufficient sequences to ensure the

complexity of the initial pool (complexity= yN, where y is the number of nucleotides

possible (4), and N is the length of the random region) is high enough that unique

sequences are present in the initial pool. This is necessary because the number of

sequences capable of significant target binding represents only 1 in 109 to 1 in 1013 of

molecules present in the starting library.77 The 5' primer region is the same sequence as

the sense primer, while the 3' primer region is the complement of 3'-5' anti-sense primer

since it will prime and extend the anti-sense template strand.

Based on the primer design guidelines, the following primers and DNA library

template were devised:

Forward primer (sense): FAM-CTC ATG GAC AGG CTG CAG AC

Reverse primer (anti-sense): Biotin-CTG TAG TGG CAT CCG AGC GT

FAM-CTC ATG GAC AGG CTG CAG AC-(N)40 -ACG CTC GGA TGC CAC TAC AG

All oligonucleotides were synthesized using standard phosphoramidite chemistry on an

AB13400 DNA/RNA synthesizer (Applied Biosystems). DNA purification was performed

with a ProStar HPLC (Varian) using a C18 column (Econosil, 5U, 250 mm x 4.6 mm)

from Alltech Associates. UV-Vis measurements to measure DNA concentration were

performed with a Cary Bio-300 UV spectrometer (Varian).

Optimization of PCR Conditions

The annealing temperature of the primers used throughout the remainder of the

selection is the highest temperature that gives efficient amplification with minimal

nonspecific binding in the desired temperature range. The higher annealing









temperature increases the likelihood that the primers will bind to their expected priming

regions, since mispriming positions will have a lower Tm.

All PCR mixtures (Takara) contained 50 mM KCI, 10 mM Tris HCI (pH 8.3), 1.5

mM MgCI2, dNTPs (2.5 mM), 0.5 pM each primer, Hot-start Taq DNA polymerase (5

units/pL), and the library concentration was 50 pM. Temperatures from 55.0C-63.0C

were compared by gel electrophoresis, with 60C indicated as optimal. The following

conditions were used for the initial test of primers and library: initial hot start

temperature of 94.0C for 3 min, repeated cycles (30 sec each) of 94.0C for

denaturation, annealing at 60C, then 72.0C for elongation, and final extension at

72.0C for 5 min followed by an indefinite hold at 4.0C (Biorad iCycler).

Selection Procedure

The first round of selection began with a very large amount (~1014 sequences) of

the initial DNA library incubated with the target. The multitude of sequences increases

the likelihood that at least one of the sequences in the pool will bind specifically to the

target. The library was heated at 95C for 5 minutes and snap cooled at 4C to ensure

the sequences were not dimerizing and were folded into their most favorable

conformations. Fifty pmol of biotinylated drug and 1 nmol of the initial library were

incubated for 30 min in 200 pL physiological buffer (PB: 25mM Tris-HCI (pH 7.5), 150

mM NaCI, 5 mM KCI, 1 mM MgCI2, 1 mM CaCI2, 1% BSA, 0.1% tRNA). One mL of

streptavidin-coated beads (Bangs Labs, 5.6 pm, polystyrene core) was centrifuged at

10,000 rpm for 3 min and washed twice with 1 mL of PB. The drug/library mixture was

then added to the washed beads and incubated for 10 min. The particles were

centrifuged and washed twice with both 1 mL of PB and 1 mL of WB (500 mL PBS

buffer and 5 mM MgCI2) in order to remove nonbinding species, then dispersed in 1 mL









of water. The binding oligonucleotides were heat-eluted for 5 min at 95C and

centrifuged to recover the supernatant liquid which contained the binding DNA

sequences. This DNA was then amplified by PCR.

PCR Amplification Procedure

The first pool is PCR amplified for the next selection round in a series of 3 steps:

a. The entire pool is amplified

b. The number of cycles is optimized

c. Preparative PCR is carried out according to (b).

However, in any subsequent rounds, only steps (b) and (c) are necessary.

PCR amplification of entire first pool

PCR serves as a step to amplify the binding sequences of each round for use in

the next round of selection. The first round requires special attention because there is

only one of each sequence present in the pool after incubation with the target.

Therefore, the entire pool is subjected to the PCR in order to amplify all possible

sequences for the next round.

Table 2-1. PCR preparation of entire pool
Reagent Volume (pL)
10X PCR Buffer 100
dNTP (2.5 mM each) 80
Primers (10 pM each) 50
DNA pool 200
DNAse Free Water 570
Taq Polymerase 3

A 1000 pL amplification reaction volume was set up as shown in Table 2-1. The

volume of water was adjusted to compensate for the total volume in instances where

the DNA pool volume varies. The PCR reaction components are mixed thoroughly and

200 pL of the solution was pipetted into 5 individual 0.5 mL PCR tubes. The mixtures









were PCR amplified using the PCR amplification program in Figure 2-2. After this initial

PCR amplification, all PCR products from individual tubes were pooled together and

used as the template for the next PCR procedure.


94C 94C
3 min 30 sec


720C 72C


30 sec


15X I 1X


Figure 2-2. PCR amplification of initial library. A) PCR amplification program; B) Cycle
optimization; C) Preparative PCR.

Cycle optimization for preparative PCR

Table 2-2. PCR cycle optimization
Reagent Volume (pL) Control (pL)
10X PCR Buffer 100 5
dNTP (2.5 mM each) 80 4
Primers (10 pM each) 50 2.5
DNA pool 200
DNAse Free Water 570 38.5
Taq Polymerase 3 0.15

In this step, the optimal number of PCR cycles is determined for the preparative

PCR that generates DNA for the next round. The purpose of this step is to perform

enough cycles to maximize the amount of DNA retrieved without nonspecific

amplification. The amplified PCR pool served as the template and accounted for 10%

of the total reaction volume. Fifty pL of the reaction mixture amplified under the

conditions in Table 2-2 was aliquoted into separate 0.5 mL PCR tubes and was

amplified by the program shown in Figure 2-2. At the end of various PCR cycles (12,


25 21 15 12 C L









15, 21, or 25 for this selection), one PCR tube was removed from the instrument. The

PCR products were analyzed by 3% agarose gel electrophoresis, and the cycle number

displaying optimal amplification (15 cycles) was utilized for the preparative PCR step

(Figure 2-2B).

Preparative PCR

This PCR amplification step is used to generate enough ssDNA for the second

round of selection. Usually ~1000 pL of total PCR reaction volume is sufficient,

although higher volumes may be used in certain cases. The reaction components for

1200 pL reaction mixture are shown in Table 2-3. Once again, the PCR reaction

components were mixed thoroughly and 200 pL of the solution was pipetted into 6

individual 0.5 mL PCR tubes. The mixtures underwent amplification using the program

shown in Figure 2-2 at the optimum cycle number (15 cycles) and were pooled together

for ssDNA preparation after analysis by gel electrophoresis (Figure 2-2C).

Table 2-3. Preparative PCR
Reagent Volume (uL) Control (uL)
10X PCR Buffer 120 20
dNTP (2.5 mM each) 96 16
Primers (10 uM each) 60 10
DNA pool 120
DNAse Free Water 804 154
Taq Polymerase 3.6 0.6

Gel electrophoresis monitoring of products

Agarose gel electrophoresis is used in a selection procedure for analysis of cycle

optimization and to ensure that satisfactory amplification has occurred following

preparative PCR. The mobility of a DNA fragment is dependent on the size, charge,

and conformation of the molecule and the strength of the electric field applied.

Additionally, gels containing different percentages of agarose will influence the









separation of fragments. For aptamers of ~80-100 bases, 3% agarose is sufficient for

separation of PCR amplicons. The intercalating UV dye ethidium bromide (EB; 0.005%

v/v) was added to a solution of the 3% agarose (w/v) in 40 mL of tris-borate-EDTA

buffer (TBE buffer) for visualization.

Ten pL of each PCR product analyzed was combined with 2 pL of 6X Blue/green

loading dye. The 25 bp ladder was prepared by adding 2 pL of 6X Blue/green Loading

dye, 1 pL of ladder, and 9 pL of deionized water. The samples were each loaded into a

different lane, and the electrophoresis conditions are set to 100 V, 2 A, for 40 min.

Preparation of ssDNA

PCR amplification generates double-stranded DNA (dsDNA), and must be

converted to ssDNA before the next round of selection, since only the sense DNA

strands have been selected to bind to the target. This is accomplished by capturing the

amplicons on streptavidin-coated beads via the biotinylated antisense primers which

have been incorporated into the antisense DNA strand following PCR. Adding NaOH

ruptures the hydrogen bonds of the strands, and the sense strands are eluted from the

beads while the antisense strands remain attached.

Streptavidin-coated Sepharose beads (GE Healthcare; 300 pL) were added to a

column (Glen Research Expedite Style) blocked by two frits (1 pm pores). The beads

were washed with 4 mL of Dulbecco's PBS buffer by syringe, and the entire PCR

product was added to the column. The plunger was loosely inserted into the syringe to

control the flow rate of the PCR product in order to allow the biotinylated strands ample

time to bind to the streptavidin. The PCR product flow-through was collected and run

through the column two additional times to ensure maximum collection of the DNA on

the beads. The column was washed with 6 mL of PBS buffer, then 0.5 mL of 0.2 M









NaOH was added to elute the ssDNA product. In each step, the syringe was removed

from the column, followed by plunger removal to minimize disturbance of the beads.

Desalting of ssDNA

The excess salt used to denature the dsDNA must be removed from the ssDNA

product before the next round of selection. A Sephadex G-25 DNA Grade NAP-5

desalting column (GE Healthcare) was washed with 3 column volumes of water. The

ssDNA product was added to the column and allowed to drain, and desalted ssDNA

was eluted and collected in 1 mL of DNAse-free water. The concentration of the ssDNA

was determined by UV absorbance at 260 nm and the ssDNA was dried by vacuum

dryer. The DNA pellet (~80 pmoles after the first round) was reconstituted in 500 pL PB

for the second round of selection.

Second and Succeeding Rounds of Selection

The SELEX procedure leading to the generation of aptamers is an evolutionary

process beginning with a large amount of initial DNA library. From beginning to end, the

population of sequences binding specifically to the target increases, until evolution of a

final pool containing a majority of binding sequences. Through the various selection

rounds, the stringency of the process is increased in order to select aptamers with the

most desirable properties. This is carried out in a variety of ways throughout the

progression of the selection including: 1) Increasing the ratio of library to target. This

increases the competition among the sequences for fewer target molecules so the

higher affinity oligonucleotides will be preferentially retained; 2) Decreasing the

incubation time. The sequences with the most favorable binding kinetics will bind the

target; 3) Increasing the wash volume. In this manner, the higher amounts of BSA and

tRNA will compete and remove weakly binding sequences from the target; 4)









Introduction of a negative selection step. This will remove the sequences that are not

specific for the target. For this work, this means removing the sequences which bind to

the streptavidin-coated beads.

The point in which a negative selection is instituted varies between researchers.

Some groups perform this task before the initial positive selection in order to

immediately subtract out any sequences that are not specific to the target.171 However,

this runs a high risk of losing potential aptamers due the number of unique sequences

present in the initial pool and the inefficiency of partitioning methods.172 Therefore, a

negative selection step is typically introduced later in the selection, when higher

numbers of specific binders are present in the pool.

Table 2-4. Conditions for initial SELEX
Amount Amount Incubation Wash Bead
Round B-Drug Library Lib:Drug Time Volume Volume
1 50 pmol 1 nmol 20.0:1 30 min 1 mL 1 mL
2 10 pmol 50 pmol 5.0:1 30 min 500 uL 200 uL
3 10 pmol 50 pmol 5.0:1 30 min 500 uL 200 uL
4 10 pmol 50 pmol 5.0:1 30 min 500 uL 200 uL
5 10 pmol 50 pmol 5.0:1 30 min 500 uL 200 uL
6 7.5 pmol 50 pmol 6.7:1 30 min 500 uL 150 uL
7 7.5 pmol 50 pmol 6.7:1 30 min 500 uL 150 uL
8 5 pmol 50 pmol 10.0:1 25 min 750 uL 150 uL
9 5 pmol 50 pmol 10.0:1 20 min 750 uL 100 uL

In the first selection, no negative selection was used. The conditions for this

selection are shown in Table 2-4. As the selection progressed from the second to the

final round, the library-to-drug ratio increased from 5:1 to 10:1, the incubation time

decreased from 30 minutes to 20 minutes, and the wash volume increased from 500 pL

to 750 pL. The bead volume was decreased in order to lower the instances of

nonspecific binding to the streptavidin and increase competition between DNA

sequences as the ratio of library-to-drug increased.









Flow Cytometry Monitoring of Selection Progression

During selection, it is necessary to monitor the progress of the pool evolution in

order to determine if conditions are appropriate and to establish when the pool appears

to be maximally enriched. One convenient method for selection monitoring is flow

cytometry, in this work facilitated by capture of the FAM-DNA/biotinylated drug

complexes by the streptavidin coated beads. Flow cytometry uses the principles of light

scattering, light excitation, and light emission of fluorophores to gather specific data on

particles from ~0.5-40 pm. Particles entering the instrument are hydrodynamically

focused by a flowing sheath of buffer to individually intersect a light source (typically an

argon ion laser). Once the particles enter the laser beam, they scatter the light, and

fluorophores present on binding sequences are excited to a higher energy state (Figure

2-3).173 This energy is released in the form of photons that emit with properties

characteristic of each dye. Many flow cytometers have an optical setup which can

detect dyes emitting at three different wavelengths or fluorescence "channels;" 525 nm

(FL-1; FAM and FITC), 575 nm (FL-2; PE), and 620 nm (FL-3; PE/Cy5 tandem). The

emitted light of each individual particle is converted into electrical pulses by the

detector, then graphically displayed to provide information on particle size, complexity,

and fluorescence intensity.

In a selection, an increase in fluorescence intensity of an oligonucleotide pool

when compared to the initial library indicates that more of the fluorescently-labeled

sequences are binding to the target, resulting in an increasingly enriched pool.

Generally pool monitoring does not begin until around the third round of selection

because pools from earlier rounds are typically not enriched enough to generate a

measurable increase in fluorescence. Also, the measurements are often performed in









batches of 2-4 rounds to ensure the conditions are comparable. As the selection

progresses past the initial rounds, one generally observes the pools gradually increase

in fluorescence, indicating that the selection conditions are appropriate for enriching the

pool. The selection is considered complete when the following conditions are met: 1)

There is a significant increase in fluorescence intensity between the unselected initial

library and the pool; 2) 2-3 consecutive rounds of selection do not further increase the

fluorescence intensity of the pool.



R. DI
ESE NSO KOR









S iLEKW \









Figure 2-3. Schematic of flow cytometry instrumentation.173

A Becton Dickinson FACScan flow cytometer was utilized for study. Each assay

involved a 30 minute incubation of 2.5 pmol of the biotinylated drug with 250 nM FAM-

aptamer pool in 50 pL total volume. Fifty pL of 5.6 pm streptavidin-coated beads were

washed twice with 50 pL of PB, then incubated for 30 min with the drug/pool mixture.

To test pools for binding to the beads, the sequences were directly incubated with
uFILTT KIHEM F( E)





Figure 2-3. Schematic of flow cytometry instrumentation.173

A Becton Dickinson FACScan flow cytometer was utilized for study. Each assay

involved a 30 minute incubation of 2.5 pmol of the biotinylated drug with 250 nM FAM-

aptamer pool in 50 uL total volume. Fifty uL of 5.6 pm streptavidin-coated beads were

washed twice with 50 uL of PB, then incubated for 30 min with the drug/pool mixture.

To test pools for binding to the beads, the sequences were directly incubated with









beads for 30 min. The beads were washed twice with 50 pL of PB, twice with 50 pL of

WB, and finally suspended in 400 pL of WB. A total of 20,000 events were counted for

each sample.

Flow cytometry for DNA blocking step

The concentration of an unlabeled random library (D3B) required to saturate the

streptavidin was determined by flow cytometry assay. Here, different molar excesses

(2X-200X) of library D3B compared to the calculated amount of streptavidin on the

beads were first incubated with the beads for 10 min. Another FAM-labeled library (D2;

50 pmole) was incubated with the beads/unlabeled library for 30 minutes and assayed

as described previously.

Flow cytometry for TCEP cleavage

To optimize the TCEP cleavage conditions, the basic flow cytometry assay

outlined above was followed with these notable changes: 1) 2.5 pmole FAM-DNA-SS-

Biotin library was substituted for the drug incubation; 2) After incubation and washing

steps, TCEP concentrations from 0.1-500 mM were added to the bead/DNA complexes

and incubated for 30 min or 1 hr on a shaker.

Results

Enrichment of Aptamer Pool

The results of flow cytometry assays for aptamer selection monitoring are typically

displayed in the form of histograms (Figure 2-4). This is a graphical plot of the

fluorescence intensity of the FL-1 (FAM) channel versus the number of events counted.

Figure 2-4A depicts a control histogram of the beads, an expected low intensity binder

(FAM-labeled initial library), and a sequence expected to demonstrate a high amount of

fluorescence (biotinylated and FAM labeled sequence). Minimal binding is observed for










the initial library, but the biotinylated sequence displays significant fluorescence (~103).

This is expected because the initial library will have few sequences binding since it is

not enriched for the target and the biotinylated/FAM sequence will show increased

binding because of the biotin binding to the streptavidin on the beads. The fluorescence

of the biotinylated sequence is commonly said to have a fluorescence "shift" (referring to

the relative positions along the x-axis) compared to the initial library.


Library
8th Round
9th Round


Library
5t Round
6" Round
71 Round


Library +
Library-
6i Round +
61 Round-

8' Round +
81 Round -


Figure 2-4. Flow cytometry data of initial selection. A) Beads, library, and positive
control; B) Rounds 5-7; C) Rounds 8-9; D) Rounds 6-8 incubated with either
beads alone (-) or drug with bead capture (+).

The shift between the initial library and each of the pools is the major area of

interest, so the particle study (beads alone) is not considered in additional assays.

Figure 2-4B is the binding assay of rounds 5-7, and a noticeable shift is observed for









round 7. Therefore, the stringency of the selection was raised for round 8 in terms of

library-to-target ratio, decreased incubation time (further decreased in round 9), and

increased wash volume. After several rounds of selection, the pool appeared to be

enriched, as evidenced by a decrease in fluorescence in round 9 (Figure 2-4C). This

fluorescence decrease between rounds 8-9 may be due to a loss of some of the lower

affinity binders due to the selection stringency.

Despite the increase in fluorescence intensity, the binding could actually be

caused by the pools binding to the streptavidin on the beads instead of the drug target.

Therefore, a flow cytometry assay comparing the shifts of each pool with (+) and without

(-) drug present was performed (Figure 2-4D). The shifts between each pool were

nearly identical, implicating the beads as the source of binding rather than the desired

drug target. Streptavidin is much larger than the drug target, which likely results in more

binding sites available to aptamers. To reduce this binding, negative selection against

the beads may subtract out sequences binding to the streptavidin.

Selection Modification #1

Table 2-5. Conditions for selection modification #1
Amount Amount Incubation Wash Bead Bead
Round B-Drug Library Lib:Drug time Volume Volume Volume -
6 7.5 pmol 50 pmol 6.7:1 30 min 500 uL 150 uL 75 uL
7 7.5 pmol 50 pmol 6.7:1 30 min 500 uL 150 uL 150 uL
8 7.5 pmol 50 pmol 6.7:1 30 min 500 uL 150 uL 150 uL
9 5 pmol 50 pmol 10.0:1 20 min 750 uL 100 uL 175 uL

The 5th round of selection demonstrated no significant shift, so a new selection

instituting a negative selection was begun from the pool obtained after five rounds of

selection. The conditions for this selection are observed in Table 2-5. The final column









shows the volume of beads used for a 30 minute negative selection step, wherein the

sequences that do not bind to the streptavidin beads are retained for further steps.

As By
Library +
Round 8B -
Round SA + Round 9B +
Roumn SA Roud 9B -







PL1.Hilght FLI.Hel ht

Figure 2-5. Flow cytometry data of selection modification #1. A) Library and round 8A
(without negative selection) incubated with either beads alone (-) or drug with
bead capture (+); B) Rounds 8-9B (with negative selection).

The flow cytometry assay of the selection progression (Figure 2-5) is divided into

two figures to clarify the results. In this assay, a small portion of the pool from round 8

was retained after the positive selection, and not subjected to the negative selection

(Figure 2-5A). This is considered round 8A, while round 8B is the pool after the

negative selection was performed (Figure 2-5B). In this manner, it was determined that

the negative step did subtract out some nontarget binders, as evidenced by the

decreased fluorescence intensity after the subtraction. But the remaining sequences

were still binding to the beads, since the shifts of round 8B with (+) and without (-) drug

were still similar. The subtraction step was also most likely insufficient to remove all of

the high affinity sequences from the beads due to the shoulder of the histograms on

Figure 2-5B. This shoulder likely corresponds to some of the sequences binding to the

streptavidin prior to the subtraction step remaining in the solution.









Selection Modification #2

For the next selection, two aspects were combined to increase the likelihood of the

selection of a target-binding sequence: 1) A random initial library was incubated with the

beads prior to the addition of the target/DNA complexes; 2) The disulfide bond

incorporated into the biotin peptide label was cleaved with the reducing agent tris(2-

carboxyethyl)phosphine (TCEP) in order to retain only sequences binding to the

peptide; 3) The library/primer set was changed to reduce possible contamination.

.The random library must be one which is not amplified by the primers used in this

selection. Random libraries D1A and D3B were tested by the PCR amplification

method described above for 30 cycles and analyzed by agarose gel electrophoresis.

The D3B library did not appreciably amplify (Figure 2-6), and was utilized for further

study.

1 2 3 4













Figure 2-6. PCR of unlabeled DNA library. 1) 25-bp ladder; 2) Negative control; 3) 500
pM library (no PCR); 4) PCR of 500 pM library. PCR did not amplify the
library.

Figure 2-7 (representative of results in triplicate) shows that the increasing

amounts of unlabeled "blocking" DNA actually increased the amount of labeled random

library binding to the surface. We hypothesize that this is due to the labeled library










actually binding to the blocking library, thus increasing the fluorescence when more

blocking library is present. Therefore, this method of attempting to nonspecifically block

the streptavidin with unlabeled DNA was abandoned, and selection was performed by

TCEP cleavage.

Unlabeled + SA

S2X Excess
10X Excess

75X Excess
100X Excess
i 200X Excess





PL1-Hslht

Figure 2-7. Streptavidin blocking with excess random library.


- Control
Lib, no biotin
25 mM TCEP
200 mM TCEP
500 mM TCEP


FL -HIliht FL1.lU- ~ht

Figure 2-8. TCEP concentration optimization. A) 30 min TCEP incubation; B) 1 hr
TCEP incubation.

TCEP concentration was optimized to selectively partition sequences binding

solely to the target. The decrease in fluorescence from the positive (+) control (no

TCEP added) as higher concentrations of TCEP were added can be observed. This

correlates to more of the DNA being cleaved from the beads. Although the signal did









not completely decrease to the levels of the negative (-) control, some of the cleaved

DNA may remain in the system and possibly bind to the streptavidin. Nonspecific

binding of a random library to the streptavidin is demonstrated in Figure 2-8B (blue line),

where the nonbiotinylated library produces appreciable binding to the beads. Based on

the results, 200 mM TCEP was used for all further studies.

Further, the effect of TCEP on either PCR amplification or biotin/streptavidin

binding was assessed. A 15-cycle PCR was performed on the initial library either with

or without 200 mM TCEP. Only the tubes devoid of TCEP were successfully amplified.

When samples containing the TCEP were desalted to remove TCEP using either a

column or centrifugal filter, the DNA was successfully amplified.

For the streptavidin/biotin interference studies (Figure 2-9), the assay conditions

were the same as the TCEP cleavage studies (using a FAM-DNA-SS-Biotin library), but

the parameters consisted of: 1) + control (2X): only beads and FAM-DNA-SS-Biotin

library- expect large shift; 2) Beads/DNA/TCEP cleavage- should have a lower intensity

than (1); 3) DNA and TCEP incubation followed by bead capture- similar to (2); 4)

Incubate beads and TCEP, followed by DNA capture- similar to (1) if the TCEP does not

effect DNA binding; 5) control: beads only- minimal fluorescence. The histogram in

Figure 2-9 shows the expected results. When TCEP is incubated with the beads first,

followed by DNA capture (#4, brown line), the streptavidin/biotin binding is the same as

the intensity of the beads directly capturing DNA without TCEP incubation (#1, green

and darker blue lines). Additionally, when the beads, DNA, and TCEP are all incubated

together (#2, purple line) or when the DNA and TCEP are incubated first, followed by









bead capture of the DNA (#3, light blue line), the cleavage is similar. Therefore, we

concluded that the TCEP does not affect biotin/streptavidin binding.


- Control
+ Control B
Beads+DNA+TCEP
Beads+TCEP, DNA


FLI1-H1lght
Figure 2-9. TCEP interference with streptavidin/biotin binding.

Preparation of new library and primers

950C 950C


720C 720C

3 sec 2 min


14X I 1X


Figure 2-10. PCR amplification program for second initial library.

To reduce the effect of contamination from previous selections, a new set of

primers and library previously designed by a coworker were synthesized and purified as

described previously.174 The PCR annealing temperature is 55C for this library set,

and the new PCR protocol is shown in Figure 2-10.


2:30 min









1X


30 sec


4C









Forward Primer (sense): 5'-FAM-ATC GTC TGC TCC GTC CAA TA

Reverse Primer (anti-sense): 5'-Biotin-GCA CGA CCT CAC ACC AAA

5'-ATC GTC TGC TCC GTC CAA TA -N45- TTT GGT GTG AGG TCG TGC

Selection

Table 2-6. Conditions for selection modification #2 with new library
Amount Amount Incubation Wash Bead Bead
Round B-Drug Library Lib:Drug Time Volume Volume Volume -
1 50 pmol 1nmol 20.0:1 30 min 1 mL 1 mL N/A
2 10 pmol 50 pmol 5.0:1 30 min 500 uL 200 uL 100 uL
3 10 pmol 50 pmol 5.0:1 30 min 500 uL 200 uL 150 uL
4 10 pmol 50 pmol 5.0:1 30 min 500 uL 200 uL 200 uL

The next selection aimed to incorporate the TCEP cleavage and negative

selection into the method. Due to the inefficiency of the TCEP cleavage, this step would

be implemented in round 5, so sequences are not lost early in the selection before the

pool has evolved. The negative selection step began in round 2, and other selection

conditions are in Table 2-6.


SLibrary +
Round 3+
Round 3-
IR Round 4-




1 1Q

FL1I-Hilght

Figure 2-11. Flow cytometry of selection modification #2. Pools were either incubated
with either beads alone (-) or drug with bead capture (+).

Flow cytometry analysis of the pools began in round 3 (Figure 2-11). The pools

containing drug (+) demonstrated at least as much fluorescence intensity as the pools









incubated directly with the beads (-). Once again, the pool was enriched for streptavidin

despite the negative selection.

Selection Modification #3

Table 2-7. Conditions for selection modification #3
Amount Amount Incubation Wash Bead Bead
Round B-Drug Library Lib:Drug time Volume Volume Volume -
2 10 pmol 50 pmol 5.0:1 30 min 500 uL 200 uL N/A
3 10 pmol 50 pmol 5.0:1 30 min 500 uL 200 uL 200 uL
4 10 pmol 50 pmol 5.0:1 30 min 500 uL 200 uL 350 uL

The procedure used to remove the sequences binding to the streptavidin clearly

needed to be more stringent. When the same round 1 conditions as shown in Table 2-6

were implemented, but a TCEP cleavage was performed in the first round, no PCR

amplification was evident even after 33 cycles. Therefore, in the next selection, the

cleavage was not implemented until round 2. This loss of sequences was believed to

be balanced against a less stringent negative selection procedure in the second round.

This selection used the round 1 selected pool from selection modification #2 to begin

the TCEP cleavage at round 2 (Table 2-7). However, beginning in the third round, a

"double negative" selection was implemented, in which the pools were incubated with

the beads for 30 min for sequence subtraction first, followed by regular positive

selection on the drug target, then a second 30 minute incubation of the binding

sequences with the beads.

Despite the modified conditions used for this selection, the flow cytometry data

(Figure 2-12) proves that the pool is still enriched for the streptavidin instead of the

drug. This data implies that a new support system without streptavidin may be better

suited to select an aptamer for the much smaller drug target.









Library
Round 3-
Round 4+


1f 101 1 I 10*
FL1-Height

Figure 2-12. Flow cytometry data of selection modification #3. Pools were either
incubated with either beads alone (-) or drug with bead capture (+).

Conclusions

The method provided in this chapter proved unsuccessful for generating an

aptamer for the drug target. The overlying explanation is believed to be due to the

streptavidin-based partitioning system that the procedure relied on for separating the

binding and nonbinding sequences. Based on size alone, streptavidin is ~25 times the

molecular weight of the peptide target. This larger size and variety of functional groups

enables the structure to form multiple binding pockets which the aptamer may find more

favorable than the limited folding of the peptide. Additionally, each streptavidin

molecule is actually a tetramer, so many favorable binding sites found on the

monomeric form are actually present 4 times (depending on how monomers assemble

into the tetramer structure, some sites may not be available), and additional binding

regions formed by the interaction of the monomers is likely. Therefore, it is not

surprising that the pool preferentially binds to the protein. Furthermore, the

tetramerization of the biotin conjugates may lead to undesirable interactions that reduce

binding of the sequences to the drug.175 The PCR may also aid in enriching the pool for









sequences that bind to the streptavidin. When the pool is PCR amplified, the vastly

greater numbers of sequences binding to the streptavidin overshadows the few

sequences with specific binding to the drug.

It is exceedingly interesting that none of the methods used to minimize the

streptavidin binders were effective. Various degrees of stringency were utilized to

wash, block, or cleave nonspecific binders from the pool. Possibly the increased

stringency only served to remove sequences weakly binding to the peptide, since small

molecule selections typically generate lower-affinity aptamers than their more sizeable

counterparts.

Regardless of the true cause of the shortcomings resulting in pool enrichment for

the beads, the consensus conclusion derived from the work is that streptavidin should

be removed from the system. However, it is possible that the pools from the selections

may contain high affinity streptavidin aptamers, which may be of use for applications

unrelated to this work. The next two chapters will describe bivalirudin selection utilizing

an alternative partitioning system composed mainly of polymeric (hydrophobic)

materials.









CHAPTER 3
MONOLITHIC SELECTION FOR BIVALIRUDIN

Introduction

Affinity columns have played an important role in generating aptamers for various

small molecule targets. Some examples include an initial SELEX study to generate

aptamers to small organic dyes by Ellington and Szostak,29 substance P and amyloid

peptide were covalently coupled to Sepharose 6B columns,55, 56 ATP was immobilized

on agarose,176 and tripeptides were bound to HiTrap NHS-activated columns.177

Binding aptamer sequences were eluted using competitive binders, cleavage of the

bonds immobilizing the target to the column matrix, EDTA addition, high salt

concentration, or column heating.

The advent of monolithic columns has introduced new technology with inherent

properties that may provide unique advantages for affinity-based column aptamer

selection. Monolithic columns are essentially constructed from one piece of material in

which the area interacting with the analyte is mainly on the surface. This allows for

mass transport by convection on the surface of the material, rather than the diffusion-

limited separations of traditional bead-based separations. This leads to a decrease in

the pressure drop across the disk, allowing for faster flow rates and simple peristaltic

pumps such as those found in low-pressure chromatography (LPC) instruments.178

Moreover, the hydrophobic backbone of the typically polymeric-based matrices has also

been reported to have low nonspecific adsorption,179 and the target molecule can be

covalently attached to the column by a variety of chemistries, minimizing instances of

target leaching.









This chapter will explore the use of monolithic columns as affinity matrices for

aptamer selection for the anticoagulant bivalirudin. First, the peptide was covalently

linked to the column via the N-terminal segment in order to facilitate aptamer binding to

the C-terminal region. A chromatographic method capable of eluting binding sequences

from the disk was then developed. Several rounds of selection were carried out and the

progress was monitored by real-time PCR. A selection on a blank disk with no peptide

immobilized was compared to that of the drug-immobilized disk as a control. The

highest-affinity DNA pool was sequenced, and individual sequences were tested for

binding to the drug.

Materials and Methods

Column Immobilization

Buffers- binding buffer (BB): PBS buffer (without MgCI2 and CaCI2) with 5 mM

MgCI2 (final concentration); elution buffer (EB): BB with 1 M NaCI. All buffers and

reagents added to the monolithic column were first filtered with a 0.22 pm cellulose

nitrate filter (Corning). Bivalirudin (Angiomax; The Medicines Company) was received

in lyophilized form as a gift from the Anesthesiology Department at the University of

Florida College of Medicine. The peptide was immobilized on a poly(glycidyl)

methacrylate-co-ethylene dimethacrylate epoxy-functionalized CIM disk (BIA

Separations) via the N-terminus as per manufacturer's instructions. Briefly, the column

was inserted into the disk housing and equilibrated with 2 mL of BB at 1 mL/min, then 2

mL of EB at 1 mL/min, and finally 2 mL BB at 2 mL/min using a LPC system (Bio-Rad

BioLogic LP). Drug immobilization began with a 2 mL wash with 0.5 M sodium

phosphate buffer (pH 8.0) at a flow rate of 2 mL/min. The drug was dissolved in 5 mL of

sodium phosphate buffer (pH 8.0) for a final concentration of 3.0 mg/mL. This solution









was introduced into the disk (1 mL), and the disk was covered and incubated with the

flow through for 48 hours at room temperature in a small dish. For the blank disk (no

drug immobilized), the same buffer without drug was incubated with the disk, and the

remainder of the protocol was followed. Following this step, the disk was washed with 2

mL 0.5 M sodium phosphate buffer (pH 8.0), and remaining epoxy groups were blocked

by flowing 4 mL of 1 M ethanolamine through the disk and incubating overnight at room

temperature in a dish. The reaction was quenched by washing the disk with 2 mL of 0.5

M sodium phosphate buffer (pH 8.0) with 1 M NaCI at 2 mL/min, and equilibrated for

use by introducing 2 mL of 0.5 M sodium phosphate buffer (pH 8.0) at 2 mL/min. The

column was washed in 2 mL BB and sealed in the column housing with the column blind

fittings for storage at 4C for usage planned less than 3 days after last usage, or

washed with 4 mL of 20% ethanol and stored at 4C in a container with 20% ethanol if

the next usage was planned for longer than three days.

The UV absorbance at Aabs= 280 nm of the peptide solution (Bio-Rad SmartSpec

Plus) was measured before and after immobilization to determine the amount of drug

immobilized on the disk. The mass of peptide remaining in the solution was calculated

and subtracted from the initial amount added for the immobilization to determine that 1

mg of peptide was immobilized on the disk. This agrees with literature values reporting

0.3-0.9 mg of peptide immobilized.180'181

Method Development

The disk was placed into the column housing with luer connections which fastened

to complementary fittings on the LPC device. The system was capable of collecting

fractions by automated fraction collector (Bio-Rad BioLogic BioFrac Fraction Collector)

with a calculated void time of 0.334 min. A gradient elution method of increasing salt









concentration was tested based on a CIM disk separation of plasmid DNA, and is

shown in Table 3-1.182 An on-line conductivity detector was available for monitoring the

increasing salt concentration of the gradient, while DNA elution was visualized by UV

detection at Aabs= 254 nm. Fractions were collected every 0.5 min beginning at 2 min.

Table 3-1. Method development conditions #1
Step # Time (min) Buffer Flow (mL/min) Total (min)
1 1 BB 2 1
2 5 0-100% EB 2 6
3 2 EB 2 8
4 5 100-0% EB 2 13
5 3 BB 2 16

Table 3-2. Method development conditions #2
Step # Time (min) Buffer Flow (mL/min) Total (min)
1 1 BB 2 1
2 5 0-100% EB 2 6
3 2 EB 2 8
4 5 100-0% EB 2 13
5 7 BB 2 20

The procedure was systematically tested with DNA incubated with: 1) Monolithic

disk; 2) Blank housing (no monolithic disk); 3) Control of the monolithic column with no

DNA injection; 4) Blank (no drug) disk. In the first trial, the conditions in Table 3-1 were

used to probe the DNA eluted from a 30 min incubation of 1 mL of 250 nM random DNA

library in BB with the disk. The first peak in the chromatogram (Figure 3-1A)

corresponds to an air bubble, while the second peak area contains DNA as confirmed

by UV absorption (previously described). For the second trial, step #5 of Table 3-1 was

extended to 5 min to allow the conductivity to level off. In the experiment, 0.7 mL of a

random 278 nM library was injected onto the housing (Figure 3-1B). In experiment #3,

the monolith underwent the 5 min wash protocol with no DNA incubation (Figure 3-1C).

Finally, in experiment #4, the final wash step was extended to 7 minutes based on the









results from the previous experiments that indicated the DNA may not be completely

eluted (Table 3-2). One mL of 100 nM DNA was incubated with the blank disk for 30

minutes and subjected to the gradient in Table 3-2 (Figure 3-1D).

The results indicate that the gradient is capable of eluting DNA binding to the

monolith, but that the DNA may bind nonspecifically to the housing or monolith matrix,

requiring a negative selection. Also, DNA was still eluting from the monolith when no

additional DNA was added (Figure 3-1C), demonstrating that the wash step should be

increased. After any incubation with DNA, from this point the disks were incubated with

2 mL of 7 M NaCI for a minimum of 30 min to remove tightly-binding sequences before

the next round of selection.

Next, the protocol in Table 3-2 was tested in triplicate to determine if the DNA

from 1 mL of 100 nM solution incubation was reproducible. Figure 3-2 is a

representative sample chromatogram showing the peaks consistently observed with the

program.

Since the program reliably showed DNA elution in the area of ~15 minutes, we

wished to decrease the slope of the gradient during that time frame in order to capture

DNA fractions in more tubes, and the fractions were collected every 0.34 min instead of

0.5 min. Therefore, the method in Table 3-3 was devised and tested to elute DNA

binding to the target after 2 mL of a 5 pM DNA solution was incubated with the disk for

30 min. The chromatogram shown in Figure 3-3 displays a prominent peak under the

SELEX-like conditions in the expected time frame. The DNA from fraction 16 was PCR

amplified using the protocol in Table 2-11 for 30 cycles, demonstrating that the UV peak

was actually detecting DNA.

















5 00-



4 oo

3 50-

4.00-





150-

1 001

oa

0.00-


0 1 2 3 4 5 6 T a 9 10 11 12 13 14 15 16 1T
Tie [minutes


B x1 E-2
5 O- /

4 50 -

4 DO -5

3 0 -


2 50-

2 DO-




0 50




1 2 3 4 5 6 7 a 9 10 11 12 13 14 15 16 17 1B 19
Tin- minutes ]



Figure 3-1. Chromatograms from method development. A) DNA incubated with drug

disk; B) DNA and column housing; C) Drug disk without DNA; D) DNA

incubated with blank disk.




















C xl E-2
500-

4 50

400-

350-

300

2 50-

2 00-

1 50-

1 lOO

050D

S00.

o so
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19
Time minutes]


Sx1 E-2


5.00 -
sa-

4.50-

4.00-

3.50

3 00-

2.50-

" 2.00o

1.50-

O ,--
1.00-

0250
D.O-

o on-


0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
Time minutes


Figure 3-1. Continued




xl E-2
50f-rr .


250









050
0 a0
os-


oo-
o o


2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21
Time fminutesl


Figure 3-2. Confirmation of LPC method.


q T T T T T p p T I j V I p I










Table 3-3. Method development conditions #3
Step # Time (min) Buffer Flow (mL/min) Total (min)
1 1 BB 2 1
2 5 0-100% EB 2 6
3 2 EB 2 8
4 5 100-25% EB 2 13
5 12 25-0% EB 2 25

x 1.E-2
5.00 -
4.50
4.00-
3.50


2.50
2.00
1.503


0.50


0 5 10 15 20 25 30 35 40
Time [rin utes]

Figure 3-3. Chromatogram under SELEX-like conditions.

Preliminary Selection

Selection-simulating conditions were utilized to compare the elution profile of the

DNA binding to the drug and blank disks. If these profiles differed, specifically if

different peaks, or peaks with a larger area occurred in the drug disk elution, there was

a strong likelihood that at least some of the sequences eluted from the peptide disk

were binding specifically to the drug instead of to the polymeric matrix.

A new set of primers and library (now with a 46-mer random region) were

synthesized as previously described with amplification using the same PCR protocol as

in Figure 2-11. An unlabeled sense primer was used to facilitate monitoring by real-time

PCR (qPCR).

Forward Primer (sense): 5'-ATC GTC TGC TCC GTC CAA TA









Reverse Primer (anti-sense): 5'-Biotin-GCA CGA CCT CAC ACC AAA (IDT)

5'-ATC GTC TGC TCC GTC CAA TA -N46- TTT GGT GTG AGG TCG TGC

For this study, 2 mL of 500 nM library DNA was incubated for 30 min with both the

drug and the blank disk. The elution protocol from Table 3-3 was administered

separately to both disks. Figure 3-4 shows an overlay of the drug and blank disks,

where the black arrow shows the peak from the drug disk, while the red arrows point to

DNA from the blank disk. A clear difference between the chromatograms is observed in

which the drug disk chromatogram has a taller and broader peak around 15 min than

the blank disk. This served as a valid proof-of-concept study to proceed with the actual

selection.








W .







Figure 3-4. Overlay of proof-of-concept studies of selections on drug (black arrow) and
blank (red arrows) disks.

Selection Conditions

The drug-immobilized or blank disk was washed for 5 minutes with both EB then

BB at 2 mL/min. Two mL of 500 nM DNA library (1 nmol) in BB was heated to 95C for

5 min and snap cooled on ice. This DNA was incubated with the disk in a small dish for

30 min. The disk was then placed in the provided column housing, and connected to









the LPC device. The elution protocol was implemented as described in Table 3-3. DNA

elution from the disk was monitored by UV detection at Aabs= 254 nm. The same

procedure was followed for the blank disk, with a separate aliquot of DNA library.

Fractions (containing ~680 pL) were analyzed by qPCR (see method below), and those

corresponding to the peak of the drug disk were combined and incubated with the blank

disk for 30 min in a small dish. Nonbinding sequences were amplified by PCR

confirmed by 3% agarose gel electrophoresis. Each disk was rinsed and soaked in a

saturated NaCI solution to remove any remaining bound sequences prior to the next

round. The selected sense DNA strands were separated from the biotinylated

antisense DNA by alkaline denaturation and affinity purification with streptavidin-coated

beads to produce ssDNA that was desalted for the second round of selection. The

methods for agarose gel electrophoresis, ssDNA production, and desalting are outlined

in Chapter 2.

Monitoring by Real-time PCR

Real-time PCR was utilized to prove the difference in DNA eluted from the drug

and blank disks. qPCR monitors an increase in fluorescence intensity of SYBR green

dye, which preferentially fluoresces when binding dsDNA, in real time as the reaction

progresses. Graphical output is typically displayed as either baseline (ARn) or

background (observed in control) subtracted fluorescence versus the cycle number

(Figure 3-5).183 A threshold is set, defined as 10 times the standard deviation of the

baseline, and the cycle time (Ct) is the number of cycles required to exceed this

threshold. A lower Ct value indicates that more DNA was initially present, as observed

for the 5-fold dilutions plotted in Figure 3-5. For our purposes, a lower Ct value for the

drug disk relative to the same fraction in the separate blank disk experiment insinuates










that some of the excess DNA in the drug fraction binds specifically to the peptide, and

not exclusively to the blank matrix.

A
2000000-
18ooooo- Plateau
--200000
1600000 50000-
-- 12500
1400000- 3125
-781
-- 195
1200000- -49
NTC
C oo00o0 Log phase



ct
600000-


200000 Threshold

Baseline 9 18 27 36 45
-200000 -


Figure 3-5. Amplification plots of IL-4 plasmid cDNA. Fivefold serial dilutions of plasmid
cDNA were amplified by qPCR (cycle number along x-axis).183

qPCR was carried out on a Bio-Rad iCycler with MyiQ software and detection

instrumentation. Centricon centrifugal filters (Millipore, YM-10) were used to desalt the

fractions prior to qPCR. PCR mixtures contained iQ SYBR Green Supermix with iTaq

DNA polymerase (Bio-Rad) and 0.5 pM of each primer. The conditions were: 95.0C for

3.0 min, 40 repeated cycles (30 sec) of 94.0C, 55.0C (30 sec), and 72.0C (15 sec),

followed by the final steps of 5 min at 72.0C, then an indefinite hold at 4.0C.

Second and Subsequent Selection Rounds

A similar protocol of selection and qPCR was carried out for the second round of

selection, except 2 mL of 100 nM DNA (200 pmoles) was now used for incubation with

each disk. Round 3 was actually repeated twice to confirm the results. In both

instances, 2 mL of 100 nM DNA (200 pmoles) were incubated with drug and blank disk.









AlphaScreen Analysis of Selected Pools

An AlphaScreen (Amplified Luminescent Proximity Homogeneous Assay) was

used to determine the pool with the highest-affinity candidates for DNA sequencing.

The method is a bead-based non-radioactive proximity assay that measures

ligand/target interactions by the output of fluorescence by binding pairs (Figure 3-6). In

the assay, biotinylated drug is immobilized on a streptavidin coated donor bead that

generates singlet oxygen upon laser excitation at 680 nm. This singlet oxygen initiates

fluorescence (520-620 nm) in anti-FAM coated acceptor beads with FAM-labeled DNA

pools immobilized on the surface only if the beads are within the limited distance (200

nm) the singlet oxygen can travel during the excited state lifetime.184 Therefore, the

beads will only generate measurable signal if the drug and pools are binding, bringing

the beads into close enough proximity for the singlet oxygen to reach the acceptor

bead.

Excitation at
680 nm
Emission
520-620 nm
102 travels no more than 9 Streptavidin
200 nm distance
SBiotinylated
Bivalirudin
phg 5'-FAM-aptamer
nor B Anti-FAM





Figure 3-6. Schematic of AlphaScreen assay.

All assayed DNA aptamer candidate pools were converted to ssDNA (described in

Chapter 2) after PCR amplification with 5'-FAM-ATC GTC TGC TCC GTC CAA TA









(IDT) to provide an immobilization label for AlphaScreen anti-FITC acceptor beads

(Perkin-Elmer). Anti-FITC beads (20 pg/mL, final) and dye-labeled DNA pools (400 nM,

final) were incubated for 1.5 hours in an OptiPlate 384-well microplate (Perkin-Elmer)

under subdued lighting. Streptavidin-coated AlphaScreen donor beads (20 pg/mL, final;

Perkin-Elmer) and biotinylated drug (100 nM and 100 pM, final) were then added to

each well and incubated for 1.5 hours. The final volume in each well was 25 pL, and

the buffer for this experiment consisted of BB with 50 mg/mL BSA to inhibit non-specific

binding. The positive control was 400 nM biotinylated FITC, substituted for the DNA in

the method described above, and the negative control was all components except DNA.

Data was read in replicate on an EnVision microplate reader (Perkin-Elmer).

Sequencing of Selected Pool

A 400 pL (total volume) 25 cycle PCR (Table 3-4) was performed according to the

protocol in Figure 2-11. This dsDNA amplified pool was prepared for sequencing by

high-fidelity (Roche Applied Science FastStart High Fidelity PCR System) PCR

amplification with fusion primers specific for 454 sequencing technology. The primer

sequences used for the PCR are a combination of the 454 fusion tag and the regular

sense and anti-sense primers previously used for amplification:

Sense: GCC TCC CTC GCG CCA TCA GAT CGT CTG CTC CGT CCA ATA

Anti-sense: GCC TCC CTC GCG CCA TCA GAT CGT CTG CTC CGT CCA ATA

Table 3-4. PCR preparation
Reagent Volume (pL)
10X PCR Buffer 40
dNTP (2.5 mM each) 32
Primers (10 pM each) 20
DNA pool 40
DNAse-free H20 268
Taq Polymerase 1.2









The fusion primer PCR also utilized the protocol from Figure 2-11, but was

prepared by adding 0.5, 1.0, or 2.0 pL of the dsDNA to 50 pL of the PCR reaction

mixture prepared from Table 3-5. Agarose gel electrophoresis (Figure 3-7) showed that

each volume of pool added amplified in a similar fashion. UV spectroscopy was used to

calculate the amount of DNA as 452 pg/mL for the 0.5 pL pool addition product.

Table 3-5. Fusion primer PCR
Reagent Volume (uL)
10X PCR Buffer 20
dNTP (10 mM each) 16
Primers (10 uM each) 8
DNA pool Varied
DNAse-free H20 152
High Fidelity Polymerase 4


1 2 3 4


Figure 3-7. Optimization of pool volume for fusion primer PCR. Lane 1= control; lane 2=
2.0 pL pool added, lane 3= 1.0 pL pool added; lane 4= 0.5 pL pool added;
lane 5= 25-bp ladder.

The dsDNA product was purified by a QiaQuick PCR Purification Kit (Qiagen) as

per manufacturer's instructions. Buffer PB (125 pL; Qiagen proprietary buffer- not the

same as in Chapter 2) was added to the PCR product, and the sample was centrifuged

at 13,000 rpm for 45 sec in a QiaQuick column. The flow through was discarded and









0.75 mL Buffer PE (Qiagen buffer) was added to the column and centrifuged for 1 min

at 13,000 rpm. This flow through was discarded as well, and the remaining product was

centrifuged 1 min at 13,000 rpm. The filter portion was placed in a clean 1.5 mL

microcentrifuge tube, and the purified product was eluted with 50 pL Qiagen buffer EB

by centrifugation (13,000 rpm, 1 min). The concentration was determined to be 24

pg/mL by UV spectroscopy, which was diluted to 4.2 pg/mL with the Qiagen EB buffer

according to 454 procedures (request samples ~5 pg/mL). Gel electrophoresis on the

product confirmed that the desired 122-bp sequence is present (Figure 3-8).

Sequencing of the selected pool was performed by 454 sequencing at the University of

Florida ICBR.

123















Figure 3-8. Agarose gel electrophoresis analysis of fusion primer-amplified, purified
pool. Lanes 1 & 3= 25-bp ladder; lane 2= purified pool.

Sequence Alignment

Following DNA sequencing, a total of 10,192 sequences were reported. The

individual DNA sequences were analyzed using the MAFFT sequence alignment

program by removing the 3' and 5' primer regions (when possible) from sequences prior












to analysis. Several of these sequences demonstrated sufficient homology, and were


synthesized with a 5'-FAM fluorescent label for further characterization according to


previously described methods. An example of the homology displayed by the alignment


procedure is shown for one aptamer in Figure 3-9.


FS26P3403C44IQ
FS26P3403CYP77
FS26P3403C48AR
FS26P3403C4FA6
FS26P3403COHVC
FS26P34-I*,31 ,E
FS26P3403C9Y1J
FS26P3403DBR1F
FS26P3403C6DN1
FS26P3403C1VFM
FS26P3403C3NQA
FS26P3403CX39T
FS26P3403C9IRA
FS26P3403C5HUH
FS26P3403CYTH2





FS26P3403DAE7D
FS26P3403CYSPU
FS26P3403DBMX3
FS26P3403C8ZA3
FS26P3403CX06Z
FS26P3403DDC60
FS26P3403CYOPP
FS26P3403CY160
FS26P3403C3KGY
FS26P3403DAW6N
FS26P3403C9B23
FS26P3403C9SLA
FS26P3403COEWF
FS26P3403C3QCS
FS26P3403C1YUD
FS26P3403C2X8X
FS26P3403DBG5B
FS26P3403COV1T
FS26P3403C98YZ
FS26P3403DAFAH


-------gg-tgaagta--------actg---------------aggcgct--agttaa
-------ca-tccaata--------ggcg--------------cgacgcgct--aggt-a
-------gg-ttccaaagtttatacaacg--------------acacgcgct--atgg--
-------gg-gcagctgatctgggcgacg---cgtaacgctagtgaggcgct--------
-------ga-tgctagtgact----------------------acggttactcctagacc
aggaca------ttaacgataatacgtatcatctacgatgg---------------ctca
gttacagataagcaaacgataatacgc--------cggtgg---------------ttct
tt--------------cgagtatacatgcgacacacggctgggt---------gcattaa
c-----------ctgacccgaatgcctcggac---tgaaagcgc---------agatgag
g----------------------------ccgcaatccggggccc---------------
g----------------------------ccgcaatccgggggcc---------------
g------------------------acatcttctacggggggcctttaacgaagtaaccg
gg---------------------ttgtatggtctccggggggcc--ggatggagacacct
cg----------------------gcgataaatacccgtgggcc--gggggcctcttttt
c--- -------------------- catacctatt --------ccttttcoatcccactca




tg-------------------------------------gcgtaacatcattttggtgtg
tgaattttgtcccggtgcatgcggctgtacgtttactatacgtggc----ctttggtgtg
c---------------tggaaacctggtagtccgagtacgtg-------actttggtgtg
--gctctt-----------ct----tcgt----tgagtaattactaggcggtttggtgtg
--tggagatta--------gc----tcta----cgtcaggtattcaggcggtttggtgtg
--cataagagt--------g--------------attacatgttcaggcggtttggtgtg
--ccttagatt--------g--------a ----ctgcttccctcgaggcggtttggtgtg
--gggactatt-----c--gc----cctt----cgaggggcaggcagaggatttggtgtg
--gggggcgag-----ac-aa----caag----tcga------ttacgcggtttggtgtg
--gaatacgtg-----a--ga---- gggg----ccgagtatctctaagcggtttggtgtg
--gtttcggct-----actac----ttag----ggagacattggtaggcggtttggtgtg
--ttgtagcgtgg---gccaa----gaca----cattgtacgagtaaacggtttggtgtg
--ttgtagcgtgg---gccaa----gaca----cattgtacgagtaaacggtttggtgtg
-------tcga--------gt----tgat----tggtatatttgggggcagtttggtgtg
-------tgcg--------cg----ttgc----tagtacg---aggggcagtttggtgtg
--gctcgccag--------ca----tgat----cagctggggcaaacgcagtttggtgtg
--tggcaggaa--------ag----ttcc----aactcgag--ggatgcagtttggtgtg
--ccagggggc--------gggcgatcct----aacccgcagcacttacggtttggtgtg
--caggtcgat--------atcgattcgg----cac---gaggctagtatgtttggtgtg
--caggtcgat--------atcgattcgg----cac---gaggctagtatgtttggtgtg


Figure 3-9. Sample alignment of JPB2 (green highlighted portion) using MAFFT.


Binding Studies


Streptavidin-coated beads (100 pL; Bangs Laboratories) were washed with 1 mL


BB and centrifuged at 10,000 rpm for 3 min. The supernatant was discarded, and the


remaining beads were resuspended in 1 mL BB. One pL of 10 pM bivalirudin was









combined with 10 pL of streptavidin-coated polystyrene beads (Bangs Labs) and

incubated for 1 hour on a shaker. The beads were centrifuged for 3 min at 10,000 rpm

and reconstituted in 1 mL BB. The beads were split into 20 samples, and each aptamer

candidate was added for a final concentration of 2.3 pM and incubated with the

bivalirudin/bead complex for 1 hour. Beads were centrifuged, washed with 1 mL of BB,

then reconstituted in 200 pL WB2 (WB2- 4.5 g/L glucose and 5 mM MgCI2 in Dulbecco's

PBS with CaCI2 (Sigma)) for flow cytometry. The fluorescence intensity of the FAM-

labeled sequences was measured with a FACScan flow cytometer by counting 20,000

events (Becton Dickinson) and analyzed using WinMDI.

Results

Selection Results

The first round of selection generated the chromatograms seen in Figure 3-10A-B.

Once again, the peak at 15 min for the drug disk appeared to have a larger area than

that of the blank disk. For confirmation, the 15th fraction eluting from both disks was

compared by qPCR (Figure 3-11A). The drug disk clearly contained more DNA, as

evidenced by the lower number of cycles required for amplification. Therefore, the

whole peak observed in the drug disk chromatogram was combined and incubated with

the blank disk to subtract out sequences binding to the column matrix. The DNA pool

remaining after the counter selection was PCR amplified and prepared for the second

round of selection.

In the second round, a chromatographic difference between the disks was also

observed (Figure 3-10C-D), but this time the drug disk did not display a noticeable peak.

This may be due to the decreased amount of DNA used from the first round to the

second round not providing enough signal to exceed the limit of detection of the system.












A similar selection using 100 nM DNA was also performed on the blank; however a


small peak was observed at 15 min for the blank disk. Groups of fractions were


combined and prepared for qPCR analysis, corresponding to fractions 80-89, and 70-79


from the drug and blank selections. Both groups of fractions displayed higher


concentrations of DNA in the drug selection than in the blank, so both groups (fractions


70-89) were combined and PCR amplified for the third round of selection (Figure 3-


11B).


A xl E-2
5 00-
4 501
4 00-
3 50-
3 00-
2 510-









0 5 10 15 20 25 30 35 40
Time [min utes]
5 00

4 50
4 00




2 50
:vv-----^_^ _^------------------------------------------------------------






















0 5 10 15 20 25 30 35 40
Time minutes]
450-


350-
300-















Figure 3-10. Chromatograms of selection rounds. A) Round 1, drug disk; B) Round 1,
blank disk; C) Round 2, drug disk; D) Round 2, blank disk; E) Round 3, drug
disk; 4) Round 3, blank disk.
disk; 4) Round 3, blank disk.



















x1 E-2



4 50-

400

3 50

300

2 50
=1
200

1 50

1 00-






-0 50 -
0



D x1 E-2
5.00 --

4.50-

4.00-




3on0-
3.50




2.50-

2.0-0-

1.50-

1 r,,-

0 50-

0.00

-0.5O- -


SiT YC S 55 LF 51,1 1. 2r f 4 & sV r


10


20
Time [min utes]


5 10 15 20 25 30 35 40
Time [min utes]


Figure 3-10. Continued

















E 1 E-2
500-

450-

400-

3 50

3 00

250.

200-

1 50-

I nn.


50
0 5 10 15 20 25 30 35 40
Time [minutes]


F 005-



0 04-



0 03-



0 002-
*<


01



:~ ~~ ~ -<---- -______

STIIIIfFiTTT ~lll TITTl T TlI TTll TTTllTl TT rTT TTIl TT TlITT l l TT lTlTl I I I!ll IITll llIrIllllirT r ITTTIllll !T i u ll
0 5 10 15 20 25 30 35 40
Time [minutes]


0

0

-0


Figure 3-10. Continued












The third and final round of selection was carried out under similar conditions to


the second round, resulting in the chromatograms in Figure 3-9E-F. Interestingly, a


small peak returns for fractions 12-20 for the drug disk, so the combination of these


fractions were assayed by qPCR along with fractions 70-79 and 80-89 (Figure 3-10C).


The differences between the drug and blank disks were minimal, indicating that the


selection was more enriched for the drug in round 2, or that some experimental error


was to blame. The round 3 selection was repeated for both disks, but qPCR analysis


yielded similar results (Figure 3-10D).


A 1400 1_ 10 B loo 0 -
I~~~~~~~~~~~tDII I I 11111I11111 011 1 11111 I I I I I I fJI1TT11117171100


1100 ---- 100
1000 1000
g 900 --- a ti -- > -- ---- 900
S800 / --- 800 d
7 700 y -- 700
S600 -- 600

100 100
-13 0 - -- 00
200 200



100 -0- -- 100
00 -- ---- -------- -- -------- 00








100 -800
200 --- -- --- ---------- 200
0 2 4 6 8 10 12 14 16 13 20 22 24 26 28 30 32 34 36 X 40 42


400 1ooo--- --------40
P300 -- -- 1- 300
1200 ---------- ---1200

100 S----- --- ---/-- 100




S60 0 4 6- 0 1- 1 16 1 C8 2 2242 8303 4 63 60 60






-igure 3-11. qPCR results for each round of
15+ cycles= 19.76; 15- cycles 26.500
4 *00 -- -- ^ 400
S300 -- ---- T F- -- -- SS





cycles=21.70; 80-89- cycles 204.3
100 --------- --- ---T- -- -- 100
-100 1-- --7F I- 100_i
0 2 4 6 8 10 12 14 16 13 20 22 24 26 28 30 32 34 36 38 40 42





15+ cycles= 19.76; 15- cycles= 26.


C) Round 3 (Threshold,102.6; 80-8
70-79+ cycles= 25.97; 70-79- cycle
cycles= 24.39); D) Round 3 repeat
80-89- cycles= 24.33; 70-79+ cycle







91


800 -- ----------------- i 00
800 -- 0 -- oo
700
600 600

400 400
300 f ^ 4 -- 000


2200 200-- ^ -- -
100 i -100
0 ------- -----" 1'- --- --- 0
-100 1-100
0 2 4 6 8 10 12 14 16 18 20 22 24 26 23 30 32 34 36 38 40 42
PoIC iplll lin tl ^ C kl^lU- Aug-O8 la3. odm
1300 100


900 -- --9--------l00












selection. A) Round 1 (Threshold=117.2;
9+ c00ycle= 27.9; 80-8- --cycl= 27.1
700 100


S900 ---- --- -- ---- ------8 -- 0















(Threshold= 80.2; 80-89+ cycles=24.71oo
s= 24.71; 70-79- cycles= 24.67).
500 ----------- ------------ ---?-------------500



0 2 4 8 10 12 14 15 18 2- 0 22 24 26 28 30 32 304 36 38 40 42


selection. A) Round 1 (Threshold=117.2;
37); B) Round 2 (Threshold= 81.7; 80-89+


); 70-79+ cycles= 23.74; 70-79-= 25.53);
9+ cycles= 27.97; 80-89- cycles= 27.12;
s= 27.26; 12-20+ cycles = 26.13; 12-20-
(Threshold= 80.2; 80-89+ cycles=24.71;
s= 24.71; 70-79- cycles= 24.67).










AlphaScreen for Pool Selection

The AlphaScreen proximity assay was employed to analyze the pools from the 3

rounds of selection and determine which one was the best candidate for sequencing.

As previously mentioned, significant signal is only generated when the drug and

aptamer are binding.


5000
5000 100 nM Drug
4500 100 pM Drug
4000
3500
3000

2500
2000
1500
1000

500

Negative Initial Library Round 1, Round 2, Round 2, Sat'd Round 2, Round 3, Frac Round 3,
Frac. 13-22 Frac. 60-69 NaCI Frac. 70-89 70-79 Frac. 80-89
Pool


Figure 3-12. AlphaScreen assay of binding of different pools with 100 nM or 100 pM
drug.

Several conclusions were drawn from the results depicted in Figure 3-12. First,

minimal signal from the negative control is measured, indicating that a binding event

must occur for a measurable response. (Fluorescence intensity values for the positive

control, 1-2x105, not shown in graph.) Also, the signal from the initial library is low, as

expected. Two concentrations of drug were used for this study, with a 1000-fold

difference between them. Therefore, it is reasoned that a pool demonstrating increased

binding at low drug concentrations will contain more high affinity aptamers, while a high

signal at high drug concentration would signify lower affinity aptamers. Several pools,









including the pool for round 1, round 2 fractions 70-89, and both round 3 pools

demonstrated significant signal at high drug concentration, but the pool from round 2

had considerably higher signal at lower drug concentration. The conclusion is that the

round 2 pool contains higher affinity aptamers, and was selected for sequencing.

The data from the AlphaScreen follows the guidelines expected for selection, and

validates the qPCR data. It was expected and confirmed, that the initial library should

have the lowest amount of binding. Also, round 1 shows a higher percentage of low

affinity aptamers, which is expected to evolve into high affinity pools as the selection

progresses. The qPCR data consistently showed, for each selection round, that the

DNA released as a consequence of the saturated NaCI wash was binding more to the

column matrix than to the drug. An example of this is the "Round 2 Sat'd NaCI" sample

that shows the lowest signal of any pool aside from the initial library. The qPCR data

also indicated that the pools from the 3rd round of selection were actually less enriched

for the peptide than for the blank. The AlphaScreen assay gives the impression that

this pool may contain low affinity aptamers, but the high affinity sequences are reduced

from round 2. The number of sequences which bind to the target as opposed to those

binding to the matrix may be so small that repeated selection rounds and PCR steps

may dwarf the population that binds to the drug, despite counter selection steps.

Sequence Alignment

The sequence alignment of the oligonucleotides obtained from 454 sequencing

revealed several sequences that displayed homology. Seven of these sequences,

displayed in Table 3-6, were synthesized for binding characterization.










Table 3-6. Probe sequences
Name Sequence
ATC GTC TGC TCC GTC CAA TAC GAG GAT GCA GAA GTT TCA ATG
JPB1 CAC TTT TGG TGT GAG GTC GTG C
ATC GTC TGC TCC GTC CAA TAC GTA ACA TCC CCG TAA TAC TAC
JPB2 TAC GGT CGT GCT GGT TTG GTG TGA GGT CGT GC
ATC GTC TGC TCC GTC CAA TAG CTG AGC AGG TAA CAA TGT GTG
JPB3 CCC AAT GTG TAT TTG GTG TGA GGT CGT GC
ATC GTC TGC TCC GTC CAA TAA AGT TAA TCC TTA GGG CTG GTA
JPB4 GGT CAT TCC GGT GGT TAT TTG GTG TGA GGT CGT GC
ATC GTC TGC TCC GTC CAA TAT ATT GTG TGA CCC CCC TCT TGT TTT
JPB5 GGT GTG AGG TCG TGC
ATC GTC TGC TCC GTC CAA TAC CAG CTA ATG TGT ATT TTG TGG
JPB6 CGG CGG ATC ATA TGA GGA GGA TTT TTG GTG TGA GGT CGT GC
ATC GTC TGC TCC GTC CAA TAG TCG GAA TAG TGA CTG TTC TTG
JPB7 TGA AAC TCA ACA CGG ATG CTG GTG TTT TGG TGT GAG GTC GTG C

Binding Studies

The assay to determine whether the sequences are binding the target is the critical

stage of post-SELEX oligonucleotide characterization; sequences that bind are officially

considered "aptamers." Nonbinding sequences represent DNA that either binds to the

target matrix, were inefficiently partitioned by the SELEX partitioning method, or are a

result of PCR bias or mutations. It is also possible that the nonbinding sequences may

actually be low affinity aptamers that may exhibit binding when the target concentration

is increased. For practical purposes, sequences which do not appear to bind with

micromolar (for small molecules, lower for protein or cellular targets) target

concentrations are not considered of interest for most studies.

Binding studies were performed on all of the sequences listed in Table 3-6 by flow

cytometry assay. Figure 3-13 shows that all of the sequences except JPB1

demonstrated an increase in fluorescence intensity upon binding to the drug.

Therefore, 6 aptamer sequences were successfully obtained as a result of this










selection. The largest fluorescence intensity shifts were demonstrated by JPB2 and

JPB5, which were subjected to further characterization in Chapter 4.


B,
i l


Beads
Initial Library


FL1-H


Beads
Initial Library
JPB3


JPB5


Beads
Initial Library
JPB2


FL1-H


Beads
Initial Library


JPB7


FL1-H FL1-H

Figure 3-13. Flow cytometry binding studies of drug and aptamer candidates. A) JPB1;
B) JPB2; C) JPB3-5; D) JPB6-7.

Conclusions

The method described for aptamer selection has resulted in several sequences

that bind to the target in only 2 rounds of selection. The conclusions of these studies

have also proven that differences in chromatographic profile of the drug and blank disk

can be validated by qPCR. Furthermore, binding was characterized to determine the


A,









optimal pool for sequencing by an AlphaScreen proximity assay. This work is

innovative in that it provides the first known studies with the goal of generating an

aptamer for a currently existing drug, and combines several exciting technologies for the

purpose.

The results described here leave room for some appealing side-studies. Of note,

the pool from round 1 may contain sequences with desirable properties due to the

highest intensity of all pools for the nanomolar target concentration study. It would be

interesting to sequence and test aptamer candidates to analyze how the affinities

compare to those from the chosen round 2 pool. Also, due to the large number of

sequences from 454 sequencing, it is feasible that many of the sequences present in

the sequencing data and not tested by the binding assay would interact with the target.

Several less-enriched sequences could be compared to those described above to

determine if the affinities are similar.

A technological advance (and the first known use of the technology for this

application) that was crucial to the success of the project is 454 sequencing. In a

traditional cloning/chain-termination sequencing protocol, the pool to be sequenced is

cloned into a vector prior to sequencing. Due to low efficiencies associated with the

cloning process, the pool must be highly evolved to ensure representation in the

sequencing data. Also, Sanger sequencing typically only generates several hundred

sequences, depending on the cloning conditions, as opposed to more than 10,000

sequences reported by 454 sequencing. Therefore, a higher percentage of sequences

present in the final pool are detected, and the large number of sequences reported

allows for a higher likelihood of homologous sequences after alignment. Thus, it is









possible that 454 sequencing permits fewer rounds of selection due to a higher

percentage of sequences detected overriding the need for a highly evolved pool.

One modification to the procedure that may produce better results would be to

implement a step gradient with a higher final salt concentration. In this manner,

fractions corresponding to each increase in salt concentration could be pooled and

assayed for binding using the AlphaScreen technique. Fractions collected later in the

process using the highest salt concentration would likely present as higher affinity

binders. Also, a more sensitive detector would aid in the process to distinguish areas

where DNA was eluting (not corresponding to peaks) from the baseline. As it was, PCR

was required for non-peak fractions to determine whether DNA was present.

As aptamer technology matures, solely generating aptamers is not considered to

be extraordinary; the novelty of the sequences lies in their applications. Despite these

promising initial results in selecting aptamers, additional testing must be carried out to

prove the aptamers can function in the desired capacity. The characterization and

testing of the aptamers for their ability to serve as an antidote to an anticoagulant is

discussed in Chapter 4.









CHAPTER 4
CHARACTERIZATION OF APTAMER AFFINITY AND APTAMER ANTIDOTE
TESTING

Introduction

A panel of 6 aptamers was identified from the selection and binding studies in

Chapter 3. After aptamers were confirmed to bind to the desired target they were

characterized in terms of their affinity for the target in this chapter. This affinity is

described by the dissociation constant (Kd) of the target/aptamer complex; lower Kd

values correspond to higher affinity between target and ligand. Some methods

proposed to determine Kd include flow cytometry,28 equilibrium dialysis,55 analytical

affinity chromatography,56 ultrafiltration,176 fluorescence anisotropy (FA),185 and surface

plasmon resonance (SPR).186

Next, the aptamers were tested to determine whether they function in the capacity

they were selected for, which is an antidote to an anticoagulant. Preliminary studies

generally take place in buffer or human plasma (in vitro), then progress to in vivo animal

models. Frequently, the aptamers are truncated after the preliminary in vitro studies to

remove regions which do not contribute to binding. Potential aptamer therapeutics are

typically truncated from initial 80-100mer sequences to 40 nucleotides or less.187 This

reduces the cost of aptamer synthesis, increases the yield, and may actually act to

stabilize the aptamer, resulting in an increased affinity for the target.188 Additionally, for

aptamers designed for in vivo applications such as targeted drug delivery, smaller

sequences have a higher depth of penetration into tissue.189 Standard methods of

aptamer truncation include radioactive labeling and fragmentation, followed by synthesis

and purification of active sequences.190 A less labor intensive method involves analysis

of the individual sequences of a homologous family obtained after pool sequencing and










alignment. However, this requires the synthesis, purification, and affinity

characterization of many sequences in a family, also a time-consuming process.

Finally, if the aptamers are susceptible to nuclease activity in plasma, additional

post-SELEX modifications are implemented to reduce the effect. Various base

modifications and end cappings such as polyethylene glycol (PEG) may be utilized, as

described in Chapter 1. Additionally, PEG end adaptors moderate renal clearance by

increasing the molecular weight of the compound, increasing the circulation time of the

probe. Figure 4-1 depicts the modifications necessary for commercialization of

Macugen, including a 40 kDa PEG linker and 3' end cap.191

U G
A C
A-U
G- U
U-A
G-U
A A

A A
A- U
G-C
rk_-^ -PEG C -Cap/ I
A H )

I --P-0







Figure 4-1. Secondary structure and modifications of the anti-VEGF aptamer
pegaptanib (Macugen). Bold gray nucleotides represent 2'-deoxy-2'-fluoro
nucleotides whereas bold black nucleotides are 2'-O-methyl nucleotides. The
two adenosine residues shown in italic black are ribonucleotides. The 5'-
position contains a 40 kDa polyethylene glycol (PEG)-linker, where n is
approximately 450. At the 3'-end, a 3'-3'-dT (Cap) structure was added.191

The work presented in this chapter describes a fluorescence anisotropy (FA)

method used to characterize the binding affinity of the aptamer/target interactions. The









highest affinity aptamer was tested as a prospective antidote to the anticoagulant

bivalirudin in both buffer and human plasma. A novel method of aptamer truncation

using a DNA microarray was studied in order to provide functional abridged aptamer

sequences in a shorter amount of time than conventional approaches.

Materials and Methods

Buffer

Binding buffer (BB): PBS buffer (without MgCI2 and CaCI2) with 5 mM MgC12.

FA Dissociation Constant Measurements

Dissociation constants are a vital measuring stick for characterizing the affinity of

an aptamer for the target. For the methods discussed, a fixed quantity of target was

incubated with various concentrations of aptamer. A binding curve was constructed

based on the increase in signal at higher aptamer concentrations until signal saturation

occurred. This curve was best-fit to a single-site saturation ligand binding equation, in

which Y is the signal intensity, Bmax represents maximum signal intensity, and X is the

concentration of ligand.

Bmax X
Y = X(4-1)
(Kd + X)

In a homogeneous solution of fluorophore-labeled molecules, each probe is

randomly oriented. When the solution is exposed to polarized light, only the

fluorophores with the transition moments oriented with the electric vector of the

polarized light are excited. Under ideal conditions, the samples will also emit light

polarized in the same direction as the excitation light, but real situations involve a

depolarization of fluorescence emission, commonly due to rotational diffusion of the


100









molecules, but also influenced by light scattering, reabsorption, and polarizer

misalignment. 192

Anisotropy (r) is a dimensionless parameter defined as the ratio of the polarized

fluorescence emission component to the total light intensity (equation 4-2). An emission

polarizer is rotated between positions detecting fluorescence intensities parallel to

polarized excitation light (III), and perpendicular to the excitation light (I-). Anisotropy is

directly proportional to the rotational correlation time of the fluorophore (9) by the

transposed form of the Perrin equation in equation 4-3, which in turn, is inversely

proportional to the diffusion constant (D) as in equation 4-4.192 Thus, higher molecular

weight molecules have lower D values, which means increased e and r values.

(Additional terms: ro is fundamental anisotropy (the anisotropy in the absence of

depolarizing processes), and T is the fluorescence lifetime.)

r= I1 (4-2)
(II +2 21)

r=ro (4-3)
1+ 0)

1
0 =- (4-4)
6D

When FA is used to measure binding between a fluorescently-labeled molecule

and a ligand, the higher molecular weight component is typically titrated into a solution

of the labeled lower molecular weight molecule. This maximizes the molecular weight

difference between the bound and unbound forms, increasing the change in anisotropy

(Ar) at each aliquot added. Based on this pertinent background, the drug target was


101









dye-labeled and the aptamer probe was titrated into the solution, and Ar was plotted as

a function of aptamer concentration

For this dissertation work, proof-of-concept work was performed using the 15-mer

thrombin aptamer of sequence 5'-FITC-GGT TGG TGT GGT TGG titrated with various

thrombin concentrations as a model system. Anisotropy measurements were

performed on a Fluoromax-4 spectrofluorometer (Horiba Jobin Yvon) with excitation and

emission polarizers in L-format. Based on the manufacturer's recommendations, the

following settings were optimized: AEx= 492 nm; AEm= 520 nm; integration time= 1 sec;

slit width= 10 nm; aptamer concentration= 61 nM; total volume= 200 pL. Thrombin

concentrations from 2-200 nM in the thrombin selection buffer (20 mM Tris-acetate, pH

7.4, 140 mM NaCI, 5 mM KCI, 1 mM CaCI2, 1 mM MgCI2) were titrated into the aptamer

solution and incubated for 1 min. A minimum of three measurements were taken with

the initial aptamer solution and for each thrombin concentration titrated. The results

were plotted using SigmaPlot as change in anisotropy versus aptamer concentration

using a best-fit one-site ligand binding model to determine the Kd. The change in

anisotropy is the average anisotropy of the initial dye-labeled aptamer subtracted from

the average anisotropy value at each thrombin concentration.

To measure the FA for our aptamers, the drug was conjugated to 5,6-NHS-TMR

(5,6-N-hydroxysuccinimide-carboxytetramethylrhodamine; AnaSpec) in a similar

manner to that described for the drug/biotin conjugation. In this case, 5 mg of TMR dye

was dissolved in PBS buffer (no MgCI2 or CaCI2 added) and dimethyl sulfoxide (DMSO),

and slowly added to 46.5 mg Angiomax in 100 pL PBS buffer for a total volume of 600

pL. The solution was stirred overnight at 4C and HPLC purified as described for the


102









biotin/drug conjugation. Anisotropy measurements were performed on a Fluoromax-4

spectrofluorometer (Horiba Jobin Yvon) with excitation and emission polarizers in L-

format. Based on the manufacturer's recommendations, the following settings were

optimized: AEx= 545 nm; AEm= 580 nm; integration time= 2 sec; slit width= 7 nm (JPB2)

or 9 nm (JPB5, TV01, and TV03); drug concentration= 1 pM; total volume= 200 pL.

DNA concentrations in the micromolar range were titrated into the peptide solution and

incubated for 1 min. A minimum of four measurements were taken for each aptamer

concentration. The results were plotted in SigmaPlot as change in anisotropy versus

aptamer concentration using a best-fit one-site ligand binding model to determine the

Kd. The change in anisotropy is the average anisotropy of the initial dye-labeled peptide

subtracted from the average anisotropy value at each DNA probe concentration.

Sequence TV01: ATCG TCT GCT CCG TCC AATA GT GCA TTG AAA CTT CTG CAT

CCT CG TTTG GTG TGA GGT CGT GC; TV03: ATCG TCT GCT CCG TCC AATA

GCG TGC ATT GGT TTA CTG CAT CCG TGA AAC TGG GCT TTG GTG TGA GGT

CGT GC.68

Clotting Experiments in Plasma

The clotting experiments functioned on an increase in light scatter over time as

thrombin cleaved soluble fibrinogen into insoluble fibrin. The terminology used to

describe the coagulation experiments is shown in Figure 4-2. The normal clotting time

expresses the time for uninhibited thrombin to cleave fibrinogen, and is expected to be

short (Figure 4-2A). When the drug is added to the system, thrombin is inhibited,

resulting in prolonged clotting times (Figure 4-2B). Ideally, when the aptamer is added,

it will prohibit the peptide from binding thrombin, allowing the protein to cleave

fibrinogen in a short time period (Figure 4-2C), similar to the normal clotting time.


103








Thrombin
A Fibrinogen
Normal": Short Clotting Time


B Fibrinogen
"Prolonged": + J Delayed Clotting Time
Peptide


C Aptamer + + Fibrinogen
Aptamer + Fib Short Clotting Time
Influence: Peptide Apta L + V C t T


Figure 4-2. Modes of coagulation. Expected results include: A) Short normal clotting
time; B) Longer clotting times for the experiments prolonged by drug inhibition
of thrombin; C) A return to short clotting times when the aptamer binds to the
drug and frees thrombin for cleavage of fibrinogen.

Bivalirudin dose response curve

The first step in the plasma studies was to determine the amount of drug that

doubled the normal clotting time. In this experiment, the reaction was carried out in a

100 pL quartz fluorescence cuvette (Starna Cells), and the light scattering was

monitored on a Fluoromax-4 spectrofluorometer (Horiba Jobin Yvon). The settings for

this experiment were as follows: A= 500 nm; slit width= 3 nm; integration time= 0.1 sec;

interval= 0.5 sec, temperature= 37C. Three different wavelengths were tested for

monitoring the reaction, and the one utilized in further studies was the wavelength which

gave the largest response when thromboplastin-L was added. Thromboplastin-L

(Pacific Hemostasis) was equilibrated to 370C in a water bath prior to experiments.

UCRP (universal coagulation reference plasma; 50 pL; Pacific Hemostasis) was added

to the cuvette and placed in the sample chamber. The plasma was incubated in the

sample chamber with different concentrations of drug (0-1000 nM) for 2 min during

which the system was equilibrated to 370C and the signal from the instrument leveled


104









off (step 1). Thromboplastin-L was mixed (50 pL) with the plasma (step 2), and the

increase in light scatter was monitored by FluorEssence software. Each drug

concentration was tested twice, then each curve of scattered light intensity versus time

was best-fit to a 4-parameter sigmoid equation using SigmaPlot to determine the

clotting time (considered as the halfway point between the minimum and maximum

intensity values). These coagulation times for each drug concentration were then

averaged, plotted versus time, and best-fit to the same 4-parameter sigmoid equation.

The drug concentration required to double the normal clotting time was considered to be

the concentration at the halfway point between the minimum and maximum clotting time

values.

Coagulation testing

The same conditions as those used in the Bivalirudin dose response curve were

retained for the aptamer-influenced measurements. Three types of measurements

were compared for this test: 1) Normal clotting time; 2) Prolonged clotting time with 352

nM drug added in step 1; 3) Addition of 5-50,000 nM aptamer sequence (or control

TV03 nonbinding sequence) and 352 nM drug in step 1. Step 2 addition of

thromboplastin-L remained the same for each type of measurement. Each aptamer

concentration was tested a minimum of two times, and each measurement plotted as a

function of light intensity versus time and best-fit to a 4-parameter sigmoid equation

using SigmaPlot to determine the clotting time (considered as the halfway point

between the minimum and maximum intensity values). The average of the clotting

times determined for each aptamer concentration represent the reported clotting time.


105









Clotting Experiments in Buffer

The clotting experiments in buffer were also based on the premise of light

scattering generated over time by thrombin cleaving soluble fibrinogen into insoluble

fibrin. The buffer experiment was a simplistic assay designed to determine whether the

aptamer could function in a model setting mainly devoid of nucleases. This assay

consisted only of thrombin, aptamer, and fibrinogen.

Optimizing fibrinogen concentration

The concentration of fibrinogen used in the assay was optimized based on a

physiologically relevant concentration of thrombin.193' 194 In this experiment, the reaction

was carried out in a 100 pL quartz fluorescence cuvette (Starna Cells), and the light

scattering was monitored on a Fluoromax-4 spectrofluorometer (Horiba Jobin Yvon).

The settings for this experiment were as follows: A= 500 nm; slit width= 3 nm;

integration time= 0.1 sec; interval= 0.5 sec, temperature= 37C. The normal clotting

time was determined by adding 50 pL of 15.5 nM thrombin (human a-thrombin;

Haematologic Technologies) in BB to the cuvette and equilibrating it in the instrument

for 2 minutes (step 1). Next, 50 pL of fibrinogen (Sigma-Aldrich) concentrations ranging

from 0.5-15 pM (final) were added and mixed well by pipette (step 2), and the increase

in light scatter was monitored using FluorEssence software. Each fibrinogen

concentration was tested twice (with the exception of 15 pM fibrinogen), and each

measurement of light intensity versus time was best-fit to a 4-parameter sigmoid

equation using SigmaPlot to determine the clotting time.

Bivalirudin dose response curve

Building the dose response curve commenced using the same settings as for the

fibrinogen concentration experiments. In these experiments, 50 pL of 15.5 nM thrombin


106









in BB and drug concentrations ranging from 0-100 nM final concentration were added to

the cuvette and equilibrated in the instrument for 2 minutes (step 1). Next, 50 pL of 0.5

pM (final) fibrinogen was added and mixed well by pipette (step 2), and the increase in

light scatter was monitored using FluorEssence software. Each drug concentration was

tested twice, then each curve of light intensity versus time was best-fit to a 4-parameter

sigmoid equation using SigmaPlot to determine the clotting time (considered as the

halfway point between the minimum and maximum intensity values). These coagulation

times at each drug concentration were then averaged, plotted versus the drug

concentration, and best-fit to the same 4-parameter sigmoid equation. The drug

concentration required to double the normal clotting time was considered to be the

concentration at the halfway point between the minimum and maximum clotting time

values.

Coagulation testing

Instrument settings similar to the fibrinogen and dose response testing in buffer

were maintained. The normal clotting time was determined by adding 50 pL of 15.5 nM

thrombin in BB to the cuvette and equilibrating it in the instrument for 2 minutes (step 1).

Next, 50 pL of 1 pM fibrinogen (500 nM final) was added and mixed well by pipette, and

the increase in light scatter was monitored using FluorEssence software (step 2). The

prolonged clotting time consisted of the addition of 1.08 pL of drug (10.8 nM final) to

step 1, followed by fibrinogen addition. For the aptamer studies, drug and 5-20 pM of

aptamer or control sequence TV03 were added in step 1, preceding fibrinogen addition.

Each aptamer concentration (with the exception of 5 pM aptamer concentration) was

tested in duplicate, and each curve of light intensity versus time was fit to a 4-parameter

sigmoid equation using SigmaPlot to determine the clotting time. In the graph, the


107









normal and aptamer (or control) induced clotting times were normalized with respect to

the prolonged clotting time, showing the effect of the DNA on the drug function.

Truncation via DNA Microarray

A DNA microarray chip may solve time and material consumption barriers

associated with aptamer truncation, as discussed in the chapter introduction. The

microarray was a commercially available platform from CombiMatrix, and contained

~12,000 sequences synthesized in precise locations on the surface. Therefore, when

fluorescently-labeled target interacted with the sequences, a standard microarray reader

imaged the array, and the position of fluorescence was correlated to the exact

sequence binding the target. Fluorescence intensity determined by imaging software

can serve as a measuring stick to estimate the affinity of the ligand for the target.

Design

A total of 8,000 of the available sequences were utilized on the chip for the

truncation experiments. Several rational truncations were performed based on the

secondary structures of JPB2 and JPB5 at 37C predicted by IDT OligoAnalyzer.

Systematic truncation of single and both primer sequences, all individual stem/loop

hairpins, combinations of hairpins, and modification of the length of linkers between

hairpins for each predicted structure was performed. Additionally, both aptamers were

truncated to every possible 20-40 mer sequence. For example, possible 20-mers of

JPB5 (Figure 4-3) would begin with the sequences at positions 1-20, the next sequence

would be bases 2-21, followed by bases 3-22, etc. until all 20-mer sequences through

bases 41-60 are represented. Similarly, 32-mers would originate with sequences at

base positions 1-32, then 2-33 and 3-34, through positions 29-60. Each truncated

sequence for both aptamers was represented 5X on the chip.


108









1 5 10 15 20 25 30 35 40 45 50 55 60
ATC GTC TGC TCC GTC CAA TAT ATT GTG TGA CCC CCC TCT TGT TTT GGT GTG AGG TCG TGC

Figure 4-3. Truncation of JPB5. Sequences in red correspond to the base position
indicated above.

Microarray studies

The chip was assembled by attaching the hybridization chamber. The chamber

was filled with BB and incubated for 10 min at 45C in order to rehybridize the probes.

The chip was then equilibrated with BB for 5 min at 37C. Next, BB supplemented with

2% BSA was added to the hybridization chamber, and incubated for 30 min at 370C.

The chip was washed 2X with nuclease free H20 and incubated for 10 min. The TMR-

labeled peptide (100 pM) in BB was added to the chip and incubated for 30 min at 370C.

The chip was washed with BB, and the buffer was removed from the hybridization

chamber. The hybridization chamber was detached from the chip, and the

semiconductor surface area was covered with Imaging Solution (CombiMatrix) and a

LifterSlip (CombiMatrix). An Axon Gene Pix 4000B was used to image the chip at

excitation wavelength 532 nm. Following usage, the chip was washed with water, and

the hybridization chamber was filled with BB and stored at 4C. Stripping of binding

target was carried out by incubating the chip in BB for 30 min at 60C. The images

were analyzed by Microarray Imager 5.9.3 software provided by CombiMatrix.

Fluorescence intensity readings are reported as the average of each reported probe

synthesized in five different locations on the chip. Background fluorescence intensity

was determined by the average of control sequences provided by CombiMatrix.


109










Results and Discussion


Kd Characterization by FA

The dissociation constant was studied using a FA method. Thrombin/thrombin

aptamer was used as a proof-of-concept model for determining Kd measurements by

FA. The 15-mer thrombin aptamer has been reported in the literature to bind exosite 1

of thrombin with Kd~ 25 nM.113 In Figure 4-4, the affinity of thrombin/thrombin aptamer

at Kd= 36.9 4.8 nM was found to be very close to the literature values. Therefore, the

technique was validated, and the system was applied to bivalirudin and the selected

aptamers.

a08









a002 -
0.004





Kd= 36.9 4.8 nM
a.0 R= 0.987
R2= 0.975

-0 002 .
0 50 100 150 200 250
[Thrombin] (nM)

Figure 4-4. Kd plot of thrombin/15-mer thrombin aptamer using FA.

JPB2 and JPB5 were chosen for further characterization due to the sequences

displaying the largest flow cytometry binding assay shifts (Chapter 3). Figure 4-5A-B

depicts the binding curves generated by JPB2 and JPB5. JPB2 demonstrated a

calculated Kd= 5.99 0.88 pM, and JPB5 displayed Kd= 5.76 1.79 pM. When these

experiments were repeated for both aptamers, estimated Kd values fell within the error


110











of the values. These affinities are on the high affinity side of normal for small

molecules/peptides, with reported KdS in the high nanomolar to high micromolar


range.22, 176, 195 Additionally, two nonbinding control sequences were tested, TV03 and

TV01 (Figure 4-5C-D) which resulted in nearly linear Kd curves under the concentrations

tested. These sequences demonstrated calculated Kd values of 502.1 and 115.6 pM

respectively under the concentrations used, ranging from approximately 20- to 85X

higher than the KdS of the aptamer sequences. This confirms the aptamer sequences

and resulting binding curves are the result of target binding instead of nonspecific DNA

interactions.

A B
0 0-5--020- O.W4 --.

**<*'- l----"'^" aw
*0015 003 -




Kd= 6.899 0.88 pM Kd= 6.76 1.79 pM
sooD FR = 0.996 o.0- R= 0.967
R 0.967
R2=0.992 R2 0.916
0 5 10 15 20 25 -0.001 .
0 5 10 15 20 25 30
[JPB2] (uM) JP (
[JPB5](uM)



0.W -004
C s a M D





0&0004

S0002 -

Kd= 602.1 pM Kd- 116.6 PM
R = 0.996 R =0.996
Oo 0 i R2 = 0.992 .o0 I R= 0.990

0 10 20 30 40 0 10 20 30
[TV03] (uM) [TV0] (M)


Figure 4-5. Kd curves of aptamers and controls. A) JPB2; B) JPB5; C) TV03 control; D)
TV01 control.


111










Clotting Studies in Plasma

JPB5 was applied to plasma in order to test the aptamer's ability to serve as an

antidote to bivalirudin. A drug dose response curve was constructed as a function of

drug concentration versus clotting time (Figure 4-6). The data was fit to a sigmoidal

curve with a high degree of accuracy, and the concentration of drug required to double

the clotting time (C2) was calculated as 352 nM bivalirudin. This concentration was

implemented into the prolonged and aptamer-mediated clotting times for plasma.

45

40

S35

3 30
I-
F 25

20
0 20
C2 = 352.2 22.1 nM
R = 0.998
R2 = 0.996

10
0 100 200 300 400 500 600 700 800 900 1000 1100 1200
[Bivalirudin] (nM)

Figure 4-6. Dose response curve of bivalirudin in human plasma.

The values were compared to control sequence TV03, which demonstrated very

weak affinity for the target. The results for JPB5 and TV03 control tested over a wide

range of concentrations (Figure 4-7) showed similar values intermediate between the

prolonged and normal clotting times. This is hypothesized to be the consequence of

aptamer degradation by nucleases present in the plasma. It has been reported in the

literature that nuclease activity, primarily DNase I, is present in human plasma which


112










can degrade oligodeoxynucleotides.196 The same group states that heating the

samples to 70C has been successful in removing the activity, but this would almost

certainly adversely affect the coagulation proteins that are the focus of this study.

Therefore, experiments commencing in buffer rather than plasma would determine

whether the aptamer can inhibit drug function in mediums devoid of nuclease activity.


34
32
30
28
Z'26
(,24
a*22
. 20
18
16
14
12
In


[Probe] (nM)


Figure 4-7. Effect of JPB5 and TV03 control sequence on clotting time in plasma.

Clotting Studies in Buffer

The fibrinogen concentration and drug dose response curve for the coagulation

tests run in buffer were optimized prior to determining the aptamer-mediated effects.

The clotting time was observed to actually decrease as the fibrinogen concentration

decreased (Figure 4-8A), which is counterintuitive to expected results, as a decrease in

substrate is generally believed to result in a decrease in product formation. One

hypothesis for the reason this phenomenon is occurring includes product inhibition.

Product inhibition is especially relevant when considering that most coagulation factors

have negative feedback activity; that is, the products can inhibit upstream reactants


113


TV 03
DJPB5











once the concentration exceeds normal limits.197 This function balances the processes


so the body remains in homeostasis. The lack of all constituents of the coagulation


cascade changes the dynamics of the system, where the different components may


stabilize this effect. Fibrin strands are known to retain binding to thrombin, so possibly


the excess product at exosite 1 inhibits binding of fibrinogen to the same site.198 The


dose response curve of bivalirudin demonstrated a good fit to the data (Figure 4-8B),


and a drug concentration of 10.8 nM was calculated to double the clotting time in the


buffer system.


A


650

550

450

350
E
250

150

50





B 800

700

600
Soo
2 500



01
1. 400-





00


0.5 1.0 10.0 15.0
[Fibrinogen] (uM)
















C2 = 10.8 2.3 nM
R= 0.976
R2 = 0.953


0 10 20 30 40 50 60 70
[Bivalirudin] (nM)


80 90 100 110 120


Figure 4-8. Optimization of conditions for buffer experiments. A) Fibrinogen
concentration; B) Bivalirudin dose response curve.


114











3
DJPB5
2.5
5 TV03 (500 uM)
2 o0O3 (100 uM)
0
1.5
N
E 1

0.5 -


5 10 20 Norrml
[Probe] (uM)



Figure 4-9. Effect of JPB5 and TV03 control on clotting time. Clotting times at each
concentration were normalized as a function of the prolonged clotting time.
(Times lower than 1 correlate to times lower than that of the prolonged
clotting time, while those higher than 1 relate to an increase from the
prolonged clotting time.)

When all components were combined for testing of the aptamer response, the

results in Figure 4-9 were obtained. The values for each DNA concentration are

normalized in terms of the prolonged clotting time to provide an illustration of the effect

of the probe on the drug. Aptamer JPB5 was able to reduce the clotting time of the

system in a dose-dependent manner, with nearly complete antidote affect at 20 pM

probe concentration. In contrast, the control TV03 probe actually served to increase the

clotting time of the system. This is likely due to dilution effects, as the TV03 results

appear similar to previous results gathered when only buffer was added. The data from

the concentrated (500 pM) and diluted (100 pM) TV03 control concentrations

demonstrate that dilution plays a role in the results, since the 20 pM concentration has a

significantly enhanced clotting time than the more dilute sample. This confirms the


115









specificity of JPB5 for bivalirudin, since the aptamer decreased the clotting time, yet the

control at the same concentration as the aptamer only diluted the sample. Thus, the

aptamer was able to serve as an antidote to the drug in buffer.

Microarray Truncation




















Figure 4-10. Image of DNA microarray target binding. The region to the right represents
a blowup of the box of the whole array on the left.

The microarray studies for aptamer truncation resulted in many positions where

fluorescence from the TMR-labeled target binding was observed. An example is shown

in Figure 4-10, where a blowup of the region in the red square revealed several

positions of fluorescence. The microarray software correlates the position of

fluorescence emission with which DNA sequence is binding the target. The program

also reports the fluorescence intensity of each binding event, which is used as a metric

for determining relative target/sequence affinities.

The experiment revealed several probes which demonstrated signal significantly

higher than the background values (Table 4-1). One 31-mer JPB5 derivative of


116









sequence TAT ATT GTG TGA CCC CCC TCT TGT TTT GGT G had a signal intensity

of 511.01, while a 20-mer derivative of JPB2 with sequence TAC GTA ACA TCC CCG

TAA TA displayed a fluorescence value of 103.49. These were higher than

fluorescence intensity values for the full versions of the probes at 63.41 for JPB5 and

61.41 for JPB2. Both truncated sequences contain primarily the random regions of the

original sequences which are expected to account for the majority of the specificity and

binding properties of individual aptamers.

Table 4-1. Microarray truncation
Sequence Intensity
JPB5 (Full) 63.41
JPB5 (31-mer) 511.01
JPB2 (Full) 61.41
JPB2 (20-mer) 103.49
Background 51.76

Conclusions

This work proved the utility of FA methods in determining affinity of aptamer/target

interactions. An important benefit of FA is that the reaction takes place in free solution,

without the necessity of capturing the complex by particles for detection (flow

cytometry), or immobilization of the target or aptamer to a surface (SPR). Two

aptamers demonstrated low micromolar dissociation constants for bivalirudin using FA,

values which are on the low (high affinity) side of the normal range for small

molecule/peptide targets. It would be interesting to validate these Kd measurements by

another technique that does not require capture or immobilization such as isothermal

calorimetry (ITC). ITC directly measures the heat released or absorbed by a

bimolecular binding event, enabling a completely label-free detection method.


117









One of the aptamers, JPB5, was successful at functioning as an antidote to

bivalirudin in buffer, restoring thrombin activity in a concentration-dependent manner.

This is believed to be the first occasion where an aptamer was generated against an

existing pharmaceutical anticoagulant. Unfortunately, the aptamer did not function in

plasma, possibly due to degradation by the presence of nucleases. This sequence is

currently under evaluation for modifications such as PEGylation to increase the lifetime

of the aptamer in the plasma.

A novel method of aptamer truncation using a DNA microarray was presented in

this work, also believed to be the first of its kind. The microarray studies revealed

abbreviated sequences from both JPB2 and JPB5 which correspond primarily to the

random region of the aptamer. The truncated JPB5 sequence is presently undergoing

affinity and antidote characterization for use in replacement of the longer JPB5 aptamer.

A PEGylated version of the truncated JPB5 is simultaneously under examination for

antidote function and nuclease-resistance study.

In summary, this work provides innovative methods and implements new

technology to generate an aptamer with exciting potential for development as an

antidote for an anticoagulant drug. This antidote will improve the safety profile of the

anticoagulant bivalirudin, which is advantageous and sometimes essential for use as an

alternative to heparin. Based on the positive results of the buffer experiments, we are

currently seeking approval for testing this drug in controlled animal studies. Ideally, the

PEGylated, truncated version of aptamer JPB5 will still retain antidote activity and will

be implemented into the in vivo studies.


118









CHAPTER 5
CANCER CELL CAPTURE USING APTAMER-IMMOBILIZED SQUARE CAPILLARY
CHANNELS: PROOF-OF-PRINCIPLE

Introduction

Separating pure cells from a complex matrix is a vital process for scientists

aspiring to gain knowledge of cellular processes, and for clinicians detecting the

presence of specific types of cells for diagnostic or therapeutic purposes. Cancer cells

are of extreme significance to diagnosticians since the probability of patient survival

increases the earlier the disease is identified. However, the cells are dispersed in

exceptionally low concentrations, usually <200 cells/mL for most types of cancer.

Therefore, devices used for cancer cell detection must be highly sensitive to capture a

significant amount of the malignant cell population.

Microfluidic devices have recently been gaining attention for the efficient

separation of cancer cells.92-94 151 While microfluidics have been successful for this

purpose in the past, the devices suffer from challenges with design and fabrication,

often requiring expensive clean rooms which are inaccessible to many laboratories.

The systems may also be difficult to interface with benchtop fluidic devices, and the

chemistries used to immobilize ligands on the typically polymeric surfaces are

nonstandard.199

In contrast, capillaries have several properties which may promote their use as cell

capture devices. Capillaries have relatively simple connection options, are

commercially available with excellent batch-to-batch reproducibility, and are easily

surface-modified using simple, well-characterized chemistries. Specifically, square

capillaries are attractive for cancer detection because flat-walled capillaries (square or

rectangular) have less optical distortion and scatter than the curved walls of circular


119









tubes.156,200 This characteristic becomes significant due to the potentially extremely low

concentrations of cancer cells present in the blood promoting direct imaging of captured

cells rather than offline quantitation.

Properties inherent to aptamers such as specificity, ease of modification, and

stability to degradation make the probes ideal ligands for cell capture. This group has

previously selected the aptamer sgc8, which selectively binds to acute lymphoblastic

leukemia CEM cells, but not to the Ramos cell line (Burkitt's lymphoma) using a

modification of the SELEX procedure known as cell-SELEX.28 This aptamer binds with

high affinity (Kd= 800 pM), implying that the aptamer/cell bond can withstand a

substantial amount of force, as discussed in Chapter 1.

The work presented in this chapter aims to combine the specificity of aptamers

with the enhanced optical properties of square capillaries to improve upon results

observed in circular capillaries, as well as to provide simple and cost-effective devices

for the selective capture of cancer cells. The simplicity of setup and capture chemistry

provide the capillary system with unique advantages with an application toward the

detection of cancer cells in initial proof-of-principle studies.

Materials and Methods

Cell Culture and Buffers

CCRF-CEM cells (CCL-119 T-cell, human acute lymphoblastic leukemia) and

Ramos cells (CRL-1596, B-cell, human Burkitt's lymphoma) were obtained from ATCC

(American Type Culture Association). The cells were cultured in RPMI medium 1640

(ATCC) supplemented with 10% FBS (heat-inactivated; GIBCO) and 100 IU/mL

penicillin-streptomycin (Cellgro). Immediately before experiments, cells were rinsed

with 2 mL of washing buffer (WB2; 4.5 g/L glucose and 5 mM MgCI2 in Dulbecco's PBS


120









with CaCI2; Sigma). Experiments using unstained cells involved direct dilution of cells to

the desired concentration with binding buffer (BB2; WB2 supplemented with yeast tRNA

(0.1 mg/mL; Sigma) and BSA (1 mg/mL; Fisher)) with 10% FBS. For cell staining

experiments, the cells were diluted to 1 x106 cells/mL in WB2, and the manufacturer's

instructions were followed by treating the cells with Vybrant Dil (AEx/AEm 549 nm/565 nm)

or Vybrant DiO (484 nm/501 nm) cell-labeling solutions (Invitrogen) for 5 min at 37 oC.

Cells were then rinsed with 1 mL WB2 and reconstituted to the desired concentration

using binding buffer with 10% FBS. All cells were stored on ice until needed.

Device Construction

A
1 2 3

Magnified
view of
capillary
B I
X Avidin
Biotin-sgc8
CEM cell

Figure 5-1. Schematic of setup and immobilization. (A) A small piece of Teflon tubing
(1) connects a square capillary (2) with an observation window (3) to a
syringe. (B) Avidin is immobilized onto the capillary walls, and 5'-FAM-sgc8-
poly(T)o1-biotin is added for capture of target CEM cells.

A 8 cm piece of 359 pm o.d./74 pm i.d. square capillary (Polymicro Technologies)

was cut from the capillary spool, and a Bunsen burner was used to form a ~2-3 cm

transparent window about 1 cm from one end (Figure 5-1A). After the window was

cleaned with ethanol to remove remaining debris, a 2 cm piece of Teflon tubing (PTFE

polytetrafluoroethylenee) tubing, regular wall, 30 Gauge; Zeus) was used to connect the


121









end of the capillary farthest from the transparent window to a 500 pL U-100 Insulin

Syringe (Becton Dickinson and Company).

DNA Synthesis

The aptamer sgc8 sequence, 5'- FAM- ATC TAA CTG CTG CGC CGC CGG GAA

AAT ACT GTA CGG TTA GAT TTT TTT TTT-3'-biotin, was synthesized using an

AB13400 DNA/RNA synthesizer (Applied Biosystems). The poly(T)10 linker was

incorporated into the structure based on previous research results showing enhanced

aptamer function by extending an immobilized aptamer away from the surface.93 DNA

synthesis reagents were purchased from Glen Research (Sterling, VA). DNA purification

was performed with a ProStar HPLC (Varian) using a C18 column (Econosil, 5U, 250

mm x 4.6 mm) from Alltech Associates. UV-Vis measurements were performed with a

Cary Bio-300 UV spectrometer (Varian) to measure DNA concentration.

Device Characterization

Table 5-1. Physical properties of capillary
Parameter Calculated Value
Volume (uL) 4.62 x 10-1
D (m2/sec) 4.90 x 10-14
t (sec) 2.95 x 104
Dh (m) 7.60 x 10-5
Re dimensionlesss) 1.96 x 10-1

The physical parameters of the capillary were calculated as described in Chapter 1

in order to characterize the properties of the device (inner diameter 76 pm, length 8 cm).

The low Re confirms laminar flow (values <2000) throughout the tube. Note the

exceptionally long cell diffusion times are counteracted by a short cell settling time due

to gravity (Chapter 1).


122









Controlling Degree of Sgc8 Immobilization

Capillaries were cleaned and activated by manually adding 50 pL of 1 M NaOH,

then rinsed with 50 pL of doubly-deionized water immediately prior to use. Avidin (5

mg/mL) was introduced at 1 pL/min for 3 minutes using a Micro4 syringe pump (World

Precision Instruments, Inc.) and incubated for 15 minutes. Excess avidin was removed

with a manual wash of 50 pL of binding buffer. Ten microliter aliquots of different

concentrations (0.5 pM, 25 pM, 50 pM, and 100 pM) of 5'-FAM-sgc8-poly(T)10-biotin

were manually drawn into separate capillaries, incubated for 30 sec, and rinsed with 50

pL of binding buffer. Several sections of each capillary were imaged using an Olympus

FV500-IX81 confocal microscope, and the raw images were analyzed in terms of

fluorescence intensity by ImageJ (NIH).

Cell Capture Assays

The chemistry involved in cell capture is presented as a schematic in Figure 5-1B.

Avidin and sgc8 immobilization was carried out with 50 pM aptamer using the method

described in the sgc8 immobilization step. To avoid cell settling, 1 mL of cell

suspension was added to a dish with a stir bar on a magnetic stir plate. A small hole in

the dish enabled capillary insertion into the cell solution, and cells were withdrawn using

the Micro4 syringe pump at either 200 nL/min (Table 5-2, Experiments 2 and 3) or 300

nL/min (Table 5-2, Experiment 1) for 10 minutes. Nonbinding cells were rinsed from the

capillary with three column volumes (1.5 pL) of binding buffer via syringe pump at either

200 nL/min or 300 nL/min (depending on flow rate of cells previously added). For the

square and round capillary images, the capillaries were imaged without cell elution. For

performance experiments, captured cells were quantified by manually introducing a

stream of air into the capillary, then collecting eluted cells in 10 pL of binding buffer, and


123









finally determining cell concentration by adding the cell solution to a hemocytometer

(Hausser Scientific).

Flow Rate Experiments

Capillaries were coated with avidin and aptamer as described in the sgc8

immobilization step. Cell capture was performed at flow rates of 300, 600, or 1200

nL/min in individual capillary devices as described in the cell capture assays section.

Non-target cells were introduced only at the lowest flow rate, 300 nL/min. Images

corresponding to different positions on each capillary window were taken, and ImageJ

was used to count the number of cells in each image. For each flow rate, the number of

cells for all positions imaged along the capillary was averaged, and the total number of

cells captured was calculated by approximating the length of capillary visible in each

image to be 1 mm. Captured target and non-target cells at the 300 nL/min flow rate

were imaged both inside the capillary and on a glass slide with a coverslip, eluted by

introduction of an air stream.

Stained Cell Imaging Using Fluorescence Microscopy

CEM cells were split into 2 samples, each stained with either Dil or DiO dye as

described above. Samples stained with each dye were manually injected into separate

square capillary devices ([cell]= 1 x107 cells/mL), then mixed ([cell]= 5x106 cells/mL of

each dye) prior to injection. Each sample was imaged using a Leica DM6000B

fluorescence microscope using either BGR or monochromatic CCD (charge-coupled

device) camera filters followed by false coloring using ImageJ.

Cell Elution Efficiency

Cells were captured, eluted, and counted using the method described in the cell

capture assays section. To determine the elution efficiency, the entire length of


124









capillary was scanned to count any cells that were not eluted by the air stream, and

these remaining cells were added to the eluted cells for the total number of cells

captured by the capillary. Eluted cells were divided by total cells and multiplied by

100% to calculate the elution efficiency of the procedure. From the data obtained from

three trials, elution efficiencies of 99.78%, 99.85%, and 99.97% were obtained, leading

to an average elution efficiency of 99.87 0.06%.

Square and Circular Capillary Cell Capture Comparison

Two square (359 pm o.d./74 pm i.d.) and two circular capillaries (358 pm o.d./76

pm i.d) were prepared following the method described in the cell capture assays

section. For this work, the initial probe concentration was 5 pM, and the cell

concentration was 5x106 cells/mL. Following cell capture, the cells retained in the two

square and two circular capillaries were combined according to geometry before

addition of the eluted cells to the hemocytometer.

Microscopy and Image Analysis

Accurate initial cell concentrations for both cell lines were obtained before each

experiment by adding the initial cell solution to a hemocytometer and imaging the device

with an Olympus FV500-IX81 confocal microscope. The cells were counted by

importing the images into ImageJ and manually counting cells from each cell line. The

cell concentration was calculated in accordance with the manufacturer's instructions.

The concentration of cells captured by the device was calculated in a similar manner;

however, nine image sets were obtained, corresponding to all units of the

hemocytometer grid. For the cell staining experiments, transmitted images, as well as

those corresponding to red fluorescent target cells and green fluorescent control cells,

were obtained for each square unit of the hemocytometer grid. Once the images were


125









imported into ImageJ, red and green cells could be individually highlighted and overlaid

on the transmitted image in order to count the number of each cell type present for both

initial and captured cells.

Square and Circular Capillary Imaging

To compare the imaging capabilities of the capillary geometries, cells were

injected into an 8 cm length of 358 pm o.d./76 pm i.d. circular ((A), 2x106 cells/mL) or

359 pm o.d./74 pm i.d. square ((B) 10x106 cells/mL) capillary and imaged using a Leica

DM6000B fluorescence microscope. Cells were then captured using the method

described in the cell capture assay portion (6.4x105 cells/mL) and imaged with an

Olympus FV500-IX81 confocal microscope for either a circular or square capillary of the

same dimensions given above.

Results and Discussion

Sgc8 Immobilization

The first step in proving the utility of the device was to show that aptamer

immobilization on the capillary surface could be controlled by varying the concentration

of sgc8 introduced into the device. In this experiment, different concentrations of 5'-

FAM-sgc8-poly(T)10-biotin were immobilized on the capillary walls coated with avidin,

and the fluorescence intensity observed at each concentration was compared (Figure 5-

2). A concentration-dependent increase in fluorescence intensity as sgc8 concentration

increased was apparent, with saturation of signal at 50 pM sgc8. The confocal images

of a representative of each aptamer concentration are presented in Figure 5-3. A

saturating sgc8 concentration (50 pM) was employed throughout the remainder of the

studies in order to maximize the number of ligands binding to the cells.


126











3500

3000

2500

2000

1500

1000

500

0


0 20 40 60 80 100
[sgc8] (uM)


Figure 5-2. Fluorescence intensity increase as aptamer concentration is increased.

















Figure 5-3. Confocal images of FAM-labeled sgc8 immobilized inside a capillary at
various aptamer concentrations. A) 0.5 pM sgc8; B) 25 pM sgc8; C) 50 pM
sgc8; D) 100 pM sgc8.

Cell Volumetric Flow Rate Trends

Basic flow rate experiments were carried out in order to investigate flow rate

trends and to help identify the proper conditions for cell capture. Target cells (CEM)

were captured in the capillaries at three different cell flow rates (300, 600, or 1200


127









nL/min). The experiments demonstrated linear velocity-dependent binding (Figure 5-4),

with the lowest flow rate (300 nL/min) capturing the highest amount of cells (1374 cells

captured), followed by intermediate flow rate (558 cells captured), and finally the highest

flow rate (240 cells captured). These results agree with published data that indicates a

decrease in flow rate will consequently decrease the linear velocity of the cell

population, leading to a higher fraction of cells with sufficient time to form enough bonds

for attachment.201 Additionally, nontarget cells (Ramos) were introduced to a capillary

device at the lowest flow rate (300 nL/min), and only 26 cells were captured.

Comparison of the amounts of target and non-target cells captured at the lowest flow

rate is better visualized by the massive amounts of cells observed upon elution of the

target cells onto a microscope slide (Figure 5-4C; large circles represent air bubbles)

versus the miniscule number of nontarget cells eluted (Figure 5-4D). Thus, even at the

lowest flow rates, target cells were captured in quantities over an order of magnitude

higher than nontarget cells.

The image in Figure 5-4E shows cells dislodged by the air/liquid interface exerting

shear forces on captured cells upon introduction of an air bubble. This air-based elution

mechanism allows the cells to remain concentrated for offline counting procedures, as

opposed to liquid shear-force strategies that dilute the captured target.166 Air-based

elution forms the basis for cell counting in a hemocytometer used in the following

studies.

Fluorescence Microscopy Imaging of Stained Cancer Cells

CEM cells were stained with Dil and DiO dyes, then introduced into separate

capillaries for imaging of each dye alone and a mixture of the two dyes. The images

obtained from capillaries containing cells with only one dye were fairly clear, but when


128









the cells were mixed and BGR (blue/green/red) light was used for excitation, all imaged

cells appeared green (DiO emission). The main cause is due to the fact that DiO

emission is much more intense than Dil, contributing to a variety of effects.202

Therefore, the dyes of the cell mixture were analyzed separately using dye-specific

filters then overlaid (Figure 5-5), which proved capable of detecting both of the dyes.

However, the microscope images were not completely resolved, which will interfere with

the signal in detection of captured cells.




















T---
.A B





















E





Figure 5-4. Confocal images of cells captured in square capillaries at different
volumetric flow rates. A) 1200 nL/min CEM; B) 600 nL/min CEM; C) 300
nL/min CEM (arrow represents cells air-eluted onto microscope slide); D) 300
nL/min Ramos (arrow represents cells air-eluted onto microscope slide); E)
An air bubble dislodging cells.


129









Based on these results, confocal microscopy was used for the remainder of the

studies unless otherwise noted to produce more resolved images. The confocal

microscope is capable of sequential laser illumination, which means each dye can be

individually excited and detected without the need for changing filter cubes. This ability

proved essential since preliminary experiments with the confocal microscope also

showed green-to-red bleedthrough issues with simultaneous excitation.










Figure 5-5. Simultaneous fluorescence microscopy imaging of cells in a square
capillary stained with two different dyes.

Cell Capture Performance

Two different groups of experiments were conducted in order to evaluate the

performance of the device, and the results are summarized in Table 1. In the first group

of experiments, solutions of pure cell types, target (CEM) or control (Ramos), were

introduced into separate capillaries to determine the extent of binding for each cell type.

Initially, cell concentrations were kept approximately equal (Table 5-2, Experiment 1),

leading to a 10 times higher retention of target cells than non-target cells. Next, the

concentration of control cells was increased to roughly double that of the target cells

(Table 5-2, Experiment 2). Even under these conditions, the number of target cells

captured exceeded that of non-target cells by a factor of 4.7-27. This experiment was

performed in duplicate, and the wide range may be due to some flow variations (lower

linear velocity) present in the capillary with the higher capture efficiency.


130









The second group of experiments was designed to more closely mimic the

complexity of the physiological environment. Prior to introduction into the device, target

and control cells were first stained with Dil and DiO dyes, respectively, then mixed

together for introduction into the same capillary device. In these experiments, target

cells were captured in numbers roughly 5 times more abundant than control cells (Table

5-2, Experiment 3). As a whole, these performance studies show that the device is

capable of selectively capturing target cells. This holds true when target cells are

administered in lower concentrations than nontarget cells or even when the cell types

are mixed.

Table 5-2. Performance of capillary system
Experiment # Cell Type Relative Cell Concentration CEM/Ramos
CEM
1 Control Equal 10
CEM
2 Control 2X Control 4.7-27

3 Mixed 2X CEM 4.2-6.5

Square and Circular Capillary Cell Capture Comparison

Additional studies were performed to directly compare the amount of target cells

captured in square and circular diameter capillaries. Under the same conditions, square

capillaries retained 33% more cells than circular capillaries of similar diameters and

lengths. A possible explanation for these results is that the higher cross-sectional area

of the square device will result in a lower linear velocity of the cells when compared to

the circular capillary, consequently increasing cell capture as previously described.201

Therefore, square capillaries of shorter length than circular capillaries of similar inner

diameter can be utilized for cancer cell study, which correlates to decreased material

consumption. However, cell capture is merely a first step in the process of determining


131









cell immunophenotype and cell morphology; further information on captured tumor cells

can be elucidated by in situ imaging analysis.

Imaging Properties of Square and Round Capillaries


I


Figure 5-6. Imaging comparison of square and round capillaries. The top panels show
CEM cells that were injected into either a circular ((A), 358 pm o.d./76 pm i.d.,
2x106 cells/mL) or square ((B) 359 pm o.d./74 pm i.d., 1 x107 cells/mL)
capillary, imaged using a fluorescence microscope. The lower panels display
cells captured in either a circular (C) or square (D) capillary (6.4x105 cells/mL)
imaged with a confocal microscope.

An optimal device would allow for straightforward interfacing to the imaging

instrumentation as well as properties which would minimize optical distortion and

scatter. The benefits of using square versus circular capillary geometry were

demonstrated by imaging cells directly injected into square and circular capillaries of

comparable diameter (Figures 5-6A-B) by fluorescence microscopy or by capturing

target cells with the aptamer and imaging by confocal microscopy (Figures 5-6C-D).

For both injection and capture, cells imaged in the square capillary are better resolved

than those in the circular device regardless of the imaging system used. The improved

resolution could have an immediate application in the area of bioimaging and

bioanalysis, specifically cancer cell detection, where inability to resolve a single cell

could result in a faulty diagnosis or indication of treatment progress.


132


111 ------- -~----~









Conclusions

This work has presented proof-in-principle studies demonstrating that a simple

square capillary device is useful for selective capture of cancer cells. The system

requires no complicated design or fabrication steps or clean room facility. Furthermore,

the inner walls of the capillary can be modified with capture probes using standard

avidin/biotin chemistry, with the degree of immobilization controlled by varying the probe

concentration. The aptamer-coated square capillaries consistently captured more target

cells than control cells, even when the cell lines were mixed prior to introduction, and

the square capillaries retained significantly more cells than the circular geometry

following a direct comparison. Elution of cells retained in the capillary by a flowing air

stream consistently eluted ~99.9% of cells. Additionally, the square capillaries are

superior to circular capillaries for imaging purposes. This is a clear benefit for cancer

cell detection, specifically for CTC analysis, since it is necessary to examine cell

morphology and immunophenotype after cell capture.

In summary, the use of a square capillary system for imaging captured cells

provides a simple and cost-effective method for counting cell subpopulations. These

studies represent proof-in-principle work to validate the concept, yet do not reflect fully

optimized conditions. In the best case for the lowest flow rate utilized for administering

the cell solutions, the capture ratios of target cells versus nontarget cells were high, but

the overall capture efficiencies of CEM cells were low, only ~5%. This is not practical

for situations when low concentrations of cancer cells are present, or for prognostic

applications that monitor patient progress as a function of blood cancer cell

concentration. Therefore, work described in Chapter 6 describes the optimization of the

system, as well as application in detecting cancer cells present in human blood


133









samples. The device is also applied to types of cancer that are not blood-borne,

demonstrating the universal function of the system.


134









CHAPTER 6
OPTIMIZATION AND APPLICATIONS OF SQUARE CAPILLARY DEVICE

Introduction

Proof-of-principle studies (Chapter 5) described the use of a simple aptamer-

coated square-capillary system to capture leukemia cells from a flowing suspension.

Target cells were retained in greater quantity even when target CEM and nontarget

Ramos cells were mixed together prior to separation. The system retained CEM cells in

higher numbers than Ramos cells due to the specificity of the sgc8 aptamer for the

target cells.28 The square capillary used for cell separation also demonstrated

enhanced imaging properties when compared to circular geometry capillaries of similar

diameter.

The work in Chapter 6 will focus on optimization of experimental parameters to

improve the capture efficiency of the capillary device. This was performed by varying

the flow rate of the cell suspension through the device, which changes the linear

velocity of the cells traveling through the system. The effect of the flow rate of the buffer

used to flush uncaptured cells from the device was also explored. The capillary system

was used to image CEM cells spiked into human blood for potential applications in

cancer diagnostics.

Since leukemia is a disease specifically affecting the blood, leukemia cells are

found in relatively high concentration in the fluid. Great interest lies in detecting cancer

cells that are not normally present in the blood in high concentration in order to monitor

treatment progress and correlate cancer cell levels with patient response or disease

stage. Therefore, conditions optimized for CEM capture were tested on two different

adherent colon cancer cell lines representing Dukes' type C colorectal adenocarcinoma


135









(DLD-1) and colorectal carcinoma (HCT 116). Aptamers demonstrating a high affinity

for these colon cancer cell lines have recently been reported by this group.203 These

aptamers were tested as cell affinity ligands in the capillary device to compare the

capture efficiency of the system for tumorigenic cancers to that of blood-based cancers.

The question of whether conditions optimized for one type of cancer cell line can be

applied to another cancer cell line is also considered.

Materials and Methods

Cell Culture, Buffers, Aptamer Synthesis, and Device Construction

CEM cell culture, buffers, sgc8 aptamer synthesis, and square capillary device

construction were described in Chapter 5. Buffers: WB2- 4.5 g/L glucose and 5 mM

MgCI2 in Dulbecco's PBS with CaCI2; BB2- WB2 supplemented with yeast tRNA (0.1

mg/mL) and BSA (1 mg/mL); red blood cell lysis buffer (LB)- 150 mM NH4CI, 10 mM

KHCO3, 0.1 mM EDTA in distilled H20 (pH 7.3) filtered with 0.45 pm filter (Nalgene);

non-enzymatic buffer (NEB)- used as provided by MP Biomedicals; sgc8 sequence: 5'-

FAM- ATC TAA CTG CTG CGC CGC CGG GAA AAT ACT GTA CGG TTA GAT TTT

TTT TTT-3'-biotin.

Flow Rate Optimization

Capillaries were cleaned and activated by manually adding 50 pL of 1 M NaOH,

then rinsed with 50 pL of doubly-deionized water immediately prior to use. Avidin (5

mg/mL) was introduced at 1 pL/min for 3 minutes using a Micro4 syringe pump (World

Precision Instruments, Inc.) and incubated for 15 minutes. Excess avidin was removed

with a manual wash of 50 pL of binding buffer. Following this step, 10 pL of 50 pM sgc8

was manually drawn into the capillary, incubated for 30 seconds, and rinsed with 50 pL

of binding buffer. To avoid cell settling, 1 mL of cell suspension was added to a dish


136









with a stir bar on a magnetic stir plate. A small hole in the dish enabled capillary

insertion into the cell solution, and cells (5x106 cells/mL) were withdrawn using the

Micro4 syringe pump at either 500 nL/min, 100 nL/min, or 20 nL/min for 10 minutes.

Nonbinding cells were rinsed from the capillary with three column volumes (~1.5 pL) of

binding buffer via syringe pump at 500 nL/min. Captured cells were quantified by

manually introducing a stream of air into the capillary, collecting eluted cells in 10 pL of

BB2, and finally determining cell concentration by adding the cell solution to a

hemocytometer (Hausser Scientific). Capture efficiency is reported as the number of

cells captured divided by the total number of cells passing through the device.

Optimization of Cell Washing Flow Rate

The protocol for the flow rate optimization studies was followed using a 20 nL/min

cell flow rate. Two cell washing flow rates were compared, 200 nL/min and 500 nL/min,

administering the same total wash volume (1.5 pL). Cell capture concentration was

quantified via hemocytometer by eluting the captured cells by flowing an air stream

through the capillary into 10 pL BB2. The amount of cells passing through the capillary

uncaptured was quantified by hemocytometer by emptying the cellular contents of the

Teflon tubing and syringe into 10 pL BB2. Capture efficiency was reported as the

number of cells captured divided by the total number of cells (captured cells added to

cells passed through into tubing and syringe) from this point forward unless otherwise

indicated.

Detecting CEM Cells in Blood

Cell staining

CEM cells (1.7x106 cells) were diluted in 2 mL WB2. CellTracker Green (2 pL;

Invitrogen) was incubated with cells for 45 min at room temperature, according to


137









manufacturer's instructions. The cells were centrifuged at 970 rpm for 3 min to remove

excess dye, resuspended in 1 mL WB2, and incubated for 30 min at 37C. The cell

solution was centrifuged at 970 rpm for 3 min, and the supernatant liquid was discarded.

The pellet was washed with 1 mL of BB2, centrifuged at 970 rpm for 3 min, then

resuspended in 100 pL BB2. This cell solution was kept on ice until needed.

Removal of red blood cells

Red blood cells were removed from human whole blood (Innovative Research) by

separation by Ficoll-Paque (GE Healthcare). Ficoll-Paque is a density gradient

centrifugation medium that separates blood into 4 main layers: the top layer consists of

the plasma, followed by the buffy coat, the Ficoll layer, and the red blood cells

concentrated at the bottom of the tube. The buffy coat contains the white blood cells

and platelets, and is the layer collected and combined with the CEM cells in this

experiment. This red blood cell removal is essential so the high concentration of red

blood cells do not clog the capillary.

The Ficoll-Paque blood separation was carried out according to manufacturer's

instructions. Briefly, 3 mL whole blood was slowly added to 3 mL Ficoll in a 15 mL

centrifuge tube. The layers were centrifuged at 1200 x g at 4C with the rotor brake off.

The buffy coat was removed and centrifuged at 1800 rpm for 5 min. The supernatant

liquid was discarded and the cells were washed with 3 mL PBS buffer. The washed cell

pellet was combined with 1 mL red blood cell lysis buffer (LB) and incubated for 15 min.

The supernatant liquid was removed and the pellet was washed twice with WB2 (all

centrifugations at 1800 rpm for 3 min). The final pellet was resuspended in 3 mL BB2 to

simulate the concentration of white blood cells in whole blood.


138









Cell capture

The cell capture experiments were performed by spiking stained CEM cells into

the white blood cell suspension at either 1 x105 or 3x104 cells/mL. The basic protocol

described in the flow rate studies was followed, with a 20 nL/min cell introduction and

500 nL/min washing flow rate. The length of the capillary window was scanned, and

fluorescent CEM cells were counted. Total cell numbers were calculated by a simple

proportion relating cells captured per length of window to entire capillary length.

Colon Cancer Cell Culture

Colorectal cancer cell lines DLD-1 (Dukes' type C colorectal adenocarcinoma) and

HCT 116 (colorectal carcinoma) were purchased from American Type Cell Culture

(ATCC). DLD-1 cells were maintained in culture with RPMI-1640 containing 10% heat-

inactivated FBS (Invitrogen) and 100 IU/mL penicillin-streptomycin (Cellgro). HCT 116

cells were maintained in McCoy's 5A culture medium containing 10% heat-inactivated

FBS and 100 Units/mL penicillin-streptomycin. All cultures were incubated at 37C

under a 5% CO2 atmosphere.

Both cell lines were grown as an adherent monolayer in 100mm x 20mm culture

dishes to >95% confluence. Cells were washed in the dish with WB2, dissociated by

trypsin treatment (2 min) and seeded into culture dishes at low concentration. Within 24

hours of seeding (<40% confluence), cells were short-time trypsinized (30-45 sec) for

capture and flow cytometry studies.

Colon Cancer Aptamer Sequences

KDED2a-3: TGC CCG CGA AAA CTG CTA TTA CGT GTG AGA GGA AAG ATC

ACG CGG GTT CGT GGA CAC GGT-biotin; KCHA10: ATC CAG AGT GAC GCA GCA

GGG GAG GCG AGA GCG CAC AAT AAC GAT GGT TGG GAC CCA ACT GTT TGG


139









ACA CGG TGG CTT AGT-biotin. Aptamer KDED2a-3 binds specifically to the DLD-1

cell line, but demonstrated minimal binding to HCT 116, while aptamer KCHA10 binds

similarly to both DLD-1 and HCT 116 cell lines. Probes were synthesized and purified

as described previously both with and without a 3'-poly(T)o1 linker.

Cell Capture of Colon Cells

Testing efficiency of aptamers (without linker) for DLD-1 capture

All steps through aptamer immobilization were the same as described in the flow

rate optimization studies. For cell administration, the cells were strained with 40 pM cell

strainers (Becton, Dickinson and Company) immediately before addition to the capillary

and were kept stirring throughout the entirety of testing. These steps proved necessary

due to cell aggregation and capillary clogging without their implementation. The cell

flow rate of 20 nL/min for 10 min and wash with 1500 pL BB2 at 500 nL/min remained

the same. Captured and unretained cells were counted for capture efficiency

calculation.

DLD-1 cell capture using aptamers with linker

The binding of aptamers with the poly(T)10 linker were compared to those without

the spacer by flow cytometry. DLD-1 cells (200 pL) were removed from a 6.5x106

cell/mL solution in BB2 and diluted to 500 pL. Aliquots of 100 pL cells were added to

each of 5 tubes. Each tube (except 1 control tube) was incubated with 250 nM DNA

(library, KDED2a-3 with and without poly(T)o0 linker, and KCHA10 with linker) for 10

min. The cells were centrifuged at 970 rpm for 3 min and washed with 1.5 mL WB2,

incubated with 100 pL streptavidin-PE-cy5.5 (channel 3) at a final dilution of 1:400 stock

solution for 10 min. The cell solution was washed with 1500 pL WB2 and the cell pellet

was resuspended in 150 pL WB2 for analysis by flow cytometry. The cell capture


140









experiments were carried out using the same method as the experiments without the

linker.

DLD-1 cell capture at low concentration using aptamer KDED2a-3

To increase the number of cells accessing the capillary, the cell capture protocol

was modified to use a lower initial cell concentration (1 x106 cells/mL) and a longer cell

administration time (increased from 10 min to 20 min). The remainder of the conditions,

including cell straining, were the same as described previously.

HCT 116 cell capture by aptamer KCHA10

The low initial cell concentration (1 x106 cells/mL), longer cell administration time

(20 min) and cell straining technique were incorporated into the protocol described in

the flow rate optimization section for HCT 116 cell capture by KCHA10 aptamer.

HCT 116 cell capture in non-enzymatic buffer by aptamer KCHA10

Avidin and probe immobilization was the same as described in the flow rate

optimization section, but the protocol for cell administration was modified. Immediately

prior to introduction into the capillary, cells were concentrated to 2X (2x106 cells/mL) in

BB2 by centrifugation at 970 rpm for 3 min. The concentrated cells (500 pL) were then

strained through a 40 pm cell strainer and combined with 450 pL non-enzymatic buffer

and 100 pL FBS with continuous agitation throughout the experiment. Flow cytometry

of aptamer binding to the cells in non-enzymatic buffer was compared to the binding of

cells without non-enzymatic buffer in the same manner as the "DLD-1 cell capture using

aptamers with linker" section to confirm that aptamer binding was not affected.

Microscopy and Image Analysis

Accurate initial cell concentrations for cell lines were obtained before each

experiment by adding the initial cell solution to a hemocytometer and imaging the device


141









with an Olympus FV500-IX81 confocal microscope. The cells were counted by

importing the images into ImageJ and manually counting cells from each cell line (if

applicable). Then the cell concentration was calculated in accordance with the

manufacturer's instructions. The concentration of cells captured by the device was

calculated in a similar manner; however, nine image sets were obtained, corresponding

to all units of the hemocytometer grid. For the cell staining in blood experiments,

transmitted images, as well as those corresponding to green fluorescent target cells

were obtained for each square unit of the hemocytometer grid to calculate the target

and nontarget cell concentration in the flow through liquid. Once the images were

imported into ImageJ, stained target cells could be individually highlighted and overlaid

on the transmitted image in order to count the number of target and nontarget cells

passed through the device.

Results and Discussion

Optimization of Flow Rate

The flow rate at which the cell solution is administered plays a major role in the

efficiency of capturing cells. The flow rate is directly proportional to the linear velocity of

the fluid through the capillary, so lower linear velocities are the consequence of lower

flow rates. Cells in a fluid will have a higher probability of binding to the ligand at lower

velocities because the contact time of the cells and ligands is higher. Higher contact

times increase the number of bonds formed during the collision (B ; equation 1-11),

which results in a stronger interaction with the ligand.

Thus, several volumetric flow rates were compared to increase the capture

efficiency of the system in relation to the proof-of-principle studies where capture

efficiency was only ~5% in the best case. Figure 6-1 shows the increase in capture


142










efficiency from 0.36% at 500 nL/min flow rate, to 2.22% at intermediate flow rate, and

92.22% at the lowest flow rate. It is expected that the decreased velocity increases the

collision duration time, which, in turn, increases the number of bonds formed during the

collision.


100
90 A
80
70
o
S60
50
LU
40
30
L 20
10
0 AA
0 100 200 300 400 500 600
Flow Rate (nL/min)


Figure 6-1. Comparison of volumetric flow rate to capture efficiency of cells.

It is possible that lower flow rates may demonstrate higher capture efficiency, but

the total volume and time limitations render further decreasing the flow rate

unreasonable. For example, under similar experimental conditions (10 min flow) a 5

nL/min flow rate would only pass 50 nL of cell solution through the -450 nL capillary. It

would take 40 min to pass the same volume of cells into the capillary as a 10 min

administration of cells at 20 nL/min. Therefore, the 20 nL/min flow rate was

implemented into further studies. When the study using 20 nL/min flow rate was

repeated in triplicate, a reproducible capture efficiency of 91.1 3.5% was obtained.

Optimization of Buffer Wash Flow Rate

The next group of studies compared capture efficiency at two different washing

flow rates to determine whether this step played a significant role in cell capture of this


143










system. Cells administered at 20 nL/min flow rate were washed with 1500 nL buffer (~3

column volumes) at either 500 nL/min or 200 nL/min. Capture efficiency is compared in

Figure 6-2 as a function of captured cells divided by the total number of cells passed

through the device. The capture efficiencies were 84.5% for the low flow rate, and

83.2% for the high flow rate, demonstrating minimal difference. The higher flow rate

was utilized in future experiments because 1500 nL buffer is passed through the system

in a shorter amount of time.


100
90
80
70
S60
= 50
S40


20
10
0
200 500
Flow Rate (nL/min)


Figure 6-2. Comparison of capture efficiency at two different cell washing flow rates.
One capillary tested with a 200 nL/min washing flow rate was damaged
during the experiment, which is why only one value is reported.

Detection of CEM Cells in Blood

To determine whether CEM cells could be detected in blood, red blood cells were

removed from human whole blood and fluorescently-stained CEM cells were spiked into

the white blood cell/platelet suspension at concentrations of either 1.02x105 or 3.00x104

cells/mL. At the higher concentration, capture efficiency of CEM was 77.2%, while at

the lower concentration the efficiency dropped to 8.1% (Figure 6-3A). Additionally, the


144










percentage of cancer cells in the flow through fluid dropped to 2.1% from an initial CEM

concentration of 20.6% of the cell population at the higher CEM initial concentration.

This demonstrates enrichment of the system for the target cells.


A









0030
S60





[Cell] (x 10^5 cellsmrL.)







Figure 6-3. Capture of CEM cells spiked in blood. A) Graphical representation of cell
capture efficiency at different cell concentrations. B) Fluorescent images of
cells captured in the square capillary. Each image represents ~1.3 mm
capillary length.

CEM cells retained in the capillaries were easily differentiated from nonspecifically

bound blood cells due to the well-resolved fluorescent signal produced from the cell

stain imaged in the square capillaries (Figure 6-3B). Therefore, nonspecific binding was

essentially irrelevant since cancerous and normal cells could be distinguished by

monitoring fluorescent light. While this is clearly not practical for patient samples, a

labeling-after-capture method administering fluorescently-labeled aptamers to cells

captured on the capillary wall would have a similar effect. The CellSearch System relies


145











on a related concept, utilizing various stained antibodies to make a distinction between

cancerous and normal blood cells.18

Colon Cancer Aptamers (without Linker) for DLD-1 Capture

The function of the system was also analyzed for cancer cells not typically found in

high concentration in the blood. A variety of aptamers have previously been selected by

this group demonstrating binding to colorectal adenocarcinoma cancer cells (DLD-1 cell

line) and/or HCT 116 colorectal carcinoma cell line. Aptamer KDED2a-3 binds

specifically to DLD-1 cell line (Kd= 29.2 nM) with minimal affinity for HCT 116, while

aptamer KCHA10 binds to both DLD-1 and HCT 116 cell lines (Kd= 21.3 nM for HCT

116).203


100.00
90.00
80.00
70.00
U
S60.00
S50.00 -
LU
1 40.00
30.00
20.00
10.00
0.00
KDED2a-3 KCHA10
Aptamer


Figure 6-4. Comparison of capture efficiency of colon cancer cell line DLD-1 with
aptamers KDED2a-3 and KCHA10.

Both aptamers (biotinylated but lacking poly(T)o1 linkers) were assessed in terms

of their capture efficiency for DLD-1 cells. In initial studies, cells aggregated in large

clumps, and were not able to pass through the capillary in significant numbers.

Implementing a cell straining step followed by constant cell agitation helped to reduce,


146









but not entirely alleviate this concern. Figure 6-4 shows that aptamer KDED2a-3

demonstrated a higher capture efficiency and lower error than aptamer KCHA10. The

aptamers were then synthesized with poly(T)10 linkers to determine if the linker had an

effect on capture efficiency.

Colon Cancer Aptamers with poly(T)1o Linker for DLD-1 Capture

The two aptamers were each synthesized with a 3'-biotin-poly(T)io linker in order

to provide greater aptamer flexibility for binding according to previous studies.93 These

aptamers were tested for binding to DLD-1 cells by flow cytometry (Figure 6-5). The

results show that the aptamers with the linker still bind to the cells with a similar affinity

to those lacking the linker.


Cells
1

2a-3 (no T10)

l 2a-3 (with T10)








FL&-Helght

Figure 6-5. Flow cytometry comparison of aptamers binding to DLD-1 with and without
poly(T)10 linker.

The aptamers with linker were then immobilized in the capillary device to asses

their ability to capture DLD-1 cells. Once again, aptamer KDED2a-3 demonstrated

higher capture efficiency (83.6 5.8%) than KCHA10 with much lower error associated


147










with the measurement (Figure 6-6). The average of 2 measurements were higher for

KDED2a-3 aptamer with 3'-poly(T)o1 linker than without the linker, 83.2% versus 74.1%,

but the measurements fell within the error of each other. The capture efficiency was

relatively high, but the low absolute number of cells passing through the device

compared to the numbers expected indicates that cell aggregation may still be

occurring. The large error associated with capture by KCHA10 may be because the

KCHA10 aptamer was actually selected for HCT 116 instead of DLD-1. This aptamer

was also tested with HCT 116 cell line to determine if this trend continued.


100

90 T

80



a 60

S50

40

0 30

20

10

0
KDED2a-3 KCHA10
Aptamer


Figure 6-6. Capture efficiency of DLD-1 aptamers with 3'-poly(T)jo linker.

HCT 116 Cell Capture by Aptamer KCHA10 in Non-enzymatic Buffer

The initial cell concentration was reduced 5-fold and the cells were administered

for a longer amount of time to discourage cell clumping and increase the amounts of

cells passing through the device. However, immediately after cell straining, the HCT

116 cells aggregated, and minimal cells were captured in the capillary. Thus, the cells


148









were diluted in non-enzymatic buffer (NEB) directly before capillary introduction to

inhibit cell-cell interactions. The binding of KDED2a-3 and KCHA10 with HCT 116 with

non-enzymatic buffer or diluted with BB2 were compared in Figure 6-7. KDED2a-3 has

previously been shown to have a much lower affinity for HCT 116 cells than for DLD-1,

as demonstrated by comparison of fluorescence intensity of Figure 6-7 and Figure 6-5.

The binding of both aptamers to the cells was not affected by cell treatment with NEB,

since cells found in NEB had a similar shift as those diluted in BB2.



Cells in BB2


Lib in NEB
KDED2a-3 in BB2
KDED2a-3 in NEB
KCHA10 in BB2
KCHA10 in NEB




PIiHal~ht

Figure 6-7. Flow cytometric comparison of aptamer binding with HCT 116 cells diluted
in BB2 or with non-enzymatic buffer (NEB).

When the capillary system was tested with KCHA10/HCT 116 binding, three

measurements resulted in an average capture efficiency of 97.2 2.8% (Figure 6-8).

This capture efficiency was higher than DLD-1 values, and even higher than CEM

capture efficiency. Also, the error in KCHA10 binding to HCT 116 was much lower than

that of the probe binding DLD-1 cells. However, the number of cells passing through

the device compared to the number expected was minimal, as was the case with DLD-1

cells. This is likely due to aggregation effects of adherent cells clogging the capillary.


149










100
90
7o




S40





CLD- H-C--116
Cell Une


Figure 6-8. Overall capture efficiency of colon cancer cell lines DLD-1 and HCT 116 in
buffer.

Conclusions

The flow rate of the CEM cells through the capillary system was optimized to give

improved capture efficiency over the proof-of-principle studies. The capture efficiency in

the proof-of-principle studies was only -5% in the best case scenario, and after

optimization the capture efficiency of CEM cells was raised to 91.1 3.5%. The flow

rate of the buffer used to wash unbound cells from the capillary did not appear to

appreciably change the capture efficiency at the flow rates tested. A wider range of flow

rates tested may have some effect, although the capture efficiency is already

considerably high. The capture efficiency is comparable and in some cases improved

upon outcomes obtained for microfluidic devices reported in the literature, emphasizing

the utility of simple capillary systems.92-94

The device was also able to capture and detect leukemia cells in blood. Cells

were detected at both 1 x 105 and 3x1 04 cells/mL concentrations when red blood cells
o 7 0 ------------- _------


50 --------




1 0---- ---------------- _-------

1.D-1 [-CT116
Cell Une


Figure 6-8. Overall capture efficiency of colon cancer cell lines DLD-1 and HCT 116 in
buffer.

Conclusions

The flow rate of the CEM cells through the capillary system was optimized to give

improved capture efficiency over the proof-of-principle studies. The capture efficiency in

the proof-of-principle studies was only ~5% in the best case scenario, and after

optimization the capture efficiency of CEM cells was raised to 91.1 3.5%. The flow

rate of the buffer used to wash unbound cells from the capillary did not appear to

appreciably change the capture efficiency at the flow rates tested. A wider range of flow

rates tested may have some effect, although the capture efficiency is already

considerably high. The capture efficiency is comparable and in some cases improved

upon outcomes obtained for microfluidic devices reported in the literature, emphasizing

the utility of simple capillary systems.92-94

The device was also able to capture and detect leukemia cells in blood. Cells

were detected at both 1 x105 and 3x104 cells/mL concentrations when red blood cells


150









were removed prior to introduction. The CEM cells were stained and spiked into the

blood, then counted by scanning for fluorescence along the capillary window. A high-

impact study would be to try labeling-after-capture techniques to mimic more realistic

conditions. For example, CEM cells (unlabeled) would be spiked into the blood and

captured in the capillary as before, but fluorescently-labeled aptamers are then flowed

through the system, specifically binding to captured target cancer cells to facilitate

fluorescent-imaging detection.

Cancer cells which are not normally found in high concentration in the blood were

also efficiently captured by the capillary system, even though the affinities for each cell

line/aptamer pair were not as high as the CEM/sgc8 system. Two colon cancer cell

lines, DLD-1 and HCT 116 were captured by aptamers previously selected by this

group. DLD-1 cells were captured by aptamer KDED2a-3 at efficiencies of 83.6 5.8%,

and aptamer KCHA10 demonstrated capture efficiencies (HCT 116 cell line) of 97.2

2.8% in buffer.

Despite the success of the capillaries in capturing cancer cells, the drawbacks

must be considered for a complete assessment of capillary potential. The major

concern with a capillary will always be throughput. Capillaries generally have microliter

volumes, so when low flow rates are used for cell capture only a very small sample

volume is examined. This property limits samples to those in which target cells are

found in high concentration, or in which a preconcentration step is performed. This is

not a step unique to capillaries, as even the commercialized CellSearch Assay requires

sample pretreatment.18 Capillaries can also be multiplexed to increase throughput and


151









introduce diverse capture ligands to the sample.95 Another challenge is the relative

ease in breaking the tubes once the coating is removed for imaging purposes.

One problem specific to the colon cancer studies was that much lower amounts of

cells were passed through the device than would be expected based on initial cell

concentration and introduction time. Two possible explanations for this phenomenon

both result in clogging of the capillary; cell aggregation and increased cell diameter.

Adherent cells are known to interact with one another to form clumps containing many

cells.204 This effect was observed by imaging the cell solution prior to introduction into

the capillary. A cell straining step preceding cell introduction increased the numbers of

cells passing through the capillary, but the numbers still were not as high as would be

expected based on the initial cell concentration (~15%). The CEM cells were passed

through the device in much higher numbers, sometimes greater than 100%, which

suggests this is a problem with DLD-1 and HCT 116 cell integration, not the system

itself. Even though the cells are strained, upon constriction in the reduced diameter of

the capillary, the cells are forced into closer contact with one another, possibly resulting

in aggregation inside the device. Also, the DLD-1 cells are visually much larger than the

~9 pm diameter of CEM cells, and actually appear larger than the HCT 116 cells, which

have a reported average diameter of 12 pm.205 206 No concrete measurement of DLD-1

cell diameter is available, but if the dimensions are ~20 pm, only 4 cells in width could

block one dimension of the capillary tube. To ease this concern, square or rectangular

capillaries of larger dimensions could be implemented (standard square tubing is

available from PolyMicro Technologies up to 200 pm x 200 pm dimensions).

Rectangular tubes are particularly attractive because capillaries only slightly larger than


152









the cells in height would increase the probability of the cell contacting the immobilized

ligands, yet the increased width would prevent cell clogging.

In summary, the aptamer-based capillary device was able to capture leukemia and

colon cancer cells in buffer in numbers comparable to literature values for standard

microfluidic devices, with a system much simpler in design, fabrication, and function.

The conditions optimized for a high capture efficiency of CEM cells were also applicable

to two different colon cancer cell lines. In addition, the capillary system could detect

stained leukemia cells which were spiked into human blood by scanning the length of

the capillary for fluorescent cells. This aptamer-immobilized device has demonstrated

many attributes which may promote its use as a technique for efficient cancer cell

detection.


153









CHAPTER 7
SUMMARY AND FUTURE WORK

Summary

Aptamers have proven versatile in application since the original successes of the

concept in 1990. A multitude of analytical and therapeutic uses of aptamers have been

reported, and the number of aptamer-based drugs progressing through clinical trials

implies this trend will continue. In this dissertation, novel methods were combined to

produce the first known aptamer antidote for a currently-existing pharmaceutical drug,

and an aptamer-immobilized capillary provided a simple yet highly efficient means for

selectively detecting cancer cells.

In the drug-SELEX project, the peptide anticoagulant bivalirudin was immobilized

on a monolithic disk stationary phase, and binding sequences were collecting using a

high salt gradient in conjunction with LPC. Fractions of eluted DNA were collected, and

the amount of sequences eluted from the drug-immobilized disk was compared to that

of a blank disk via qPCR. The basic assumption was that more DNA binding to the

drug-containing disk would indicate that at least some fraction of the sequences were

specifically binding to the drug instead of solely to the stationary phase matrix. After

only 2 rounds, an AlphaScreen assay was performed to select the highest-affinity pool

for sequencing. Six of the aptamer candidates tested demonstrated binding to the

target, and 2 of the sequences characterized by FA generated Kd values in the low

micromolar range, while control sequences displayed Kd values at least 20-fold higher.

The antidote potential of aptamer JPB5 was evaluated, establishing a dose-dependent

decrease in anticoagulant activity, with a concentration of 20 pM effectively reversing

the anticoagulant effect. A DNA microarray was developed to truncate the sequence to


154









<40-mer while still retaining binding activity, resulting in abbreviated JPB5 31-mer and

JPB2 21-mer candidates. The conclusion from this project is that DNA aptamers can

serve as therapeutic antidotes to pharmaceutical drugs.

The sgc8 aptamer specifically binding to CEM leukemia cells was evaluated in

proof-of-principle studies to act as a capillary affinity chromatography stationary phase

ligand. The amount of aptamer immobilized on the inner wall of the square capillary

could be controlled by varying the concentration of sgc8 administered. Target cells

were captured in higher amounts than nontarget Ramos cells even when the cell lines

were mixed prior to introduction into the capillary. Additionally, square capillaries

provided imaging properties superior to standard circular diameter capillaries, an

important characteristic for detecting cancer cells. When the flow rate of cell

introduction was optimized, the capture efficiency of the system was raised from ~5% to

>90%, enabling detection of CEM cells in human blood. Two colon cancer cell lines and

aptamers were incorporated into the system using the optimized CEM flow rates, and

the colon cancer cells also demonstrated high capture efficiency. This system was

much simpler in design, fabrication, and application than standard microfluidic devices,

yet provided similar capture efficiencies. Thus, aptamers can function as a highly-

specific stationary phase for cancer cell detection.

Future Work

Aptamer Antidote Project

The Kd values for this experiment were obtained by labeling the target and

measuring the anisotropy as increasing amounts of aptamer was titrated into the

solution. However, label-free detection methods are highly desirable for this purpose

because labeling aptamer or target may change the binding properties of the system.


155









Therefore, comparing the FA Kd values to those obtained from a label-free method such

as ITC may have great significance. Ideally, the affinities and antidote activity of all

aptamers would be reported, and crystallography studies of the aptamer/bivalirudin

structures would be beneficial as well.

One effect that was observed but not discussed is related to the effect of Ca2+ on

aptamer binding. When Kd studies were carried out in buffer that contained calcium, the

anisotropy demonstrated a time-dependent increase followed by a plateau. We

hypothesize that the time-dependence is related to the higher affinity of Ca2+ for DNA

than Mg2+.207 The Mg2+ competes with the Ca2+ over time, resulting in a more flexible

DNA structure capable of increased target binding.208 Studying this effect could provide

important insights into the role of metal ion binding for the aptamers studied. However,

it is important to note that experiments in vivo may not show this effect due to the

complex internal environment balancing the process more than the closed buffer

system.

Ongoing work includes study of the truncated sequences for antidote activity using

the buffer screening assay. Also of interest is PEGylating the truncated and full-length

aptamers to determine whether this can decrease nuclease degradation in both plasma

and in vivo models. Incorporating modified bases such as LNA into the aptamer

structure could be carried out with the same goal. Currently, this project is in the

approval stage for small-scale in vivo animal studies. The results of these experiments

will be a main factor in determining the ultimate future of the selected aptamers.

Cell Affinity Project

Interesting studies for this project would be to examine the effect of different

capillary internal diameters on the capture efficiency of the system. Increasing the


156









cross-sectional area of the tubes will decrease the effective linear velocity of the cells

traveling through the system. This will likely result in higher capture efficiencies for

larger diameter tubes, and may also allow for higher volumetric flow rates implemented

into the system (increased throughput). As discussed earlier, experiments performed in

rectangular tubes may be beneficial for increasing capture efficiency and decreasing

capillary clogging of adherent cell lines by increasing cell/ligand interactions in one

dimension yet maintaining an increased width for cell passage.

Other valuable studies would be to compare the capture efficiencies obtained with

a pulsed flow to those of continuous flow. In theory, pulsed flow would allow cell

receptors an increased contact time with the aptamers, increasing the amount of

cell/ligand bonds formed. Higher throughput may be possible using this technique since

increased flow rates could be implemented to fill the capillary at each pulse.

A pretreatment step using magnetic EpCAM-coated beads (similar to the

CellSearch pretreatment)18 compared to the Ficoll-Paque enrichment step for whole

blood samples would be a study of interest. Also of importance is study of a technique

to discriminate between cancerous and noncancerous cells captured from blood. The

method in this work involved prestaining cancer cells and spiking them into blood

samples, clearly not applicable to real patient samples. One possible method to

generate a similar effect would be to implement a label-after-capture step in which

fluorescently-labeled aptamers are introduced to selectively bind cancer cells captured

on the capillary walls. The length of the capillary could be imaged to count only the

fluorescent cells, effectively eliminating concerns of nonspecific binding.


157









LIST OF REFERENCES

1. B. N. Dominy, Curr. Pharm. Biotechnol., 2008, 9, 87-95.

2. T. Schoneberg, M. Hofreiter, A. Schulz and H. Rompler, Trends Pharmacol. Sci.,
2007, 28, 117-121.

3. S. Henikoff and K. Ahmad, Annu. Rev. Cell Dev. Biol., 2005, 21, 133-153.

4. E. A. MacGregor, S. Janecek and B. Svensson, Biochimica Et Biophysica Acta-
Protein Structure and Molecular Enzymology, 2001, 1546, 1-20.

5. A. Serganov, A. Polonskaia, A. T. Phan, R. R. Breaker and D. J. Patel, Nature,
2006, 441, 1167-1171.

6. N. Hirokawa, Science, 1998, 279, 519-526.

7. P. Stenlund, A. Frostell-Karlsson and O. P. Karlsson, Anal. Biochem., 2006, 353,
217-225.

8. H. Nordin, M. Jungnelius, R. Karlsson and O. P. Karlsson, Anal. Biochem., 2005,
340, 359-368.

9. Y. D. Ivanov, I. P. Kanaeva, Karuzina, II, A. I. Archakov, G. H. B. Hoa and S. G.
Sligar, Arch. Biochem. Biophys., 2001, 391, 255-264.

10. F. M. Huennekens, Advances in Enzyme Regulation, Vol 34, 1994, 34, 397-419.

11. F. R. Cockerill and R. S. Edson, Mayo Clinic Proceedings, 1991, 66, 1260-1269.

12. D. Fourmy, M. I. Recht, S. C. Blanchard and J. D. Puglisi, Science, 1996, 274,
1367-1371.

13. A. Sin, S. K. Murthy, A. Revzin, R. G. Tompkins and M. Toner, Biotechnol.
Bioeng., 2005, 91, 816-826.

14. C. L. Sawyers, Nature, 2008, 452, 548-552.

15. A. Beck, T. Wurch, C. Bailly and N. Corvaia, Nat. Rev. Immunol., 2010, 10, 345-
352.

16. J. Wojciechowski, D. Danley, J. Cooper, N. Yazvenko and C. R. Taitt, Sensors,
2010, 10, 3351-3362.

17. Y. Chen, S. Tang, S. Ding, F. He and X. Xiao, Anal. Lett., 2009, 42, 2793-2806.

18. M. Naoe, Y. Ogawa, J. Morita, K. Omori, K. Takeshita, T. Shichijyo, T. Okumura,
A. Igarashi, A. Yanaihara, S. Iwamoto, T. Fukagai, A. Miyazaki and H. Yoshida,
Cancer, 2007, 109, 1439-1445.


158









19. D. R. Burton, Nat. Rev. Immunol., 2002, 2, 706-713.

20. 0. Leavy, Nat. Rev. Immunol., 2010, 10, 297-297.

21. J. M. Reichert, MAbs, 2010 2, 84-100.

22. R. Stoltenburg, C. Reinemann and B. Strehlitz, Biomol. Eng., 2007, 24, 381-403.

23. S. M. Nimjee, C. P. Rusconi and B. A. Sullenger, Annu. Rev. Med., 2005, 56, 555-
583.

24. S. D. Jayasena, Clinical chemistry, 1999, 45, 1628-1650.

25. J. R. de Oliveira, F. B. Nunes, M. G. Simoes Pires, D. A. da Silva Melo and M. V.
Fagundes Donadio, Anti-Infect. Agents Med. Chem., 2008, 7, 249-257.

26. T. Sampson, World Patent Information, 2003, 25, 123-129.

27. E. Luzi, M. Minunni, S. Tombelli and M. Mascini, Trac-Trends in Analytical
Chemistry, 2003, 22, 810-818.

28. D. Shangguan, Y. Li, Z. W. Tang, Z. H. C. Cao, H. W. Chen, P. Mallikaratchy, K.
Sefah, C. Y. J. Yang and W. H. Tan, Proc. Natl. Acad. Sci. U. S. A., 2006, 103,
11838-11843.

29. A. D. Ellington and J. W. Szostak, Nature, 1990, 346, 818-822.

30. C. Tuerk and L. Gold, Science, 1990, 249, 505-510.

31. J. Ciesiolka, J. Gorski and M. Yarus, Rna-a Publication of the Rna Society, 1995,
1,538-550.

32. H. P. Hofmann, S. Limmer, V. Hornung and M. Sprinzl, Rna-a Publication of the
Rna Society, 1997, 3, 1289-1300.

33. D. Mann, C. Reinemann, R. Stoltenburg and B. Strehlitz, Biochem. Biophys. Res.
Commun., 2005, 338, 1928-1934.

34. D. Grate and C. Wilson, Bioorg. Med. Chem., 2001, 9, 2565-2570.

35. A. D. Ellington and J. W. Szostak, Nature, 1992, 355, 850-852.

36. J. J. Tang, J. W. Xie, N. S. Shao and Y. Yan, Electrophoresis, 2006, 27, 1303-
1311.

37. C. Mannironi, A. DiNardo, P. Fruscoloni and G. P. Tocchini-Valentini,
Biochemistry, 1997, 36, 9726-9734.


159









38. T. Kato, T. Takemura, K. Yano, K. Ikebukuro and I. Karube, Biochimica Et
Biophysica Acta-Gene Structure and Expression, 2000, 1493, 12-18.

39. M. Meli, J. Vergne, J. L. Decout and M. C. Maurel, J. Biol. Chem., 2002, 277,
2104-2111.

40. M. Sassanfar and J. W. Szostak, Nature, 1993, 364, 550-553.

41. D. Kiga, Y. Futamura, K. Sakamoto and S. Yokoyama, Nucleic Acids Res., 1998,
26, 1755-1760.

42. C. T. Lauhon and J. W. Szostak, J. Am. Chem. Soc., 1995, 117, 1246-1257.

43. P. Burgstaller and M. Famulok, Angewandte Chemie-lnternational Edition in
English, 1994, 33, 1084-1087.

44. C. Wilson, J. Nix and J. Szostak, Biochemistry, 1998, 37, 14410-14419.

45. J. H. Ko, B. Cho, J. K. Ahn, Y. Lee and I. Park, Bull. Korean Chem. Soc., 1999, 20,
1335-1339.

46. C. Boiziau, E. Dausse, L. Yurchenko and J. J. Toulme, J. Biol. Chem., 1999, 274,
12730-12737.

47. A. Geiger, P. Burgstaller, H. vonderEltz, A. Roeder and M. Famulok, Nucleic Acids
Res., 1996, 24, 1029-1036.

48. M. Famulok and J. W. Szostak, J. Am. Chem. Soc., 1992, 114, 3990-3991.

49. I. Majerfeld, D. Puthenvedu and M. Yarus, J. Mol. Evol., 2005, 61, 226-235.

50. E. Fukusaki, T. Kato, H. Maeda, N. Kawazoe, Y. Ito, A. Okazawa, S. Kajiyama and
A. Kobayashi, Bioorg. Med. Chem. Lett., 2000, 10, 423-425.

51. C. Srisawat, I. J. Goldstein and D. R. Engelke, Nucleic acids research, 2001, 29,
1-5.

52. S. M. Lato, A. R. Boles and A. D. Ellington, Chem. Biol., 1995, 2, 291-303.

53. S. T. Wallace and R. Schroeder, Rna-a Publication of the Rna Society, 1998, 4,
112-123.

54. C. Berens, A. Thain and R. Schroeder, Bioorg. Med. Chem., 2001, 9, 2549-2556.

55. D. Nieuwlandt, M. Wecker and L. Gold, Biochemistry, 1995, 34, 5651-5659.

56. F. Ylera, R. Lurz, V. A. Erdmann and J. P. Furste, Biochemical and biophysical
research communications, 2002, 290, 1583-1588.


160









57. R. White, C. Rusconi, E. Scardino, A. Wolberg, J. Lawson, M. Hoffman and B.
Sullenger, Mol. Ther., 2001, 4, 567-573.

58. A. Tahiri-Alaoui, L. Frigotto, N. Manville, J. Ibrahim, P. Romby and W. James,
Nucleic Acids Res., 2002, 30.

59. T. W. Wiegand, P. B. Williams, S. C. Dreskin, M. H. Jouvin, J. P. Kinet and D.
Tasset, J. Immunol., 1996, 157, 221-230.

60. L. S. Green, D. Jellinek, R. Jenison, A. Ostman, C. H. Heldin and N. Janjic,
Biochemistry, 1996, 35, 14413-14424.

61. D. Jellinek, L. S. Green, C. Bell and N. Janjic, Biochemistry, 1994, 33, 10450-
10456.

62. D. A. Daniels, H. Chen, B. J. Hicke, K. M. Swiderek and L. Gold, Proc. Natl. Acad.
Sci. U. S. A., 2003, 100, 15416-15421.

63. Y. J. Lee and S. W. Lee, Journal of microbiology and biotechnology, 2006, 16,
1149-1153.

64. Z. W. Tang, D. Shangguan, K. M. Wang, H. Shi, K. Sefah, P. Mallikratchy, H. W.
Chen, Y. Li and W. H. Tan, Anal. Chem., 2007, 79, 4900-4907.

65. H. W. Chen, C. D. Medley, K. Sefah, D. Shangguan, Z. W. Tang, L. Meng, J. E.
Smith and W. H. Tan, ChemMedChem, 2008, 3, 991-1001.

66. K. Sefah, Z. W. Tang, D. H. Shangguan, H. Chen, D. Lopez-Colon, Y. Li, P.
Parekh, J. Martin, L. Meng, J. A. Phillips, Y. M. Kim and W. H. Tan, Leukemia,
2009, 23, 235-244.

67. D. H. Shangguan, L. Meng, Z. H. C. Cao, Z. Y. Xiao, X. H. Fang, Y. Li, D.
Cardona, R. P. Witek, C. Liu and W. H. Tan, Anal. Chem., 2008, 80, 721-728.

68. Z. Tang, P. Parekh, P. Turner, R. W. Moyer and W. Tan, Clinical chemistry, 2009,
55, 813-822.

69. W. H. Pan, R. C. Craven, Q. Qiu, C. B. Wilson, J. W. Wills, S. Golovine and J. F.
Wang, Proc. Natl. Acad. Sci. U. S. A., 1995, 92, 11509-11513.

70. J. G. Bruno and J. L. Kiel, Biosens. Bioelectron., 1999, 14, 457-464.

71. C. L. A. Hamula, H. Q. Zhang, L. L. Guan, X. F. Li and X. C. Le, Anal. Chem.,
2008, 80, 7812-7819.

72. S. Tombelli, A. Minunni and A. Mascini, Biosens. Bioelectron., 2005, 20, 2424-
2434.


161









73. L. S. Green, D. Jellinek, C. Bell, L. A. Beebe, B. D. Feistner, S. C. Gill, F. M.
Jucker and N. Janjic, Chem. Biol., 1995, 2, 683-695.

74. A. Rhodes, A. Deakin, J. Spaull, B. Coomber, A. Aitken, P. Life and S. Rees, J.
Biol. Chem., 2000, 275, 28555-28561.

75. J. Ruckman, L. S. Green, J. Beeson, S. Waugh, W. L. Gillette, D. D. Henninger, L.
Claesson-Welsh and N. Janjic, J. Biol. Chem., 1998, 273, 20556-20567.

76. M. Petersen and J. Wengel, Trends Biotechnol., 2003, 21, 74-81.

77. S. C. B. Gopinath, Anal. Bioanal. Chem., 2007, 387, 171-182.

78. C. Ravelet, C. Grosset and E. Peyrin, Journal of Chromatography A, 2006, 1117,
1-10.

79. T. S. Romig, C. Bell and D. W. Drolet, Journal of Chromatography B-Analytical
Technologies in the Biomedical and Life Sciences, 1999, 731, 275-284.

80. Q. Deng, I. German, D. Buchanan and R. T. Kennedy, Anal. Chem., 2001, 73,
5415-5421.

81. M. A. Rehder-Silinski and L. B. McGown, Journal of Chromatography A, 2003,
1008, 233-245.

82. J. Ruta, C. Ravelet, I. Baussanne, J. L. Decout and E. Peyrin, Anal. Chem., 2007,
79, 4716-4719.

83. M. Michaud, E. Jourdan, A. Villet, A. Ravel, C. Grosset and E. Peyrin, J. Am.
Chem. Soc., 2003, 125, 8672-8679.

84. M. Liss, B. Petersen, H. Wolf and E. Prohaska, Anal. Chem., 2002, 74, 4488-4495.

85. M. Lee and D. R. Walt, Anal. Biochem., 2000, 282, 142-146.

86. T. G. McCauley, N. Hamaguchi and M. Stanton, Anal. Biochem., 2003, 319, 244-
250.

87. R. Kirby, E. J. Cho, B. Gehrke, T. Bayer, Y. S. Park, D. P. Neikirk, J. T. McDevitt
and A. D. Ellington, Anal. Chem., 2004, 76, 4066-4075.

88. C. A. Savran, S. M. Knudsen, A. D. Ellington and S. R. Manalis, Anal. Chem.,
2004, 76, 3194-3198.

89. S. D. Jhaveri, R. Kirby, R. Conrad, E. J. Maglott, M. Bowser, R. T. Kennedy, G.
Glick and A. D. Ellington, J. Am. Chem. Soc., 2000, 122, 2469-2473.

90. R. Yamamoto and P. K. R. Kumar, Genes Cells, 2000, 5, 389-396.


162









91. K. Wang, M. K. Marshall, G. Garza and D. Pappas, Anal. Chem., 2008, 80, 2118-
2124.

92. Y. Xu, J. A. Phillips, J. Yan, Q. Li, Z. H. Fan and W. Tan, Analytical chemistry,
2009, 81, 7436-7442.

93. J. A. Phillips, Y. Xu, Z. Xia, Z. H. Fan and W. Tan, Analytical chemistry, 2009, 81,
1033-1039.

94. U. Dharmasiri, S. Balamurugan, A. A. Adams, P. I. Okagbare, A. Obubuafo and S.
A. Soper, Electrophoresis, 2009, 30, 3289-3300.

95. D. Pappas and K. Wang, Analytica chimica acta, 2007, 601, 26-35.

96. M. Blank, T. Weinschenk, M. Priemer and H. Schluesener, J. Biol. Chem., 2001,
276, 16464-16468.

97. S. Fredriksson, M. Gullberg, J. Jarvius, C. Olsson, K. Pietras, S. M. Gustafsdottir,
A. Ostman and U. Landegren, Nat. Biotechnol., 2002, 20, 473-477.

98. L. S. Green, C. Bell and N. Janjic, BioTechniques, 2001, 30, 1094-1110.

99. C. D. Pieve, A. C. Perkins and S. Missailidis, Nucl. Med. Biol., 2009, 36, 703-710.

100. J. Charlton, J. Sennello and D. Smith, Chem. Biol., 1997, 4, 809-816.

101. S. M. Taghdisi, K. Abnous, F. Mosaffa and J. Behravan, J. Drug Targeting, 2010,
18, 277-281.

102. Y. Wu, K. Sefah, H. Liu, R. Wang and W. Tan, Proceedings of the National
Academy of Sciences of the United States of America, 2010, 107, 5-10.

103. G. J. Tong, S. C. Hsiao, Z. M. Carrico and M. B. Francis, J. Am. Chem. Soc.,
2009, 131, 11174-11178.

104. D. Proske, M. Blank, R. Buhmann and A. Resch, Appl. Microbiol. Biotechnol.,
2005, 69, 367-374.

105. K. W. Thiel and P. H. Giangrande, Oligonucleotides, 2009, 19, 209-222.

106. B. A. Sullenger, H. F. Gallardo, G. E. Ungers and E. Gilboa, Cell, 1990, 63, 601-
608.

107. C. M. Browning, L. Cagnon, P. D. Good, J. Rossi, D. R. Engelke and D. M.
Markovitz, J. Virol., 1999, 73, 5191-5195.

108. C. P. Rusconi, J. D. Roberts, G. A. Pitoc, S. M. Nimjee, R. R. White, G. Quick, E.
Scardino, W. P. Fay and B. A. Sullenger, Nat. Biotechnol., 2004, 22, 1423-1428.


163









109. S. Santulli-Marotto, S. K. Nair, C. Rusconi, B. Sullenger and E. Gilboa, Cancer
Res., 2003, 63, 7483-7489.

110. Y. Kim, Z. Cao and W. Tan, Proc. Natl. Acad. Sci. U. S. A., 2008, 105, 5664-5669.

111. P. J. Bates, J. B. Kahlon, S. D. Thomas, J. O. Trent and D. M. Miller, J. Biol.
Chem., 1999, 274, 26369-26377.

112. B. J. Hicke, C. Marion, Y. F. Chang, T. Gould, C. K. Lynott, D. Parma, P. G.
Schmidt and S. Warren, J. Biol. Chem., 2001, 276, 48644-48654.

113. L. C. Bock, L. C. Griffin, J. A. Latham, E. H. Vermaas and J. J. Toole, Nature,
1992, 355, 564-566.

114. C. P. Rusconi, E. Scardino, J. Layzer, G. A. Pitoc, T. L. Ortel, D. Monroe and B. A.
Sullenger, Nature, 2002, 419, 90-94.

115. C. P. Rusconi, A. Yeh, H. K. Lyerly, J. H. Lawson and B. A. Sullenger, Thromb.
Haemostasis, 2000, 84, 841-848.

116. S. K. Buddai, J. M. Layzer, G. Lu, C. P. Rusconi, B. A. Sullenger, D. M. Monroe
and S. Krishnaswamy, J. Biol. Chem., 2010, 285, 5212-5223.

117. S. W. Gal, S. Amontov, P. T. Urvil, D. Vishnuvardhan, F. Nishikawa, P. K. R.
Kumar and S. Nishikawa, Eur. J. Biochem., 1998, 252, 553-562.

118. S. Oney, S. M. Nimjee, J. Layzer, N. Que-Gewirth, D. Ginsburg, R. C. Becker, G.
Arepally and B. A. Sullenger, Oligonucleotides, 2007, 17, 265-274.

119. S. Oney, R. T. S. Lam, K. M. Bompiani, C. M. Blake, G. Quick, J. D. Heidel, J. Y.
C. Liu, B. C. Mack, M. E. Davis, K. W. Leong and B. A. Sullenger, Nat. Med. (N.
Y., NY, U. S.), 2009, 15, 1224-1228.

120. A. Heckel, M. C. R. Buff, M. S. L. Raddatz, J. Muller, B. Potzsch and G. Mayer,
Angewandte Chemie-lnternational Edition, 2006, 45, 6748-6750.

121. Y. Kim, J. A. Phillips, H. Liu, H. Kang and W. Tan, Proceedings of the National
Academy of Sciences of the United States of America, 2009, 106, 6489-6494.

122. A. Joachimi, G. Mayer and J. S. Hartig, J. Am. Chem. Soc., 2007, 129, 3036-7.

123. J. Fanikos, J. L. Cina, S. Baroletti, K. Fiumara, L. Matta and S. Z. Goldhaber,
American Journal of Cardiology, 2007, 100, 1465-1469.

124. D. Monroe, M. Hoffman and H. R. Roberts, Arteriosclerosis Thrombosis and
Vascular Biology, 2002, 22, 1381-1389.

125. T. E. Warkentin, Best Practice & Research Clinical Haematology, 2004, 17, 105-
125.


164









126. J. I. Weitz and M. Crowther, Thromb. Res., 2002, 106, V275-V284.

127. E. Di Cera, Chest, 2003, 124, 11S-17S.

128. W. Bode, Blood Cells Molecules and Diseases, 2006, 36, 122-130.

129. M. S. Conners III and S. R. Money, The Ochsner Journal, 2002, 4, 41-47.

130. P. C. A. Kam, N. Kaur and C. L. Thong, Anaesthesia, 2005, 60, 565-574.

131. T. M. Sciulli and V. F. Mauro, Annals of Pharmacotherapy, 2002, 36, 1028-1041.

132. M. Salzet, Curr. Pharm. Des., 2002, 8, 493-503.

133. L. A. Linkins and J. I. Weitz, Curr. Pharm. Des., 2005, 11, 3877-3884.

134. N. W. Shammas, Cardiovascular Drug Reviews, 2005, 23, 345-360.

135. M. A. A. Parry, J. M. Maraganore and S. R. Stone, Biochemistry, 1994, 33, 14807-
14814.

136. J. I. Witting, P. Bourdon, D. V. Brezniak, J. M. Maraganore and J. W. Fenton,
Biochem. J., 1992, 283, 737-743.

137. J. M. Maraganore, P. Bourdon, J. Jablonski, K. L. Ramachandran and J. W.
Fenton, Biochemistry, 1990, 29, 7095-7101.

138. R. De Cristofaro and E. De Candia, Journal of Thrombosis and Thrombolysis,
2003, 15, 151-163.

139. E. Skrzypczak-Jankun, V. E. Carperos, K. G. Ravichandran, A. Tulinsky, M.
Westbrook and J. M. Maraganore, J. Mol. Biol., 1991, 221, 1379-1393.

140. W. Bode, I. Mayr, U. Baumann, R. Huber, S. R. Stone and J. Hofsteenge, EMBO
J., 1989, 8, 3467-3475.

141. D. Mark, S. Haeberle, G. Roth, F. von Stetten and R. Zengerle, Chem. Soc. Rev.,
2010, 39, 1153-1182.

142. J. M. Hicks and M. losefsohn, N. Engl. J. Med., 1989, 320, 320-321.

143. R. Pacific, M. Farre, S. Pichini, J. Ortuno, P. N. Roset, P. Zuccaro, J. Segura and
R. de la Torre, J. Anal. Toxicol., 2001, 25, 144-146.

144. J. C. Stachowiak, E. E. Shugard, B. P. Mosier, R. F. Renzi, P. F. Caton, S. M.
Ferko, J. L. V. de Vreugde, D. D. Yee, B. L. Haroldsen and V. A. VanderNoot,
Anal. Chem., 2007, 79, 5763-5770.


165









145. R. J. Meagher, A. V. Hatch, R. F. Renzi and A. K. Singh, Lab Chip, 2008, 8, 2046-
2053.

146. L. Zhao, T. Wu, J.-P. Lefevre, I. Leray and J. A. Delaire, Lab Chip, 2009, 9, 2818-
2923.

147. C. Zhai, W. Qiang, J. Sheng, J. Lei and H. Ju, Journal of Chromatography A,
1217, 785-789.

148. E. Leclerc, K. S. Furukawa, F. Miyata, Y. Sakai, T. Ushida and T. Fujii,
Biomaterials, 2004, 25, 4683-4690.

149. A. M. Elizarov, Lab Chip, 2009, 9, 1326-1333.

150. S. Upadhyaya and P. R. Selvaganapathy, Lab Chip, 2010, 10, 341-348.

151. A. A. Adams, P. I. Okagbare, J. Feng, M. L. Hupert, D. Patterson, J. Gottert, R. L.
McCarley, D. Nikitopoulos, M. C. Murphy and S. A. Soper, J. Am. Chem. Soc.,
2008, 130, 8633-8641.

152. S. Nagrath, L. V. Sequist, S. Maheswaran, D. W. Bell, D. Irimia, L. Ulkus, M. R.
Smith, E. L. Kwak, S. Digumarthy, A. Muzikansky, P. Ryan, U. J. Balis, R. G.
Tompkins, D. A. Haber and M. Toner, Nature, 2007, 450, 1235-1239.

153. K. Pantel, R. H. Brackenhoff and B. Brandt, Nat. Rev. Cancer, 2008, 8, 329-340.

154. K. A. Herold, Rasooly, A., Lab-on-a-Chip Technology, Caister Academic Press,
Norfolk, UK, 2009.

155. T. Tsuda, J. V. Sweedler and R. N. Zare, Anal. Chem., 1990, 62, 2149-2152.

156. A. Cifuentes and H. Poppe, Chromatographia, 1994, 39, 391-404.

157. Polym icro-Technologies, The Book on the Technologies of Polymicro, Polym icro
Technologies, LLC, 2004.

158. M. C. Peoples and H. T. Karnes, Journal of Chromatography B-Analytical
Technologies in the Biomedical and Life Sciences, 2008, 866, 14-25.

159. Z. Guo, R. A. Guilfoyle, A. J. Thiel, R. F. Wang and L. M. Smith, Nucleic Acids
Res., 1994, 22, 5456-5465.

160. A. Bogusiewicz, N. I. Mock and D. M. Mock, Anal. Biochem., 2004, 327, 156-161.

161. I. N. Levine, Physical Chemistry, 5 edn., McGraw-Hill, New York, 2002.

162. D. A. Hammer and D. A. Lauffenburger, Biophys. J., 1987, 52, 475-487.


166









163. L. A. Olivier, J. Yen, W. M. Reichert and G. A. Truskey, Biotechnol. Prog., 1999,
15, 33-42.

164. G. I. Bell, Science, 1978, 200, 618-627.

165. C. M. Hertz, D. J. Graves, D. A. Lauffenburger and F. T. Serota, Biotechnol.
Bioeng., 1985, 27, 603-612.

166. S. Barkley, H. Johnson, R. Eisenthal and J. Hubble, Biotechnol. Appl. Biochem.,
2004, 40, 145-149.

167. M. F. Caron and G. R. McKendall, American Journal of Health-System Pharmacy,
2003, 60, 1841-1849.

168. S. C. B. Gopinath, Thromb. Res., 2008, 122, 838-847.

169. T. D. Gladwell, Clinical Therapeutics, 2002, 24, 38-58.

170. M. Wilchek, E. A. Bayer and 0. Livnah, Immunol. Lett., 2006, 103, 27-32.

171. M. Ellingham, D. H. J. Bunka, D. J. Rowlands and N. J. Stonehouse, Rna-a
Publication of the Rna Society, 2006, 12, 1970-1979.

172. M. Djordjevic, Biomol. Eng., 2007, 24, 179-189.

173. M. Rieseberg, C. Kasper, K. F. Reardon and T. Scheper, Appl. Microbiol.
Biotechnol., 2001, 56, 350-360.

174. Z. W. Tang, P. Parekh, P. Turner, R. W. Moyer and W. H. Tan, Clinical chemistry,
2009, 55, 813-822.

175. M. Howarth, D. J. F. Chinnapen, K. Gerrow, P. C. Dorrestein, M. R. Grandy, N. L.
Kelleher, A. EI-Husseini and A. Y. Ting, Nat. Methods, 2006, 3, 267-273.

176. D. E. Huizenga and J. W. Szostak, Biochemistry, 1995, 34, 656-665.

177. W. Z. Niu, N. Jiang and Y. H. Hu, Anal. Biochem., 2007, 362, 126-135.

178. N. Ishizuka, H. Kobayashi, H. Minakuchi, K. Nakanishi, K. Hirao, K. Hosoya, T.
Ikegami and N. Tanaka, Journal of Chromatography A, 2002, 960, 85-96.

179. M. W. Roberts, C. M. Ongkudon, G. M. Forde and M. K. Danquah, J. Sep. Sci.,
2009, 32, 2485-2494.

180. E. G. Vlakh, G. A. Platonova, G. P. Vlasov, C. Kasper, A. Tappe, G. Kretzmer and
T. B. Tennikova, Journal of Chromatography A, 2003, 992, 109-119.

181. E. G. Vlakh, A. Tappe, C. Kasper and T. B. Tennikova, Journal of Chromatography
B-Analytical Technologies in the Biomedical and Life Sciences, 2004, 810, 15-23.


167









182. R. Giovannini, R. Freitag and T. B. Tennikova, Analytical chemistry, 1998, 70,
3348-3354.

183. A. Giulietti, L. Overbergh, D. Valckx, B. Decallonne, R. Bouillon and C. Mathieu,
Methods, 2001, 25, 386-401.

184. J. F. Glickman, X. Wu, R. Mercuri, C. Illy, B. R. Bowen, Y. He and M. Sills, Journal
of Biomolecular Screening, 2002, 7, 3-10.

185. J. J. Tang, T. Yu, L. Guo, J. W. Xie, N. S. Shao and Z. K. He, Biosens.
Bioelectron., 2007, 22, 2456-2463.

186. S. C. B. Gopinath, Y. Shikamoto, H. Mizuno and P. K. R. Kumar, Thromb.
Haemostasis, 2006, 95, 767-771.

187. R. C. Becker, C. Rusconi and B. Sullenger, Thromb. Haemostasis, 2005, 93,
1014-1020.

188. C. Wilson and J. W. Szostak, Chem. Biol., 1998, 5, 609-617.

189. T. O. Pangburn, M. A. Petersen, B. Waybrant, M. M. Adil and E. Kokkoli, Journal
of Biomechanical Engineering-Transactions of the Asme, 2009, 131, 1-20.

190. D. Shangguan, Z. W. Tang, P. Mallikaratchy, Z. Y. Xiao and W. H. Tan,
Chembiochem, 2007, 8, 603-606.

191. M. Famulok, J. S. Hartig and G. Mayer, Chem. Rev. (Washington, DC, U. S.),
2007, 107, 3715-3743.

192. J. R. Lakowicz, Principles of Fluorescence Spectroscopy, 2 edn., Kluwer
Academic/Plenum, New York, 1999.

193. X. M. Liu, G. J. Cao, H. M. Ding, D. J. Zhang, G. Yang, N. L. Liu, M. Fan, B. F.
Shen and N. S. Shao, FEBS Lett., 2004, 562, 125-128.

194. K. Brummel-Ziedins, C. Y. Vossen, F. R. Rosendaal, K. Umezaki and K. G. Mann,
Journal of Thrombosis and Haemostasis, 2005, 3, 1472-1481.

195. J. B. H. Tok and N. O. Fischer, Chem. Commun. (Cambridge, U. K.), 2008, 1883-
1885.

196. M. von Kockritz-Blickwede, O. A. Chow and V. Nizet, Blood, 2009, 114, 5245-
5246.

197. V. A. Holmes and J. M. W. Wallace, Biochem. Soc. Trans., 2005, 33, 428-432.

198. J. I. Weitz, M. Hudoba, D. Massel, J. Maraganore and J. Hirsh, J. Clin. Invest.,
1990, 86, 385-391.


168









199. M. Armani, R. Probst and B. Shapiro, Lab-on-a-chip Technology: Fabrication and
Microfluidics, Caister Academic Press, Norfolk, UK, 2009.

200. A. Cifuentes and H. Poppe, Journal of chromatography, 1994, 680, 321-340.

201. F. Ming, R. Eisenthal, W. J. D. Whish and J. Hubble, Enzyme Microb. Technol.,
2000, 26, 216-221.

202. Molecular Probes, Package Insert, 2001, 1-5.

203. K. Sefah, University of Florida, 2009.

204. D. Glaves, Br. J. Cancer, 1983, 48, 665-673.

205. F. Q. Schafer and G. R. Buettner, Photochem. Photobiol., 1999, 70, 858-867.

206. C. Soncini, P. Carpinelli, L. Gianellini, D. Fancelli, P. Vianello, L. Rusconi, P.
Storici, P. Zugnoni, E. Pesenti, V. Croci, R. Ceruti, M. L. Giorgini, P. Cappella, D.
Ballinari, F. Sola, M. Varasi, R. Bravo and J. Moll, Clinical Cancer Research, 2006,
12, 4080-4089.

207. N. Korolev, A. P. Lyubartsev, A. Rupprecht and L. Nordenskiold, Biophys. J.,
1999, 77, 2736-2749.

208. N. Sundaresan, C. K. S. Pillai and C. H. Suresh, J. Phys. Chem. A, 2006, 110,
8826-8831.


169









BIOGRAPHICAL SKETCH

Jennifer Anne Martin was born in Clearfield, PA in October of 1982. After

graduating from Clearfield Area High School, she attended Misericordia University to

pursue a degree in chemistry. Following a fellowship at the National Institutes of Health

(National Institute of Allergy and Infectious Diseases) she relocated Gainesville, FL to

attend the University of Florida in 2005. While at the University of Florida, she joined

the research group of Dr. Weihong Tan for her Ph.D. work in the area of aptamers for

analytical and therapeutic applications.


170





PAGE 1

1 DNA APTAMERS FOR BIOANALYTICAL APPLICATIONS By JENNIFER ANNE MARTIN A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORID A IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2010

PAGE 2

2 2010 Jennifer Anne Martin

PAGE 3

3 To my family

PAGE 4

4 ACKNOWLEDGMENTS I thank my advisor Dr. Weihong T an for his support and encouragement throughout my graduate studies at the University of Florida. I also thank the many past and present members of Tan Group for their aid in variou s aspects of the research process. In particular, I appreciate t he patience and support from Dr. Kwame Sefah and Dr. Youngmi Kim. Dr. Parag Parekh has been invaluable in collaborating on the SELEX project, while Dr. Joseph Phillips was irrepl aceable in his assistance in the cell capture studies. In addition, I am deeply grateful for the many discussions and friends hips cultivated during my tenure in the group, including Dimitri Van Simaeys, Elizabeth Jimenez, Dalia Lopez Colon, Meghan ODonoghue, Dr. Karen Ma rtinez, Dr. Josh Smith, Dr. Colin Medley, Dr. M.-Carmen Estevez, and Dr. Prabodhika Mallikarachy. Angela Bojarski, Dr. Kathryn Williams, and Lori Clark have been of exceptional aid as well. I wish to recognize my committee mem bers, Dr. Nicolo Omenetto, Dr. Nicole Horenstein, Dr. David Powell, Dr. Charles Cao, and Dr. Donn Dennis, who have guided my development with constructive dialogue. Dr. Donn Dennis, along with Dr. Tim Morey, Dr. Mark Rice, and Dr. Nikolas Gr avenstein has been instrumental in the development of the SELEX project thr ough collaboration with the medical school. Above all, I have relied on the support of my family to achieve all successes throughout my life. My parents, Jim and Deborah Martin have encouraged me in all pursuits, and my brothers, James and Just in have been right there beside them. Without the inspiration from my family, my new Gainesville friends such as Megan and Dan, fellow group members, and my crews fr om college and home, I would not be in the position I am in today, and for t hat I am eternally grateful.

PAGE 5

5 TABLE OF CONTENTS page ACKNOWLEDG MENTS .................................................................................................. 4 LIST OF TABLES............................................................................................................ 9 LIST OF FIGURES ........................................................................................................ 10 LIST OF ABBR EVIATIONS ........................................................................................... 13 ABSTRACT................................................................................................................... 16 CHA PTER 1 INTRODUCTION.................................................................................................... 18 Molecular Re cognition ............................................................................................ 18 Antibod ies......................................................................................................... 19 Aptamers.......................................................................................................... 20 Comparison to anti bodies .......................................................................... 21 SELEX pr ocedure ...................................................................................... 22 Analytical applicat ions of aptamer s............................................................ 23 Biomedical/therapeutic applic ations of apta mers....................................... 25 Introduction to Specific Dissertation Projects .......................................................... 27 Selection of an Aptamer Antidote to an Anticoagulant ..................................... 27 Heparin, protamine, and the importance of thrombin ................................. 28 Direct thrombin inhibi t ors and bi valirud in................................................... 29 Square Capillaries as Simple Micr ofluidic Channels for Bioanalysis ................ 33 Microfluidic devices .................................................................................... 33 Antibodies and aptamers as stati onary phas e ligan ds............................... 37 Theory of ce ll capt ure ................................................................................ 38 Overview of t he Disser t ation................................................................................... 43 2 METHODOLOGY AND DEVELOPMENT OF BEAD-BASED APTAMER SELECTION ........................................................................................................... 44 Introduc tion ............................................................................................................. 44 Anticoagulation wit h Bivali rudin ........................................................................ 44 Aptamer Deve lopment ...................................................................................... 45 Materials and Methods............................................................................................ 46 Experimental Design ........................................................................................ 46 Design of PCR Pr imers and Li brary ................................................................. 48 Optimization of PCR Conditions ....................................................................... 49 Selection Procedure ......................................................................................... 50 PCR Amplificat ion Proc edure ........................................................................... 51 PCR amplification of entire fi rst pool .......................................................... 51

PAGE 6

6 Cycle optimization fo r preparativ e PCR ..................................................... 52 Preparativ e PCR ........................................................................................ 53 Gel electrophoresis m onitoring of products ................................................ 53 Preparation of ssDNA ....................................................................................... 54 Desalting of ssDNA .......................................................................................... 55 Second and Succeeding R ounds of Se lection .................................................. 55 Flow Cytometry Monitoring of Selection Progression ....................................... 57 Flow cytometry for DNA blocki ng step ....................................................... 59 Flow cytometry fo r TCEP cleavage............................................................ 59 Results.................................................................................................................... 59 Enrichment of Aptame r Pool............................................................................. 59 Selection Modi fication #1.................................................................................. 61 Selection Modi fication #2.................................................................................. 63 Preparation of New Li brary and Pr imers........................................................... 66 Selection.................................................................................................... 67 Selection Modi fication #3.................................................................................. 68 Conclusi ons ............................................................................................................ 69 3 MONOLITHIC SELECTION FOR BIVAL IRUDIN .................................................... 71 Introduc tion ............................................................................................................. 71 Materials and Methods............................................................................................ 72 Column Imm obilization ..................................................................................... 72 Method Deve lopment ....................................................................................... 73 Preliminary Select ion ........................................................................................ 78 Selection C onditions ......................................................................................... 79 Monitoring by R eal-time PCR ........................................................................... 80 Second and Subsequent Selecti on Rounds ..................................................... 81 AlphaScreen Analysis of Selected Pools .......................................................... 82 Sequencing of Selected Pool ........................................................................... 83 Sequence A lignment ........................................................................................ 85 Binding St udies ................................................................................................ 86 Results.................................................................................................................... 87 Selection Results .............................................................................................. 87 AlphaScreen for P ool Sele ction ........................................................................ 92 Sequence A lignment ........................................................................................ 93 Binding St udies ................................................................................................ 94 Conclusi ons ............................................................................................................ 95 4 CHARACTERIZATION OF APTAMER AFFINITY AND APTAMER ANTIDOTE TESTIN G ................................................................................................................ 98 Introduc tion ............................................................................................................. 98 Materials and Methods.......................................................................................... 100 Buffer.............................................................................................................. 100 FA Dissociation Cons tant Measur ements....................................................... 100 Clotting Experiment s in Pl asma ...................................................................... 103

PAGE 7

7 Bivalirudin dose re sponse curve .............................................................. 104 Coagulation testing .................................................................................. 105 Clotting Experiment s in Bu ffer ........................................................................ 106 Optimizing fibri nogen concent ration ......................................................... 106 Bivalirudin dose re sponse curve .............................................................. 106 Coagulation testing .................................................................................. 107 Truncation via DNA Microarray ...................................................................... 108 Design...................................................................................................... 108 Microarray studies .................................................................................... 109 Results and Discussion......................................................................................... 110 Kd Characteriza tion by FA.............................................................................. 110 Clotting Studies in Plasma.............................................................................. 112 Clotting Studies in Buffer ................................................................................ 113 Microarray Tr uncation .................................................................................... 116 Conclusi ons .......................................................................................................... 117 5 CANCER CELL CAPTURE USING APT AMER-IMM OBILIZED SQUARE CAPILLARY CHANNELS: PROOF-OF-P RINCIPL E............................................. 119 Introduc tion ........................................................................................................... 119 Materials and Methods.......................................................................................... 120 Cell Cultur e and Bu ffers ................................................................................. 120 Device Construction ....................................................................................... 121 DNA Synthesis ............................................................................................... 122 Device Charac teriza tion ................................................................................. 122 Controlling Degree of Sgc8 Immo bi lization..................................................... 123 Cell Captur e Assays ....................................................................................... 123 Flow Rate Experim ents .................................................................................. 124 Stained Cell Imaging Using Fl uor escence Mi croscopy................................... 124 Cell Elution Effici ency ..................................................................................... 124 Square and Circular Capillary Cell Capture Comparison ............................... 125 Microscopy and Image Analysis ..................................................................... 125 Square and Circular Capillary Imaging ........................................................... 126 Results and Discussion......................................................................................... 126 Sgc8 Immobi lization ....................................................................................... 126 Cell Volumetric Flow Rate Trends .................................................................. 127 Fluorescence Microscopy Imagi ng of Stained C ancer Cells ........................... 128 Cell Capture Performa nce.............................................................................. 130 Square and Circular Capillary Cell Capture Comparison ............................... 131 Imaging Properties of Squar e and Round C apillar ies ..................................... 132 Conclusi ons .......................................................................................................... 133 6 OPTIMIZATION AND APPLICATIONS OF SQUARE CAPILLARY DEVICE ........ 135 Introduc tion ........................................................................................................... 135 Materials and Methods.......................................................................................... 136 Cell Culture, Buffers, Aptamer Sy nthesis, and Device Construction .............. 136

PAGE 8

8 Flow Rate Op timiza tion .................................................................................. 136 Optimization of Cell Washing Fl ow Rate ........................................................ 137 Detecting CEM Cells in Blood ........................................................................ 137 Cell staining............................................................................................. 137 Removal of r ed blood cells ....................................................................... 138 Cell capture.............................................................................................. 139 Colon Cancer Cell Cult ure .............................................................................. 139 Colon Cancer Ap tamer Sequences ................................................................ 139 Cell Capture of Colon Cells ............................................................................ 140 Testing efficiency of aptamers (wit hout linker ) for DLD-1 capture............ 140 DLD-1 cell capture using aptamers wit h linker ......................................... 140 DLD-1 cell capture at low conc entration using aptamer KDED2a-3 ......... 141 HCT 116 cell capture by aptamer KCHA10.............................................. 141 HCT 116 cell capture in non-enzymatic buffer by aptamer KCHA10....... 141 Microscopy and Image Analysis ..................................................................... 141 Results and Discussion......................................................................................... 142 Optimization of Flow Rate .............................................................................. 142 Optimization of Buffe r Wash Fl ow Rate.......................................................... 143 Detection of CEM Cells in Bl ood .................................................................... 144 Colon Cancer Aptamers (without Linker) for DLD-1 Capture.......................... 146 Colon Cancer Aptamers with poly(T)10 Linker for DL D-1 Capture.................. 147 HCT 116 Cell Capture by Aptamer K CHA10 in Non-enzymatic Buffe r........... 148 Conclusi ons .......................................................................................................... 150 7 SUMMARY AND FU TURE WORK ....................................................................... 154 Summary.............................................................................................................. 154 Future Wor k.......................................................................................................... 155 Aptamer Anti dote Pr oject................................................................................ 155 Cell Affinity Project ......................................................................................... 156 LIST OF RE FERENCES ............................................................................................. 158 BIOGRAPHICAL SKETCH .......................................................................................... 170

PAGE 9

9 LIST OF TABLES Table page 1-1 Aptamers in t he clinical pipeline ......................................................................... 25 1-2 Diffusion constants and times for samples of differ ent diam eters ....................... 39 2-1 PCR preparati on of ent ire pool ........................................................................... 51 2-2 PCR cycle optimization ....................................................................................... 52 2-3 Preparat ive PCR ................................................................................................. 53 2-4 Conditions fo r initial SELEX ................................................................................ 56 2-5 Conditions for se lection modificati on #1 ............................................................. 61 2-6 Conditions for selection modification #2 wit h new li brary .................................... 67 2-7 Conditions for se lection modificati on #3 ............................................................. 68 3-1 Method developm ent conditi ons #1 .................................................................... 74 3-2 Method developm ent conditi ons #2 .................................................................... 74 3-3 Method developm ent conditi ons #3 .................................................................... 78 3-4 PCR Pr eparation ................................................................................................ 83 3-5 Fusion Pr imer PCR ............................................................................................. 84 3-6 Probe Sequences ............................................................................................... 94 4-1 Microarra y Truncat ion ....................................................................................... 117 5-1 Physical properti es of capillary ......................................................................... 122 5-2 Performance of capillar y system ....................................................................... 131

PAGE 10

10 LIST OF FIGURES Figure page 1-1 Secondary struct ure o f aptam er sgc8................................................................. 21 1-2 Structur e o f LNA................................................................................................. 22 1-3 The SELEX proc edure. ....................................................................................... 24 1-4 Aptamer Ch-9.3t and cDNA strand fo r use as anticoagulant/antidote pair for factor IX .............................................................................................................. 26 1-5 Thrombin structure (left), heparin binding (center), and steric constraints associated with heparin binding fibrinbound thrombin prior to antithrombin (AT) (ri ght) .......................................................................................................... 28 1-6 Bivalirudin peptide sequence an d binding/cleavage by thrombin. ...................... 30 1-7 Comparison of square and ci rcular capillary geometries. ................................... 36 1-8 Antibody immobilization to stat io nary phase....................................................... 37 1-9 Processes occu rring in ce ll adhesion. ................................................................ 41 2-1 Mechanisms of aptamer/peptide inte ractions for restoring anticoagulant activity ................................................................................................................ 45 2-2 PCR amplification of initial library.. ..................................................................... 52 2-3 Schematic of flow cyt o metry instru mentation...................................................... 58 2-4 Flow cytometry data of initial se lection.. ............................................................. 60 2-5 Flow cytometry data of se lection modifica tion #1. ........................................... 62 2-6 PCR of unlabeled DNA lib rary.. .......................................................................... 63 2-7 Streptavidin blocking with ex cess r andom libra ry............................................... 64 2-8 TCEP concentra tion optimiz ation. ...................................................................... 64 2-9 TCEP interference with st reptavidin/bio tin bi nding. ............................................ 66 2-10 PCR amplification program for se cond init ial library........................................... 66 2-11 Flow cytometry of se lection modi ficati on #2...................................................... 67 2-12 Flow cytometry data of selection modifi cation #3.. ............................................. 69

PAGE 11

11 3-1 Chromatograms fr om method dev elopment.. ..................................................... 76 3-2 Confirmation of LPC method. ............................................................................. 77 3-3 Chromatogram under SELEX-like c onditions .................................................... 78 3-4 Overlay of proof-of-co ncept studies of selections on drug and blank disks. ....... 79 3-5 Amplification plots of IL-4 plas mid cDNA. ........................................................... 81 3-6 Schematic of AlphaScreen assay. ...................................................................... 82 3-7 Optimization of pool volume for fusion prim er P CR............................................ 84 3-8 Agarose gel electrophor esis analysis of fusion primer-amplified, purified pool ... 85 3-9 Sample alignment of JPB2 using MAFFT. .......................................................... 86 3-10 Chromatograms of selection rounds.. ................................................................. 88 3-11 qPCR results for eac h round of se lection.. ......................................................... 91 3-12 AlphaScreen assay of binding of different pools wit h 100 nM or 100 pM drug. .. 92 3-13 Flow cytometry binding studies of drug and aptamer can didates....................... 95 4-1 Secondary structure and modifica tions of the anti-VEG F aptamer pegaptanib (Macugen) ......................................................................................................... 99 4-2 Modes of c oagulat ion....................................................................................... 104 4-3 Truncation of J PB5. .......................................................................................... 109 4-4 Kd plot of thrombin/15-mer th rombin aptamer using FA.................................... 110 4-5 Kd curves of aptam ers and cont rols.................................................................. 111 4-6 Dose response curve of bivalirudin in human plasma. ..................................... 112 4-7 Effect of JPB5 and TV03 control sequence on clotting ti me in plasma. ............ 113 4-8 Optimization of conditi ons for buffer ex perim ents............................................. 114 4-9 Effect of JPB5 and TV 03 control on cl otting ti me. ............................................. 115 4-10 Image of DNA micr oarray target bindi ng. ......................................................... 116 5-1 Schematic of set up and immobiliz ation.. .......................................................... 121 5-2 Fluorescence intensity increase as aptamer concentration is increased. ......... 127

PAGE 12

12 5-3 Confocal images of FAM-labeled sgc8 immobilized ins ide a capillary at various aptamer concentra tions........................................................................ 127 5-4 Confocal images of cells captured in square capillaries at different volumetric flow ra tes.. ........................................................................................................ 129 5-5 Simultaneous fluorescence microscopy imaging of cells in a square capillary stained wit h two di fferent dyes.......................................................................... 130 5-6 Imaging comparison of square and round capi llaries.. ..................................... 132 6-1 Comparison of volumetric flow ra te to capture e fficiency of cells...................... 143 6-2 Comparison of capture efficiency at two different cell wa shing flow rates. ....... 144 6-3 Capture of CEM cells spiked in blood. .............................................................. 145 6-4 Comparison of capture efficiency of colon cancer cell lin e DLD-1 with aptamers KDED2a-3 and KCHA 10................................................................... 146 6-5 Flow cytometry comparison of aptamers binding to DLD-1 with and without poly(T)10 linker.................................................................................................. 147 6-6 Capture efficiency of DLD-1 aptamers with 3-poly(T)10 linker.......................... 148 6-7 Flow cytometric comparison of aptam er binding with HCT 116 cells diluted in BB2 or with non-enzymat ic buffe r (NEB).......................................................... 149 6-8 Overall capture efficiency of col on cancer cell lines DLD-1 and HCT 116 in buffer. ............................................................................................................... 150

PAGE 13

13 LIST OF ABBREVIATIONS AT antithrombin ATP adenosine triphosphate BB binding buffer BB2 binding buffer #2 BSA bovine serum albumin CAC cell affinity chromatography CCD charge-coupled device CD cluster of differentiation cDNA complementary DNA CIM convective interaction media CTC circulating tumor cell Cy5 cyanine derivative 5 DFP diisopropyl fluorophosphate DHFR dihydrofolate reductase protein DMSO dimethyl sulfoxide DNA deoxyribonucleic acid DNAse deoxyribonuclease dNTP deoxyribonucl eotide triphosphate dsDNA double-stranded DNA DTI direct thrombin inhibitor EB elution buffer EDTA ethylenediaminet etraacetic acid ELISA enzyme-linked immunosorbent assay EpCAM epithelial cell adhesion molecule

PAGE 14

14 FA fluorescence anisotropy FACS fluorescence activated cell sorting FAM 5/6-carboxyfluorescein FBS fetal bovine serum FDA Food and Drug Administration FITC fluorescein isothiocyanate HIT heparin-induced thrombocytopenia HPLC high performance liquid chromatography ICBR Interdisciplinary Center for Biotechnology Research IDT Integrated DNA Technologies IgG immunoglobulin G ITC isothermal calorimetry IU international unit LB red blood cell lysis buffer LMWH low molecular weight heparin LPC low-pressure chromatography MACS magnetic activated cell sorting MALDI/TOF matrix-assisted laser desorption ionization/time-of-flight spectrometry NEB non-enzymatic buffer NHS N-hydroxysuccinimide PB physiological buffer PBS phosphate buffered saline PCR polymerase chain reaction PE R-phycoerythrin PEG polyethylene glycol

PAGE 15

15 PF4 platelet factor 4 PTFE polytetrafluoroethylene qPCR quantitative P CR (real-time PCR) RNA ribonucleic acid SELEX systematic evolution of ligands by exponential enrichment S-Hir53-64 sulfated exosite 1 bi nding region of hirudin ssDNA single-stranded DNA TBE tris-borate-EDTA TCEP tris(2-carboxyethyl)phosphine TFA trifluoroacetic acid TMR tetramethylrhodamine tris tris(hydroxymethyl)aminomethane tRNA transfer RNA UV ultraviolet WB washing buffer WB2 washing buffer #2 Rn baseline subtracted fluorescence

PAGE 16

16 Abstract of Dissertation Pr esented to the Graduate School of the University of Florida in Partial Fulf illment of the Requirements for t he Degree of Doctor of Philosophy DNA APTAMERS FOR BIOANALYTICAL APPLICATIONS By Jennifer Anne Martin August 2010 Chair: Weihong Tan Major: Chemistry Aptamers are oligonucleotides selected by the SELEX (Systematic Evolution of Ligands by EXponential enrichment) method that bind to a target with high affinity and specificity based on the thr ee dimensional conformation they adopt. Selected for targets ranging from small molecules to proteins or whole cells, aptamers may similarly assume diverse roles such as drug delivery vehicles, decoys, inhibitory therapeutics, and affinity separation probes. This work inve stigates aptamers as inhibitory drugs and affinity ligands in two bioanalytical applications. In the first set of studies, an aptamer was selected to the peptide anticoagulant bivalirudin to serve as an antidote to the dr ug in instances of severe patient bleeding. The drug was immobilized on a monolithic co lumn and binding sequences were eluted by salt gradient. The elution profile was co mpared to that of a blank column (no drug), and fractions with a chromatographic differ ence between drug-immobilized and blank counterparts were analyzed via real-time PCR (polymerase chain reaction) and used for further selection. Sequences were ident ified which demonstrat ed low micromolar dissociation constants through fluorescenc e anisotropy after only two rounds of selection. One aptamer displayed a dosedependent reduction of the clotting time in

PAGE 17

17 buffer, with a 20 M aptamer concentration achieving a nearly complete antidote effect. A DNA microarray was used to truncate two of the aptamers to dec rease the cost of aptamer synthesis for medical applications. The second group of experiments tested a simple aptamer-based square capillary cell affinity chromatography system fo r selective capture of target cancer cells from a flowing suspension. Aptamers were immobilized on the inner surface of the capillary through biotin-avidin chemistry, the extent of which was controlled by adjusting the aptamer concentration. The device capt ured target leukemia cells in higher amounts than nontarget cells und er a variety of conditions, and the capture efficiency was optimized to retain >90% of cells. In addition, the system was used to capture two colon cancer cell lines by their respective ap tamers at high effici ency. The capillary system could also detect stained cancer cells spiked in blood by imaging the length of the tubing and counting the fl uorescent captured cells.

PAGE 18

18 CHAPTER 1 INTRODUCTION Molecular Recognition Molecular recognition between two biomol ec ules has long been recognized as a cornerstone of specificity and control of complex biological processes.1 A number of factors mediate this development, including the shape of the two binding surfaces, hydrogen bonding, ionic, dipole, and solvent interactions, and van der Waals forces.1 These binding events may occur between prot eins and small molecules to induce signaling processes,2 amongst proteins and nucleic acids as in DNA/histone interactions, and in catalysis of various substrates by enzymes.3 Enzymes catalyzing reactions in sugars (protein/carbohydrate) are well-known interactions, for example the -amylase enzyme family recognizing starch.4 RNA/small molecule interactions are also of interest, such as those displayed by riboswitches.5 Additional interactions of significance include protein/pr otein binding, for example dynein or kinesin interacting with microtubule proteins.6 Researchers have taken advantage of molecula r recognition principles to form the foundation of current target detection me thods and modern diagnosis/treatment of disease. At the molecular level, biosens ors have been designed for detection and study of different properties of small molecule /protein and protein/protein interactions.7-9 Molecular recognition has also been utilized in designing novel therapeutic drugs. Examples include targeting of dihydrofol ate reductase protein (DHFR) by the chemotherapy drug Methotrexate10 and the antibiotic Trimethoprim.11 Also, the aminoglycoside paromomycin was found to bind to specific RNA in E. coli to exhibit antibiotic effects.12

PAGE 19

19 Despite the term molecular recognition t he phrase also refers to specific cellular recognition events resulting fr om the affinity between a li gand and cell surface marker. Analytically, ligands immobilized on a surface have been used to detect and/or purify cellular targets.13 Furthermore, biomarker expression has been manipulated in disease detection, or targeted fo r pharmacological response.14 Molecular recognition principles have been extremely effective for both analytical and pharmaceutical applications regardless of the target. In particular, antibodies raised for targets either expressed on cell surfaces or found in the blood have made tremendous progress as detection reagents and therapeutics for various diseases.15 Antibodies Antibodies are well-known molec ular recognition agents deserving special attention due to rapid development in the last 25 years.15 Antibodies can be immobilized as a stationary phase for sele ctive detection or capture of target molecules.16, 17 The proteins are also used to bot h preconcentrate cellular samples and stain cancer cells for immunocytochemical detection.18 Additionally, vaccines utilizing antibody/antigen response have resulted in protection against influenza and the eradication of smallpox, a disease which was responsible for hundreds of millions of deaths in the twentieth century.19 More recently, the utility of antibodies used in disease treatment has been investigated with a high degree of success. Since the first CD3specific antibody was approved fo r the treatment of acute tr ansplant rejection in 1986,20 more than 30 antibodies have been approved for use in various indications.15 Of the $15 billion in revenue, 78% derives from mo noclonal antibodies targeting cancer and autoimmune diseases.20 Some FDA-approved antibodies include tumor-necrosis factor specific antibodies inflix imab and adalimumab, the human epidermal growth factor

PAGE 20

20 receptor 2 antibody trastuzumab, the va scular endothelial growth factor targeting antibody bevacizumab, and the CD20-specific rituximab antibody. Phase III trials are underway on 26 more antibody drug candidates, 9 of which are for diseases designated orphan diseases, or rare diseases a fflicting only a small number of patients.15, 21 Despite the tremendous achievements of anti bodies, several challenges remain. Antibody generation requires a biological syst em, limiting the conditions for optimization to physiological conditions, possibly reducing their utility for in vitro applications such as detection platforms.22, 23 Due to the biological require ment, antibodies cannot be raised against biological toxins whic h would damage the host system.24 Antibodies also suffer from significant batch-to-batch va riability in function. Furthermore, it is difficult to modify antibodies with labels to facilitate capture or detection while still retaining function, and the proteins are susceptible to degradati on upon temperature changes and prolonged storage. Another concern is the time and expense associated with the antibody screening process.25 In addition, a limited number of available antibodies,26 difficulties with cross-species reactivity (i.e. applyi ng a murine-generated antibody to detect the human version of the target), and gener ation of an immune response upon administration24 also present serious challenges for antibody applications.23 Molecules designed to address these limitat ions would be highly attracti ve candidates for various target/ligand binding functions. Aptamers Aptamers are DNA or RNA oligonucleotides t hat bind specifically to a target based on their three dimensional confor mation. This structure is the result of a combination of hydrogen bonding, electrostatic and van der Waals interactions, Watson-Crick base pairing, and stacking of aromatic rings.22, 27 Aptamers usually adopt stem/loop

PAGE 21

21 secondary structures, as seen in Figure 1-1. The loop regions are primarily responsible for target interactions, wh ile the remainder of the sequences aid in stabilizing the structure. Figure 1-1. Secondary stru cture of aptamer sgc8.28 Aptamers were first described in 1990 by two separate groups targeting either an organic dye29 or a DNA polymerase.30 Since the inception of the technology, aptamers have been selected for a vast range of targets including metal ions,31, 32 small organic molecules,29, 33-38 nucleotides and derivatives,39-41 cofactors,42-44 nucleic acids,45, 46 amino acids,47-49 carbohydrates,50, 51 antibiotics,52-54 peptides,55, 56 proteins,57-61 whole cells,28, 62-67 viruses and virus infected cells,68, 69 and bacteria.70, 71 Comparison to antibodies When compared with antibodies, aptamers demonstrate sim ilar bin ding affinities (pM-nM), yet have the following advantages: 1) They are chemically synthesized, and can therefore be selected for virtually any target, including biological toxins; 2) Chemical synthesis eliminates batch-to-batch vari ability and reduces cost of production; 3)

PAGE 22

22 Experimental conditions can be manipulated to select aptamers with specific properties; 4) Aptamers can be easily modified with labels to facilitate target purification or detection by fluorescent dyes; 5) The seque nces are relatively stable to degradation from high temperatures or long storage times; 6) No evidence of immunogenicity has been observed; 7) Animals are not required for aptamer selection.22-24, 72 One major drawback of aptamers is degradation by nucleases when in vivo applications are desired. Since most polym erases are unable to incorporate modified bases in the PCR, several groups have pr oposed different post-SELEX modifications to minimize the effects. For RNA, typicall y attacked by RNAse present in nearly all biological samples,26 the 2-OH groups can be substituted with 2-F, -NH2, or -OCH3 groups, which have proven successful at inhibiting degradation.73-75 Substitution of pyramidines at the 5-posit ion with I, Br, Cl, NH3, or N3 has also been shown to improve stability.26 Additionally, 3and/or 5-end capping (ex. polyethylene glycol) is able to protect sequences against exonuclease activity.22 Incorporating locked nucleic acids (LNA) consisting of a methylene bridge from the O2 to C4 of the s ugar into the aptamer structure is also useful in stabilizing the structure of oligonucleotides (Figure 1-2).76 Figure 1-2. Structure of LNA.76 SELEX procedure Aptamers are selected using a method know n as SELEX (Systematic Evolution of Ligands by EXponential enrichm ent). The SE LEX method is depicted in Figure 1-3, and

PAGE 23

23 certain aspects are described in more detail in future chapters. The SELEX procedure begins with 1013-1015 unique sequences from a chemically synthesized, randomized oligonucleotide library competing for binding to the target. Nonbinding sequences are partitioned from binding oligonuc leotides, and binders are eluted from the target. This partitioning step is the main determinant of the efficiency of the selection. The binding sequences are PCR-amplified and converted to single-st randed DNA (ssDNA) for the next round of selection. Af ter the first round, researchers may institute a counterselection step, in which sequences binding to a control (such as support matrix) are removed from solution. Sequences that do not bind to the control are retained for future SELEX rounds. The process is repeated in cyclical fashion until the final pool is enriched for sequences binding to the target The number of cycles varies depending on parameters such as selection stringency and the efficiency of the partitioning method; however, a typical selection require s an average of 12 cycles and a timeline of 2-3 months.24, 77 Following enrichment, the oligonuc leotide pool is sequenced, and the oligonucleotides are grouped according to intermolecular homology using programs such as ClustalX or MAFFT. Several seque nces from each family are synthesized to test for target binding, and su ccessful candidates are characterized in terms of binding affinity, target specificity, etc. Analytical applications of aptamers Antibod ies are frequently utilized as imm unoaffinity column ligands, but their linkage to the column is nontrivial, resulting in couplings that ar e not uniform. This reduces the binding capacity and in some ca ses leads to leaching of the separation phase from the column.78 In contrast, aptamers are ea sily chemically modified for linkage to a column, and are smaller in size than antibodies, producing a higher ligand

PAGE 24

24 density for chromatography stationary phas es. Aptamers have been employed as affinity ligands for small molecules and proteins,79-81 and to discriminate between chiral molecules in a separation due to thei r exquisite specificity properties.82, 83 Aptamers have also been used as biosensors for specific targets due to their high ligand density and ease of chemical modification. The pro bes have been chemically modified to bind to a surface for acoustic sensing,84 optical sensing,85-87 cantilever-based biosensors enabling label-free detection,88 and fluorescent signal sensing.89, 90 Figure 1-3. The SELEX procedure.28 At the cellular level, aptamers hav e found use as ligands for separation, enrichment, and/or detection. For example, aptamers immobilized in a microfluidic channel are capable of select ively capturing target cells from buffer or blood.91-94 Cell separations have important implications in areas such as CD4+ T lymphocyte counting, blood cell isolation, bacterial ident ification, rare cell capture, and cancer cell detection.95 Cell separation plays a central role in the work completed for this dissertation, as discussed later in this chapter.

PAGE 25

25 Biomedical/therapeutic applications of aptamers The unique properties of aptamers present exciting opportunities for biomedical diagnostic and therapeutic applic ations. Aptamers have been selected to serve as histological detection reagents for glioblastoma96 or in proximity-dependent assays (ELISA),97 and as throughput validation ligands for drug discovery.98 The probes have also been developed for in vivo imaging99, 100 and have demonstrated promising results for reducing side-effects by targeted drug delivery.101-103 Table 1-1. Aptamers in the clinical pipeline104 Aptamer, Target Company Disease Stage of Development Pegaptanib (Macugen), VEGF-165 Eyetech Pharmaceuticals/ Pfizer AMD, diabetic retinopathy FDA-approved AMD 2004; Phase II for diabetic retinopathy: improved vision, decreased edema REG-1 (RB006/RB007), Factor IXa Regado Biosciences PCI, CABGPhase II trial completed AS1411, Nucleolin Antisoma RCC, AML Phase II underway NU172, Thrombin Nuvelo/Archemix PCI, CABGPhase II planned ARC1779, von Willebrand factor Archemix TMA, TTP, CEA Phase II underway Abbreviations: AMD, age-related macular degeneration; PCI, percutaneous coronary intervention; CABG, coronary artery bypass graft; RCC, renal cell carcinoma; AML, acute myeloid leukemia; TMA, th rombotic microangiopathies; TTP, thrombocytopenic purpura; CEA, carotid endarterectomy. The FDA approval of the aptamer dr ug Macugen targeting vascular endothelial growth factor (VEGF) in 2004 was a breakthrough for aptamer pharmaceuticals, renewing interest in aptamer based therapeutics.105 Therapeutic aptamers can be divided into several categories104 including decoy-like aptamer s targeting proteins that naturally bind nucleic acids,106, 107 regulatable aptamers which e ffectively serve as their own antidote,108 multivalent aptamers to improve aptamer binding,109, 110 and inhibitory aptamers that target proteins, transcription factors, cell-surfac e markers, etc. in order to

PAGE 26

26 treat human disease.111, 112 Aptamer drug candidates current ly in the clinical pipeline are summarized in Table 1-1. Several of these aptamers are progressing to Phase II studies, and many of the target s are coagulation factors. Inhibitory aptamers have shown particular promise in the role of anticoagulants, targeting various points of the coagulat ion cascade. Aptamers to thrombin,113 factor IX,114 factor VII,115 factor X,116 protein C,117 and von Willebrand factor118 have been shown to successfully modulate thrombus formation, with cDNA antidotes able to restore normal activity. In particular, Rusconi and coworkers developed the factor IX aptamer (Table 1-1) that is now in the clinical pipeline. The aptamer was modified with a 5-cholesterol and a 3-inverted deoxythymidi ne, and consists of pyramidines modified at the 2-position with fluorine (Figure 1-4).108 Upon addition of 2O -methylated cDNA corresponding to the contiguous stem r egion at the 5-end of the aptamer, the anticoagulant effect was neutralized and clotti ng activity commenced in animal models. Aptamer/cDNA development is relatively co st-prohibitive, so studies have been carried out to find different methods to reverse anticoagulant binding.119 Figure 1-4. Aptamer Ch-9.3t and cDNA str and for use as anticoagulant/antidote pair for factor IX.108

PAGE 27

27 Recent work has investigated the use of light to photoregulate aptamer antidote activity by introducing a caged structure into an aptamer,120 or an azobenzene moiety to reverse anticoagulation.121 Clearly the use of light to i nhibit internal bleeding is not feasible, since the light would not penetrate through the skin, but the design may have topical applications. Researchers have used aptamers and cationic porphyrins to serve as an anticoagulant/antidote pair, however this method will only work for aptamers known to form a G-quartet.122 A polymeric antidote has shown a promising universal antidote response for various anticoagulant aptamers, but the mechanism of action is still unclear and the research is still in the preliminary phase.119 Therefore, developing a DNA aptamer to a currently available pharmaceutical agent will be another major theme for this dissertation work. Introduction to Specific Dissertation Projects Selection of an Aptamer Antidote to an Anticoagulant Anticoagulants are one of the drug classe s with the highest instances of adverse drug reactions and medication errors.123 These actions dire ctly correlate to an increased occurrence of complications such as severe bleeding th at increase patient morbidity and mortality.119 Additionally, blood transfusions are required for 5-10% of patients with severe bleeding, at an esti mated cost of $8,000$12,000 per incident.108 An ideal anticoagulant would exhibit minima l adverse drug reactions, and would have a safe and effective antidote readily availabl e to reverse severe patient bleeding. Heparin and protamine are the only anticoagulant/anti dote pair commonly used in clinics, but both drugs have considerable risk associated with their use. Heparin demonstrates an unpredictable anticoagulant response, and some patients cannot be administered heparin in any circumstance due to severe physiological responses. A

PAGE 28

28 variety of synthetic anticoagu lant drugs were developed to improve upon the challenges created by heparin use. Bivalirudin is one su ch synthetic peptide anticoagulant that has several advantages over heparin, including a more predictable drug response, yet does not currently have an antidote available. Therefore, the objective of this work was to provide an antidote to bivalirudin to intr oduce a safe and reliable anticoagulant/antidote pair. Heparin, protamine, and th e importance of thrombin Thrombin is a common target for anticoagu lant drugs due to its position at the juncture of the coagulation ca scade and the variety of roles it play s in the clotting process.124 Most importantly, thrombin activation of soluble fibrinogen to insoluble fibrin and fibrin crosslinking by thrombin-activation of factor XIII is a critical step of the clotting process. Figure 1-5. Thrombin structure (left), hepar in binding (center), and steric constraints associated with heparin binding fibrinbound thrombin prior to antithrombin (AT) (right).125 The binding sites of thrombin consist of two positively charged exosites and the active site where catalytic acti vity takes place (Figure 1-5).125-128 Heparin indirectly

PAGE 29

29 targets the protein thrombin by binding to both antithrombin (AT; a naturally occurring thrombin inhibitor) and exosite 2 of thrombin.129 One problem with heparin is that it cannot inhibit fibrin bound thrombin, possibly due to steric constraints. If heparin docks to exosite 2 without previously binding antit hrombin, it can fo rm a bond with thrombinbound fibrin, actually strengtheni ng the clot (Figure 1-5).126 Heparin also binds to certain plasma proteins in the blood, resulting in an unpre dictable anticoagulant response and increased patient monitoring. H eparin is neutralized by platelet factor 4 (PF4) a product of activated platelets.130 A major challenge to heparin use is that heparin complexed with PF4 or other plasma proteins can stimulate a response called heparin-induced thrombocytopenia (HIT), which can cause severe reactions in some patients. The use of low molecular weight heparin (LMWH) alleviates several, but not all of these concerns. The antidote to heparin, protamine, also has serious side effects associated with its administ ration, including increased pulmonary artery pressure, decreased systolic and diastolic blood pr essure, impaired myocardial oxygen consumption, and reduced cardiac outpu t, heart rate, and systemic vascular resistence.119 Therefore, there is an obvious need for alternativ e anticoagulant/antidote pairs with a safer t herapeutic profile. Direct thrombin inhibitors and bivalirudin In order to improve the safety profile of antico agulants, synthetic thrombin inhibitors which directly intera ct with thrombin, termed direct thrombin inhibitors (DTI), were designed. One such drug, bivalirudin, is a promis ing alternative to heparin anticoagulation. In comparison to heparin, bivalirudin generates a more predictable anticoagulant response because it does not bind to other plasma proteins. It also binds both fibrin-bound and free thrombin, is not inac tivated in the pres ence of PF4, and does

PAGE 30

30 not induce HIT.126, 131 In addition, bivalirudin can be c hemically synthesized using solid state chemistry, while heparin is obtained from animals. Bivalirudin is a 20 amino acid peptide (Figure 1-6) used in the treatment of coronary angioplasty in unstable angina patients. The drug is a derivative of the leach anticoagulant hirudin that demonstrates essentially irreversible binding to thrombin at (inhibition constant) Ki= 231 fM.132 Femtomolar inhibiti on constants are generally considered to represent irreversible binding. The engineering of hi rudin to bivalirudin has produced a reversible anticoagulant with Ki= 2 nM.125 The peptide has a molecular weight of 2180 Daltons, and forms a 1:1 complex with thrombin that is slowly cleaved at the Arg3-Pro4 bond with a half-life of 25 minute s, but currently no antidote exists for rapid reversal in the instance of severe bleeding.133 Figure 1-6. Bivalirudin peptide s equence and binding/cleavage by thrombin.125 The N-terminus of bivalirudin binds to the active site of thrombin, and the Cterminus docks to exosite 1 (Figure 1-6). In itially, bivalirudin co mpetes with fibrinogen for access to the exosite 1.132 Binding to thrombin can be considered to occur in 4

PAGE 31

31 steps: 1) The C-terminus binds to exosite 1 with a Kd of 0.75 M; 2) A conformational change occurs in thrombin; 3) Arg3 of biva lirudin interacts with the active site; 4) Another thrombin conformational change oc curs, with a rate constant of 30 s-1.134-136 Steps 3 and 4 result in a stable inhibitor-t hrombin complex. After cleavage of the Arg3/Pro4 bond, the N-terminus is released fr om the active site and the C-terminal fragment interaction at exosite 1 becomes a low-affinity weakly-competitive inhibitor of fibrinogen.131 Fibrinogen easily displaces the fr agment from the binding site and resumes conversion to fibrin. The mechanism of action of the pepti de is important to consider for future work on select ing an aptamer for the peptide. In initial studies, the drug marketed today as bivalirudin (Angiomax) was referred to as hirulog-1, short for hirudin analog 1. Hirulog-1 was designed based on linking the exosite 1 and active site (with slight modi fications to make the binding reversible) binding domains of hirudin with a (Gly)4 linker.137 Binding tests comparing hirulog-1 with (Gly)6 and (Gly)8 linkers produced very similar Ki values (2.1-3.0 nM) of which the (Gly)4 was the lowest. Interestingly, when the DPhe1 of hirulog-1 was replaced by L-Phe1, the inhibition constant was increased from 2.1 nM to 156.0 nM, demonstrating the specificity of D-Phe for thrombin binding. These studies also showed that the active site-binding region and the exosite 1 binding domain separately lacked the ability to significantly inhibit thrombin, as both inhibition constants were >2 M. Instead, the combination of these domains results in the anticoagulant properties of the drug. Furthermore, the group studied which terminus of hirulog-1 occupies the binding site by testing the ability of hirulog-1 to block a modification of Ser195 by [14C] DFP (diisopropyl fluorophosphate). The results show that a 3and 30-fold excess of drug to thrombin

PAGE 32

32 blocked the modification, while much higher concentrations of S-Hir53-64 (sulfated exosite 1 binding region) failed to block the sa me modification. This highlights the fact that the N-terminus of hirulog-1 occupies the active site of thrombin. Further studies in the same literature proved the s pecificity of the interaction of hirulog-1 with thrombin by showing that it does not affect human plasmin or bovine trypsin activity. Additional studies on hirulog-1 probed the binding of the drug to various forms of thrombin.136 Nanomolar Kis were reported for human and -thrombin, but no inhibition was observed for -thrombin or trypsin with 7.85 M hirulog-1 concentrations. The binding of the drug to -thrombin can actually be considered to be a positive result, since of all of the proteolytic cleavage products of prothrombin ( -, -, -, -, and thrombin), only and -thrombin retain the ability to significantly interact with fibrinogen. However, the other forms ar e still capable of cleav ing small synthetic molecules and proteins, which include factor XIII, AT, and prothrombin.136, 138 Therefore, hirulog-1 is a spec ific inhibitor of the two reported fibrinogen-catalyzing forms of human thrombin. Final studies relevant to the scope of this project report on the crystal structure of hirulog-1 binding to thrombin.139 Only 8 of the 20 amino acid residues were observed in the electron density map, corresponding to t he first 3 residues of the N-terminus, and the Asp11-Ile15 of the C-terminus. The disorder of Pro4-Gly10 is believed to be a result of the cleavage of the Arg3Pro4 bond. The residues adjacent to the C-terminus and not involved in the electron density map, Glu17-Tyr20, coil into one turn of a 310 helix. Additionally, the side-chains of D-Phe1-Arg3 play a major role in stabilization inside the

PAGE 33

33 active site.128, 140 This structural information will form the basis for the decisions made regarding the methodology described in Chapters 2-3. Square Capillaries as Simple Microf lu idic Channels for Bioanalysis Cell separations are important in a variet y of applications for both basic scientists and clinicians. Standard microf luidic devices have been successful in cell capture, yet key drawbacks in design and fabrication ma y promote the use of capillary-based systems. This work explores the utility of aptamers as stat ionary phase ligands in simple capillary affinity chromatography syst ems for the selective capture of target cancer cells. Microfluidic devices Microfluidic technology has emerged as a ma jor research area in the last 10 years, as evidenced by publication of nearly 10,000 papers on the topic.141 The blanket term microfluidics refers to technology ranging from dipsticks for simple measurements (for example pH testing), to more complex lateral flow te sts and lab-on-a-chip devices with intricate microvalves and pumps integr ated into the device. Some attractive features of microfluidic pla tforms encouraging the growth in research are: 1) Portability; 2) High sensitivity; 3) Low cost per te st; 4) Short data acqui sition time; 5) Less laboratory space required. Additional advantages incurred by microfluidic minuturization include lower sample/reagent consumption, consistent and controlled laminar flow properties, and the possib ility to multiplex measurements. Thus far, microfluidic systems have been administer ed in a wide variety of applications. Analytical devices for detecting pregnancy142 and drug use143 are particularly well known to the general population. Detection of biowarfare agents has become increasingly more important as we ll, with devices capable of recognizing

PAGE 34

34 bacterial144 and molecular targets such as ricin, Shiga toxin I, and Staphylococcal enterotoxin B.145 Microfluidic systems have al so been applied to detection of contaminants in food and water.146, 147 Biochemical applications are also of ma jor interest for microfluidic devices, observed in research in microfluidic bioreactors,148 radiopharmaceutical synthesis,149 and high throughput screening for drug discovery.150 Particularly interesting are the possibilities of microfluidic systems for ce ll capture and detection, a process known as cell affinity chromatography (CAC). Cell se parations have thus far been utilized for applications such as CD4+ T lymphocyte c ounting, blood cell isolation, bacterial identification, rare cell capture, and cancer cell detection.95 The development of devices capable of selectively capt uring cancer cells may enable advancement in cancer diagnosis and monitoring treatment progr ess, which significantly enhances the outcomes for cancer patients.92-94, 151, 152 Either physical or chemical methods may be used to sort cell subpopulations. Physical methods such as dielectrophoresis and sedimentation have previously been successful, but these methods rely on signi ficant differences between the physical properties of the target cells and the matrix cells.95 Chemical methods were therefore designed to separate cells based on their uni que surface chemistry. Most commonly, fluorescence activated cell sorting (F ACS) is utilized for cell separation.92, 95 While the method tends to reproducibly generat e highly pure subpopulations, FACS can be expensive to operate, requires highly trained personnel, and along with a similar technique termed MACS (magnetic activated ce ll sorting) necessitates preprocessing the cells with dyes or fluorescently-labeled antibodies, dramatically increasing the

PAGE 35

35 sorting time of the sample. Similarly, in circulating tumor cell (CTC) analysis the prevailing method is the commercially-avai lable CellSearch technology, which also employs immunological assays to different iate tumor cells from normal cells, in conjunction with traditional cell imm unophenotyping and morphologic analysis.153 Alternatively, microarrays do not require a preproce ssing step and are able to simultaneously detect multiple targets, but tend to be low throughput, and cells cannot be eluted without mixing t he different subpopulations.91 In contrast, CAC is a relatively simple process that can be multiplexed to perform the same function as an array; however, CAC allows for the elution of ce lls, enabling culture and fu rther study of the population. Due to these advantages, CA C was chosen as the method of cell separation used in this work. Despite the many successes of microf luidic technology, many drawbacks are associated with their use whic h may promote the use of si mple capillary devices for certain CAC applications. Mo st importantly, capillary syst ems remove the difficulties associated with the design and fabr ication of microfluidic devices, which typically require the use of expensive clean rooms, not easily accessibl e for many laboratories.154 Microfluidic channels may also be more comp licated to interface with a benchtop fluidic system for sample flow rate control than a c apillary. In contrast to microfluidics, capillaries often have simple, commercia lly available connection options, and are available in a variety of internal and exte rnal sizes and geometries, depending on the desired application. Capillaries are commercially manufactured to have consistent properties throughout, thus eliminating conc erns about batch-to-batch variability, a frequent problem when fabricating microfluid ic devices. Also, capillaries can easily

PAGE 36

36 undergo surface modification using standard chem istries to facilitate the immobilization of cell capture ligands. The mo st prominent disadvantage of capillaries lies in their low throughput, which results in challenges for detecting targets present in low concentration. However, capilliaries can be a rranged either in parallel or in series to increase throughput or to facilitat e the capture of multiple ta rgets, with the possibility of simultaneous detection of several types of cancer.91 Also, a preconcentration step to enrich the sample for target cells may a lleviate this concern. The commercialized CellSearch technology pla tform for CTC detection preconcentrates the sample for epithelial cells usi ng the EpCAM antibody. Figure 1-7. Comparison of square and circular capillary geometries.155 Specifically, square capillaries have properties with distinct benefits when compared to circular tubes. They have a higher surface area and volume than circular capillaries of the same dimensions,156 leading to more capture probe immobilized on the walls and a higher sample throughput. Additi onally, the sensitivity of path lengthdependent detection methods, such as absorpti on and fluorescence, can be enhanced by an order of magnitude over circular c apillaries. Most importantly, flat-walled capillaries (square or rectangular) have le ss optical distortion and scatter than the curved walls of circular tubes.156, 157 Figure 1-7 shows this effect, as collimated light

PAGE 37

37 passes directly through the square capillary, but light in the ci rcular capillary is scattered at the curved walls of each interface.155 This is especially significant when direct observation is used for detection, as may be advisable in the case of potentially extremely low numbers of cancer cells or CTCs present in the blood. Antibodies and aptamers as stationary phase ligands Figure 1-8. Antibody immobilization to stationary phase. In a) no antigen recognition sites are available for binding. In b) and c) either 1 or 2 sites are oriented correctly (as designated with arrows) for antigen binding.158 An ideal stationary phas e should possess the followi ng characteristics: 1) Chemical stability; 2) Low nonspecific binding; 3) Mechanical stabilit y for favorable flow rates; 4) Sufficient surface area for binding.158 As previously discussed, antibodies are not robust in terms of stability, as changes in temperature and buffer conditions can reduce binding. Antibodies may also bind an epitope of a protein found on many cell surfaces, depending on the conditions used fo r their generation (f or example, EpCAM antibody will bind to cancerous and normal epithelial cells). However, aptamers can be selected to bind to one specific type of canc er, or even a specific cell line within the

PAGE 38

38 disease. Aptamer synthesis is more stra ightforward than antibody generation, and an affinity label such as biotin facilitates imm obilization to a (strept)avidin-coated surface. Linkers such as polyT are often incorporated to provide distance from the solid support and flexibility for aptamer binding.159 Antibodies can be labeled by various chemistries to promote immobilization, but this requires an extra purification step and often it is uncertain exactly where on the protein the tags end up. This can cause the protein to be immobilized in a way that renders their antigen bind ing sites unavailable, as demonstrated in Figure 1-8.158 Also, it may interfere with antigen binding, or reduce the stability of the protein, as was the case in immunoglobulin G (IgG) biotinylation.160 Therefore, aptamer s are immobilized on capillary walls as CAC stationary phases in this work. Theory of cell capture The theory dictating flow in square capill aries is similar for square and circular capillaries, although the equations may va ry, especially when area is used as a parameter. Fluid flow is characterized by either exhibiting laminar or turbulent properties; laminar flow is desired because it is an ordered flow of fluid throughout the tubing, while turbulent flow is ch aotic. The Reynolds number (Re) is calculated to determine whether the flow is laminar or tu rbulent (equation 1-1) Reynolds numbers below ~2000 dictate that laminar flow will occur within the system.161 vd eR [1-1] In equation 1-1, describes the density of the fluid (kg/m3), v is the fluid velocity (m/s), d is the diameter of the tube (m), and is the viscosity of the flui d (Pa*s). For fluid flow in a square or rectangular (noncircular) tube, the diameter is actually defined as hydraulic

PAGE 39

39 diameter (Dh), calculated via equation 12, where L and W represent length and width of the cross-sectional area of the capillary in meters. WL 2LW Dh [1-2] After determining that the system will operat e by laminar flow guidelines, it is important to take into consi deration the diffusion times of t he target so proper incubation times can be utilized. Stokes Law (equation 1-3) describes the force exerted on spherical objects (such as cells) at low Reynolds number through a fluid.161 rvFfr 6 [1-3] Ffr represents the frictional force (N), r is the cell radius (m), and v is the fluid velocity (m/s). When Stokes Law holds (as in conditions of low Re), the Einstein relation can be used to calculate the diffusion const ant of the target (equation 1-4). r kT D6 [1-4] T is the temperature (K) and k depicts Boltzmans constant (J/K*mol). The diffusion constant is then used to calculate the diffusi on time (t) by equation 1-5, where x is the average distance (m). D x t 22 [1-5] Table 1-2. Diffusion constants and time s for samples of di fferent diameters Sample r (m) D (m2/s) t (s) ATP 1.00 x 10-9 2.45 x 10-10 2.9 Thrombin 3.00 x 10-9 8.17 x 10-11 8.6 Nanoparticle 5.00 x 10-9 4.90 x 10-11 14.3 Cell 5.00 x 10-6 4.90 x 10-14 14341.6

PAGE 40

40 The diffusion times of targets of several different sizes are listed in Table 1-2 for a capillary with an inner diameter of 75 m. The value of x is assumed to be approximately half of the diameter, 37.5 m. Note that for a ce ll, the diffusion time is close to 4 hours. However, the effects of gravity on a particl e as large as a cell must be considered. When Stokes Law applies, the settling velocity (Vs) of a cell can be calculated by equation 1-6. 9 22gr Vfp s [1-6] represents the density of the particles ( p) or fluid ( f) (kg/m3), and g is the gravitational acceleration constant (m/s2). Using a dynamic viscosity of 9.0 x 10-4 Pa*s, Vs is calculated as 5.51 x 10-5 m/s. This value is then used to calculate the settling time for a cell (ts) as 0.68 sec for a 75 m diameter capillary by equation 1-7. s sV x t [1-7] Therefore, cells/affinity ligand incubation ti mes should be set with consideration of Vs, since the system is dominated by sedimentation forces, not diffusion forces. Figure 1-9 represents several crucial parameters deter mining whether a cell will adhere to a surface.162 When the cell closely appr oaches (100-300 ) the ligandimmobilized surface, substant ial pressure builds up on the r egion of the cell closest to the binding surface, and that area of the ce ll flattens. This area becomes the contact area between the cell and surface. The cell is assumed to possess receptors on the surface which bind to the capture ligand. U pon binding, the inherent diffusional mobility of the receptors in the cell membrane allow fo r additional contacts to be made. Once

PAGE 41

41 bound, shear forces of the flowing liquid act on the cell, with the strongest forces acting on the cell apex.163 Figure 1-9. Processes occurring in cell adhesion.162 The cell will bind to the surface if the adhesion forces are greater than the shear forces (equation 1-3) acting to break the ligand/cell bonds. The adhesion force (FA) is given by equation 1-8,151 and the critical force (fc) required to rupt ure the ligand/cell bond is given by equation 1-9.164 sccACAfF [1-8] Ac is the cell contact area with the ligand-bou nd surface (~1 m2 for cell radius of 5 m),165 and Cs is the number of receptors on the ligand-bound surface within the cell contact area. 0r kT fc c [1-9] r0 represents the separation distance between receptors at the mini mum breaking force, and c= KCs, where K is the equilibrium constant between the cell receptor and ligand.

PAGE 42

42 When substituting KCs in for the c term and Kd= 1/K for the K term, Kd is inversely proportional to fc, meaning a lower Kd (high affinity) will require a stronger force to break the ligand/cell bond.164 Another similar way to predict cell adhes ion is by comparing the bonds formed during a collision (B*) to the number of bonds required to resist the forces acting against the bonds (Bc). When B*>Bc, cell adhesion to the surface will occur.165 Collisions of cells with the surface will commence duri ng a specified amount of time called the collision duration ( c), in which a is the radius of t he contact area (equation 1-10). This is applied to equation 1-11 to determine B*, in which case B is the density of bonds formed. Bc is determined by the Stokes force and t he critical force required to break the ligand/receptor bond (equation 1-12).95 v ac [1-10] BABcc* [1-11] c fr cf F B [1-12] Following cell capture, the ce lls can be eluted from the capillary and quantified by off-line measurements using tools such as a hemocytometer, or retained for additional studies. Traditional CAC elution methods include addition of a competitive agent or implementing high mechanica l shear rates, however, bubble-induced detachment has recently shown attractive attributes.91 The air/liquid interface exerts shear forces on the captured cells, breaking the cell/ligand bonds, and the cells can be eluted directly without further dilution.166 The bubble-induced detachment method also demonstrates a

PAGE 43

43 higher elution efficiency than previous elut ion methods, and higher cell viability for applications where the eluted cells are cultured.91 Overview of the Dissertation The data presented in this dissertati on will d emonstrate the utility of DNA aptamers for two different bioanalytical appli cations. Chapters 2-4 will detail the work performed to generate an aptamer which can se rve as an antidote to the anticoagulant bivalirudin. Chapter 2 f eatures some in depth general discussion of several SELEX steps with a bead-based selectio n strategy, while Chapters 3-4 describe an aptamer selection and characterization using a mono lithic column. In C hapter 5, the focus switches to proof-of-concept studies for the aptamer-based capture of leukemia cells in a square capillary. Chapter 6 illustrates the improvements in the system operation upon optimization, particularly highlighting an increase in capture efficiency, leading to detection of leukemia cells in blood samples. The capture of two different colon cancer cell lines by the specific aptamer for each is also described with this system. The concluding chapter summarizes the global si gnificance of this work, and outlines prospective directions of the research.

PAGE 44

44 CHAPTER 2 METHODOLOGY AND DEVELOPMENT OF BEAD-BASED APTAMER SELECTION Introduction Anticoagulation with Bivalirudin Bivalirudin is an anticoagulant used as an alternative to heparin in patients with unstable angina undergoing perc ut aneous coronary intervention (PCI).167 The drug is a bivalent direct thrombin i nhibitor, binding to both exos ite 1 and the active site of thrombin with an inhibition constant of 2 nM. The 20 amino acid peptide has a molecular weight of 2180 Da, and is cleaved at the Arg3-Pro4 bond with a half life of approximately 25 min. Among the most important advantages of using bivalirudin in place of heparin (described in detail in Chapter 1) are that it inhibits soluble and clotbound thrombin and it will not instigate HIT. Approximately 600,000 (5%) patients out of an annual total of 12 million receiving he parin develop HIT and can no longer continue heparin administration.168 Despite the many advantages, one reason why the use of bivalirudin has not become more widespread is the co st of the treatment. Admi nistration of the drug could reach upwards of $2000, while a 24 hour dose of heparin is only $3.50.131, 132, 169 However, these estimates do not account fo r possible cost savings due to decreased bleeding and lower need for monitoring from bivalirudin, which have been reported as ~$400 lower with bivalirudin than other inhibitors.134 Perhaps the most significant explanati on of why bivalirudin use is not more prevalent is because an antidote to bivalirudin is not currently available. This critical issue is the inspiration for much of the work performed in this di ssertation. For this project, an aptamer antidot e will be selected for bivalir udin which can restore

PAGE 45

45 coagulation activity in the instance of severe patient bleeding. Th e aptamer is expected to function according to the schematic depicted in Figure 2-1.128 The selected aptamer may compete with bound thrombin for peptide binding, releasing the drug from the complex and restori ng coagulation, and/or inhibit fr ee peptide from interacting with thrombin. High-affinity aptamers (similar affini ty to the drug/thrombin complex) will likely display both functions, while lower affini ty aptamers may still be able to operate according to the second mechanism. Aptamer ThrombinPeptideAptamer interacts with thrombinbound peptide, releasing it from complex Aptamer binds free peptide, inhibiting binding to thrombin Figure 2-1. Mechanisms of apt amer/peptide interactions for restoring coagulation. The aptamer may compete with thrombin for peptide binding, releasing the drug from the complex (left) and/or bind to t he free peptide, render ing it unavailable to interact with thrombin (right).128 Aptamer Development Aptamers are generated via an in vitro process known as SELEX, depicted in Figure 1-3. The critical step determining the efficiency of the selection is the partitioning of the binding sequences from the nonbinding sequences. A vari ety of methods have proven effective for this purpose, including ni trocellulose membranes, affinity capture, column immobilization, cross-linking, gel electrophoresis, i mmunoprecipitation, centrifugation, and capillary electrophoresis.77 However, many of these methods rely on

PAGE 46

46 a significant difference between the bound a nd unbound forms of aptamer/target, which is problematic with small molecules such as peptides because the difference between the bound and free DNA is frequently insuffici ent for separation. For example, nitrocellulose membranes employ a molecular weight cutoff in which the disparity between the free aptamer and bound complex is not large enough for the nonbinding sequences to be partitioned by the membrane Similarly, gel electrophoresis and capillary electrophoresis separate samples based on a change in electrophoretic mobility between the bound and unbound sequences. This difference may not be enough for separation with small molecule targets depending on the properties of the system. This work focuses on using affinity tags to capture binding sequences with micrometer-sized beads, followed by centri fugal separation from nonbinding sequences. Specifically, the work presented in this chapter outlines the SELEX steps and uses a streptavidin-coated bead-based partitioning method to se parate DNA/biotinylated complexes from nonbinding sequences to select an aptamer for bivalirudin. The SELEX process is divided into severa l steps which are generally applicable to any selection, although the exact met hods used may change. These subcategories include: 1) Experiment al design; 2) Design of PCR prim ers and library; 3) Optimization of PCR conditions; 4) SELEX procedure; 5) PCR procedures; 6) Monitoring selection progress; 7) Sequencing and post-selection c haracterization studies (discussed later). Materials and Methods Experimental Design Partitioning of the binding sequences from nonbinding s equences was achieved by immobilization of the DNA/biotinylated tar get to streptavidin-coated micrometer-sized

PAGE 47

47 beads, then centrifuging the beads and disc arding the supernatant liquid. The streptavidin/biotin method was chosen for furt her study because the interaction exhibits a femtomolar (~10-15 M) dissociation constant, enabling an efficient capture of biotintagged species.170 The peptide was biotinylated at the N-terminus by an overnight incubation with Sulfo-NHS-SS-biotin (Pierce) in PBS at a 1:0.8 drug-to-biotin ratio. The limiting amount of biotin was to assure the drug would not be doubly biotinylated, possibly altering the properties. The N-terminus was utilized for capture (and immobilizat ion in Chapter 3) due to the drug structure and mechanism of bind ing discussed in Chapter 1. Since the drug binds to exosite 1 via the C-terminus prio r to active site bindi ng, aptamer inhibition of C-terminus binding will most likely be effe ctive in reducing the anticoagulant effect. Also, most of the tertiary structure of bivalirudin whic h presents a probable area for aptamer binding is observed at the C-terminus. Therefore, this re gion should remain free for aptamer binding. Next, the biotinylated drug wa s purified on an analytical scale using reverse-phase HPLC (1.0 mL/min; C18 column; 4.6 mm ID; 25 cm l ength; a gradient beginning with water/TFA (100/0.1), and continuing to acetoni trile/TFA (100/0.1), 10 -80% over 32 min, monitoring at 214 nm). The samples corre sponding to the major peaks were analyzed by MALDI/TOF ( -cyano-4-hydroxycinnamic acid matr ix) in order to determine which fraction contained the correct mass-to-charge ratio (m/z ~2569). A negative ninhydrin test, (a test for a free -amino group) indicated that t he N-terminus was biotinylated instead of amino acid side-chains. Finally the process was repeated on the preparative scale to generate enough product for the ensui ng research. All peptide purifications

PAGE 48

48 and mass spectrometry were perfo rmed at the Interdisciplinar y Center for Biotechnology Research (ICBR) at the Universi ty of Florida (Gainesville, FL) Design of PCR Primers and Library The design of a functional set of primer s and DNA library is a cruci al step in determining the efficiency of the PCR, in order to optimize the yield and avoid mispriming. One consideration in primer design is nucleoti de length, which is typically around 20 nucleotides. This length is long enough to specifically recognize the priming region of the template, yet short enough to bind at annealing tem peratures from 5260C. This temperature is optimal because it is sufficiently lower than the polymerase extension temperature of 72C, so ext ension will not begin until the primers have annealed. Also, the GC content must be care fully controlled in order to maintain a desirable melting temperature (Tm), which is approximately C of the annealing temperature used. The primers are designed to have similar Tm (within 1C), and avoid stretches of repeated sequences or base runs (such as AAAAAA or CGCGCG) to encourage the proper prim er annealing positions. Finally, the primers should not demonstrate significant secondary struct ure formation (hairpin) near the Tm, or form either homoor heterodimers. These intera ctions will lower the yield of the PCR by decreasing the amount of primers available for annealing to the te mplate strand. The primers and library were designed using IDT OligoAnalyzer software. The forward (sense) primer was labeled with 5-FAM (5/6-carboxyfluorescein) for monitoring the progress of the selection by flow cytometry. The reverse (anti-sense) primer was 5-biotinylated for capture by streptavidin coated beads used for preparing ssDNA for subsequent selection rounds after PCR amplification.

PAGE 49

49 The library template is constructed based on the sequences of the primers. The library consists of two PCR primer regi ons flanking a random region of ~30-50 sequences. This random region generally cont ains sufficient sequences to ensure the complexity of the initial pool (complexity= yN, where y is the num ber of nucleotides possible (4), and N is the length of the random region) is high enough that unique sequences are present in the initial pool. This is necessary because the number of sequences capable of significant tar get binding represents only 1 in 109 to 1 in 1013 of molecules present in the starting library.77 The 5 primer region is the same sequence as the sense primer, while the 3 primer region is the complement of 3-5 anti-sense primer since it will prime and extend t he anti-sense template strand. Based on the primer desi gn guidelines, the followin g primers and DNA library template were devised: Forward primer (sense): FAM-CTC ATG GAC AGG CTG CAG AC Reverse primer (anti-sense): Biot in-CTG TAG TGG CAT CCG AGC GT FAM-CTC ATG GAC A GG CTG CAG AC-(N)40 -ACG CTC GGA TGC CAC TAC AG All oligonucleotides were synthesized us ing standard phosphoramidite chemistry on an ABI3400 DNA/RNA synthesizer (A pplied Biosystems). DNA pur ification was performed with a ProStar HPLC (Varian) using a C 18 column (Econosil, 5U, 250 mm 4.6 mm) from Alltech Associates. UV-Vis measur ements to measure DNA concentration were performed with a Cary Bio-300 UV spectrometer (Varian). Optimization of PCR Conditions The annealing temperature of the primers used throughout the remainder of the selection is the highest temperature that gives efficient amplification with minimal nonspecific binding in the desired temperature range. T he higher annealing

PAGE 50

50 temperature increases the likelihood that t he primers will bind to their expected priming regions, since mispriming positions will have a lower Tm. All PCR mixtures (Takara) contained 50 mM KCl, 10 mM Tris HCl (pH 8.3), 1.5 mM MgCl2, dNTPs (2.5 mM), 0.5 M each primer, Hot-start Taq DNA polymerase (5 units/L), and the library concentration was 50 pM. Temperatures from 55.0C-63.0C were compared by gel electrophoresis, with 60C indicated as optimal. The following conditions were used for the initial test of primers and library: initial hot start temperature of 94.0C for 3 min, repeat ed cycles (30 sec each) of 94.0C for denaturation, annealing at 60C, then 72.0C for elongation, and final extension at 72.0C for 5 min followed by an indefinit e hold at 4.0C (Biorad iCycler). Selection Procedure The first round of selection bega n with a very large amount (~1014 sequences) of the initial DNA library incubated with the ta rget. The multitude of sequences increases the likelihood that at least one of the sequences in the pool will bind specifically to the target. The library was heated at 95C for 5 minutes and snap cooled at 4C to ensure the sequences were not dimerizing and we re folded into their most favorable conformations. Fifty pmol of biotinylated drug and 1 nmol of the initial library were incubated for 30 min in 200 L physiological buffer (PB: 25mM Tris-HCl (pH 7.5), 150 mM NaCl, 5 mM KCl, 1 mM MgCl2, 1 mM CaCl2, 1% BSA, 0.1% t RNA). One mL of streptavidin-coated beads (Bangs Labs, 5.6 m, polystyrene core) was centrifuged at 10,000 rpm for 3 min and washed twice with 1 mL of PB. The drug/library mixture was then added to the washed beads and incubated for 10 min. The particles were centrifuged and washed twice with both 1 mL of PB and 1 mL of WB (500 mL PBS buffer and 5 mM MgCl2) in order to remove nonbinding species, then dispersed in 1 mL

PAGE 51

51 of water. The binding oligonucleotides were heat-eluted for 5 min at 95C and centrifuged to recover t he supernatant liquid which contained the binding DNA sequences. This DNA was then amplified by PCR. PCR Amplification Procedure The first pool is PCR amplified for the nex t selection round in a series of 3 steps: a. The entire pool is amplified b. The number of cycles is optimized c. Preparative PCR is carried out according to (b). However, in any subsequent rounds, only steps (b) and (c) are necessary. PCR amplification of entire first pool PCR serves as a step to amplify the bi nding sequences of each round for use i n the next round of selection. The first round requires specia l attention because there is only one of each sequence present in the pool after incubation with the target. Therefore, the entire pool is subjected to the PCR in order to amplify all possible sequences for the next round. Table 2-1. PCR preparation of entire pool Reagent Volume ( L) 10X PCR Buffer dNTP (2.5 mM each) Primers (10 M each) DNA pool DNAse Free Water Taq Polymerase 100 80 50 200 570 3 A 1000 L amplification reaction volume was set up as shown in Table 2-1. The volume of water was adjusted to compensate for the total volume in instances where the DNA pool volume varies. The PCR reaction components are mixed thoroughly and 200 L of the solution was pipetted into 5 i ndividual 0.5 mL PCR tubes. The mixtures

PAGE 52

52 were PCR amplified using the P CR amplification program in Figur e 2-2. After this initial PCR amplification, all PCR products from individual tubes were pooled together and used as the template for the next PCR procedure. 94C94C 60C 72C72C 4C 3 min30 sec 30 sec 30 sec5 min1X 15X1X AB CCL 12 15 21 2515 L Figure 2-2. PCR amplification of initial library. A) PCR amplification pr ogram; B) Cycle optimization; C) Preparative PCR. Cycle optimization fo r preparative PCR Table 2-2. PCR cycle optimization Reagent Volume ( L) Control ( L) 10X PCR Buffer 100 5 dNTP (2.5 mM each) 80 4 Primers (10 M each) 50 2.5 DNA pool 200 DNAse Free Water 570 38.5 Taq Polymerase 3 0.15 In this step, the optimal number of P CR cycles is determined for the preparative PCR that generates DNA for the next round. The purpose of this step is to perform enough cycles to maximize the amount of DNA retrieved without nonspecific amplification. The amplif ied PCR pool served as the template and accounted for 10% of the total reaction volume. Fifty L of the reaction mixt ure amplified under the conditions in Table 2-2 was aliquoted into separate 0.5 mL PCR tubes and was amplified by the program shown in Figure 2-2. At the end of various PCR cycles (12,

PAGE 53

53 15, 21, or 25 for this selection), one PCR tube was removed from the instrument. The PCR products were analyzed by 3% agarose gel electrophoresis, and the cycle number displaying optimal amplificat ion (15 cycles) was utilized fo r the preparative PCR step (Figure 2-2B). Preparative PCR This PCR amplification step is used to generate enough ssDNA for the second round of selection. Usually ~1000 L of tota l PCR reaction volume is sufficient, although higher volumes may be used in certai n cases. The reaction components for 1200 L reaction mixture are shown in Table 2-3. Once again, the PCR reaction components were mixed thoroughly and 200 L of the solution was pipetted into 6 individual 0.5 mL PCR tubes. The mixtures underwent amplificatio n using the program shown in Figure 2-2 at the optimum cycle number (15 cycles) and were pooled together for ssDNA preparation after analysis by gel electrophoresis (Figure 2-2C). Table 2-3. Preparative PCR Reagent Volume (uL) Control (uL) 10X PCR Buffer 120 20 dNTP (2.5 mM each) 96 16 Primers (10 uM each) 60 10 DNA pool 120 DNAse Free Water 804 154 Taq Polymerase 3.6 0.6 Gel electrophoresis monitoring of products Agarose gel electrophoresis is used in a selection procedure for analysis of cycle optimization and to ensure that satisfacto ry amplification has occurred following preparative PCR. The mobility of a DNA fragment is dependent on the size, charge, and conformation of the molecule and the str ength of the electric field applied. Additionally, gels containing different percentages of agarose will influence the

PAGE 54

54 separation of fragments. Fo r aptamers of ~80-100 bases, 3% agarose is sufficient for separation of PCR amplicons. The intercalat ing UV dye ethidium bromide (EB; 0.005% v/v) was added to a solution of the 3% agar ose (w/v) in 40 mL of tris-borate-EDTA buffer (TBE buffer) for visualization. Ten L of each PCR product analyzed was combined with 2 L of 6X Blue/green loading dye. The 25 bp ladder was prepared by adding 2 L of 6X Blue/green Loading dye, 1 L of ladder, and 9 L of deionized water. The samples were each loaded into a different lane, and the electrophoresis conditi ons are set to 100 V, 2 A, for 40 min. Preparation of ssDNA PCR amplification gener ates double-stranded DNA (dsDNA), and must be converted to ssDNA before the next round of selection, since only the sense DNA strands have been selected to bind to the targe t. This is accomplished by capturing the amplicons on streptavidin-coated beads via the biotiny l ated antisense primers which have been incorporated into the antisense DN A strand following PCR. Adding NaOH ruptures the hydrogen bonds of the strands, and the sense st rands are eluted from the beads while the antisense strands remain attached. Streptavidin-coated Sepharose beads (G E Healthcare; 300 L) were added to a column (Glen Research Expedite Style) bloc ked by two frits (1 m pores). The beads were washed with 4 mL of Dulbeccos PBS buffer by syringe, and the entire PCR product was added to the column. The plunger was loosely inserted into the syringe to control the flow rate of the PCR product in order to allow the biotinylated strands ample time to bind to the streptavidin. The PCR product flow-through was collected and run through the column two additional times to ensure maximum collection of the DNA on the beads. The column was washed with 6 mL of PBS buffer, then 0.5 mL of 0.2 M

PAGE 55

55 NaOH was added to elute the ssDNA product. In each step, the syringe was removed from the column, followed by plunger remo val to minimize disturbance of the beads. Desalting of ssDNA The exc ess salt used to denature the ds DNA must be removed from the ssDNA product before the next round of selecti on. A Sephadex G-25 DNA Grade NAP-5 desalting column (GE Healthcare) was was hed with 3 column volumes of water. The ssDNA product was added to the column and allowed to drain, and desalted ssDNA was eluted and collected in 1 mL of DNAse-fr ee water. The concentration of the ssDNA was determined by UV absorbance at 260 nm and the ssDNA was dried by vacuum dryer. The DNA pellet (~80 pmoles after the first round) was reconstituted in 500 L PB for the second round of selection. Second and Succeeding Rounds of Selection The SELEX procedure leading to the generation of aptamers is an evolutionary process beginning wit h a large am ount of initial DNA library. From beginning to end, the population of sequences binding specifically to the target increases, until evolution of a final pool containing a majori ty of binding sequences. Through the various selection rounds, the stringency of the process is increas ed in order to select aptamers with the most desirable properties. This is carried out in a variety of ways throughout the progression of the selection including: 1) Incr easing the ratio of libra ry to target. This increases the competition am ong the sequences for fewer target molecules so the higher affinity oligonucleotides will be preferentially retained; 2) Decreasing the incubation time. The sequences with the most favorable binding kinetics will bind the target; 3) Increasing the wash volume. In this manner, the higher amounts of BSA and tRNA will compete and remove weakly bi nding sequences from the target; 4)

PAGE 56

56 Introduction of a negative selection step. Th is will remove the sequences that are not specific for the target. For this work, this means removing the sequences which bind to the streptavidin-coated beads. The point in which a negative selection is instituted varies betw een researchers. Some groups perform this task before the initial positive selection in order to immediately subtract out any sequences t hat are not specific to the target.171 However, this runs a high risk of losing potential aptamers due the number of unique sequences present in the initial pool and the inefficiency of partitioning methods.172 Therefore, a negative selection step is typically introduced later in the selection, when higher numbers of specific binders are present in the pool. Table 2-4. Conditions for initial SELEX Round Amount B-Drug Amount Library Lib:Drug Incubation Time Wash Volume Bead Volume 1 50 pmol 1 nmol 20.0:1 30 min 1 mL 1 mL 2 10 pmol 50 pmol 5.0:1 30 min 500 uL 200 uL 3 10 pmol 50 pmol 5.0:1 30 min 500 uL 200 uL 4 10 pmol 50 pmol 5.0:1 30 min 500 uL 200 uL 5 10 pmol 50 pmol 5.0:1 30 min 500 uL 200 uL 6 7.5 pmol 50 pmol 6.7:1 30 min 500 uL 150 uL 7 7.5 pmol 50 pmol 6.7:1 30 min 500 uL 150 uL 8 5 pmol 50 pmol 10.0:1 25 min 750 uL 150 uL 9 5 pmol 50 pmol 10.0:1 20 min 750 uL 100 uL In the first selection, no negative select ion was used. The conditions for this selection are shown in Table 2-4. As the selection progressed from the second to the final round, the library-to-drug ratio increased from 5:1 to 10:1, the incubation time decreased from 30 minutes to 20 minutes, a nd the wash volume increased from 500 L to 750 L. The bead volume was decreased in order to lower the instances of nonspecific binding to the streptavidin and increase com petition between DNA sequences as the ratio of library-to-drug increased.

PAGE 57

57 Flow Cytometry Monitoring of Selection Progressi on During selection, it is necessary to m onitor the progress of the pool evolution in order to determine if conditions are appropr iate and to establish when the pool appears to be maximally enriched. O ne convenient method for selection monitoring is flow cytometry, in this work facilitated by c apture of the FAM-DNA/biotinylated drug complexes by the streptavidin coated beads. Fl ow cytometry uses the principles of light scattering, light excitation, and light emission of fluorophores to gat her specific data on particles from ~0.5-40 m. Particles ent ering the instrument are hydrodynamically focused by a flowing sheath of buffer to individually intersect a light source (typically an argon ion laser). Once the particles enter the laser beam, they scatter the light, and fluorophores present on binding sequences are excited to a higher energy state (Figure 2-3).173 This energy is released in the form of photons that em it with properties characteristic of each dye. Many flow cytometers have an optical setup which can detect dyes emitting at three different wavelengths or fluorescence channels; 525 nm (FL-1; FAM and FITC), 575 nm (FL-2; PE), and 620 nm (FL-3; PE/Cy5 tandem). The emitted light of each individual particle is converted into electrical pulses by the detector, then graphically displayed to prov ide information on particle size, complexity, and fluorescence intensity. In a selection, an increase in fluorescenc e intensity of an oligonucleotide pool when compared to the in itial library indicates that mo re of the fluorescently-labeled sequences are binding to the target, resulting in an increasingly enriched pool. Generally pool monitoring does not begin until around the third round of selection because pools from earlier rounds are ty pically not enriched enough to generate a measurable increase in fluorescence. Also the measurements are often performed in

PAGE 58

58 batches of 2-4 rounds to ensure the conditi ons are comparable. As the selection progresses past the initial rounds, one genera lly observes the pools gradually increase in fluorescence, indicating that the selection conditions are appropriate for enriching the pool. The selection is considered complete when the following conditions are met: 1) There is a significant increase in fluoresc ence intensity between the unselected initial library and the pool; 2) 2-3 consecutive rounds of selection do not further increase the fluorescence intensity of the pool. Figure 2-3. Schematic of flow cytometry instrumentation.173 A Becton Dickinson FACScan flow cytomet er was utilized for study. Each assay involved a 30 minute incubation of 2.5 pmol of the biotinylated drug with 250 nM FAMaptamer pool in 50 L total volume. Fifty L of 5.6 m streptavidin-coated beads were washed twice with 50 L of PB, then incubated for 30 min with the dr ug/pool mixture. To test pools for binding to the beads, the sequences were directly incubated with

PAGE 59

59 beads for 30 min. The beads were washed twice with 50 L of PB, twice with 50 L of WB, and finally suspended in 400 L of WB. A total of 20,000 events were counted for each sample. Flow cytometry for DNA blocking step The concentration of an unl abeled random library (D3B) required to saturate the streptavidin was determined by flow cytometry assay. Here, diffe rent molar excesses (2X-200X) of library D3B compared to the calculated amount of streptavidin on the beads wer e first incubated with the beads for 10 min. Another FAM-labeled library (D2; 50 pmole) was incubated with the beads/unlabel ed library for 30 minutes and assayed as described previously. Flow cytometry for TCEP cleavage To optimize the TCEP cleavage conditions, the basic flow cytometry assay outlined above was followed with these not able changes: 1) 2.5 pmole FAM-DNA-SSBiotin library was substituted for the drug incubation; 2) After incubation and washing steps, TCEP concentrations from 0.1-500 mM were added to the bead/DNA complexes and incubated for 30 min or 1 hr on a shaker. Results Enrichment of Aptamer Pool The results of flow cytometry assays for apt amer selection monitoring are typically displayed in the form of histograms (Figur e 2-4). This is a graphical plot of the fluorescenc e intensity of the FL-1 (FAM) ch annel versus the number of events counted. Figure 2-4A depicts a control histogram of the beads, an expec ted low intensity binder (FAM-labeled initial library), and a sequence expected to demonstrate a high amount of fluorescence (biotinylated and FAM labeled sequence). Minimal binding is observed for

PAGE 60

60 the initial library, but the biotinylated sequence displays significant fluorescence (~103). This is expected because the initial library will have few sequences binding since it is not enriched for the target and the biotinylated/FAM s equence will show increased binding because of the biotin binding to the streptavidin on the beads. The fluorescence of the biotinylated sequence is commonly said to have a fluorescence shift (referring to the relative positions along the x-ax is) compared to the initial library. Beads Biotin-DNA-FAM Library Library 8th Round 9th Round Library + Library 6thRound + 6thRound 7thRound + 7thRound 8thRound + 8thRound Library 5thRound 6thRound 7thRound C A B D Figure 2-4. Flow cytometry dat a of initial selection. A) Beads, library, and positive control; B) Rounds 5-7; C) Rounds 8-9; D) Rounds 6-8 incubated with either beads alone (-) or drug with bead capture (+). The shift between the initial library and each of the pools is the major area of interest, so the particle study (beads alone) is not consi dered in additional assays. Figure 2-4B is the binding assay of rounds 5-7, and a noticeable shift is observed for

PAGE 61

61 round 7. Therefore, the stri ngency of the selection was raised for round 8 in terms of library-to-target ratio, decreased incubation time (further decreased in round 9), and increased wash volume. After several roun ds of selection, the pool appeared to be enriched, as evidenced by a decrease in fluo rescence in round 9 (Figure 2-4C). This fluorescence decrease between rounds 8-9 may be due to a loss of some of the lower affinity binders due to the selection stringency. Despite the increase in fluorescence in tensity, the binding could actually be caused by the pools binding to the streptavidin on the beads instead of the drug target. Therefore, a flow cytometry assay comparing the shifts of each pool with (+) and without (-) drug present was performed (Figure 2-4D). The shifts between each pool were nearly identical, implicating the beads as the source of binding rather than the desired drug target. Streptavidin is much larger t han the drug target, which likely results in more binding sites available to aptamers. To reduce this binding, negative selection against the beads may subtract out sequences binding to the streptavidin. Selection Modification #1 Table 2-5. Conditions fo r selection modification #1 Round Amount B-Drug Amount Library Lib:Drug Incubation time Wash Volume Bead Volume Bead Volume 6 7.5 pmol 50 pmol 6.7:1 30 min 500 uL 150 uL 75 uL 7 7.5 pmol 50 pmol 6.7:1 30 min 500 uL 150 uL 150 uL 8 7.5 pmol 50 pmol 6.7:1 30 min 500 uL 150 uL 150 uL 9 5 pmol 50 pmol 10.0:1 20 min 750 uL 100 uL 175 uL The 5th round of selection demon strated no significant sh ift, so a new selection instituting a negative selection was begun from the pool obtained after five rounds of selection. The conditions for this selection are observed in Table 2-5. The final column

PAGE 62

62 shows the volume of beads used for a 30 mi nute negative selection step, wherein the sequences that do not bind to the streptavidin beads are re tained for further steps. A B Figure 2-5. Flow cytometry data of selection modification #1. A) Library and round 8A (without negative selection) incubated with either beads alone (-) or drug with bead capture (+); B) Rounds 89B (with negative selection). The flow cytometry assay of the selection progression (Figure 2-5) is divided into two figures to clarify the result s. In this assay, a small po rtion of the pool from round 8 was retained after the positive selection, and not subjected to the negative selection (Figure 2-5A). This is c onsidered round 8A, while round 8B is the pool after the negative selection was performed (Figure 2-5B). In this manner, it was determined that the negative step did subtract out some nontarget binde rs, as evidenced by the decreased fluorescence intensity after the s ubtraction. But the remaining sequences were still binding to the beads, since the shifts of round 8B with (+) and without (-) drug were still similar. The subtra ction step was also most likely insufficient to remove all of the high affinity sequences from the beads due to the shoulder of the histograms on Figure 2-5B. This shoulder likely corresponds to some of the sequences binding to the streptavidin prior to the subtraction step remaining in the solution.

PAGE 63

63 Selection Modification #2 For the next selection, two aspects were combined to inc rease the likelihood of the selection of a target-binding sequence: 1) A random initial library was incubated with the beads prior to the addition of the target/DNA complexes; 2) The disulfide bond incorporated into the biotin peptide label was cleaved with the reducing agent tris (2carboxyethyl)phosphine (TCEP) in order to retain only sequences binding to the peptide; 3) The library/primer set was c hanged to reduce possible contamination. 1 The random library must be one which is not amplified by the primers used in this selection. Random libraries D1A and D3B were tested by the PCR amplification method described above for 30 cycles and analyzed by agarose gel electrophoresis. The D3B library did not appreciably amplify (Figure 2-6), and was utilized for further study. 1234 Figure 2-6. PCR of unlabeled DNA library. 1) 25-bp ladder; 2) Negative control; 3) 500 pM library (no PCR); 4) PCR of 500 pM library. PCR did not amplify the library. Figure 2-7 (representative of results in triplicate) shows that the increasing amounts of unlabeled blocking DNA actually increased the amount of labeled random library binding to the surface. We hypothesize that this is due to the labeled library

PAGE 64

64 actually binding to the blocking library, thus increasing the fluorescence when more blocking library is present. Therefore, this method of attempting to nonspecifically block the streptavidin with unlabeled DNA wa s abandoned, and selection was performed by TCEP cleavage. Figure 2-7. Streptavidin blo cking with excess random library. 1 2 5 1 1 2 -Control + Control 0.1 mMTCEP 1.0 mMTCEP 25 mMTCEP 50 mMTCEP 100 mMTCEP 150 mMTCEP 200 mMTCEP + Control Lib, no biotin 25 mMTCE P 100 mMTC E 200 mMTC E 500 mMTC E -Control + Control Lib, no biotin 25 mMTCEP 100 mMTCEP 200 mMTCEP 500 mMTCEPA B Figure 2-8. TCEP concentration optimization. A) 30 min TCEP incubation; B) 1 hr TCEP incubation. TCEP concentration was optimized to selectively partition sequences binding solely to the target. The decrease in fl uorescence from the positive (+) control (no TCEP added) as higher concentrations of TCEP were added can be observed. This correlates to more of the DNA being cleav ed from the beads. Although the signal did Unlabeled + SA Labeled + SA 2X Excess 10X Excess 25X Excess 75X Excess 100X Excess 200X Excess

PAGE 65

65 not completely decrease to the levels of the negative (-) control, some of the cleaved DNA may remain in the system and possibly bind to the streptavidin. Nonspecific binding of a random library to the streptavidin is demonstrated in Figure 2-8B (blue line), where the nonbiotinylated library produces appreciable binding to the beads. Based on the results, 200 mM TCEP was used for all further studies. Further, the effect of TCEP on either PCR amplificatio n or biotin/streptavidin binding was assessed. A 15-cycle PCR was per formed on the initial library either with or without 200 mM TCEP. Only the tubes devoid of TCEP were successfully amplified. When samples containing the TCEP were des alted to remove TCEP using either a column or centrifugal filter, t he DNA was successfully amplified. For the streptavidin/biotin interference studies (Figure 2-9), the assay conditions were the same as the TCEP cleavage studies (using a FAM-DNA-SS-Biotin library), but the parameters consisted of: 1) + control (2X): only beads and FAM-DNA-SS-Biotin libraryexpect large shift; 2) Beads/DNA/T CEP cleavageshould have a lower intensity than (1); 3) DNA and TCEP incubation followed by bead capturesimilar to (2); 4) Incubate beads and TCEP, followed by DNA capturesimilar to (1) if the TCEP does not effect DNA binding; 5) control: beads onl yminimal fluorescence. The histogram in Figure 2-9 shows the expected results. When TCEP is incubated with the beads first, followed by DNA capture (#4, brown line), the streptavidin/b iotin binding is the same as the intensity of the beads di rectly capturing DNA withou t TCEP incubation (#1, green and darker blue lines). Additionally, when the beads, DNA, and TCEP are all incubated together (#2, purple line) or when the DNA and TCEP are incubated first, followed by

PAGE 66

66 bead capture of the DNA (#3, light blue line), the cleavage is similar. Therefore, we concluded that the TCEP does not a ffect biotin/streptavidin binding. Beads DNA + Beads -Control + Control A + Control B Beads+DNA+TCEP DNA+TCEP, Beads Beads+TCEP, DNA Figure 2-9. TCEP interference wit h streptavidin/biotin binding. Preparation of new library and primers Figure 2-10. PCR amplification program for second in itial library. To reduce the effect of contamination from previous selections, a new set of primers and library previously designed by a coworker were synthesized and purified as described previously.174 The PCR annealing temperature is 55C for this library set, and the new PCR protocol is shown in Figure 2-10. 95C 95C 55C 72C 72C 4C 2:30 min 30 sec 30 sec 30 sec 2 min 1X 14X 1X

PAGE 67

67 Forward Primer (sense): 5-FAM-ATC GTC TGC TCC GTC CAA TA Reverse Primer (anti-sense): 5' -Biotin-GCA CGA CCT CAC ACC AAA 5'-ATC GTC TGC TCC GTC CAA TA -N45TTT GGT GTG AGG TCG TGC Selection Table 2-6. Conditions for select ion modification #2 with new library Round Amount B-Drug Amount Library Lib:Drug Incubation Time Wash Volume Bead Volume Bead Volume 1 50 pmol 1nmol 20.0:1 30 min 1 mL 1 mL N/A 2 10 pmol 50 pmol 5.0:1 30 min 500 uL 200 uL 100 uL 3 10 pmol 50 pmol 5.0:1 30 min 500 uL 200 uL 150 uL 4 10 pmol 50 pmol 5.0:1 30 min 500 uL 200 uL 200 uL The next selection aimed to incor porate the TCEP cleavage and negative selection into the method. D ue to the inefficiency of the TCEP cleavage, this step would be implemented in round 5, so s equences are not lost early in the selection before the pool has evolved. The negative selection step began in round 2, and other selection conditions are in Table 2-6. Poo Po o Po o Po o Library + Library Round 3+ Round 3Round 4+ Round 4Figure 2-11. Flow cytometry of selection modification #2. Pools were either incubated with either beads alone (-) or drug with bead capture (+). Flow cytometry analysis of the pools began in round 3 (Figure 2-11). The pools containing drug (+) demonstrated at least as much fluorescence intensity as the pools

PAGE 68

68 incubated directly with the beads (-). Once agai n, the pool was enriched for streptavidin despite the negative selection. Selection Modification #3 Table 2-7. Conditions fo r selection modification #3 Round Amount B-Drug Amount Library Lib:Drug Incubation time Wash Volume Bead Volume Bead Volume 2 10 pmol 50 pmol 5.0:1 30 min 500 uL 200 uL N/A 3 10 pmol 50 pmol 5.0:1 30 min 500 uL 200 uL 200 uL 4 10 pmol 50 pmol 5.0:1 30 min 500 uL 200 uL 350 uL The procedure used to remove the sequences binding to the streptavidin clearly needed to be more stringent. When the same round 1 conditions as shown in Table 2-6 were implemented, but a TCEP cleavage was performed in the fi rst round, no PCR amplification was evident even after 33 cycles. Therefore, in the next selection, the cleavage was not implemented unt il round 2. This loss of sequences was believed to be balanced against a less stringent negative se lection procedure in the second round. This selection used the round 1 selected pool from selection modification #2 to begin the TCEP cleavage at round 2 (T able 2-7). However, beginn ing in the third round, a double negative selection was implemented, in which the pools were incubated with the beads for 30 min for sequence subtraction first, followed by regular positive selection on the drug target, then a se cond 30 minute incubation of the binding sequences with the beads. Despite the modified conditions used for this selection, the flow cytometry data (Figure 2-12) proves that the pool is still enriched for the streptavidin instead of the drug. This data implies that a new support system without streptavidin may be better suited to select an aptamer for the much smaller drug target.

PAGE 69

69 P oo l 3+ Pool 3Pool 4 + Pool 4Library Round 3+ Round 3Round 4+ Round 4Figure 2-12. Flow cytometry data of selection modifica tion #3. Pools were either incubated with either beads alone (-) or drug with bead capture (+). Conclusions The method provided in this chapte r proved unsuccessful for generating an aptamer for the drug target. The overlying explanation is believed to be due to the streptavidin-based partitioning system that the procedure relied on for separating the binding and nonbinding sequences. Based on si ze alone, streptavidin is ~25 times the molecular weight of the peptide target. This larger size and variety of functional groups enables the structure to form multiple binding pockets which the aptamer may find more favorable than the limited folding of the peptide. Additionally each streptavidin molecule is actually a tetramer, so many favorable binding sites found on the monomeric form are actually present 4 times (depending on how monomers assemble into the tetramer structure, some sites may not be ava ilable), and additional binding regions formed by the interaction of the m onomers is likely. Therefore, it is not surprising that the pool preferentially bi nds to the protein. Furthermore, the tetramerization of the biotin conjugates may lead to undes irable interactions that reduce binding of the s equences to the drug.175 The PCR may also aid in enriching the pool for

PAGE 70

70 sequences that bind to the streptavidin. W hen the pool is PCR amplified, the vastly greater numbers of sequences binding to the streptav idin overshadows the few sequences with specific binding to the drug. It is exceedingly interesting that none of the methods used to minimize the streptavidin binders were effective. Va rious degrees of stringency were utilized to wash, block, or cleave nonspecific binders from the pool. Possibly the increased stringency only served to remove sequences w eakly binding to the peptide, since small molecule selections typically generate lower-affinity aptamers than their more sizeable counterparts. Regardless of the true cause of the shortcomings result ing in pool enrichment for the beads, the consensus conclusion derived from the work is that streptavidin should be removed from the system. However, it is possible that the pools from the selections may contain high affinity streptavidin aptam ers, which may be of use for applications unrelated to this work. The next two chapter s will describe bivalirudin selection utilizing an alternative partitioning syst em composed mainly of polymeric (hydrophobic) materials.

PAGE 71

71 CHAPTER 3 MONOLITHIC SELECTION FOR BIVALIRUDIN Introduction Affinity columns have played an important role in gener ating aptamers for various small molecule targets. Some examples include an intial SEL EX study to generate aptamers to small organic dyes by Ellington and Szostak,29 substance P and amyloid peptide were covalently coupled to Sepharose 6B columns,55, 56 ATP was immobilized on agarose,176 and tripeptides were bound to HiTrap NHS-activated columns.177 Binding aptamer sequences were eluted us ing competitive binder s, cleavage of the bonds immobilizing the target to the column matrix, EDTA addition, high salt concentration, or column heating. The advent of monolithic columns has introduced new technology with inherent properties that may provide unique advant ages for affinity-based column aptamer selection. Monolithic columns are essentially constructed from one piece of material in which the area interacting with the analyte is mainly on the surface. This allows for mass transport by convection on the surface of the material, rat her than the diffusionlimited separations of traditional bead-based s eparations. This leads to a decrease in the pressure drop across the disk, allowing fo r faster flow rates and simple peristaltic pumps such as those found in low-pre ssure chromatography (LPC) instruments.178 Moreover, the hydrophobic backb one of the typically polymeric-based matrices has also been reported to have low nonspecific adsorption,179 and the target molecule can be covalently attached to the column by a variet y of chemistries, minimizing instances of target leaching.

PAGE 72

72 This chapter will explore the use of monolit hic columns as affinity matrices for aptamer selection for the anticoagulant bivalirudin. Firs t, the peptide was covalently linked to the column via the N-terminal segment in order to facilitate aptamer binding to the C-terminal region. A chromatographic me thod capable of elutin g binding sequences from the disk was then developed. Several r ounds of selection were carried out and the progress was monitored by real-time PCR. A selection on a blank disk with no peptide immobilized was compared to that of the drug-immobilized disk as a control. The highest-affinity DNA pool was sequenced, a nd individual sequences were tested for binding to the drug. Materials and Methods Column Immobilization Buffersbinding buffer (BB): PBS buffer (without MgCl2 and CaCl2) with 5 mM MgCl2 (final concentration); elution buffer (EB): BB with 1 M NaCl. All buffers and reagents added to the monolithic column were first filtered with a 0.22 m cellulose nitrate filter (Corning). Bi valirudin (Angiomax; The Medicines Company) was received in lyophilized form as a gift fr om the Anesthesiology Departm ent at the University of Florida College of Medicine. The pept ide was immobilized on a poly(glycidyl) methacrylateco -ethylene dimethacrylate epoxyfunctionalized CIM disk (BIA Separations) via the N-terminus as per manufact urers instructions. Briefly, the column was inserted into the disk housing and equilibrat ed with 2 mL of BB at 1 mL/min, then 2 mL of EB at 1 mL/min, and finally 2 mL BB at 2 mL/min using a LPC system (Bio-Rad BioLogic LP). Drug immobilization began with a 2 mL wash with 0.5 M sodium phosphate buffer (pH 8.0) at a flow rate of 2 mL/min. The drug was dissolved in 5 mL of sodium phosphate buffer (pH 8.0) for a final concentration of 3.0 mg/mL. This solution

PAGE 73

73 was introduced into the disk (1 mL), and t he disk was covered and incubated with the flow through for 48 hours at room temperature in a small di sh. For the blank disk (no drug immobilized), the same buffer without drug was incubated wit h the disk, and the remainder of the protocol was followed. Fo llowing this step, the disk was washed with 2 mL 0.5 M sodium phosphate buffer (pH 8.0), and remaining epoxy groups were blocked by flowing 4 mL of 1 M ethanolamine through the disk and incubating overnight at room temperature in a dish. The reaction was que nched by washing the disk with 2 mL of 0.5 M sodium phosphate buffer (pH 8.0) with 1 M NaCl at 2 mL /min, and equilibrated for use by introducing 2 mL of 0.5 M sodium phosphate buffer (p H 8.0) at 2 mL/min. The column was washed in 2 mL BB and sealed in the column housing with the column blind fittings for storage at 4C for usage planne d less than 3 days after last usage, or washed with 4 mL of 20% ethanol and stored at 4C in a container with 20% ethanol if the next usage was planned for longer than three days. The UV absorbance at abs= 280 nm of the peptide solu tion (Bio-Rad SmartSpec Plus) was measured before and after immobilization to determine the amount of drug immobilized on the disk. The mass of peptide remaining in the solution was calculated and subtracted from the initial amount added fo r the immobilization to determine that 1 mg of peptide was immobilized on the disk. This agrees with literat ure values reporting 0.3-0.9 mg of pept ide immobilized.180, 181 Method Development The disk was placed into the column housing with luer connections which fastened to complementary fittings on the LPC device. The system was capable of collecting fractions by automated fraction collector (Bio -Rad BioLogic BioFrac Fraction Collector) with a calculated void time of 0.334 min. A gradient elution met hod of increasing salt

PAGE 74

74 concentration was tested based on a CIM di sk separation of plasmid DNA, and is shown in Table 3-1.182 An on-line conductivity detector was available for monitoring the increasing salt concentration of the gradient, while DNA el ution was visualized by UV detection at abs= 254 nm. Fractions were collected every 0.5 min beginn ing at 2 min. Table 3-1. Method development conditions #1 Step # Time (min) Buffer Fl ow (mL/min) Total (min) 1 1 BB 2 1 2 5 0-100% EB 2 6 3 2 EB 2 8 4 5 100-0% EB 2 13 5 3 BB 2 16 Table 3-2. Method development conditions #2 Step # Time (min) Buffer Fl ow (mL/min) Total (min) 1 1 BB 2 1 2 5 0-100% EB 2 6 3 2 EB 2 8 4 5 100-0% EB 2 13 5 7 BB 2 20 The procedure was systematically tested with DNA incubated wit h: 1) Monolithic disk; 2) Blank housing (no monolithic disk); 3) Control of the monolithic column with no DNA injection; 4) Blank (no drug) disk. In the first trial, the conditions in Table 3-1 were used to probe the DNA eluted from a 30 min in cubation of 1 mL of 250 nM random DNA library in BB with the disk. The first peak in the chromatogram (Figure 3-1A) corresponds to an air bubble, while the sec ond peak area contains DNA as confirmed by UV absorption (previously described). For the second trial, step #5 of Table 3-1 was extended to 5 min to allow the conductivity to level off. In the expe riment, 0.7 mL of a random 278 nM library was injected onto the hous ing (Figure 3-1B). In experiment #3, the monolith underwent t he 5 min wash protocol with no DNA incubation (Figure 3-1C). Finally, in experiment #4, the final wash step was extended to 7 minutes based on the

PAGE 75

75 results from the previous experiments that indicated the DNA may not be completely eluted (Table 3-2). One mL of 100 nM DNA was incubated with the blank disk for 30 minutes and subjected to the gradient in Table 3-2 (Figure 3-1D). The results indicate that the gradient is capable of eluting DNA binding to the monolith, but that the DNA may bind nonspecifically to the housing or monolith matrix, requiring a negative selection. Also, DNA was still eluting from the monolith when no additional DNA was added (Figure 3-1C), de monstrating that t he wash step should be increased. After any incubation with DNA, fr om this point the disks were incubated with 2 mL of 7 M NaCl for a minimum of 30 min to remove tightly-binding sequences before the next round of selection. Next, the protocol in Table 3-2 was te sted in triplicate to determine if the DNA from 1 mL of 100 nM solu tion incubation was reproducible. Figure 3-2 is a representative sample chromatogram showin g the peaks consistently observed with the program. Since the program reliably showed DNA el ution in the area of ~15 minutes, we wished to decrease the slope of the gradient during that time frame in order to capture DNA fractions in more tubes, and the fraction s were collected every 0.34 min instead of 0.5 min. Therefore, the method in T able 3-3 was devised and tested to elute DNA binding to the target after 2 mL of a 5 M DNA solution was incubated with the disk for 30 min. The chromatogram shown in Figur e 3-3 displays a prominent peak under the SELEX-like conditions in the expected time fr ame. The DNA from fraction 16 was PCR amplified using the protocol in Table 2-11 for 30 cycles, demonstrating that the UV peak was actually detecting DNA.

PAGE 76

76 A B Figure 3-1. Chromatograms from method development. A) DNA incubated with drug disk; B) DNA and column housing; C) Drug disk withou t DNA; D) DNA incubated with blank disk.

PAGE 77

77 Figure 3-1. Continued Figure 3-2. Confirmation of LPC method. C D

PAGE 78

78 Table 3-3. Method development conditions #3 Step # Time (min) Buffer Fl ow (mL/min) Total (min) 1 1 BB 2 1 2 5 0-100% EB 2 6 3 2 EB 2 8 4 5 100-25% EB 2 13 5 12 25-0% EB 2 25 Figure 3-3. Chromatogram under SELEX-like conditions. Preliminary Selection Selection-simulating conditions were utiliz ed to compare the elut ion profile of the DNA binding to the drug and blank disks. If these profiles differed, specifically if different peaks, or peaks with a larger area o ccurred in the drug disk elution, there was a strong likelihood that at least some of the sequences eluted from the peptide disk were binding specifically to the dr ug instead of to the polymeric matrix. A new set of primers and library ( now with a 46-mer random region) were synthesized as previously described with amp lification using the same PCR protocol as in Figure 2-11. An unlabeled sense primer was used to fac ilitate monitoring by real-time PCR (qPCR). Forward Primer (sense): 5-A TC GTC TGC TCC GTC CAA TA

PAGE 79

79 Reverse Primer (anti-sense): 5'-Biotin-GCA CGA CCT CAC ACC AAA (IDT) 5'-ATC GTC TGC TCC GTC CAA TA -N46TTT GGT GTG AGG TCG TGC For this study, 2 mL of 500 nM library DNA was incubated for 30 min with both the drug and the blank disk. The elution prot ocol from Table 3-3 was administered separately to both disks. Figure 3-4 shows an overlay of the drug and blank disks, where the black arrow shows the peak from t he drug disk, while the red arrows point to DNA from the blank disk. A clear differenc e between the chromatograms is observed in which the drug disk chromatogram has a taller and broader peak around 15 min than the blank disk. This served as a valid proof -of-concept study to proceed with the actual selection. Figure 3-4. Overlay of proof-o f-concept studies of selections on drug (black arrow) and blank (red arrows) disks. Selection Conditions The drug-immobilized or blank disk was washed for 5 minutes with both EB then BB at 2 mL/min. Two mL of 500 nM DNA library (1 nmol) in BB was heated to 95C for 5 min and snap cooled on ice. This DNA was incubated with the disk in a small dish for 30 min. The disk was then placed in the provided column housing, and connected to

PAGE 80

80 the LPC device. The elution protocol was implemented as described in Table 3-3. DNA elution from the disk was m onitored by UV detection at abs= 254 nm. The same procedure was followed for the blank disk, wit h a separate aliquot of DNA library. Fractions (containing ~680 L) were analyzed by qPCR (see method below), and those corresponding to the peak of the drug disk were combined and incubated with the blank disk for 30 min in a small dish. Nonbi nding sequences were amplified by PCR confirmed by 3% agarose gel electrophoresis. Each disk was rinsed and soaked in a saturated NaCl solution to remove any remaining bound sequences prior to the next round. The selected sense DNA strands were separated from the biotinylated antisense DNA by alkaline denatur ation and affinity purificat ion with streptavidin-coated beads to produce ssDNA that was desalted fo r the second round of selection. The methods for agarose gel electrophoresis, ssDNA production, and desalting are outlined in Chapter 2. Monitoring by Real-time PCR Real-time PCR was u tilized to prove the difference in DNA eluted from the drug and blank disks. qPCR monitors an increase in fluorescence intensity of SYBR green dye, which preferentially fluoresces when bind ing dsDNA, in real time as the reaction progresses. Graphical output is typica lly displayed as either baseline (Rn) or background (observed in control) subtra cted fluorescence versus the cycle number (Figure 3-5).183 A threshold is set, defined as 10 times the standard deviation of the baseline, and the cycle time (Ct) is the number of cycles required to exceed this threshold. A lower Ct value indicates that more DNA wa s initially present, as observed for the 5-fold dilutions plotted in Fi gure 3-5. For our purposes, a lower Ct value for the drug disk relative to the same fraction in the separate blank disk experiment insinuates

PAGE 81

81 that some of the excess DNA in the drug fraction binds spec ifically to the peptide, and not exclusively to the blank matrix. Figure 3-5. Amplification plots of IL-4 plasmid cDNA. Fivefold serial dilutions of plasmid cDNA were amplified by qPCR (cycle number along x-axis).183 qPCR was carried out on a Bio-Rad iCycler with MyiQ software and detection instrumentation. Centricon c entrifugal filters (Millipore, Y M-10) were used to desalt the fractions prior to qPCR. PCR mixtures contained iQ SYBR Gr een Supermix with iTaq DNA polymerase (Bio-Rad) and 0.5 M of each primer. The conditions were: 95.0C for 3.0 min, 40 repeated cycles (30 sec) of 94.0C, 55.0C (30 sec), and 72.0C (15 sec), followed by the final steps of 5 min at 72.0C, then an indefin ite hold at 4.0C. Second and Subsequent Selection Rounds A similar protocol of selection and qP CR was carried out for the second round of selection, except 2 mL of 100 nM DNA ( 200 pmoles) was now used for incubation with each disk. Round 3 was actually repeated twice to confirm the results. In both instances, 2 mL of 100 nM DNA (200 pmoles ) were incubated with drug and blank disk.

PAGE 82

82 AlphaScreen Analysis of Selected Pools An AlphaScreen (Amplifie d Luminescent Proximity Homogeneous Assay) was used to determine the pool with the highest-a ffinity candidates for DNA sequencing. The method is a bead-based non-radioactive proximity assay that measures ligand/target interactions by the output of fluorescence by bi nding pairs (Figure 3-6). In the assay, biotinylated drug is immobilized on a streptavidin coated donor bead that generates singlet oxygen upon laser excitation at 680 nm. This singlet oxygen initiates fluorescence (520-620 nm) in anti-FAM c oated acceptor beads with FAM-labeled DNA pools immobilized on the surfac e only if the beads are within the limited distance (200 nm) the singlet oxygen can travel during the excit ed state lifetime.184 Therefore, the beads will only generate measurabl e signal if the drug and pools are binding, bringing the beads into close enough proximity for the singlet oxygen to reach the acceptor bead. AlphaScreen Donor Bead AlphaScreen Acceptor Bead Excitation at 680 nm1O2travels no more than 200 nm distance Biotinylated Bivalirudin 5-FAM-aptamer Streptavidin Anti-FAMEmission 520-620 nm Figure 3-6. Schematic of AlphaScreen assay. All assayed DNA aptamer candidate pools were converted to ssDNA (described in Chapter 2) after PCR amplification with 5-FAM-ATC GTC TGC TCC GTC CAA TA

PAGE 83

83 (IDT) to provide an immobilization label for AlphaScreen anti-FITC acceptor beads (Perkin-Elmer). Anti-FITC beads (20 g/mL, final) and dye-labeled DNA pools (400 nM, final) were incubated for 1.5 hours in an Op tiPlate 384-well microplate (Perkin-Elmer) under subdued lighting. Streptavidin -coated AlphaScreen donor beads (20 g/mL, final; Perkin-Elmer) and biotinylated drug (100 nM and 100 pM, final) were then added to each well and incubated for 1.5 hours. The final volume in each well was 25 L, and the buffer for this experiment consisted of BB with 50 mg/mL BSA to inhibit non-specific binding. The positive contro l was 400 nM biotinylated FITC, substituted for the DNA in the method described above, and the negative control was all co mponents except DNA. Data was read in replicate on an EnVi sion microplate reader (Perkin-Elmer). Sequencing of Selected Pool A 400 L (total volume) 25 cycle PCR (Table 34) was performed according to the protocol in Figure 2-11. This dsDNA amplified pool was pr epared for sequencing by high-fidelity (Roche Applied Science Fa stStart High Fidelity PCR System) PCR amplification with fusion prim ers specific for 454 sequencing technology. The primer sequences used for the PCR are a combinati on of the 454 fusion tag and the regular sense and anti-sense primers previ ously used for amplification: Sense: GCC TCC CTC GCG CCA TCA GAT CGT CTG CTC CGT CCA ATA Anti-sense: GCC TCC CTC GCG CCA TCA GAT CGT CTG CTC CGT CCA ATA Table 3-4. PCR preparation Reagent Volume ( L) 10X PCR Buffer 40 dNTP (2.5 mM each) 32 Primers (10 M each) 20 DNA pool 40 DNAse-free H2O 268 Taq Polymerase 1.2

PAGE 84

84 The fusion primer PCR also utilized t he protocol from Figure 2-11, but was prepared by adding 0.5, 1.0, or 2.0 L of the dsDNA to 50 L of the PCR reaction mixture prepared from T able 3-5. Agarose gel electrophor esis (Figure 3-7) showed that each volume of pool added amplified in a simila r fashion. UV spectroscopy was used to calculate the amount of DNA as 452 g/mL for the 0.5 L pool addition product. Table 3-5. Fusion primer PCR Reagent Volume (uL) 10X PCR Buffer 20 dNTP (10 mM each) 16 Primers (10 uM each) 8 DNA pool Varied DNAse-free H2O 152 High Fidelity Polymerase 4 Figure 3-7. Optimization of pool volume for fusion primer PCR. Lane 1= control; lane 2= 2.0 L pool added, lane 3= 1.0 L pool added; lane 4= 0.5 L pool added; lane 5= 25-bp ladder. The dsDNA product was purified by a QiaQuick PCR Purification Kit (Qiagen) as per manufacturers instructions. Buffer PB (125 L; Qiagen proprietary buffernot the same as in Chapter 2) was added to the P CR product, and the sample was centrifuged at 13,000 rpm for 45 sec in a QiaQuick colu mn. The flow through was discarded and 1 2 4 3 5

PAGE 85

85 0.75 mL Buffer PE (Qiagen buffer) was added to the column and centrifuged for 1 min at 13,000 rpm. This flow through was di scarded as well, and the remaining product was centrifuged 1 min at 13,000 rpm. The filter portion was placed in a clean 1.5 mL microcentrifuge tube, and the purified product was eluted with 50 L Qiagen buffer EB by centrifugation (13,000 rpm, 1 min). The concentration was determined to be 24 g/mL by UV spectroscopy, which was diluted to 4.2 g/mL with the Qiagen EB buffer according to 454 procedures (request samples ~5 g/mL). Gel electrophoresis on the product confirmed that the desired 122bp sequence is present (Figure 3-8). Sequencing of the selected pool was perform ed by 454 sequencing at the University of Florida ICBR. 123 Figure 3-8. Agarose gel electrophoresis an alysis of fusion primer-amplified, purified pool. Lanes 1 & 3= 25-bp l adder; lane 2= purified pool. Sequence Alignment Following DNA sequencing, a total of 10,192 sequenc es were reported. The individual DNA sequences were analyze d using the MAFFT sequence alignment program by removing the 3 and 5 primer regions (when possible) from sequences prior

PAGE 86

86 to analysis. Several of these sequences de monstrated sufficient homology, and were synthesized with a 5-FAM fluorescent label for further characteri zation according to previously described methods. An example of the homology displayed by the alignment procedure is shown for one aptamer in Figure 3-9. Figure 3-9. Sample alignm ent of JPB2 (green highlight ed portion) using MAFFT. Binding Studies Streptavidin-coated beads (100 L; Bangs Laboratories) were washed with 1 mL BB and centrifuged at 10,000 rpm for 3 min. The supernatant was discarded, and the remaining beads were resuspended in 1 mL BB. One L of 10 M bivalirudin was

PAGE 87

87 combined with 10 L of streptavidin-coated pol ystyrene beads (Bangs Labs) and incubated for 1 hour on a shaker. The beads were centrifuged for 3 min at 10,000 rpm and reconstituted in 1 mL BB. The beads were split into 20 samples, and each aptamer candidate was added for a final concentration of 2.3 M and incubated with the bivalirudin/bead complex for 1 hour. Beads were centrifuged, washed with 1 mL of BB, then reconstituted in 200 L WB2 (WB24.5 g/L glucose and 5 mM MgCl2 in Dulbeccos PBS with CaCl2 (Sigma)) for flow cytom etry. The fluorescence intensity of the FAMlabeled sequences was meas ured with a FACScan flow cyt ometer by counting 20,000 events (Becton Dickinson) and analyzed using WinMDI. Results Selection Results The first round of selection gener ated the ch romatograms seen in Figure 3-10A-B. Once again, the peak at 15 min for the drug disk appeared to have a larger area than that of the blank disk. For confirmation, the 15th fraction eluting from both disks was compared by qPCR (Figure 3-11A). The drug disk clearly contained more DNA, as evidenced by the lower number of cycles requi red for amplification. Therefore, the whole peak observed in the drug disk chroma togram was combined and incubated with the blank disk to subtract out sequences bind ing to the column matrix. The DNA pool remaining after the counter selection was PCR amplified and prepared for the second round of selection. In the second round, a chromatographic difference between the disks was also observed (Figure 3-10C-D), but this time the drug disk did not display a noticeable peak. This may be due to the decreased amount of DNA used from t he first round to the second round not providing enough signal to exceed the limit of detection of the system.

PAGE 88

88 A similar selection using 100 nM DNA was also performed on the blank; however a small peak was observed at 15 min for the blank disk. Groups of fractions were combined and prepared for qPCR analysis, corre sponding to fractions 80-89, and 70-79 from the drug and blank selections. Both groups of fractions displayed higher concentrations of DNA in the drug selection than in the blank, so both groups (fractions 70-89) were combined and PCR amplified fo r the third round of selection (Figure 311B). A B Figure 3-10. Chromatograms of selection rounds. A) Round 1, drug disk; B) Round 1, blank disk; C) Round 2, drug disk; D) R ound 2, blank disk; E) Round 3, drug disk; 4) Round 3, blank disk.

PAGE 89

89 C D Figure 3-10. Continued

PAGE 90

90 E F Figure 3-10. Continued

PAGE 91

91 The third and final round of selection wa s carried out under similar conditions to the second round, resulting in the chromatogr ams in Figure 3-9E-F. Interestingly, a small peak returns for fractions 12-20 for t he drug disk, so the co mbination of these fractions were assayed by qPCR along with fractions 70-79 and 80-89 (Figure 3-10C). The differences between the drug and blank di sks were minimal, indicating that the selection was more enriched for the drug in round 2, or that some experimental error was to blame. The round 3 selection wa s repeated for both disks, but qPCR analysis yielded similar results (Figure 3-10D). 15Negative 15+ 15NegativeA 80-89+ 80-8970-79+ 70-79-B 12-20+ 12-2080-89+ 80-8970-79+ 70-79-C 80-89+ 80-8970-79+ 70-79-D Figure 3-11. qPCR results for each round of selection. A) Round 1 (Threshold=117.2; 15+ cycles= 19.76; 15cycles= 26.37); B) Round 2 (Threshold= 81.7; 80-89+ cycles=21.70; 80-89cycles= 24.39; 7079+ cycles= 23.74; 70-79-= 25.53); C) Round 3 (Threshold,102.6; 80-89+ cycles= 27.97; 80-89cycles= 27.12; 70-79+ cycles= 25.97; 70-79cycles= 27. 26; 12-20+ cycles = 26.13; 12-20cycles= 24.39); D) Round 3 repeat (Thr eshold= 80.2; 80-89+ cycles=24.71; 80-89cycles= 24.33; 70-79+ cycles = 24.71; 70-79cycles= 24.67).

PAGE 92

92 AlphaScreen for Pool Selection The AlphaScreen proximity assay was em ployed to analyze the pools from the 3 rounds of selection and determine which on e was the best candidate for sequencing. As previously mentioned, signific ant si gnal is only generated when the drug and aptamer are binding. 0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000 NegativeInitial LibraryRound 1, Frac. 13-22 Round 2, Frac. 60-69 Round 2, Sat'd NaCl Round 2, Frac. 70-89 Round 3, Frac 70-79 Round 3, Frac. 80-89PoolIntensity 100 nM Drug 100 pM Drug Figure 3-12. AlphaScreen assay of binding of different pools with 100 nM or 100 pM drug. Several conclusions were drawn from the results depicted in Figure 3-12. First, minimal signal from the negative control is measured, indicating that a binding event must occur for a measurable response. (Fl uorescence intensity values for the positive control, 1-2x105, not shown in graph.) Also, the signal from the initial library is low, as expected. Two concentrations of drug were used for this study, with a 1000-fold difference between them. Therefore, it is reasoned that a pool demonstrating increased binding at low drug concentrations will contain more high affinity aptamers, while a high signal at high drug concentration would signify lower affinity aptamers. Several pools,

PAGE 93

93 including the pool for round 1, round 2 fractions 70-89, and both round 3 pools demonstrated significant signal at high drug concentration, but the pool from round 2 had considerably higher signal at lower drug c oncentration. The conclusion is that the round 2 pool contains higher affinity apt amers, and was selected for sequencing. The data from the AlphaScreen follows the guidelines expected for selection, and validates the qPCR data. It wa s expected and confirmed, that the initial library should have the lowest amount of binding. Also, round 1 shows a higher percentage of low affinity aptamers, which is expected to evolve into high affinity pools as the selection progresses. The qPCR data consistently show ed, for each selection round, that the DNA released as a consequence of the satura ted NaCl wash was binding more to the column matrix than to the drug. An example of this is the Round 2 Satd NaCl sample that shows the lowest signal of any pool asid e from the initial library. The qPCR data also indicated that the pools from the 3rd round of selection were actually less enriched for the peptide than for the blank. The AlphaScreen assay gives the impression that this pool may contain low affinity aptamers, but the high affinity sequences are reduced from round 2. The number of sequences which bind to the ta rget as opposed to those binding to the matrix may be so small t hat repeated selection rounds and PCR steps may dwarf the population that binds to the drug, despite counter selection steps. Sequence Alignment The sequence alignment of the oligonucleot ides obtained from 454 sequencing revealed s everal sequences that display ed homology. Seven of these sequences, displayed in Table 3-6, were synthesized for binding characterization.

PAGE 94

94 Table 3-6. Probe sequences Name Sequence JPB1 ATC GTC TGC TCC GTC CAA TAC GAG GAT GCA GAA GTT TCA ATG CAC TTT TGG TGT GAG GTC GTG C JPB2 ATC GTC TGC TCC GTC CAA TAC GTA ACA TCC CCG TAA TAC TAC TAC GGT CGT GCT GGT TTG GTG TGA GGT CGT GC JPB3 ATC GTC TGC TCC GTC CAA TAG CTG AGC AGG TAA CAA TGT GTG CCC AAT GTG TAT TTG GTG TGA GGT CGT GC JPB4 ATC GTC TGC TCC GTC CAA TAA AGT TAA TCC TTA GGG CTG GTA GGT CAT TCC GGT GGT TAT TTG GTG TGA GGT CGT GC JPB5 ATC GTC TGC TCC GTC CAA TAT ATT GTG TGA CCC CCC TCT TGT TTT GGT GTG AGG TCG TGC JPB6 ATC GTC TGC TCC GTC CAA TAC CAG CTA ATG TGT ATT TTG TGG CGG CGG ATC ATA TGA GGA GGA TTT TTG GTG TGA GGT CGT GC JPB7 ATC GTC TGC TCC GTC CAA TAG TCG GAA TAG TGA CTG TTC TTG TGA AAC TCA ACA CGG ATG CTG GTG TTT TGG TGT GAG GTC GTG C Binding Studies The assay to determine whether the sequence s are binding the target is the critical stage of post-SELEX oligonucleoti de characterization; sequences that bind are officially considered aptamers. N onbinding sequences represent DNA that either binds to the target matrix, were inefficiently partitioned by the SELEX partitioning method, or are a result of PCR bias or mutations. It is also possible that the nonbinding sequences may actually be low affinity aptamers that may ex hibit binding when the target concentration is increased. For practical purposes, sequences which do not appear to bind with micromolar (for small molecules, lower fo r protein or cellular targets) target concentrations are not considered of interest for most studies. Binding studies were performed on all of the sequences listed in Table 3-6 by flow cytometry assay. Figure 3-13 shows t hat all of the sequences except JPB1 demonstrated an increase in fluorescence intensity upon binding to the drug. Therefore, 6 aptamer sequences were successfully obtained as a result of this

PAGE 95

95 selection. The largest fluorescence intensity shifts were demonstrated by JPB2 and JPB5, which were subjected to furt her characterization in Chapter 4. Beads Initial Library JPB2 Beads Initial Library JPB1 Beads Initial Library JPB6 JPB7 Beads Initial Library JPB3 JPB4 JPB5AB BD Figure 3-13. Flow cytometry binding studies of drug and aptamer c andidates. A) JPB1; B) JPB2; C) JPB3-5; D) JPB6-7. Conclusions The method described for aptamer selection has resulted in several sequenc es that bind to the target in only 2 rounds of se lection. The conclusions of these studies have also proven that differences in chro matographic profile of the drug and blank disk can be validated by qPCR. Furthermore, binding was characterized to determine the

PAGE 96

96 optimal pool for sequencing by an AlphaSc reen proximity assay. This work is innovative in that it provides the first known studies with the goal of generating an aptamer for a currently existing drug, and comb ines several exciting technologies for the purpose. The results described here leave room for some appealing side-studies. Of note, the pool from round 1 may contain sequences with desirable properties due to the highest intensity of all pools for the nanomolar ta rget concentration study. It would be interesting to sequence and test aptamer ca ndidates to analyze how the affinities compare to those from the chosen round 2 pool. Also, due to the large number of sequences from 454 sequencing, it is feasible that many of the sequences present in the sequencing data and not tested by the binding assay would interact with the target. Several less-enriched sequences could be compared to those described above to determine if the affinities are similar. A technological advance (and the first known use of the technology for this application) that was crucial to the success of the project is 454 sequencing. In a traditional cloning/chain-terminat ion sequencing protocol, the pool to be sequenced is cloned into a vector prior to sequencing. Due to low efficiencies associated with the cloning process, the pool must be highly evolved to ensure representation in the sequencing data. Also, Sanger sequencing typically only generates several hundred sequences, depending on the cloning conditi ons, as opposed to more than 10,000 sequences reported by 454 sequenc ing. Therefore, a high er percentage of sequences present in the final pool are detected, and the large number of sequences reported allows for a higher likelihood of homologous sequences after alignment. Thus, it is

PAGE 97

97 possible that 454 sequencing permits fewer rounds of selection due to a higher percentage of sequences detected overriding the need for a highly evolved pool. One modification to the procedure that may produce better results would be to implement a step gradient with a higher final salt concentration. In this manner, fractions corresponding to each increase in salt concentration could be pooled and assayed for binding using the AlphaScreen tec hnique. Fractions collected later in the process using the highest salt concentration would likely present as higher affinity binders. Also, a more sensitive detector w ould aid in the process to distinguish areas where DNA was eluting (not co rresponding to peaks) from the bas eline. As it was, PCR was required for non-peak fractions to determine whether DNA was present. As aptamer technology matures, solely ge nerating aptamers is not considered to be extraordinary; the novelty of the sequences lies in their applications. Despite these promising initial results in selecting aptamer s, additional testing must be carried out to prove the aptamers can function in the des ired capacity. The characterization and testing of the aptamers for t heir ability to serve as an antidote to an anticoagulant is discussed in Chapter 4.

PAGE 98

98 CHAPTER 4 CHARACTERIZATION OF APTAMER AFFINITY AND APTAMER ANTIDOTE TESTING Introduction A panel of 6 aptamers was i dentified from the selecti on and binding studies in Chapter 3. After aptamers were confirmed to bind to t he desired target they were characterized in terms of their affi nity for the target in this chapter. This affinity is described by the disso ciation constant (Kd) of the target/aptam er complex; lower Kd values correspond to higher affinity betw een target and ligand. Some methods proposed to determine Kd include flow cytometry,28 equilibrium dialysis,55 analytical affinity chromatography,56 ultrafiltration,176 fluorescence anisotropy (FA),185 and surface plasmon resonance (SPR).186 Next, the aptamers were tested to determi ne whether they function in the capacity they were selected for, which is an anti dote to an anticoagulant. Preliminary studies generally take place in buffer or human plasma ( in vitro ), then progress to in vivo animal models. Frequently, the aptamers ar e truncated after the preliminary in vitro studies to remove regions which do not contribute to bi nding. Potential aptam er therapeutics are typically truncated from initial 80-100mer sequences to 40 nucleotides or less.187 This reduces the cost of aptamer synthesis, incr eases the yield, and may actually act to stabilize the aptamer, resu lting in an increased affinity for the target.188 Additionally, for aptamers designed for in vivo applications such as targ eted drug delivery, smaller sequences have a higher depth of penetration into tissue.189 Standard methods of aptamer truncation include radi oactive labeling and fragmentati on, followed by synthesis and purification of active sequences.190 A less labor intensive method involves analysis of the individual sequences of a homologous family obtained after pool sequencing and

PAGE 99

99 alignment. However, this requires t he synthesis, purification, and affinity characterization of many sequences in a fa mily, also a time-consuming process. Finally, if the aptamers are susceptible to nuclease activity in plasma, additional post-SELEX modifications are implemented to reduce the effect. Various base modifications and end cappings such as pol yethylene glycol (PEG) may be utilized, as described in Chapter 1. Additionally, PEG end adaptors moderate renal clearance by increasing the molecular weight of the comp ound, increasing the circulation time of the probe. Figure 4-1 depicts the modificati ons necessary for commercialization of Macugen, including a 40 kDa PEG linker and 3 end cap.191 Figure 4-1. Secondary structure and m odifications of th e anti-VEGF aptamer pegaptanib (Macugen). Bold gray nucleot ides represent 2 -deoxy-2-fluoro nucleotides whereas bold black nucleotid es are 2-O-methyl nucleotides. The two adenosine residues shown in italic black are ribonucleotides. The 5position contains a 40 kDa polyethylene glycol (PEG)-linker, where n is approximately 450. At the 3-end, a 3-3-dT (Cap) structure was added.191 The work presented in this chapter de scribes a fluorescence anisotropy (FA) method used to characterize the binding affini ty of the aptamer/target interactions. The

PAGE 100

100 highest affinity aptamer was tested as a pr ospective antidote to the anticoagulant bivalirudin in both buffer and human plasma. A novel method of aptamer truncation using a DNA microarray was studied in orde r to provide functional abridged aptamer sequences in a shorter amount of time than conventional approaches. Materials and Methods Buffer Binding buf fer (BB): PBS buffer (without MgCl2 and CaCl2) with 5 mM MgCl2. FA Dissociation Constant Measurements Dissoc iation constants are a vi tal measuring stick for charac terizing the affinity of an aptamer for the target. Fo r the methods discussed, a fi xed quantity of target was incubated with various concentrations of apt amer. A binding cu rve was constructed based on the increase in signal at higher apt amer concentrations until signal saturation occurred. This curve was best-fit to a singl e-site saturation ligand binding equation, in which Y is the signal intensity, Bmax represents maximum signal intensity, and X is the concentration of ligand. XK XB Yd max (4-1) In a homogeneous solution of fluorophorelabeled molecules, each probe is randomly oriented. When the solution is exposed to polarized light, only the fluorophores with the transiti on moments oriented with the electric vector of the polarized light are excited. Under ideal co nditions, the samples will also emit light polarized in the same direction as the exci tation light, but real situations involve a depolarization of fluorescence emission, commonly due to ro tational diffusion of the

PAGE 101

101 molecules, but also influenced by light scattering, reabsorption, and polarizer misalignment.192 Anisotropy (r) is a dimensionless paramet er defined as the ra tio of the polarized fluorescence emission component to the total li ght intensity (equation 4-2). An emission polarizer is rotated between positions detecting fluorescence intensities parallel to polarized excitation light (I), and perpendicular to the excitation light (I). Anisotropy is directly proportional to the rotational correlation time of the fluorophore ( ) by the transposed form of the Perrin equation in eq uation 4-3, which in turn, is inversely proportional to the diffusion c onstant (D) as in equation 4-4.192 Thus, higher molecular weight molecules have lower D values, which means increased and r values. (Additional terms: r0 is fundamental anisotropy (the anisotropy in the absence of depolarizing processes), and is the fluorescence lifetime.) II II rll ll2 (4-2) 10r r (4-3) D6 1 (4-4) When FA is used to measure binding betw een a fluorescently-labeled molecule and a ligand, the higher molecular weight component is typically titrated into a solution of the labeled lower molecular weight molecu le. This maximizes the molecular weight difference between the bound and unbound forms, increasing the change in anisotropy ( r) at each aliquot added. Based on this pertinent background, the drug target was

PAGE 102

102 dye-labeled and the aptamer probe was titrated into the solution, and r was plotted as a function of aptamer concentration For this dissertation work, proof-of-concept work was performed using the 15-mer thrombin aptamer of sequence 5-FITC-GGT TGG TGT GGT TGG titrated with various thrombin concentrations as a model syst em. Anisotropy measurements were performed on a Fluoromax-4 spectrofluorometer (Horiba Jobi n Yvon) with excitation and emission polarizers in L-format. Bas ed on the manufacturers recommendations, the following settings were optimized: Ex= 492 nm; Em= 520 nm; integration time= 1 sec; slit width= 10 nm; aptamer concentration= 61 nM; total volume= 200 L. Thrombin concentrations from 2-200 nM in the thrombin selection buffer (20 mM Tris-acetate, pH 7.4, 140 mM NaCl, 5 mM KCl, 1 mM CaCl2, 1 mM MgCl2) were titrated into the aptamer solution and incubated for 1 min. A mini mum of three measurem ents were taken with the initial aptamer solution and for each thro mbin concentration titrated. The results were plotted using SigmaPlot as change in anisotropy versus aptamer concentration using a best-fit one-site ligand bi nding model to determine the Kd. The change in anisotropy is the average anisotropy of the initial dye-labeled aptamer subtracted from the average anisotropy value at each thrombin concentration. To measure the FA for our aptamers, the drug was conjugated to 5,6-NHS-TMR (5,6-N-hydroxysuccinimidecarboxytetramethylrhodamine; AnaSpec) in a similar manner to that described for the drug/biotin c onjugation. In this case, 5 mg of TMR dye was dissolved in PBS buffer (no MgCl2 or CaCl2 added) and dimethyl sulfoxide (DMSO), and slowly added to 46.5 mg Angiomax in 100 L PBS buffer for a total volume of 600 L. The solution was stirred overnight at 4C and HPLC purified as described for the

PAGE 103

103 biotin/drug conjugation. Anisotropy m easurements were performed on a Fluoromax-4 spectrofluorometer (Horiba Jobin Yvon) with excitation and emission polarizers in Lformat. Based on the manufacturers recommendations, the following settings were optimized: Ex= 545 nm; Em= 580 nm; integration time= 2 sec; slit width= 7 nm (JPB2) or 9 nm (JPB5, TV01, and TV 03); drug concentration= 1 M; total volume= 200 L. DNA concentrations in the micromolar range were titrated into the peptide solution and incubated for 1 min. A minimum of four measurements were taken for each aptamer concentration. The results were plotted in SigmaPlot as change in anisotropy versus aptamer concentration using a best-fit onesite ligand binding model to determine the Kd. The change in anisotropy is the average ani sotropy of the initia l dye-labeled peptide subtracted from the average anisotropy value at each DNA probe concentration. Sequence TV01: ATCG TCT GCT CCG TCC AATA GT GCA TTG AAA CTT CTG CAT CCT CG TTTG GTG TGA GGT CGT GC; TV03: ATCG TCT GCT CCG TCC AATA GCG TGC ATT GGT TTA CTG CAT CCG TGA AAC TGG GCT TTG GTG TGA GGT CGT GC.68 Clotting Experiments in Plasma The clotting experiments functioned on an in crease in light scatter over time as thrombin cleaved soluble fibrinogen into in soluble fibrin. The terminology used to describe the coagulation experiments is shown in Figur e 4-2. The normal clotting time expresses the time for uninhib ited thrombin to cleave fibrin ogen, and is expected to be short (Figure 4-2A). When the drug is added to the system, thrombin is inhibited, resulting in prolonged clotting times (Figure 42B). Ideally, when the aptamer is added, it will prohibit the peptide from binding th rombin, allowing the protein to cleave fibrinogen in a short time period (Figure 42C), similar to the normal clotting time.

PAGE 104

104 Normal: FibrinogenShort Clotting Time Prolonged:Thrombin Peptide+ FibrinogenDelayed Clotting Time Aptamer Influence: Peptide+ + Aptamer + FibrinogenShort Clotting TimeA B C Figure 4-2. Modes of coagulation. Expected results incl ude: A) Short normal clotting time; B) Longer clotting ti mes for the experiments prol onged by drug inhibition of thrombin; C) A return to short clo tting times when the aptamer binds to the drug and frees thrombin for cleavage of fibrinogen. Bivalirudin dose response curve The first step in the plasma studies wa s to determine the amount of drug that doubled the normal clotting time. In this expe riment, the reaction was carried out in a 100 L quartz fluorescence cuvette (Star na Cells), and the light scattering was monitored on a Fluoromax-4 spectrofluorometer (Horiba Jobin Yv on). The settings for this experiment were as follows: = 500 nm; slit width= 3 nm; integration time= 0.1 sec; interval= 0.5 sec, temperature= 37C. Th ree different wavelengths were tested for monitoring the reaction, and t he one utilized in further studi es was the wavelength which gave the largest response when thrombopl astin-L was added. Thromboplastin-L (Pacific Hemostasis) was equilib rated to 37C in a water bath prior to experiments. UCRP (universal coagulat ion reference plasma; 50 L; Pacific Hemostasis) was added to the cuvette and placed in the sample ch amber. The plasma wa s incubated in the sample chamber with different concentrati ons of drug (0-1000 nM) for 2 min during which the system was equilibrate d to 37C and the signal from the instrument leveled

PAGE 105

105 off (step 1). Thromboplastin-L was mixed (50 L) with the plasma (step 2), and the increase in light scatter was monitored by FluorEssence software. Each drug concentration was tested twice, then each curve of scattered light intensity versus time was best-fit to a 4-parameter sigmoid equation using SigmaPlot to determine the clotting time (considered as the halfway point between the minimum and maximum intensity values). These coagulation times for each drug concentration were then averaged, plotted versus time, and best-fit to the same 4-parameter sigmoid equation. The drug concentration required to double the normal clotting time was considered to be the concentration at the halfway point betwe en the minimum and maxi mum clotting time values. Coagulation testing The same conditions as those used in the Bivalirudin dose response curve were retained for the aptamer-influenc ed measur ements. Three type s of measurements were compared for this test: 1) Normal clotting time; 2) Pr olonged clotting time with 352 nM drug added in step 1; 3) Addition of 5-50,000 nM aptamer sequence (or control TV03 nonbinding sequence) and 352 nM drug in step 1. Step 2 addition of thromboplastin-L remained the same for eac h type of measurement. Each aptamer concentration was tested a minimum of two times, and each measurement plotted as a function of light intensity versus time and best-fit to a 4-paramet er sigmoid equation using SigmaPlot to determine the clotting time (considered as the halfway point between the minimum and maximum intensity values). The average of the clotting times determined for each aptamer concentration represent the reported clotting time.

PAGE 106

106 Clotting Experiments in Buffer The clotting experiments in buffer were also based on the premise of light scattering generated over time by thrombin cl eaving soluble fibrinogen into ins oluble fibrin. The buffer experiment was a simplistic assay designed to determine whether the aptamer could function in a model setting mainly devoid of nucleases. This assay consisted only of thromb in, aptamer, and fibrinogen. Optimizing fibri nogen concentration The concentration of fibrinogen used in the assay was optimized b ased on a physiologically relevant c oncentration of thrombin.193, 194 In this experiment, the reaction was carried out in a 100 L quartz fluorescence cuvette (Starna Cells), and the light scattering was monitored on a Fluoromax-4 s pectrofluorometer (Horiba Jobin Yvon). The settings for this experiment were as follows: = 500 nm; slit width= 3 nm; integration time= 0.1 sec; interval= 0.5 sec, temperature= 37C. The normal clotting time was determined by adding 50 L of 15.5 nM thrombin (human -thrombin; Haematologic Technologies) in BB to the cu vette and equilibrating it in the instrument for 2 minutes (step 1). Next, 50 L of fibrinogen (Sigma-Aldrich) concentrations ranging from 0.5-15 M (final) were added and mixed well by pipette (step 2), and the increase in light scatter was monitored using Fl uorEssence software. Each fibrinogen concentration was tested twice (with the exception of 15 M fibrinogen), and each measurement of light intensity versus ti me was best-fit to a 4-parameter sigmoid equation using SigmaPlot to determine the clotting time. Bivalirudin dose response curve Building the dose response curve commenced using the same settings as for the fibrinogen concentration experiments. In these experiments, 50 L of 15.5 nM thrombi n

PAGE 107

107 in BB and drug concentrations ranging from 0-100 nM final concentration were added to the cuvette and equilibrated in the instrum ent for 2 minutes (step 1). Next, 50 L of 0.5 M (final) fibrinogen was added and mixed well by pipette (step 2), and the increase in light scatter was monitored using FluorEssence software. Each drug concentration was tested twice, then each curve of light intensit y versus time was best-fit to a 4-parameter sigmoid equation using SigmaP lot to determine the clotti ng time (considered as the halfway point between the minimum and maximum intensity values). These coagulation times at each drug concentration were then averaged, plotted versus the drug concentration, and best-fit to the same 4parameter sigmoid equation. The drug concentration required to double the normal clotting time was considered to be the concentration at the halfway point between the minimum and maximum clotting time values. Coagulation testing Instrument settings similar to the fi brinogen and dose response testing in buffer were maintained. The normal clo tting time w as determined by adding 50 L of 15.5 nM thrombin in BB to the cuvette and equilibrating it in the instrument fo r 2 minutes (step 1). Next, 50 L of 1 M fibrinogen (500 nM final) was added and mixed well by pipette, and the increase in light scatter was monitored using FluorEssence software (step 2). The prolonged clotting time consis ted of the addition of 1.08 L of drug (10.8 nM final) to step 1, followed by fibri nogen addition. For the aptamer studies, drug and 5-20 M of aptamer or control sequence TV 03 were added in step 1, preceding fibrinogen addition. Each aptamer concentration (with the exception of 5 M aptamer concentration) was tested in duplicate, and each curve of light int ensity versus time was fit to a 4-parameter sigmoid equation using SigmaP lot to determine the clotting time. In the graph, the

PAGE 108

108 normal and aptamer (or contro l) induced clotting times were normalized with respect to the prolonged clotting time, showing the effect of the DNA on the drug function. Truncation via DNA Microarray A DNA mic roarray chip may solve time and material consumption barriers associated with aptamer truncat ion, as discussed in the chapter introduction. The microarray was a commercially available pl atform from CombiM atrix, and contained ~12,000 sequences synthesized in precise locations on the surface. Therefore, when fluorescently-labeled target interacted wit h the sequences, a standard microarray reader imaged the array, and the position of fluorescence was correlated to the exact sequence binding the target. Fluorescence intensity determined by imaging software can serve as a measuring stick to estimate the affinity of th e ligand for the target. Design A total of 8,000 of the available s equenc es were utilized on the chip for the truncation experiments. Several rational truncations were performed based on the secondary structures of JPB2 and JPB5 at 37C predicted by IDT OligoAnalyzer. Systematic truncation of single and both primer sequenc es, all individual stem/loop hairpins, combinations of hairpins, and m odification of the length of linkers between hairpins for each predicted structure was perfo rmed. Additionally, both aptamers were truncated to every possible 20-40 mer sequenc e. For example, possible 20-mers of JPB5 (Figure 4-3) would begin with the sequenc es at positions 1-20, the next sequence would be bases 2-21, followed by bases 322, etc. until all 20-mer sequences through bases 41-60 are represented. Similarly, 32 -mers would originate with sequences at base positions 1-32, then 2-33 and 3-34, through positions 29-60. Each truncated sequence for both aptamers was represented 5X on the chip.

PAGE 109

109 A TC G T C TGC T CC GT C CAA T A T ATT G TG TG A CCC CC CTCT T GT TT T GGT G T G AGG T CG TG C151015 25 20 35 30 4045505560 Figure 4-3. Truncation of JPB5. Sequences in red correspond to the base position indicated above. Microarray studies The chip was assembled by attaching t he hybridization cham ber. The chamber was fille d with BB and incubated for 10 min at 45C in order to rehybridize the probes. The chip was then equilibrated with BB for 5 min at 37C. Next, BB supplemented with 2% BSA was added to the hybridization cham ber, and incubated for 30 min at 37C. The chip was washed 2X with nuclease free H2O and incubated for 10 min. The TMRlabeled peptide (100 M) in BB was added to the chip and in cubated for 30 min at 37C. The chip was washed with BB, and the buffe r was removed from the hybridization chamber. The hybridization chamber was detached from the chip, and the semiconductor surface area was covered wit h Imaging Solution (CombiMatrix) and a LifterSlip (CombiMatrix). An Axon Gene Pix 4000B was used to image the chip at excitation wavelength 532 nm. Following usage, the chip was washed with water, and the hybridization chamber wa s filled with BB and st ored at 4C. Stripping of binding target was carried out by inc ubating the chip in BB for 30 min at 60C. The images were analyzed by Microarray Imager 5.9.3 software provided by CombiMatrix. Fluorescence intensity readings are repor ted as the average of each reported probe synthesized in five different locations on the chip. Background fluorescence intensity was determined by the average of contro l sequences provided by CombiMatrix.

PAGE 110

110 Results and Discussion Kd Characterization by FA The dissociation constant was studied us ing a FA method. Thrombin/thrombin aptamer was used as a proof-of-concept model for determining Kd measurements by FA. The 15-mer thrombin aptamer has been repor ted in the literature to bind exosite 1 of thrombin with Kd~ 25 nM.113 In Figure 4-4, the affinity of thrombin/thrombin aptamer at Kd= 36.9 4.8 nM was found to be very close to the literature values Therefore, the technique was validated, and the system was applied to biva lirudin and the selected aptamers. Kd= 36.9 4.8 nM R= 0.987 R2= 0.975 Figure 4-4. Kd plot of thrombin/15-mer thrombin aptamer using FA. JPB2 and JPB5 were chosen for further characterization due to the sequences displaying the largest flow cytometry binding assay shifts (Chapter 3). Figure 4-5A-B depicts the binding curves generated by JPB2 and JPB5. JPB2 demonstrated a calculated Kd= 5.99 0.88 M, and JPB5 displayed Kd= 5.76 1.79 M. When these experiments were repeated for both aptamers, estimated Kd values fell within the error

PAGE 111

111 of the values. These affinities are on the high affinity side of normal for small molecules/peptides, with reported Kds in the high nanomolar to high micromolar range.22, 176, 195 Additionally, two nonbinding control sequences were tested, TV03 and TV01 (Figure 4-5C-D) which resulted in nearly linear Kd curves under the concentrations tested. These sequences demonstrated calculated Kd values of 502.1 and 115.6 M respectively under the concentrations used, ranging from approxim ately 20to 85X higher than the Kds of the aptamer sequences. This confirms the aptamer sequences and resulting binding curves ar e the result of target bindi ng instead of nonspecific DNA interactions. Kd= 5.99 0.88 M R = 0.996 R2= 0.992A Kd= 502.1 M R = 0.996 R2= 0.992C DKd= 115.6 M R = 0.995 R2= 0.990 Kd= 5.76 1.79 M R = 0.957 R2= 0.915B Figure 4-5. Kd curves of aptamers and controls. A) JPB2; B) JPB5; C) TV03 control; D) TV01 control.

PAGE 112

112 Clotting Studies in Plasma JPB5 was applied to plasma in o rder to te st the aptamers ability to serve as an antidote to bivalirudin. A drug dose respons e curve was constructed as a function of drug concentration versus clotting time (Figure 4-6). The data was fit to a sigmoidal curve with a high degree of accuracy, and th e concentration of drug required to double the clotting time (C2) was calculated as 352 nM bivalirudin. This concentration was implemented into the prolonged and aptame r-mediated clotting times for plasma. [Bivalirudin] (nM) 0100200300400500600700800900100011001200 Clotting Time (sec) 10 15 20 25 30 35 40 45 C2= 352.2 22.1 nM R = 0.998 R2= 0.996 Figure 4-6. Dose response curve of bivalirudin in human plasma. The values were compared to control sequence TV03, which demonstrated very weak affinity for the target. The results for JPB5 and TV03 control tested over a wide range of concentrations (Figure 4-7) show ed similar values intermediate between the prolonged and normal clotting times. This is hypothesized to be the consequence of aptamer degradation by nucleases present in the plasma. It has been reported in the literature that nuclease activity, primarily DNase I, is present in human plasma which

PAGE 113

113 can degrade oligodeoxynucleotides.196 The same group states that heating the samples to 70C has been successful in removing the activity, but this would almost certainly adversely affect the coagulation proteins that are the focus of this study. Therefore, experiments comm encing in buffer rather than plasma would determine whether the aptamer c an inhibit drug function in medium s devoid of nuclease activity. 10 12 14 16 18 20 22 24 26 28 30 32 34Prolong e d 5 50 10 0 250 50 0 1 0 0 0 5 000 10000 20 0 0 0 35000 50000 N o r m al[Probe] (nM)Time (sec) TV03 JPB5 Figure 4-7. Effect of JPB5 and TV03 c ontrol sequence on clotting time in plasma. Clotting Studies in Buffer The fibrinogen concentration and drug dose response curve for the coagula tion tests run in buffer were optimized prior to determining the aptamer -mediated effects. The clotting time was observed to actually decrease as the fibrinogen concentration decreased (Figure 4-8A), which is counterintuitive to expect ed results, as a decrease in substrate is generally believed to result in a decrease in product formation. One hypothesis for the reason this phenomenon is occurring includes product inhibition. Product inhibition is especially relevant when considering that mo st coagulation factors have negative feedback activity; that is, the products can inhibit upstream reactants

PAGE 114

114 once the concentration exceeds normal limits.197 This function balances the processes so the body remains in homeostasis. The lack of all constituents of the coagulation cascade changes the dynamics of the system, where the different components may stabilize this effect. Fibrin strands are know n to retain binding to thrombin, so possibly the excess product at exosite 1 inhibits binding of fibrinogen to the same site.198 The dose response curve of bivalirudin demonstr ated a good fit to the data (Figure 4-8B), and a drug concentration of 10.8 nM was calc ulated to double the clotting time in the buffer system. [Bivalirudin] (nM) 0102030405060708090100110120 Clotting Time (sec) 0 100 200 300 400 500 600 700 800 C2= 10.8 2.3 nM R = 0.976 R2= 0.953 50 150 250 350 450 550 6500.5 1.0 10.0 15.0[Fibrinogen] (uM)Time (sec) A B Figure 4-8. Optimization of conditions for buffer experiment s. A) Fibrinogen concentration; B) Bivalirudin dose response curve.

PAGE 115

115 0 0.5 1 1.5 2 2.5 3 51020Normal [Probe] (uM)Normalization Ratio JPB5 TV03 (500 uM) TV03 (100 uM) Figure 4-9. Effect of JPB5 and TV03 contro l on clotting time. Clotting times at each concentration were normalized as a functi on of the prolonged clotting time. (Times lower than 1 correlate to times lower than that of the prolonged clotting time, while those higher than 1 relate to an increase from the prolonged clotting time.) When all components were combined for te sting of the aptamer response, the results in Figure 4-9 were obtained. T he values for each DNA concentration are normalized in terms of the prolonged clotting ti me to provide an illustration of the effect of the probe on the drug. Aptamer JPB5 wa s able to reduce the clotting time of the system in a dose-dependent m anner, with nearly complete antidote affect at 20 M probe concentration. In contrast the control TV03 pr obe actually served to increase the clotting time of the system. This is likely due to dilution effects, as the TV03 results appear similar to previous results gathered w hen only buffer was added. The data from the concentrated (500 M) and diluted (100 M) TV03 control concentrations demonstrate that dilution plays a ro le in the results, since the 20 M concentration has a significantly enhanced clotting ti me than the more dilute sa mple. This confirms the

PAGE 116

116 specificity of JPB5 for bivalirudin, since t he aptamer decreased the clotting time, yet the control at the same concentration as the aptamer only diluted the sample. Thus, the aptamer was able to serve as an antidote to the drug in buffer. Microarray Truncation Figure 4-10. Image of DNA micr oarray target binding. The regi on to the right represents a blowup of the box of the whole array on the left. The microarray studies for aptamer truncat ion resulted in many positions where fluorescence from the TMR-l abeled target binding was observ ed. An example is shown in Figure 4-10, where a blowup of the r egion in the red square revealed several positions of fluorescence. The microa rray software correlates the position of fluorescence emission with which DNA sequence is binding the target. The program also reports the fluorescence intensity of each binding event, which is used as a metric for determining relative ta rget/sequence affinities. The experiment revealed several probes which demonstrated signal significantly higher than the background values (Table 4-1). One 31-mer J PB5 derivative of

PAGE 117

117 sequence TAT ATT GTG TGA CCC CCC TCT TGT TTT GGT G had a signal intensity of 511.01, while a 20-mer derivative of JPB2 with sequence TAC GTA ACA TCC CCG TAA TA displayed a fluorescence value of 103.49. These were higher than fluorescence intensity values for the full vers ions of the probes at 63.41 for JPB5 and 61.41 for JPB2. Both truncated sequences cont ain primarily the random regions of the original sequences which are expected to acc ount for the majority of the specificity and binding properties of individual aptamers. Table 4-1. Microarray truncation Sequence Intensity JPB5 (Full) 63.41 JPB5 (31-mer) 511.01 JPB2 (Full) 61.41 JPB2 (20-mer) 103.49 Background 51.76 Conclusions This work proved the utility of FA methods in determining affinity of aptamer/target interactions. An important benefit of FA is that the reaction takes place in free solution, without the necessity of capt uring the complex by particles for detection (flow cytometry), or immobilization of the target or aptamer to a surface (SPR). Two aptamers demonstrated low micr omolar dissociation constants for bivalirudin using FA, values which are on the low (high affinity) side of the normal range for small molecule/peptide targets. It woul d be interesting to validate these Kd measurements by another technique that does not require captur e or immobilization such as isothermal calorimetry (ITC). ITC directly meas ures the heat released or absorbed by a bimolecular binding event, enabling a co mpletely label-free detection method.

PAGE 118

118 One of the aptamers, JPB5, was successful at functioning as an antidote to bivalirudin in buffer, restor ing thrombin activity in a concentration-dependent manner. This is believed to be the first occasi on where an aptamer wa s generated against an existing pharmaceutical anticoagulant. Unfortunately, the aptamer did not function in plasma, possibly due to degradation by the pres ence of nucleases. This sequence is currently under evaluation for modifications su ch as PEGylation to increase the lifetime of the aptamer in the plasma. A novel method of aptamer truncation us ing a DNA microarray was presented in this work, also believed to be the first of its kind. The microarray studies revealed abbreviated sequences from both JPB2 and JPB5 which correspond primarily to the random region of the aptamer The truncated JPB5 sequence is presently undergoing affinity and antidote characterization for use in replacement of the longer JPB5 aptamer. A PEGylated version of the truncated JPB5 is simultaneously under examination for antidote function and nuclease-resistance study. In summary, this work provides innovative methods and implements new technology to generate an aptamer with ex citing potential for development as an antidote for an anticoagulant drug. This antidot e will improve the safety profile of the anticoagulant bivalirudin, whic h is advantageous and sometimes essential for use as an alternative to heparin. Based on the positive results of the buffer experiments, we are currently seeking approval for te sting this drug in controlled an imal studies. Ideally, the PEGylated, truncated version of aptamer JPB5 will still retain antidote activity and will be implemented into the in vivo studies.

PAGE 119

119 CHAPTER 5 CANCER CELL CAPTURE USING APTAMER -IMMOBILIZED SQUARE CAPILL ARY CHANNELS: PROOF-OF-PRINCIPLE Introduction Separating pure cells from a complex matrix is a vital process for scientists aspiring to gain knowledge of cellular processes, and for clinicians detecting the presence of specific types of cells for di agnostic or therapeutic pur poses. Cancer cells are of extreme significance to diagnosticians since the probability of patient survival increases the earlier the diseas e is identified. However, the cells are dispersed in exceptionally low concentrati ons, usually <200 cells/mL for most types of cancer. Therefore, devices used for c ancer cell detection must be hi ghly sensitive to capture a significant amount of the malignant cell population. Microfluidic devices have recently been gaining attention for the efficient separation of cancer cells.92-94, 151 While microfluidics have been successful for this purpose in the past, the devices suffer from challenges with desi gn and fabrication, often requiring expensive clean rooms which are inaccessible to many laboratories. The systems may also be difficult to inte rface with benchtop flui dic devices, and the chemistries used to immobilize ligands on the typically polymeric surfaces are nonstandard.199 In contrast, capillaries have several proper ties which may promote their use as cell capture devices. Capillaries have relati vely simple connection options, are commercially available with excellent batch-to-batch reproducibility, and are easily surface-modified using simple, well-characterized chemistries. Specifically, square capillaries are attractive fo r cancer detection because flat-w alled capillaries (square or rectangular) have less optical distortion and sca tter than the curved walls of circular

PAGE 120

120 tubes.156, 200 This characteristic becomes signific ant due to the potent ially extremely low concentrations of cancer cells present in t he blood promoting direct imaging of captured cells rather than o ffline quantitation. Properties inherent to aptam ers such as specificity, ease of modification, and stability to degradation make the probes ideal ligands for cell capture. This group has previously selected the aptam er sgc8, which selectively binds to acute lymphoblastic leukemia CEM cells, but not to the Ramo s cell line (Burkitts lymphoma) using a modification of the SELEX procedure known as cell-SELEX.28 This aptamer binds with high affinity (Kd= 800 pM), implying that the aptamer/cell bond can withstand a substantial amount of force, as discussed in Chapter 1. The work presented in this chapter aims to combine the specificity of aptamers with the enhanced optical properties of square capillaries to improve upon results observed in circular capillaries, as well as to provide simple and cost-effective devices for the selective capture of cancer cells. The simplicity of setup and capture chemistry provide the capillary system with unique advantages with an application toward the detection of cancer cells in init ial proof-of-principle studies. Materials and Methods Cell Culture and Buffers CCRF-CEM cells (CCL-119 T-cell, human acute lymphoblastic le ukemia) and Ramos cells (CRL-1596, B-cell, human Burk itts lymphoma) were obtained from ATCC (American Type Culture Association). The cells were cultured in RPMI medium 1640 (ATCC) supplemented with 10% FBS (heat-inactivated; GIBCO) and 100 IU/mL penicillin-streptomycin (Cellgro ). Immediately before exper iments, cells were rinsed with 2 mL of washing buffer (WB2; 4.5 g/L glucose and 5 mM MgCl2 in Dulbeccos PBS

PAGE 121

121 with CaCl2; Sigma). Experiments using unstained cells involved direct dilution of cells to the desired concentration with binding buffe r (BB2; WB2 supplemented with yeast tRNA (0.1 mg/mL; Sigma) and BSA (1 mg/mL; Fi sher)) with 10% FBS. For cell staining experiments, the cells were diluted to 16 cells/mL in WB2, and the manufacturers instructions were followed by treating the cells with Vybrant DiI ( Ex/ Em 549 nm/565 nm) or Vybrant DiO (484 nm /501 nm) cell-label ing solutions (Invitrogen) for 5 min at 37 C. Cells were then rinsed with 1 mL WB2 and re constituted to the desired concentration using binding buffer with 10% FBS. All cells were stored on ice until needed. Device Construction 3 12 Magnified view of capillary Avidin Biotin-sgc8 CEM cell A B Figure 5-1. Schematic of se tup and immobilization. (A) A small piece of Teflon tubing (1) connects a square capillary (2) with an observation window (3) to a syringe. (B) Avidin is immobilized onto the capillary walls, and 5-FAM-sgc8poly(T)10-biotin is added for capture of target CEM cells. A 8 cm piece of 359 m o.d./74 m i.d. square capillary (Polymicro Technologies) was cut from the capillary spool, and a Bunsen burner was used to form a ~2-3 cm transparent window about 1 cm from one end (Figure 5-1A). After the window was cleaned with ethanol to remove remaining debr is, a 2 cm piece of Teflon tubing (PTFE (polytetrafluoroethylene) tubi ng, regular wall, 30 Gauge; Zeus) was used to connect the

PAGE 122

122 end of the capillary farthest from the transparent window to a 500 L U-100 Insulin Syringe (Becton Dickinson and Company). DNA Synthesis The aptamer sgc8 sequence, 5 FAMATC TAA CTG CTG CGC CGC CGG GAA AAT ACT GTA CGG TTA GAT TTT TTT TTT-3 -biotin, was synthesized using an ABI3400 DNA/RNA synthesizer (Appl ied Biosystems). The poly(T)10 linker was incorporated into the structure based on pr evious research results showing enhanced aptamer function by extending an imm obilized aptamer away from the surface.93 DNA synthesis reagents were purchas ed from Glen Research (Ster ling, VA). DNA purification was performed with a ProStar HPLC (Varian) using a C18 column (Econosil, 5U, 250 mm 4.6 mm) from Alltech Associates. UV-Vis measurements were performed with a Cary Bio-300 UV spectrometer (Vari an) to measure DNA concentration. Device Characterization Table 5-1. Physical pr operties of capillary Parameter Calculated Value Volume (uL) 4.62 x 10-1 D (m2/sec) 4.90 x 10-14 t (sec) 2.95 x 104 Dh (m) 7.60 x 10-5 Re (dimensionless) 1.96 x 10-1 The physical parameters of t he capillary were calculated as described in Chapter 1 in order to characterize the properti es of the device (i nner diameter 76 m, length 8 cm). The low Re confirms laminar flow (values < 2000) throughout the tube. Note the exceptionally long cell diffusi on times are counteracted by a short cell settling time due to gravity (Chapter 1).

PAGE 123

123 Controlling Degree of Sgc8 Immobilization Capillaries were cleaned and activated by manually adding 50 L of 1 M NaOH, then rinsed with 50 L of doubly-deionized water immediatel y prior to use. Avidin (5 mg/mL) was introduced at 1 L/min for 3 minutes using a Micro4 syringe pump (World Precision Instruments, Inc.) and i ncubated fo r 15 minutes. Excess avidin was removed with a manual wash of 50 L of binding buffer. Ten micro liter aliquots of different concentrations (0.5 M, 25 M, 50 M, and 100 M) of 5-FAM-sgc8-poly(T)10-biotin were manually drawn into separate capillaries, incubated for 30 sec, and rinsed with 50 L of binding buffer. Several sections of each capillary were imaged using an Olympus FV500-IX81 confocal microscope, and the raw images were analyzed in terms of fluorescence intensity by ImageJ (NIH). Cell Capture Assays The chemistry involved in cell capture is pr esented as a schematic in Figure 5-1B. Avidin and sgc8 immobilization was carried out with 50 M aptamer using the method described in the sgc8 immobilization step. To avoid cell settling, 1 mL of cell suspension was added to a dish with a stir bar on a magnetic stir plate. A small hole in the dish enabled capillary insertion into the cell solution, and cells were withdrawn using the Micro4 syringe pump at ei ther 200 nL/min (Table 5-2, Experiments 2 and 3) or 300 nL/min (Table 5-2, Experiment 1) for 10 minut es. Nonbinding cells were rinsed from the capillary with three column volumes (1.5 L) of binding buffer via syringe pump at either 200 nL/min or 300 nL/min (depending on flow rate of cells previously added). For the square and round capillary images, the capillari es were imaged without cell elution. For performance experiments, captured cells we re quantified by manually introducing a stream of air into the capillary, then collecting eluted cells in 10 L of binding buffer, and

PAGE 124

124 finally determining cell concentration by adding the cell solution to a hemocytometer (Hausser Scientific). Flow Rate Experiments Capillaries were coated with avidin and aptamer as described in the sgc8 immobiliz ation step. Cell capture was perfo rmed at flow rates of 300, 600, or 1200 nL/min in individual capillary devices as de scribed in the cell capture assays section. Non-target cells were introduced only at the lowest flow rate, 300 nL/min. Images corresponding to different positions on each capillary window were taken, and ImageJ was used to count the number of cells in each image. For each flow rate, the number of cells for all positions imaged along the capi llary was averaged, and the total number of cells captured was calculated by approximating the length of capillary visible in each image to be 1 mm. Captured target and non-tar get cells at the 300 nL/min flow rate were imaged both inside the c apillary and on a glass slide with a coverslip, eluted by introduction of an air stream. Stained Cell Imaging Using Fluorescence Microscopy CEM cells were split into 2 samples, eac h stained with either DiI or DiO dye as described above. Samples stained with eac h dy e were manually injected into separate square capillary dev ices ([cell]= 17 cells/mL), then mixed ([cell]= 56 cells/mL of each dye) prior to injection. Each sample was imaged using a Leica DM6000B fluorescence microscope using either BGR or monochromatic CCD (charge-coupled device) camera filters followed by false coloring using ImageJ. Cell Elution Efficiency Cells were captured, eluted, and counted using the method described in the cell capture assays section. To determine the el ution efficiency, the entire length of

PAGE 125

125 capillary was scanned to count any cells t hat were not eluted by the air stream, and these remaining cells were added to the el uted cells for the total number of cells captured by the capillary. Eluted cells were divided by total cells and multiplied by 100% to calculate the elution efficiency of the procedure. From the data obtained from three trials, elution efficiencies of 99.78% 99.85%, and 99.97% were obtained, leading to an average elution efficiency of 99.87 0.06%. Square and Circular Capillary Cell Capture Comparison Two squar e (359 m o.d./74 m i.d.) and two circular capillaries (358 m o.d./76 m i.d) were prepared following the met hod described in the cell capture assays section. For this work, the initial probe concentration was 5 M, and the cell concentration was 56 cells/mL. Following cell capture, the cells retained in the two square and two circular capillaries were combined according to geometry before addition of the eluted cells to the hemocytometer. Microscopy and Image Analysis Accurate initial cell concentrations for both cell lines were obtained before each experiment by adding the initial c ell solution to a hemocytometer a nd imaging the device with an Olympus FV500-IX81 confocal microscope. The cells were counted by importing the images into ImageJ and manually counting cells from each cell line. The cell concentration was calculated in accordan ce with the manufacturers instructions. The concentration of cells captured by t he device was calculated in a similar manner; however, nine image sets were obtained, corresponding to all units of the hemocytometer grid. For the cell staining experiments, transmitted images, as well as those corresponding to red fluorescent target cells and green fluorescent control cells, were obtained for each square unit of the hemo cytometer grid. Once the images were

PAGE 126

126 imported into ImageJ, red and gr een cells could be individually highlighted and overlaid on the transmitted image in order to count the number of each cell type present for both initial and captured cells. Square and Circular Capillary Imaging To compare the imaging capabilit ies of t he capillary geometries, cells were injected into an 8 cm length of 358 m o.d./76 m i.d. circular ((A), 26 cells/mL) or 359 m o.d./74 m i.d. square ((B) 106 cells/mL) capillary and imaged using a Leica DM6000B fluorescence microscope. Cells were then captured using the method described in the cell capture assay portion (6.45 cells/mL) and imaged with an Olympus FV500-IX81 confocal microscope for eit her a circular or square capillary of the same dimensions given above. Results and Discussion Sgc8 Immobilization The first step in proving the utility of the device was to show that aptamer immobiliz ation on the capillary surface coul d be controlled by varying the concentration of sgc8 introduced into the device. In this experiment, different concentrations of 5FAM-sgc8-poly(T)10-biotin were immobilized on the c apillary walls coated with avidin, and the fluorescence intensity observed at ea ch concentration was compared (Figure 52). A concentration-dependent increase in fluor escence intensity as sgc8 concentration increased was apparent, with sa turation of signal at 50 M sgc8. The confocal images of a representative of each aptamer concentration are presented in Figure 5-3. A saturating sgc8 concentration (50 M) was employed throughout the remainder of the studies in order to maximize the num ber of ligands binding to the cells.

PAGE 127

127 0 500 1000 1500 2000 2500 3000 3500 020406080100120 [sgc8] (uM)Fluorescent Signal Figure 5-2. Fluorescence intensity increas e as aptamer concentration is increased. A B CD Figure 5-3. Confocal images of FAM-labeled sgc8 immobilized inside a capillary at various aptamer concentrations. A) 0.5 M sgc8; B) 25 M sgc8; C) 50 M sgc8; D) 100 M sgc8. Cell Volumetric Flow Rate Trends Basic flow rate experiments were carried out in order to investigat e flow rate trends and to help identify the proper conditions for cell capture. Target cells (CEM) were captured in the capillaries at three different cell flow rates (300, 600, or 1200

PAGE 128

128 nL/min). The experiments demonstrated linear velocity-dependent binding (Figure 5-4), with the lowest flow rate (300 nL/min) capt uring the highest amount of cells (1374 cells captured), followed by intermedi ate flow rate (558 cells capt ured), and finally the highest flow rate (240 cells captured) These results agree with pub lished data that indicates a decrease in flow rate will consequently decrease the linear ve locity of the cell population, leading to a higher fraction of cells with sufficient time to form enough bonds for attachment.201 Additionally, nontarget cells (Ram os) were introduced to a capillary device at the lowest flow rate (300 nL/ min), and only 26 cells were captured. Comparison of the amounts of target and non-ta rget cells captured at the lowest flow rate is better visualized by the massive amounts of cells observed upon elution of the target cells onto a microscope slide (Figure 5-4C; large circles represent air bubbles) versus the miniscule number of nontarget ce lls eluted (Figure 5-4D). Thus, even at the lowest flow rates, target cells were captured in quantities over an order of magnitude higher than nontarget cells. The image in Figure 5-4E shows cells dislodg ed by the air/liquid interface exerting shear forces on captured cells upon introduction of an air bubble. This air-based elution mechanism allows the cells to remain concentrated for offline counting procedures, as opposed to liquid shear-force strategies that dilute the captured target.166 Air-based elution forms the basis for cell counting in a hemocytometer used in the following studies. Fluorescence Microscopy Imaging of Stained Cancer Cells CEM cells were stained with DiI and DiO dyes, then introduced int o separate capillaries for imaging of each dye alone and a mixture of the tw o dyes. The images obtained from capillaries containing cells with only one dye were fairly clear, but when

PAGE 129

129 the cells were mixed and BGR (blue/green/red) light was used for excitation, all imaged cells appeared green (DiO emi ssion). The main cause is due to the fact that DiO emission is much more intense than DiI, contributing to a variety of effects.202 Therefore, the dyes of the cell mixture were analyzed sepa rately using dye-specific filters then overlaid (Figure 5-5), which proved capable of detecting both of the dyes. However, the microscope images were not comp letely resolved, which will interfere with the signal in detection of captured cells. D E C B A Figure 5-4. Confocal images of cells capt ured in square capillaries at different volumetric flow rates. A) 1200 nL/mi n CEM; B) 600 nL/min CEM; C) 300 nL/min CEM (arrow represents cells air-e luted onto microscope slide); D) 300 nL/min Ramos (arrow represents cells ai r-eluted onto microscope slide); E) An air bubble dislodging cells.

PAGE 130

130 Based on these results, confocal microscopy was used for the remainder of the studies unless otherwise noted to produce more resolved images. The confocal microscope is capable of sequential laser illumination, which means each dye can be individually excited and detect ed without the need for changing filter cubes. This ability proved essential since preliminary experim ents with the confocal microscope also showed green-to-red bleedthrough issues with simultaneous excitation. Figure 5-5. Simultaneous fluorescence mi croscopy imaging of cells in a square capillary stained with two different dyes. Cell Capture Performance Two different groups of experim ents we re conducted in order to evaluate the performance of the device, and the results are summarized in Table 1. In the first group of experiments, solutions of pure cell types, target (CEM) or control (Ramos), were introduced into separate capillaries to determine the extent of binding for each cell type. Initially, cell concentrations were kept appr oximately equal (Table 5-2, Experiment 1), leading to a 10 times higher retention of ta rget cells than non-target cells. Next, the concentration of control cells was increased to roughly double that of the target cells (Table 5-2, Experiment 2). Even under these conditions, t he number of target cells captured exceeded that of non-tar get cells by a factor of 4.7-27. This experiment was performed in duplicate, and the wide range may be due to some flow variations (lower linear velocity) present in the capill ary with the higher ca pture efficiency.

PAGE 131

131 The second group of experiments was designed to more closely mimic the complexity of the physiological environment. Prior to introduc tion into the device, target and control cells were first stained with DiI and DiO dyes, respectively, then mixed together for introduction into the same capill ary device. In thes e experiments, target cells were captured in number s roughly 5 times more abundant than control cells (Table 5-2, Experiment 3). As a whole, these per formance studies show that the device is capable of selectively capturi ng target cells. This holds true when target cells are administered in lower concentrations than nontarget cells or even when the cell types are mixed. Table 5-2. Performance of capillary system Experiment # Cell Type Relative Cell Concentration CEM/Ramos CEM 1 Control Equal 10 CEM 2 Control 2X Control 4.7-27 3 Mixed 2X CEM 4.2-6.5 Square and Circular Capillary Cell Capture Comparison Additional studies were per formed to directly compare the amount of target cells captured in square and circular di ameter capillaries. Under the same conditions, square capillaries retained 33% more cells than circ ular capillaries of similar diameters and lengths. A possible explanation for these results is that the higher cross-sectional area of the square device will result in a lower linea r velocity of the cells when compared to the circular capillary, consequently incr easing cell capture as previously described.201 Therefore, square capill aries of shorter length than circular capillaries of similar inner diameter can be utilized for cancer cell study which correlates to decreased material consumption. However, cell capture is merely a first step in the process of determining

PAGE 132

132 cell immunophenotype and cell mor phology; further information on captured tumor cells can be elucidated by in situ imaging analysis. Imaging Properties of Squa re and Round Capillaries A B C D Figure 5-6. Imaging comparison of square and round capillaries. The top panels show CEM cells that were injected in to either a circular ((A), 358 m o.d./76 m i.d., 26 cells/mL) or square ((B) 359 m o.d./74 m i.d., 17 cells/mL) capillary, imaged using a fluorescence microscope. The lower panels display cells captured in either a circular (C) or square (D) capillary (6.45 cells/mL) imaged with a confocal microscope. An optimal device would allow for strai ghtforward interfacing to the imaging instrumentation as well as properties whic h would minimize optical distortion and scatter. The benefits of using square vers us circular capillary geometry were demonstrated by imaging cells directly injected into square and circular capillaries of comparable diameter (Figures 5-6A-B) by fluorescence microscopy or by capturing target cells with the aptamer and imaging by c onfocal microscopy (Figures 5-6C-D). For both injection and capture, cells imaged in the square capillary are better resolved than those in the circular device regardle ss of the imaging system used. The improved resolution could have an immediate applicat ion in the area of bioimaging and bioanalysis, specifically cancer cell detection, where inability to resolve a single cell could result in a faulty diagnosis or indication of treatment progress.

PAGE 133

133 Conclusions This work has presented proof-in-principl e studies dem onstrating that a simple square capillary device is useful for select ive capture of cancer cells. The system requires no complicated design or fabrication steps or clean room facility. Furthermore, the inner walls of the capillary can be modified with capture probes using standard avidin/biotin chemistry, with the degree of immobilization c ontrolled by varying the probe concentration. The aptamer-coated square capill aries consistently captured more target cells than control cells, even when the cell lines were mixed prior to introduction, and the square capillaries retained significantly more cells than the circular geometry following a direct comparison. Elution of cell s retained in the capillary by a flowing air stream consistently eluted ~ 99.9% of cells. Additionally the square capillaries are superior to circular capillarie s for imaging purposes. This is a clear benefit for cancer cell detection, specifically for CTC analysis, since it is necessary to examine cell morphology and immunophenotype after cell capture. In summary, the use of a square cap illary system for imaging captured cells provides a simple and cost-effective method for counting cell subpopulations. These studies represent proof-in-principle work to validate the concept, yet do not reflect fully optimized conditions. In the best case for th e lowest flow rate utilized for administering the cell solutions, the capture ratios of target cells versus nontarget cells were high, but the overall capture efficiencies of CEM cells we re low, only ~5%. This is not practical for situations when low concentrations of c ancer cells are present, or for prognostic applications that monitor patient progress as a func tion of blood cancer cell concentration. Therefore, work described in Chapter 6 describes the optimization of the system, as well as application in detect ing cancer cells present in human blood

PAGE 134

134 samples. The device is also applied to types of cancer that are not blood-borne, demonstrating the universal function of the system.

PAGE 135

135 CHAPTER 6 OPTIMIZATION AND APPLICATIONS OF SQUARE CAPILLARY DEVICE Introduction Proof-of-pri nciple studies (Chapter 5) described the use of a simple aptamercoated square-capillary system to capture leuk emia cells from a flowing suspension. Target cells were retained in greater quant ity even when target CEM and nontarget Ramos cells were mixed together prior to se paration. The system retained CEM cells in higher numbers than Ramos cells due to the s pecificity of the sg c8 aptamer for the target cells.28 The square capillary used for ce ll separation also demonstrated enhanced imaging properties when compared to circular geometry capi llaries of similar diameter. The work in Chapter 6 will focus on optimization of experimental parameters to improve the capture efficiency of the capillary device. This was performed by varying the flow rate of the cell suspension through the device, which changes the linear velocity of the cells traveling through the system. The effect of the fl ow rate of the buffer used to flush uncaptured cells from the device was also explored. The capillary system was used to image CEM cells spiked into hu man blood for potential applications in cancer diagnostics. Since leukemia is a disease specifically affecting the blood, leukemia cells are found in relatively high concentration in the flui d. Great interest lies in detecting cancer cells that are not normally present in the bl ood in high concentration in order to monitor treatment progress and correlate cancer cell levels with patient response or disease stage. Therefore, conditions optimized for CEM capture were tested on two different adherent colon cancer cell lines representin g Dukes' type C colorectal adenocarcinoma

PAGE 136

136 (DLD-1) and colorectal carcinoma (HCT 116). Aptamers demonstrating a high affinity for these colon cancer cell lines have recently been reported by this group.203 These aptamers were tested as cell affinity ligand s in the capillary device to compare the capture efficiency of the system for tumorigenic cancers to that of blood-based cancers. The question of whether conditions optimiz ed for one type of cancer cell line can be applied to another cancer cell line is also considered. Materials and Methods Cell Culture, Buffers, Aptamer Sy nthesis, and Device Construction CEM cell c ulture, buffers, sgc8 aptame r synthesis, and square capillary device construction were described in Chapter 5. Buffers: WB24.5 g/L glucose and 5 mM MgCl2 in Dulbeccos PBS with CaCl2; BB2WB2 supplemented with yeast tRNA (0.1 mg/mL) and BSA (1 mg/mL); red blood cell lysis buffer (LB)150 mM NH4Cl, 10 mM KHCO3, 0.1 mM EDTA in distilled H2O (pH 7.3) filt ered with 0.45 m filter (Nalgene); non-enzymatic buffer (NEB)used as provided by MP Biomedicals; sgc8 sequence: 5 FAMATC TAA CTG CTG CGC CGC CGG GAA AAT ACT GTA CGG TTA GAT TTT TTT TTT-3 -biotin. Flow Rate Optimization Capillaries were cleaned and activated by manually adding 50 L of 1 M NaOH, then rinsed with 50 L of doubly-deionized water immediatel y prior to use. Avidin (5 mg/mL) was introduced at 1 L/min for 3 minutes using a Micro4 syringe pump (World Precision Instruments, Inc.) and i ncubated fo r 15 minutes. Excess avidin was removed with a manual wash of 50 L of binding buffer. Following this step, 10 L of 50 M sgc8 was manually drawn into the capillary, incubated for 30 seconds, and rinsed with 50 L of binding buffer. To avoid cell settling, 1 mL of cell suspension was added to a dish

PAGE 137

137 with a stir bar on a magnetic stir plate. A small hole in the dish enabled capillary insertion into the cell solution, and cells (56 cells/mL) were withdrawn using the Micro4 syringe pump at either 500 nL/min, 1 00 nL/min, or 20 nL/min for 10 minutes. Nonbinding cells were rinsed from the capillary with three column volumes (~1.5 L) of binding buffer via syringe pump at 500 nL/min. Captured cells were quantified by manually introducing a stream of air into t he capillary, collecti ng eluted cells in 10 L of BB2, and finally determining cell concentration by adding the cell solution to a hemocytometer (Hausser Scientif ic). Capture efficiency is reported as the number of cells captured divided by the total num ber of cells passing through the device. Optimization of Cell Washing Flow Rate The protocol for the flow rate optimizat ion studies was followed using a 20 nL/min cell flow rate. Two cell washing flow ra tes were compared, 200 nL/min and 500 nL/min, administering the sam e total wash volume (1.5 L). Cell capture concentration was quantified via hemocytometer by eluting the captured cells by flowing an air stream through the capillary into 10 L BB2. The amount of cells passing through the capillary uncaptured was quantified by hem ocytometer by emptying the cellular contents of the Teflon tubing and syringe into 10 L BB2. Capture efficien cy was reported as the number of cells captured divi ded by the total number of cells (captured cells added to cells passed through into tubing and syringe) from this point forward unless otherwise indicated. Detecting CEM Cells in Blood Cell staining CEM cells (1.76 cells) were diluted in 2 mL WB2. CellTracker Green (2 L; Invitrogen) was incubated with cells for 45 mi n at room temperature, according to

PAGE 138

138 manufacturers instructions. The cells were centrifuged at 970 rpm for 3 min to remove excess dye, resuspended in 1 mL WB2, and in cubated for 30 min at 37C. The cell solution was centrifuged at 970 rpm for 3 min, and the supernatant liquid was discarded. The pellet was washed with 1 mL of BB2, centrifuged at 970 rp m for 3 min, then resuspended in 100 L BB2. This cell solution was kept on ice until needed. Removal of red blood cells Red blood cells were removed from human whole blood (Innovative Research) by separation by Ficoll-Paque (GE Healthcare ). Ficoll-Paque is a density gradient centrifugation medium that separates blood into 4 main layers: the top layer consists of the plasma, followed by the buffy coat, t he Fi coll layer, and the red blood cells concentrated at the bottom of the tube. The buffy coat contains the white blood cells and platelets, and is the la yer collected and combined wit h the CEM cells in this experiment. This red blood cell removal is essential so the high concentration of red blood cells do not clog the capillary. The Ficoll-Paque blood separation was carried out according to manufacturers instructions. Briefly, 3 mL whole blood wa s slowly added to 3 mL Ficoll in a 15 mL centrifuge tube. The layers were centrifuged at 1200 g at 4C with the rotor brake off. The buffy coat was removed and centrifuged at 1800 rpm for 5 min. The supernatant liquid was discarded and the cells were washed with 3 mL PBS buffer. The washed cell pellet was combined with 1 mL red blood cell ly sis buffer (LB) and incubated for 15 min. The supernatant liquid was removed and t he pellet was washed twice with WB2 (all centrifugations at 1800 rpm for 3 min). The final pellet was resuspended in 3 mL BB2 to simulate the concentration of white blood cells in whole blood.

PAGE 139

139 Cell capture The cell capture exper iments were performe d by spiking stained CEM cells into the white blood cell sus pension at either 15 or 34 cells/mL. The basic protocol described in the flow rate studies was follo wed, with a 20 nL/min cell introduction and 500 nL/min washing flow rate. The length of the capillary window was scanned, and fluorescent CEM cells were counted. Total cell numbers were calculated by a simple proportion relating cells capt ured per length of window to entire capillary length. Colon Cancer Cell Culture Colorectal cancer cell lines DLD1 (Duk es' type C colorectal adenocarcinoma) and HCT 116 (colorectal carcinoma) were purc hased from American Type Cell Culture (ATCC). DLD-1 cells were maintained in culture with RPMI-1640 containing 10% heatinactivated FBS (Invitrogen) and 100 IU/mL peni cillin-streptomycin (Cellgro). HCT 116 cells were maintained in McCoys 5A cu lture medium containing 10% heat-inactivated FBS and 100 Units/mL penicillin -streptomycin. All cultures were incubated at 37C under a 5% CO2 atmosphere. Both cell lines were grown as an adherent monolayer in 100mm x 20mm culture dishes to >95% confluence. Cells were wa shed in the dish with WB2, dissociated by trypsin treatment (2 min) and seeded into culture dishes at low concentration. Within 24 hours of seeding (<40% confluence), cells were short-time trypsinized (30-45 sec) for capture and flow cytometry studies. Colon Cancer Aptamer Sequences KDED2a-3: TGC CCG CGA AAA CTG CTA TTA CGT GTG AGA GGA AAG ATC ACG CGG GTT CGT GGA CAC GGT-biotin; KCHA10: ATC CAG AGT GAC GCA GCA GGG GAG GCG AGA GCG CAC AAT AAC GAT GGT TGG GAC CCA ACT GTT TGG

PAGE 140

140 ACA CGG TGG CTT AGT-biotin. Aptamer KDED2a-3 binds specifically to the DLD-1 cell line, but demonstrated minimal binding to HCT 116, while aptamer KCHA10 binds similarly to both DLD-1 and HCT 116 cell lines. Probes were synthesized and purified as described previously both with and without a 3-poly(T)10 linker. Cell Capture of Colon Cells Testing efficiency of aptamers (w ithout linker) for DLD-1 capture All steps through aptamer immobilization were the same as described in the flow rate optimization studies. For cell adminis tration, the cells were strained with 40 M cell strainers (Becton Dickinson and Company) immediately before addition to the capillary and were kept stirring throughout the entirety of test ing. These steps proved necessary due to cell aggregation and capillary clogging without their implementation. The cell flow rate of 20 nL/min for 10 min and wash with 1500 L BB2 at 500 nL/min remained the same. Captured and unr etained cells were count ed for capture efficiency calculation. DLD-1 cell capture using aptamers with linker The binding of aptamers with the poly(T)10 linker were compared to those without the spacer by flow cytom etry. DLD-1 cells (200 L) were removed from a 6.5106 cell/mL solution in BB2 and diluted to 500 L. Aliquots of 100 L cells were added to each of 5 tubes. Each tube (except 1 c ontrol tube) was inc ubated with 250 nM DNA (library, KDED2a-3 with and without poly(T)10 linker, and KCHA10 with linker) for 10 min. The cells were centrifuged at 970 rp m for 3 min and washed with 1.5 mL WB2, incubated with 100 L streptavid in-PE-cy5.5 (channel 3) at a final dilution of 1:400 stock solution for 10 min. The cell solution wa s washed with 1500 L WB2 and the cell pellet was resuspended in 150 L WB2 for analysis by flow cytometry. The cell capture

PAGE 141

141 experiments were carried out using the same method as the experiments without the linker. DLD-1 cell capture at low concentration using aptamer KDED2a-3 To increas e the number of cells accessing t he capillary, the ce ll capture protocol was modified to use a lower initial cell concentration (16 cells/mL) and a longer cell administration time (increased from 10 min to 20 min). The remainder of the conditions, including cell straining, were the same as described previously. HCT 116 cell capture by aptamer KCHA10 The low in itial cell concentration (16 cells/mL), longer cell administration time (20 min) and cell straining technique were incorporated into the protocol described in the flow rate optimization section for HCT 116 cell capture by KCHA10 aptamer. HCT 116 cell capture in non-enzymatic buffer by aptamer KCHA10 Avidin and probe immobilization was the sa me as described in the flow rate optimization section, but the protocol for cell administration was modified. Immediately prior to introduction int o the capillary, cells were concentrated to 2X (26 cells/mL) in BB2 by centrifugation at 970 rpm for 3 min. The concentrated cells (500 L) were then strained through a 40 m cell strainer and combined with 450 L non-enzymatic buffer and 100 L FBS with continuous agitation throughout the experiment. Flow cytometry of aptamer binding to the cells in non-enzymatic buffer wa s compared to the binding of cells without non-enzymatic buffer in the same manner as the DLD-1 cell capture using aptamers with linker section to confirm t hat aptamer binding was not affected. Microscopy and Image Analysis Accurate initial cell concentrations for cell lines were obtained before each experiment by adding the initial c ell solution to a hemocytometer a nd imaging the device

PAGE 142

142 with an Olympus FV500-IX81 confocal microscope. The cells were counted by importing the images into ImageJ and manually counting cells from each cell line (if applicable). Then the cell concentration wa s calculated in accordance with the manufacturers instructions. The concentra tion of cells captured by the device was calculated in a similar manner; however, nine image sets were obtained, corresponding to all units of the hemocytometer grid. For the cell staining in blood experiments, transmitted images, as well as those corresponding to green fluorescent target cells were obtained for each square unit of the hemo cytometer grid to ca lculate the target and nontarget cell concentration in the flow through liquid. Once the images were imported into ImageJ, stained tar get cells could be individually highlighted and overlaid on the transmitted image in order to count the number of target and nontarget cells passed through the device. Results and Discussion Optimization of Flow Rate The flow rate at which the cell solution is administered plays a major role in the efficiency of capturing cells. The flow rate is directly proportional to the linear velocity of the fluid through the capillary, so lower li near velocities are the consequence of lower flow rates. Cells in a fluid will have a hi gher probability of binding to the ligand at lower velocities because the contact time of the cells and ligands is higher. Higher contact times increase the number of bonds formed during the collision (B*; equation 1-11), which results in a stronger interaction with the ligand. Thus, several volumetric flow rates were compared to increase the capture efficiency of the system in relation to the proof-of-principle st udies where capture efficiency was only ~5% in the best case. Figure 6-1 shows the increase in capture

PAGE 143

143 efficiency from 0.36% at 500 nL/min flow rate to 2.22% at intermediate flow rate, and 92.22% at the lowest flow rate It is expected that the decreased velocity increases the collision duration time, which, in turn, in creases the number of bonds formed during the collision. 0 10 20 30 40 50 60 70 80 90 100 0100200300400500600Flow Rate (nL/min)Capture Efficiency (%) Figure 6-1. Comparison of volumetric fl ow rate to capture efficiency of cells. It is possible that lower flow rates may demonstrate higher capture efficiency, but the total volume and time limitations r ender further decreasing the flow rate unreasonable. For example, under similar experimental conditions (10 min flow) a 5 nL/min flow rate would only pass 50 nL of cell solution through the ~450 nL capillary. It would take 40 min to pass the same volume of cells into the capillary as a 10 min administration of cells at 20 nL/min. T herefore, the 20 nL/min flow rate was implemented into further studies. When t he study using 20 nL/min flow rate was repeated in triplicate, a repr oducible capture efficiency of 91.1 3.5% was obtained. Optimization of Buffer Wash Flow Rate The next group of studies compared capture efficiency at two different washin g flow rates to determine whether this step played a significant role in cell capture of this

PAGE 144

144 system. Cells administered at 20 nL/min flow rate were washed with 1500 nL buffer (~3 column volumes) at either 500 nL/min or 200 nL/min. Capture efficiency is compared in Figure 6-2 as a function of captured cells divided by the total number of cells passed through the device. The capture efficiencie s were 84.5% for the low flow rate, and 83.2% for the high flow rate, demonstrating minimal differenc e. The higher flow rate was utilized in future experiments becaus e 1500 nL buffer is passed through the system in a shorter amount of time. 0 10 20 30 40 50 60 70 80 90 100 200500Flow Rate (nL/min)Capture Efficiency (%) Figure 6-2. Comparison of capture efficien cy at two different cell washing flow rates. One capillary tested with a 200 nL/min washing flow rate was damaged during the experiment, which is why only one value is reported. Detection of CEM Cells in Blood To determine whether CEM cells could be detected in blood, red blood cells were removed from human whole bloo d and fluorescently-stained CEM cells were spiked into the white blood cell/platelet suspension at concentrations of either 1.025 or 3.004 cells/mL. At the higher conc entration, capture efficiency of CEM was 77.2%, while at the lower concentration the efficiency dropped to 8.1% (Figure 6-3A). Additionally, the

PAGE 145

145 percentage of cancer cells in the flow through fluid dropped to 2.1% from an initial CEM concentration of 20.6% of the cell population at the higher CEM initial concentration. This demonstrates enrichment of the system for the target cells. A B 0 10 20 30 40 50 60 70 80 90 1000.30 1.02[Cell] (x 10^5 cells/mL) Capture Efficiency (%) Figure 6-3. Capture of CEM cells spiked in blood. A) Graphical representation of cell capture efficiency at different cell c oncentrations. B) Fluorescent images of cells captured in the square capilla ry. Each image represents ~1.3 mm capillary length. CEM cells retained in the capillaries were easily differentiated from nonspecifically bound blood cells due to the well-resolved fl uorescent signal produced from the cell stain imaged in the square capillaries (Figure 6-3B). Therefore, nonspecific binding was essentially irrelevant since cancerous and normal cells could be distinguished by monitoring fluorescent light. While this is clearly not practical for patient samples, a labeling-after-capture method administering fluorescently-labeled aptamers to cells captured on the capillary wall wo uld have a similar effect. The CellSearch System relies

PAGE 146

146 on a related concept, utilizing various stained antibodies to make a distinction between cancerous and normal blood cells.18 Colon Cancer Aptamers (witho ut Linker) for DLD-1 Capture The function of the system was also analyzed for cancer cells not typically found in high concentration in the blood. A variety of aptamers have previously been selected by this group demonstrating binding to colorect al adenocarcinoma cancer cells (DLD-1 cell line) and/or HCT 116 colorectal carcinoma cell line. Aptamer KDED2a-3 binds specifically to DLD-1 cell line (Kd= 29.2 nM) with minimal affinity for HCT 116, while aptamer KCHA10 binds to both DLD-1 and HCT 116 cell lines (Kd= 21.3 nM for HCT 116).203 0.00 10.00 20.00 30.00 40.00 50.00 60.00 70.00 80.00 90.00 100.00 KDED2a-3 KCHA10AptamerCapture Efficiency (%) Figure 6-4. Comparison of capture effici ency of colon cancer cell line DLD-1 with aptamers KDED2a-3 and KCHA10. Both aptamers (biotinylated but lacking poly(T)10 linkers) were assessed in terms of their capture efficiency for DLD-1 cells. In initial studies, ce lls aggregated in large clumps, and were not able to pass through t he capillary in significant numbers. Implementing a cell straining step followed by constant cell agitation helped to reduce,

PAGE 147

147 but not entirely alleviate this concern. Figure 6-4 shows that aptamer KDED2a-3 demonstrated a higher capture efficiency and lower error than aptamer KCHA10. The aptamers were then synthesized with poly(T)10 linkers to determine if the linker had an effect on capture efficiency. Colon Cancer Aptamers with poly(T)10 Linker for DLD-1 Capture The two aptamers were each synthesized with a 3-biotin-poly(T)10 linker in order to provide greater aptamer flexibility for binding according to previous studies.93 These aptamers were tested for binding to DLD-1 ce lls by flow cytometry (Figure 6-5). The results show that the aptamers with the linker still bind to the cells with a similar affinity to those lacking the linker. Cells Library 2a-3 (no T10) 2a-3 (with T10) A10 (with T10) Figure 6-5. Flow cytometry comparison of aptamers binding to DL D-1 with and without poly(T)10 linker. The aptamers with linker were then immobili zed in the capillary device to asses their ability to capture DLD-1 cells. On ce again, aptamer KDED2a-3 demonstrated higher capture efficiency (83.6 5.8%) t han KCHA10 with much lo wer error associated

PAGE 148

148 with the measurement (F igure 6-6). The average of 2 measurements were higher for KDED2a-3 aptamer with 3-poly(T)10 linker than without the linker, 83.2% versus 74.1%, but the measurements fell within the error of each other. The capture efficiency was relatively high, but the low absolute num ber of cells passing through the device compared to the numbers expected indicates that cell aggregation may still be occurring. The large error associated wit h capture by KCHA10 may be because the KCHA10 aptamer was actually selected for HCT 116 instead of DLD1. This aptamer was also tested with HCT 116 cell line to determine if this trend continued. 0 10 20 30 40 50 60 70 80 90 100 KDED2a-3 KCHA10AptamerCapture Efficiency (%) Figure 6-6. Capture efficiency of DLD-1 aptamers with 3-poly(T)10 linker. HCT 116 Cell Capture by Aptamer KCHA10 in Non-enzymatic Buffer The initial cell concentration was reduced 5-fold and the cells were administer ed for a longer amount of time to discourage cell clumping and in crease the amounts of cells passing through the device. However, immediately after cell straining, the HCT 116 cells aggregated, and minimal cells were c aptured in the capillary. Thus, the cells

PAGE 149

149 were diluted in non-enzymatic buffer (NEB) directly before capillary introduction to inhibit cell-cell interactions. The bind ing of KDED2a-3 and KCHA10 with HCT 116 with non-enzymatic buffer or diluted with BB2 we re compared in Figure 6-7. KDED2a-3 has previously been shown to have a much lower affinity for HCT 116 cells than for DLD-1, as demonstrated by comparison of fluorescence intensity of Figure 6-7 and Figure 6-5. The binding of both aptamers to the cells was not affected by cell treatment with NEB, since cells found in NEB had a simila r shift as those diluted in BB2. Cells in BB2 Cells in NEB Lib in BB2 Lib in NEB KDED2a-3 in BB2 KDED2a-3 in NEB KCHA10 in BB2 KCHA10 in NEB Figure 6-7. Flow cytometric comparison of aptamer binding with HCT 116 cells diluted in BB2 or with non-enzymatic buffer (NEB). When the capillary system was tested with KCHA10/HCT 116 binding, three measurements resulted in an average capture efficiency of 97. 2 2.8% (Figure 6-8). This capture efficiency was higher than DLD-1 values, and even higher than CEM capture efficiency. Also, the error in K CHA10 binding to HCT 116 was much lower than that of the probe binding DLD-1 cells. However, the number of cells passing through the device compared to the num ber expected was minimal, as was the case with DLD-1 cells. This is likely due to aggregation effe cts of adherent cells clogging the capillary.

PAGE 150

150 0 10 20 30 40 50 60 70 80 90 100 DLD-1HCT 116Cell LineCapture Efficiency ( ( % ) Figure 6-8. Overall capture efficiency of colon cancer cell lines DLD-1 and HCT 116 in buffer. Conclusions The flow rate of the CEM cells through t he capillary system was optimized to give improved capture efficiency over the proof-of-principle st udies. The capture efficiency in the proof-of-principle studies was only ~5 % in the best case scenario, and after optimization the capture effici ency of CEM cells was raised to 91.1 3.5%. The flow rate of the buffer used to wash unbound cell s from the capillary did not appear to appreciably change the capture effi ciency at the flow rates tested. A wider range of flow rates tested may have some effect, although the capture efficiency is already considerably high. The capture efficiency is comparable and in some cases improved upon outcomes obtained for microfluidic devices reported in the literature, emphasizing the utility of simple capillary systems.92-94 The device was also able to capture and detect leukemia cells in blood. Cells were detected at both 15 and 34 cells/mL concentrations when red blood cells

PAGE 151

151 were removed prior to introduction. The CE M cells were stained and spiked into the blood, then counted by scanning for fluoresc ence along the capillary window. A highimpact study would be to try labeling-after-cap ture techniques to mi mic more realistic conditions. For example, CEM cells (unl abeled) would be spiked into the blood and captured in the capillary as before, but fluorescently-lab eled aptamers are then flowed through the system, specifically binding to capt ured target cancer cells to facilitate fluorescent-imaging detection. Cancer cells which are not normally found in high concentration in the blood were also efficiently captured by the capillary system, even though the affinities for each cell line/aptamer pair were not as high as the CEM/sgc8 system. Two colon cancer cell lines, DLD-1 and HCT 116 were captured by aptamers previously selected by this group. DLD-1 cells were captured by aptamer KDED2a-3 at efficiencies of 83.6 5.8%, and aptamer KCHA10 demonstrat ed capture efficiencies (HCT 116 cell line) of 97.2 2.8% in buffer. Despite the success of the capillaries in capturing cancer cells, the drawbacks must be considered for a complete assessm ent of capillary potential. The major concern with a capillary will always be throughpu t. Capillaries generally have microliter volumes, so when low flow rates are used for cell capture only a very small sample volume is examined. This property limits sa mples to those in which target cells are found in high concentration, or in which a pr econcentration step is performed. This is not a step unique to capillaries, as even the commercialized CellSearch Assay requires sample pretreatment.18 Capillaries can also be mult iplexed to increase throughput and

PAGE 152

152 introduce diverse capture ligands to the sample.95 Another challenge is the relative ease in breaking the tubes once the coating is removed for imaging purposes. One problem specific to the colon cancer studies was that much lower amounts of cells were passed through the device than would be expected based on initial cell concentration and introduction time. Two possible explanations for this phenomenon both result in clogging of the capillary; cell aggregation an d increased cell diameter. Adherent cells are known to interact with one another to form clumps containing many cells.204 This effect was observed by imaging the cell solution prior to introduction into the capillary. A cell straining step precedi ng cell introduction increased the numbers of cells passing through the capill ary, but the number s still were not as high as would be expected based on the initial cell concentra tion (~15%). The CEM cells were passed through the device in much higher number s, sometimes greater than 100%, which suggests this is a problem with DLD-1 and HCT 116 cell integratio n, not the system itself. Even though the cells are strained, upon constriction in the reduced diameter of the capillary, the cells are forced into closer contact with one another, possibly resulting in aggregation inside the device. Also, the DLD1 cells are visually much larger than the ~9 m diameter of CEM cells, and actually appear larger than the HCT 116 cells, which have a reported average diameter of 12 m.205, 206 No concrete measurement of DLD-1 cell diameter is available, but if the dimensions are ~20 m, only 4 cells in width could block one dimension of the capillary tube. To ease this concern, square or rectangular capillaries of larger dimensions coul d be implemented (standar d square tubing is available from PolyMicro Technologies up to 200 m 200 m dimensions). Rectangular tubes are particularl y attractive because capillar ies only slightly larger than

PAGE 153

153 the cells in height would incr ease the probability of the cell contacting the immobilized ligands, yet the increased width would prevent cell clogging. In summary, the aptamer-based capillary de vice was able to capture leukemia and colon cancer cells in buffer in numbers co mparable to literature values for standard microfluidic devices, with a system much simpler in design, fabrication, and function. The conditions optimized for a high capture efficiency of CEM cells were also applicable to two different colon cancer cell lines. In addition, the capillary system could detect stained leukemia cells which were spiked into human blood by scanning the length of the capillary for fluorescent cells. This aptamer-immobilized device has demonstrated many attributes which may promote its us e as a technique for efficient cancer cell detection.

PAGE 154

154 CHAPTER 7 SUMMARY AND FUTURE WORK Summary Aptamers have proven versatile in applicat ion since the original successes of the concept in 1990. A multitude of analytical and therapeutic uses of aptamers have been reported, and the number of aptamer-based drugs progressing through clinical trials implies this trend will continue. In this di ssertation, novel methods were combined to produce the first known aptamer antidote for a currently-exi sting pharmaceutical drug, and an aptamer-immobilized capillary provided a simple yet highly efficient means for selectively detecting cancer cells. In the drug-SELEX project, the peptide anticoagulant bivalirudin was immobilized on a monolithic disk stationar y phase, and binding sequences were collecting using a high salt gradient in conjunction with LPC. Fractions of eluted DNA were collected, and the amount of sequences elut ed from the drug-immobilized disk was compared to that of a blank disk via qPCR. The basic assu mption was that more DNA binding to the drug-containing disk would indica te that at least some fraction of the sequences were specifically binding to the dr ug instead of solely to the st ationary phase matrix. After only 2 rounds, an AlphaScreen assay was perfo rmed to select the highest-affinity pool for sequencing. Six of the aptamer candi dates tested demonstrated binding to the target, and 2 of the sequences characterized by FA generated Kd values in the low micromolar range, while control sequences displayed Kd values at least 20-fold higher. The antidote potential of aptam er JPB5 was evaluated, establishing a dose-dependent decrease in anticoagulant activi ty, with a concentration of 20 M effectively reversing the anticoagulant effect. A DNA microarray was developed to truncate the sequence to

PAGE 155

155 <40-mer while still retaining binding activi ty, resulting in abbreviated JPB5 31-mer and JPB2 21-mer candidates. The conclusion from this project is that DNA aptamers can serve as therapeutic antidotes to pharmaceutical drugs. The sgc8 aptamer specifically binding to CEM leukemia cells was evaluated in proof-of-principle studies to act as a capillary affinity chromatography stationary phase ligand. The amount of aptamer immobilized on the inner wall of the square capillary could be controlled by varying the concentrati on of sgc8 administered. Target cells were captured in higher amounts than nontarget Ramos cells even when the cell lines were mixed prior to introduction into the capillary. Additionally, square capillaries provided imaging properties superior to st andard circular diameter capillaries, an important characteristic for detecting canc er cells. When the flow rate of cell introduction was optimized, the capture effici ency of the system was raised from ~5% to >90%, enabling detection of CEM cells in hum an blood. Two colon cancer cell lines and aptamers were incorporated into the system using the optimized CEM flow rates, and the colon cancer cells also demonstrated high capture efficiency. This system was much simpler in design, fabrication, and application than standard microfluidic devices, yet provided similar capture efficiencies. Thus, aptamers can function as a highlyspecific stationary phase for cancer cell detection. Future Work Aptamer Antidote Project The Kd values for this experiment were obtained by labeling the target and measuring the anisotropy as increasing am ounts of aptamer was titrated into the solution. However, label-free detection met hods are highly desirable for this purpose because labeling aptamer or target may cha nge the binding properties of the system.

PAGE 156

156 Therefore, com paring the FA Kd values to those obtained fr om a label-free method such as ITC may have great significance. Ideally, the affinities and antidote activity of all aptamers would be repor ted, and crystallography studies of the aptamer/bivalirudin structures would be beneficial as well. One effect that was observ ed but not discussed is rela ted to the effect of Ca2+ on aptamer binding. When Kd studies were carried out in buffe r that contained calcium, the anisotropy demonstrated a time-dependent in crease followed by a plateau. We hypothesize that the timedependence is related to the higher affinity of Ca2+ for DNA than Mg2+.207 The Mg2+ competes with the Ca2+ over time, resulting in a more flexible DNA structure capable of increased target binding.208 Studying this effect could provide important insights into the ro le of metal ion binding for t he aptamers studied. However, it is important to note that experiments in vivo may not show this effect due to the complex internal environment balancing t he process more than the closed buffer system. Ongoing work includes study of the truncated sequences for antidote activity using the buffer screening assay. Also of interest is PEGylating the truncated and full-length aptamers to determine whether this can decrease nuclease degradation in both plasma and in vivo models. Incorporating modified bas es such as LNA into the aptamer structure could be carried out with the same goal. Currently, this project is in the approval stage for small-scale in vivo animal studies. The results of these experiments will be a main factor in determining the ul timate future of the selected aptamers. Cell Affinity Project Interesting studies for this project woul d be to examine t he effect of different capillary internal diameters on the captur e efficiency of the syst em. Increasing the

PAGE 157

157 cross-sectional area of the t ubes will decrease the effective linear velocity of the cells traveling through the system. This will likely result in higher capture efficiencies for larger diameter tubes, and may also allow for higher volumetric flow rates implemented into the system (increased throughput). As discussed earlier, experiments performed in rectangular tubes may be beneficial for increasing capture efficiency and decreasing capillary clogging of adherent cell lines by increasing cell/ligand interactions in one dimension yet maintaining an increased width for cell passage. Other valuable studies would be to compare the capture efficiencies obtained with a pulsed flow to those of continuous flow. In theory, pulsed fl ow would allow cell receptors an increased contact time with the aptamers, increas ing the amount of cell/ligand bonds formed. Higher throughput ma y be possible using this technique since increased flow rates could be implemented to fill the capillary at each pulse. A pretreatment step usi ng magnetic EpCAM-coated b eads (similar to the CellSearch pretreatment)18 compared to the Ficoll-Paqu e enrichment step for whole blood samples would be a study of interest. Al so of importance is study of a technique to discriminate between cancerous and noncanc erous cells captured from blood. The method in this work involved prestaining cancer cells and spiking them into blood samples, clearly not applicable to real pat ient samples. One possible method to generate a similar effect would be to implement a label-after-capture step in which fluorescently-labeled aptamers are introduced to selectively bind cancer cells captured on the capillary walls. The length of the capillary could be imaged to count only the fluorescent cells, effectively eliminati ng concerns of nonspecific binding.

PAGE 158

158 LIST OF REFERENCES 1. B. N. Dominy, Curr. Pharm. Biotechnol. 2008, 9 87-95. 2. T. Schoneberg, M. Hofreite r, A. Schulz and H. Rompler, Trends Pharmacol. Sci. 2007, 28, 117-121. 3. S. Henikoff and K. Ahmad, Annu. Rev. Cell Dev. Biol. 2005, 21, 133-153. 4. E. A. MacGregor, S. Janecek and B. Svensson, Biochimica Et Biophysica ActaProtein Structure and Molecular Enzymology 2001, 1546 1-20. 5. A. Serganov, A. Polonskaia, A. T. Phan, R. R. Breaker and D. J. Patel, Nature 2006, 441 1167-1171. 6. N. Hirokawa, Science 1998, 279 519-526. 7. P. Stenlund, A. FrostellKarlsson and O. P. Karlsson, Anal. Biochem. 2006, 353 217-225. 8. H. Nordin, M. Jungnelius, R. Karlsson and O. P. Karlsson, Anal. Biochem. 2005, 340 359-368. 9. Y. D. Ivanov, I. P. Kanaeva, Karuzina, II, A. I. Archakov, G. H. B. Hoa and S. G. Sligar, Arch. Biochem. Biophys. 2001, 391 255-264. 10. F. M. Huennekens, Advances in Enzyme Regulation, Vol 34 1994, 34, 397-419. 11. F. R. Cockerill and R. S. Edson, Mayo Clinic Proceedings 1991, 66, 1260-1269. 12. D. Fourmy, M. I. Recht, S. C. Blanchard and J. D. Puglisi, Science 1996, 274 1367-1371. 13. A. Sin, S. K. Murthy, A. Re vzin, R. G. Tompkins and M. Toner, Biotechnol. Bioeng. 2005, 91, 816-826. 14. C. L. Sawyers, Nature 2008, 452 548-552. 15. A. Beck, T. Wurch, C. Bailly and N. Corvaia, Nat. Rev. Immunol. 2010, 10, 345352. 16. J. Wojciechowski, D. Danley, J. Cooper, N. Yazvenko and C. R. Taitt, Sensors 2010, 10, 3351-3362. 17. Y. Chen, S. Tang, S. Ding, F. He and X. Xiao, Anal. Lett. 2009, 42, 2793-2806. 18. M. Naoe, Y. Ogawa, J. Mo rita, K. Omori, K. Takeshita, T. Shichijyo, T. Okumura, A. Igarashi, A. Yanaihara, S. Iwamoto, T. Fukagai, A. Miyazaki and H. Yoshida, Cancer 2007, 109 1439-1445.

PAGE 159

159 19. D. R. Burton, Nat. Rev. Immunol. 2002, 2 706-713. 20. O. Leavy, Nat. Rev. Immunol. 2010, 10, 297-297. 21. J. M. Reichert, MAbs 2010 2 84-100. 22. R. Stoltenburg, C. Re inemann and B. Strehlitz, Biomol. Eng. 2007, 24, 381-403. 23. S. M. Nimjee, C. P. Rusconi and B. A. Sullenger, Annu. Rev. Med. 2005, 56, 555583. 24. S. D. Jayasena, Clinical chemistry 1999, 45, 1628-1650. 25. J. R. de Oliveira, F. B. Nunes, M. G. Simoes Pires, D. A. da Silva Melo and M. V. Fagundes Donadio, Anti-Infect. Agents Med. Chem. 2008, 7 249-257. 26. T. Sampson, World Patent Information, 2003, 25, 123-129. 27. E. Luzi, M. Minunni, S. Tombelli and M. Mascini, Trac-Trends in Analytical Chemistry 2003, 22, 810-818. 28. D. Shangguan, Y. Li, Z. W. Tang, Z. H. C. Cao, H. W. Chen, P. Mallikaratchy, K. Sefah, C. Y. J. Yang and W. H. Tan, Proc. Natl. Acad. Sci. U. S. A. 2006, 103 11838-11843. 29. A. D. Ellington and J. W. Szostak, Nature 1990, 346 818-822. 30. C. Tuerk and L. Gold, Science 1990, 249 505-510. 31. J. Ciesiolka, J. Gorski and M. Yarus, Rna-a Publication of the Rna Society 1995, 1 538-550. 32. H. P. Hofmann, S. Limmer, V. Hornung and M. Sprinzl, Rna-a Publication of the Rna Society 1997, 3 1289-1300. 33. D. Mann, C. Reinemann, R. Stoltenburg and B. Strehlitz, Biochem. Biophys. Res. Commun. 2005, 338 1928-1934. 34. D. Grate and C. Wilson, Bioorg. Med. Chem., 2001, 9 2565-2570. 35. A. D. Ellington and J. W. Szostak, Nature 1992, 355 850-852. 36. J. J. Tang, J. W. Xi e, N. S. Shao and Y. Yan, Electrophoresis 2006, 27, 13031311. 37. C. Mannironi, A. DiNardo, P. Frus coloni and G. P. Tocchini-Valentini, Biochemistry 1997, 36 9726-9734.

PAGE 160

160 38. T. Kato, T. Takemura, K. Yano, K. Ikebukuro and I. Karube, Biochimica Et Biophysica Acta-Gene St ructure and Expression 2000, 1493 12-18. 39. M. Meli, J. Ve rgne, J. L. Decout and M. C. Maurel, J. Biol. Chem. 2002, 277 2104-2111. 40. M. Sassanfar and J. W. Szostak, Nature 1993, 364 550-553. 41. D. Kiga, Y. Futamura, K. Sakamoto and S. Yokoyama, Nucleic Acids Res. 1998, 26, 1755-1760. 42. C. T. Lauhon and J. W. Szostak, J. Am. Chem. Soc. 1995, 117 1246-1257. 43. P. Burgstaller and M. Famulok, Angewandte Chemie-Inter national Edition in English, 1994, 33, 1084-1087. 44. C. Wilson, J. Nix and J. Szostak, Biochemistry 1998, 37, 14410-14419. 45. J. H. Ko, B. Cho, J. K. Ahn, Y. Lee and I. Park, Bull. Korean Chem. Soc. 1999, 20, 1335-1339. 46. C. Boiziau, E. Dausse, L. Yurchenko and J. J. Toulme, J. Biol. Chem. 1999, 274 12730-12737. 47. A. Geiger, P. Burgstaller, H. vonderEltz, A. Roeder and M. Famulok, Nucleic Acids Res. 1996, 24, 1029-1036. 48. M. Famulok and J. W. Szostak, J. Am. Chem. Soc. 1992, 114 3990-3991. 49. I. Majerfeld, D. Puthenvedu and M. Yarus, J. Mol. Evol. 2005, 61, 226-235. 50. E. Fukusaki, T. Kato, H. Maeda, N. Kawazoe, Y. Ito, A. Okazawa, S. Kajiyama and A. Kobayashi, Bioorg. Med. Chem. Lett. 2000, 10, 423-425. 51. C. Srisawat, I. J. Go ldstein and D. R. Engelke, Nucleic acids research, 2001, 29, 1-5. 52. S. M. Lato, A. R. Boles and A. D. Ellington, Chem. Biol. 1995, 2 291-303. 53. S. T. Wallace and R. Schroeder, Rna-a Publication of the Rna Society 1998, 4 112-123. 54. C. Berens, A. T hain and R. Schroeder, Bioorg. Med. Chem. 2001, 9 2549-2556. 55. D. Nieuwlandt, M. Wecker and L. Gold, Biochemistry 1995, 34, 5651-5659. 56. F. Ylera, R. Lurz, V. A. Erdmann and J. P. Furste, Biochemical and biophysical research communications 2002, 290 1583-1588.

PAGE 161

161 57. R. White, C. Rusconi, E. Scardino, A. Wolberg, J. Lawson, M. Hoffman and B. Sullenger, Mol. Ther., 2001, 4 567-573. 58. A. Tahiri-Alaoui, L. Frigotto, N. Manvil le, J. Ibrahim, P. Romby and W. James, Nucleic Acids Res. 2002, 30. 59. T. W. Wiegand, P. B. William s, S. C. Dreskin, M. H. Jouvin, J. P. Kinet and D. Tasset, J. Immunol. 1996, 157 221-230. 60. L. S. Green, D. Jellinek, R. Jenison, A. Ostman, C. H. Heldin and N. Janjic, Biochemistry 1996, 35 14413-14424. 61. D. Jellinek, L. S. Gree n, C. Bell and N. Janjic, Biochemistry 1994, 33, 1045010456. 62. D. A. Daniels, H. Chen, B. J. Hicke, K. M. Swiderek and L. Gold, Proc. Natl. Acad. Sci. U. S. A. 2003, 100 15416-15421. 63. Y. J. Lee and S. W. Lee, Journal of microbiol ogy and biotechnology, 2006, 16, 1149-1153. 64. Z. W. Tang, D. Shangguan, K. M. Wang, H. Shi, K. Sefah, P. Mallikratchy, H. W. Chen, Y. Li and W. H. Tan, Anal. Chem. 2007, 79, 4900-4907. 65. H. W. Chen, C. D. Medley, K. Sefah, D. Shangguan, Z. W. T ang, L. Meng, J. E. Smith and W. H. Tan, ChemMedChem 2008, 3 991-1001. 66. K. Sefah, Z. W. Tang, D. H. Shangguan, H. Chen, D. Lopez-Colon, Y. Li, P. Parekh, J. Martin, L. Meng, J. A. Ph illips, Y. M. Kim and W. H. Tan, Leukemia 2009, 23, 235-244. 67. D. H. Shangguan, L. Meng, Z. H. C. Cao, Z. Y. Xiao, X. H. Fang, Y. Li, D. Cardona, R. P. Witek, C. Liu and W. H. Tan, Anal. Chem. 2008, 80 721-728. 68. Z. Tang, P. Parekh, P. Tur ner, R. W. Moyer and W. Tan, Clinical chemistry, 2009, 55, 813-822. 69. W. H. Pan, R. C. Craven, Q. Qiu, C. B. Wilson, J. W. Wills, S. Golovine and J. F. Wang, Proc. Natl. Acad. Sci. U. S. A. 1995, 92, 11509-11513. 70. J. G. Bruno and J. L. Kiel, Biosens. Bioelectron. 1999, 14 457-464. 71. C. L. A. Hamula, H. Q. Zhang, L. L. Guan, X. F. Li and X. C. Le, Anal. Chem. 2008, 80, 7812-7819. 72. S. Tombelli, A. Minunni and A. Mascini, Biosens. Bioelectron., 2005, 20, 24242434.

PAGE 162

162 73. L. S. Green, D. Jellinek, C. Bell, L. A. Beebe, B. D. Fe istner, S. C. Gill, F. M. Jucker and N. Janjic, Chem. Biol. 1995, 2 683-695. 74. A. Rhodes, A. Deakin, J. Spaull, B. Coomber, A. Aitken, P. Life and S. Rees, J. Biol. Chem., 2000, 275 28555-28561. 75. J. Ruckman, L. S. Green, J. Beeson, S. Wa ugh, W. L. Gillette, D. D. Henninger, L. Claesson-Welsh and N. Janjic, J. Biol. Chem. 1998, 273 20556-20567. 76. M. Petersen and J. Wengel, Trends Biotechnol. 2003, 21, 74-81. 77. S. C. B. Gopinath, Anal. Bioanal. Chem. 2007, 387 171-182. 78. C. Ravelet, C. Grosset and E. Peyrin, Journal of Chromatography A 2006, 1117, 1-10. 79. T. S. Romig, C. Bell and D. W. Drolet, Journal of Chromatography B-Analytical Technologies in the Biomedical and Life Sciences 1999, 731 275-284. 80. Q. Deng, I. German, D. Buchanan and R. T. Kennedy, Anal. Chem. 2001, 73 5415-5421. 81. M. A. Rehder-Silinski and L. B. McGown, Journal of Chromatography A 2003, 1008 233-245. 82. J. Ruta, C. Ravelet, I. Baussa nne, J. L. Decout and E. Peyrin, Anal. Chem. 2007, 79, 4716-4719. 83. M. Michaud, E. Jourdan, A. Villet, A. Ravel, C. Grosset and E. Peyrin, J. Am. Chem. Soc., 2003, 125 8672-8679. 84. M. Liss, B. Petersen, H. Wolf and E. Prohaska, Anal. Chem. 2002, 74, 4488-4495. 85. M. Lee and D. R. Walt, Anal. Biochem. 2000, 282 142-146. 86. T. G. McCauley, N. Hamaguchi and M. Stanton, Anal. Biochem. 2003, 319 244250. 87. R. Kirby, E. J. Cho, B. Gehrke, T. Bayer, Y. S. Park, D. P. Neikirk, J. T. McDevitt and A. D. Ellington, Anal. Chem. 2004, 76, 4066-4075. 88. C. A. Savran, S. M. Knudsen, A. D. Ellin gton and S. R. Manalis, Anal. Chem. 2004, 76, 3194-3198. 89. S. D. Jhaveri, R. Kirby, R. Conrad, E. J. Maglott, M. Bowser, R. T. Kennedy, G. Glick and A. D. Ellington, J. Am. Chem. Soc. 2000, 122 2469-2473. 90. R. Yamamoto and P. K. R. Kumar, Genes Cells, 2000, 5 389-396.

PAGE 163

163 91. K. Wang, M. K. Mars hall, G. Garza and D. Pappas, Anal. Chem. 2008, 80, 21182124. 92. Y. Xu, J. A. Phillips, J. Y an, Q. Li, Z. H. Fan and W. Tan, Analytical chemistry 2009, 81, 7436-7442. 93. J. A. Phillips, Y. Xu, Z. Xia, Z. H. Fan and W. Tan, Analytical chemistry 2009, 81, 1033-1039. 94. U. Dharmasiri, S. Balamurugan, A. A. A dams, P. I. Okagbare, A. Obubuafo and S. A. Soper, Electrophoresis 2009, 30, 3289-3300. 95. D. Pappas and K. Wang, Analytica chimica acta 2007, 601 26-35. 96. M. Blank, T. Weinschenk, M. Priemer and H. Schluesener, J. Biol. Chem. 2001, 276 16464-16468. 97. S. Fredriksson, M. Gullberg, J. Jarvius, C. Olsson, K. Pietras, S. M. Gustafsdottir, A. Ostman and U. Landegren, Nat. Biotechnol., 2002, 20, 473-477. 98. L. S. Green, C. Bell and N. Janjic, BioTechniques 2001, 30, 1094-1110. 99. C. D. Pieve, A. C. Perkins and S. Missailidis, Nucl. Med. Biol. 2009, 36, 703-710. 100. J. Charlton, J. Sennello and D. Smith, Chem. Biol. 1997, 4 809-816. 101. S. M. Taghdisi, K. Abnous F. Mosaffa and J. Behravan, J. Drug Targeting 2010, 18, 277-281. 102. Y. Wu, K. Sefah, H. Liu, R. Wang and W. Tan, Proceedings of the National Academy of Sciences of t he United States of America 2010, 107 5-10. 103. G. J. Tong, S. C. Hsiao, Z. M. Carrico and M. B. Francis, J. Am. Chem. Soc. 2009, 131 11174-11178. 104. D. Proske, M. Blank R. Buhmann and A. Resch, Appl. Microbiol. Biotechnol. 2005, 69, 367-374. 105. K. W. Thiel and P. H. Giangrande, Oligonucleotides 2009, 19, 209-222. 106. B. A. Sullenger, H. F. Galla rdo, G. E. Ungers and E. Gilboa, Cell 1990, 63, 601608. 107. C. M. Browning, L. Cagnon, P. D. Good, J. Rossi, D. R. Engelke and D. M. Markovitz, J. Virol. 1999, 73, 5191-5195. 108. C. P. Rusconi, J. D. Roberts G. A. Pitoc, S. M. Nimjee, R. R. White, G. Quick, E. Scardino, W. P. Fa y and B. A. Sullenger, Nat. Biotechnol. 2004, 22, 1423-1428.

PAGE 164

164 109. S. Santulli-Marotto, S. K. Nair, C. Rusconi, B. Sullenger and E. Gilboa, Cancer Res. 2003, 63, 7483-7489. 110. Y. Kim, Z. Cao and W. Tan, Proc. Natl. Acad. Sci. U. S. A. 2008, 105 5664-5669. 111. P. J. Bates, J. B. Kahlon, S. D. Thomas, J. O. Trent and D. M. Miller, J. Biol. Chem. 1999, 274 26369-26377. 112. B. J. Hicke, C. Marion, Y. F. Chang, T. Gould, C. K. Lynott, D. Parma, P. G. Schmidt and S. Warren, J. Biol. Chem. 2001, 276 48644-48654. 113. L. C. Bock, L. C. Griffin, J. A. Latham, E. H. Vermaas and J. J. Toole, Nature 1992, 355 564-566. 114. C. P. Rusconi, E. Scardino, J. Layzer, G. A. Pitoc, T. L. Orte l, D. Monroe and B. A. Sullenger, Nature 2002, 419 90-94. 115. C. P. Rusconi, A. Yeh, H. K. Ly erly, J. H. Lawson and B. A. Sullenger, Thromb. Haemostasis 2000, 84, 841-848. 116. S. K. Buddai, J. M. Layzer, G. Lu, C. P. Rusconi, B. A. Sullenger, D. M. Monroe and S. Krishnaswamy, J. Biol. Chem. 2010, 285 5212-5223. 117. S. W. Gal, S. Amontov, P. T. Urvil, D. Vishnuvardhan, F. Nishikawa, P. K. R. Kumar and S. Nishikawa, Eur. J. Biochem. 1998, 252 553-562. 118. S. Oney, S. M. Nimjee, J. Layzer, N. Que-Gewirth, D. Ginsburg, R. C. Becker, G. Arepally and B. A. Sullenger, Oligonucleotides 2007, 17, 265-274. 119. S. Oney, R. T. S. Lam, K. M. Bompiani, C. M. Blake, G. Quick, J. D. Heidel, J. Y. C. Liu, B. C. Mack, M. E. Da vis, K. W. Leong and B. A. Sullenger, Nat. Med. (N. Y., NY, U. S.) 2009, 15, 1224-1228. 120. A. Heckel, M. C. R. Buff, M. S. L. R addatz, J. Muller, B. Potzsch and G. Mayer, Angewandte Chemie-International Edition 2006, 45, 6748-6750. 121. Y. Kim, J. A. Phillips, H. Liu, H. Kang and W. Tan, Proceedings of the National Academy of Sciences of t he United States of America 2009, 106 6489-6494. 122. A. Joachimi, G. Mayer and J. S. Hartig, J. Am. Chem. Soc. 2007, 129 3036-7. 123. J. Fanikos, J. L. Cina, S. Baroletti, K. Fiumara, L. Matta and S. Z. Goldhaber, American Journal of Cardiology 2007, 100 1465-1469. 124. D. Monroe, M. Hoffman and H. R. Roberts, Arteriosclerosis Thrombosis and Vascular Biology, 2002, 22, 1381-1389. 125. T. E. Warkentin, Best Practice & Research Clinical Haematology 2004, 17, 105125.

PAGE 165

165 126. J. I. Weitz and M. Crowther, Thromb. Res. 2002, 106 V275-V284. 127. E. Di Cera, Chest 2003, 124 11S-17S. 128. W. Bode, Blood Cells Molecules and Diseases 2006, 36, 122-130. 129. M. S. Conners II I and S. R. Money, The Ochsner Journal 2002, 4 41-47. 130. P. C. A. Kam, N. Kaur and C. L. Thong, Anaesthesia 2005, 60, 565-574. 131. T. M. Sciulli and V. F. Mauro, Annals of Pharmacotherapy 2002, 36, 1028-1041. 132. M. Salzet, Curr. Pharm. Des. 2002, 8 493-503. 133. L. A. Linkins and J. I. Weitz, Curr. Pharm. Des. 2005, 11, 3877-3884. 134. N. W. Shammas, Cardiovascular Drug Reviews 2005, 23, 345-360. 135. M. A. A. Parry, J. M. Maraganore and S. R. Stone, Biochemistry 1994, 33, 1480714814. 136. J. I. Witting, P. Bourdon, D. V. Br ezniak, J. M. Maraganor e and J. W. Fenton, Biochem. J., 1992, 283 737-743. 137. J. M. Maraganore, P. Bourdon, J. Jablonski, K. L. Ramachandran and J. W. Fenton, Biochemistry, 1990, 29, 7095-7101. 138. R. De Cristofaro and E. De Candia, Journal of Thrombosis and Thrombolysis 2003, 15, 151-163. 139. E. Skrzypczak-Jankun, V. E. Carperos K. G. Ravichandr an, A. Tulinsky, M. Westbrook and J. M. Maraganore, J. Mol. Biol. 1991, 221 1379-1393. 140. W. Bode, I. Mayr, U. Baumann, R. Huber, S. R. Stone and J. Hofsteenge, EMBO J. 1989, 8 3467-3475. 141. D. Mark, S. Haeberle, G. Roth, F. von Stetten and R. Zengerle, Chem. Soc. Rev. 2010, 39, 1153-1182. 142. J. M. Hicks and M. Iosefsohn, N. Engl. J. Med. 1989, 320 320-321. 143. R. Pacifici, M. Farre, S. Pichini, J. Ortuno, P. N. Roset, P. Zuccaro, J. Segura and R. de la Torre, J. Anal. Toxicol. 2001, 25, 144-146. 144. J. C. Stachowiak, E. E. Shugard, B. P. Mosier, R. F. Renzi, P. F. Caton, S. M. Ferko, J. L. V. de Vreugde, D. D. Yee, B. L. Haroldsen and V. A. VanderNoot, Anal. Chem. 2007, 79 5763-5770.

PAGE 166

166 145. R. J. Meagher, A. V. Hatch, R. F. Renz i and A. K. Singh, Lab Chip 2008, 8 20462053. 146. L. Zhao, T. Wu, J.-P. Lefevr e, I. Leray and J. A. Delaire, Lab Chip 2009, 9 28182923. 147. C. Zhai, W. Qiang, J. Sheng, J. Lei and H. Ju, Journal of Chromatography A 1217 785-789. 148. E. Leclerc, K. S. Furukawa, F. Miya ta, Y. Sakai, T. Ushida and T. Fujii, Biomaterials 2004, 25, 4683-4690. 149. A. M. Elizarov, Lab Chip 2009, 9 1326-1333. 150. S. Upadhyaya and P. R. Selvaganapathy, Lab Chip 2010, 10, 341-348. 151. A. A. Adams, P. I. Oka gbare, J. Feng, M. L. Hupert, D. Patterson, J. Gottert, R. L. McCarley, D. Nikitopoulos, M. C. Murphy and S. A. Soper, J. Am. Chem. Soc. 2008, 130 8633-8641. 152. S. Nagrath, L. V. Sequis t, S. Maheswaran, D. W. Bell, D. Irimia, L. Ulkus, M. R. Smith, E. L. Kwak, S. Digum arthy, A. Muzikansky, P. Ryan, U. J. Balis, R. G. Tompkins, D. A. Haber and M. Toner, Nature 2007, 450 1235-1239. 153. K. Pantel, R. H. Brackenhoff and B. Brandt, Nat. Rev. Cancer 2008, 8 329-340. 154. K. A. Herold, Rasooly, A., Lab-on-a-Chip Technology Caister Academic Press, Norfolk, UK, 2009. 155. T. Tsuda, J. V. Sw eedler and R. N. Zare, Anal. Chem. 1990, 62, 2149-2152. 156. A. Cifuentes and H. Poppe, Chromatographia 1994, 39, 391-404. 157. Polymicro-Technologies, The Book on the Technologies of Polymicro Polymicro Technologies, LLC, 2004. 158. M. C. Peoples and H. T. Karnes, Journal of Chromatography B-Analytical Technologies in the Biomedical and Life Sciences 2008, 866 14-25. 159. Z. Guo, R. A. Guilfoy le, A. J. Thiel, R. F. Wang and L. M. Smith, Nucleic Acids Res. 1994, 22, 5456-5465. 160. A. Bogusiewicz, N. I. Mock and D. M. Mock, Anal. Biochem. 2004, 327 156-161. 161. I. N. Levine, Physical Chemistry 5 edn., McGraw-Hill, New York, 2002. 162. D. A. Hammer and D. A. Lauffenburger, Biophys. J., 1987, 52, 475-487.

PAGE 167

167 163. L. A. Olivier, J. Yen, W. M. Reicher t and G. A. Truskey, Biotechnol. Prog. 1999, 15, 33-42. 164. G. I. Bell, Science 1978, 200 618-627. 165. C. M. Hertz, D. J. Graves, D. A. Lauffenburger and F. T. Serota, Biotechnol. Bioeng. 1985, 27, 603-612. 166. S. Barkley, H. Johnson, R. Eisenthal and J. Hubble, Biotechnol. Appl. Biochem. 2004, 40, 145-149. 167. M. F. Caron and G. R. McKendall, American Journal of Health-System Pharmacy 2003, 60, 1841-1849. 168. S. C. B. Gopinath, Thromb. Res., 2008, 122 838-847. 169. T. D. Gladwell, Clinical Therapeutics 2002, 24, 38-58. 170. M. Wilchek, E. A. Bayer and O. Livnah, Immunol. Lett. 2006, 103 27-32. 171. M. Ellingham, D. H. J. Bunka, D. J. Rowlands and N. J. Stonehouse, Rna-a Publication of the Rna Society 2006, 12, 1970-1979. 172. M. Djordjevic, Biomol. Eng. 2007, 24, 179-189. 173. M. Rieseberg, C. Kasper, K. F. Reardon and T. Scheper, Appl. Microbiol. Biotechnol. 2001, 56, 350-360. 174. Z. W. Tang, P. Parekh, P. Tu rner, R. W. Moyer and W. H. Tan, Clinical chemistry 2009, 55, 813-822. 175. M. Howarth, D. J. F. Chinnapen, K. Gerro w, P. C. Dorrestein, M. R. Grandy, N. L. Kelleher, A. El-Husseini and A. Y. Ting, Nat. Methods 2006, 3 267-273. 176. D. E. Huizenga and J. W. Szostak, Biochemistry 1995, 34, 656-665. 177. W. Z. Niu, N. Jiang and Y. H. Hu, Anal. Biochem. 2007, 362 126-135. 178. N. Ishizuka, H. Kobayashi, H. Minakuchi K. Nakanishi, K. Hirao, K. Hosoya, T. Ikegami and N. Tanaka, Journal of Chromatography A 2002, 960 85-96. 179. M. W. Roberts, C. M. Ongkudon, G. M. Forde and M. K. Danquah, J. Sep. Sci. 2009, 32, 2485-2494. 180. E. G. Vlakh, G. A. Platonova, G. P. Vlasov, C. Kasper, A. Tappe, G. Kretzmer and T. B. Tennikova, Journal of Chromatography A 2003, 992 109-119. 181. E. G. Vlakh, A. Tappe, C. Kasper and T. B. Tennikova, Journal of Chromatography B-Analytical Technologies in t he Biomedical and Life Sciences 2004, 810 15-23.

PAGE 168

168 182. R. Giovannini, R. Freitag and T. B. Tennikova, Analytical chemistry, 1998, 70, 3348-3354. 183. A. Giulietti, L. Overbergh, D. Valckx, B. Decallonne, R. Boui llon and C. Mathieu, Methods 2001, 25, 386-401. 184. J. F. Glickman, X. Wu, R. Mercuri, C. Illy, B. R. Bowen, Y. He and M. Sills, Journal of Biomolecular Screening 2002, 7 3-10. 185. J. J. Tang, T. Yu, L. Guo, J. W. Xie, N. S. Shao and Z. K. He, Biosens. Bioelectron. 2007, 22, 2456-2463. 186. S. C. B. Gopinath, Y. Shikamoto, H. Miz uno and P. K. R. Kumar, Thromb. Haemostasis 2006, 95, 767-771. 187. R. C. Becker, C. Rusconi and B. Sullenger, Thromb. Haemostasis 2005, 93, 1014-1020. 188. C. Wilson and J. W. Szostak, Chem. Biol., 1998, 5 609-617. 189. T. O. Pangburn, M. A. Petersen, B. Waybrant, M. M. Adil and E. Kokkoli, Journal of Biomechanical Engineering-Transactions of the Asme, 2009, 131 1-20. 190. D. Shangguan, Z. W. T ang, P. Mallikaratchy, Z. Y. Xiao and W. H. Tan, Chembiochem 2007, 8 603-606. 191. M. Famulok, J. S. Hartig and G. Mayer, Chem. Rev. (Washington, DC, U. S.) 2007, 107 3715-3743. 192. J. R. Lakowicz, Principles of Fluorescence Spectroscopy 2 edn., Kluwer Academic/Plenum, New York, 1999. 193. X. M. Liu, G. J. Cao, H. M. Ding, D. J. Zhang, G. Yang, N. L. Liu, M. Fan, B. F. Shen and N. S. Shao, FEBS Lett. 2004, 562 125-128. 194. K. Brummel-Ziedins, C. Y. Vossen, F. R. Rosendaal, K. Umezaki and K. G. Mann, Journal of Thrombosis and Haemostasis 2005, 3 1472-1481. 195. J. B. H. Tok and N. O. Fischer, Chem. Commun. (Cambridge, U. K.) 2008, 18831885. 196. M. von Kockritz-Blickwede, O. A. Chow and V. Nizet, Blood 2009, 114 52455246. 197. V. A. Holmes and J. M. W. Wallace, Biochem. Soc. Trans. 2005, 33, 428-432. 198. J. I. Weitz, M. Hudoba, D. Massel, J. Maraganore and J. Hirsh, J. Clin. Invest. 1990, 86, 385-391.

PAGE 169

169 199. M. Armani, R. Pr obst and B. Shapiro, Lab-on-a-chip Technology: Fabrication and Microfluidics Caister Academic Press, Norfolk, UK, 2009. 200. A. Cifuentes and H. Poppe, Journal of chromatography 1994, 680 321-340. 201. F. Ming, R. Eisenthal, W. J. D. Whish and J. Hubble, Enzyme Microb. Technol. 2000, 26, 216-221. 202. Molecular Probes, Package Insert 2001, 1-5. 203. K. Sefah, Universi ty of Florida, 2009. 204. D. Glaves, Br. J. Cancer 1983, 48 665-673. 205. F. Q. Schafer and G. R. Buettner, Photochem. Photobiol. 1999, 70, 858-867. 206. C. Soncini, P. Carpinelli, L. Gianellini, D. Fancelli, P. Vianello, L. Rusconi, P. Storici, P. Zugnoni, E. Pesenti, V. Croci, R. Ceruti, M. L. Giorgini, P. Cappella, D. Ballinari, F. Sola, M. Va rasi, R. Bravo and J. Moll, Clinical Cancer Research 2006, 12, 4080-4089. 207. N. Korolev, A. P. Lyubartsev A. Rupprecht and L. Nordenskiold, Biophys. J. 1999, 77, 2736-2749. 208. N. Sundaresan, C. K. S. Pillai and C. H. Suresh, J. Phys. Chem. A 2006, 110 8826-8831.

PAGE 170

170 BIOGRAPHICAL SKETCH Jennifer Anne Martin was born in Clearfield, PA in October of 1982. After graduating from Clearfield Area High School, she attended Misericordia University to pursue a degree in chemistry. Following a fellowship at the National Institutes of Health (National Institute of Allergy and Infectious Diseases) she relocated Gainesville, FL to attend the University of Flori da in 2005. While at the University of Florida, she joined the research group of Dr. Weihong Tan for he r Ph.D. work in the area of aptamers for analytical and therapeutic applications.