Citation
Therapeutically modified hollow fiber dialyzers

Material Information

Title:
Therapeutically modified hollow fiber dialyzers
Creator:
Branham, Micheal Lee, 1955-
Publication Date:
Language:
English
Physical Description:
xv, 163 leaves : illustrations. ; 29 cm.

Subjects

Subjects / Keywords:
Research ( mesh )
Renal Dialysis -- instrumentation ( mesh )
Renal Dialysis -- methods ( mesh )
Antigens ( mesh )
Membranes, Artificial ( mesh )
Department of Medicinal Chemistry thesis Ph. D ( mesh )
Medicinal Chemistry thesis, Ph. D
Dissertations, Academic -- College of Pharmacy -- Department of Medicinal Chemistry -- UF ( mesh )
Dissertations, Academic -- Medicinal Chemistry -- UF
Genre:
Academic theses ( lcgft )
Academic theses ( fast )
bibliography ( marcgt )
non-fiction ( marcgt )

Notes

Thesis:
Thesis (Ph. D.)--University of Florida, 1998.
Bibliography:
Includes bibliographical references (leaves 155-162).
General Note:
Typescript.
General Note:
Vita.
Statement of Responsibility:
by Micheal Lee Branham.

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
The University of Florida George A. Smathers Libraries respect the intellectual property rights of others and do not claim any copyright interest in this item. This item may be protected by copyright but is made available here under a claim of fair use (17 U.S.C. §107) for non-profit research and educational purposes. Users of this work have responsibility for determining copyright status prior to reusing, publishing or reproducing this item for purposes other than what is allowed by fair use or other copyright exemptions. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder. The Smathers Libraries would like to learn more about this item and invite individuals or organizations to contact the RDS coordinator (ufdissertations@uflib.ufl.edu) with any additional information they can provide.
Resource Identifier:
029898200 ( ALEPH )
51620228 ( OCLC )

Downloads

This item has the following downloads:


Full Text
THERAPEUTICALLY MODIFIED HOLLOW FIBER DIALYZERS
By
MICHEL LEE BRANHAM
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
1998


This work is dedicated to my parents Landon Frank Branham Jr.
and Norma Elizabeth Branham, who through love and guidance
have inspired in me the joy of learning.


ACKNOWLEDGEMENTS
I would like to express my gratitude and admiration to
Professor Ian R. Tebbett for his guidance, patience and
encouragement throughout the course of this research program.
Also, I thank the very talented members of my supervisory
committee, Drs. Christopher Batich Ph.D., Edward Ross M.D.,
Kenneth Sloan Ph.D., Roger Tran-Son-Tay Ph.D. and Donna
Wielbo Ph.D. I do greatly appreciate their advice and support
during my graduate study. This project would not have been
successful without the financial support of the Florida
Education Fund and RW Johnson Graduate Research Fellowships
for which I am both honored and grateful.
For their endearing kinship and personal support I owe
special thanks to Dr. Betty Parker-Smith (Florida Education
Fund) and Mr. Robert L. Woods (University of Florida Graduate
Minority Programs). Finally, since much of the laboratory
work was done in the immunology laboratory of the Department
of Pharmacy Practice; I thank Dr. Janet Karlix for all her
assistance and the use of those facilities.


TABLE OF CONTENTS
page
ACKNOWLEDGEMENTS iii
LIST OF TABLES vii
LIST OF FIGURES ix
LIST OF ABBREVIATIONS xiii
ABSTRACT xiv
CHAPTERS
1 INTRODUCTION
1.1 Statement of Problem 1
1.2 Background Review 7
1.2.1 Clinical Implications of Middle Molecules in
Extracorporeal Therapy 7
1.2.2 The Effects of Dialysate Additives on the
Transport Properties of Donor Solutes across
Hollow Fiber Membranes 9
1.2.3 Molecular Pathogensis of Lupus Nephritis... 12
1.2.4 Anti-(DNA-Histone) Antibodies in Active Lupus
Nephritis 14
1.2.5 Surface Modification of PMMA and PSF Dialysis
Membranes 17
1.2.6 Surface-modified polymethylmethacrylate
(PMMA) 20
1.2.7 Surface-modified polysulfone (PSF) 25
1.2.8 Immunoreactivity of Immobilized Histones
Characterized using the BIACORE 2000 26
1.2.9 Surface Analysis of Surface-Modified Hollow
Fibers 28
1.2.10 Extracorporeal Immunoadsorption Therapies
in SLE 31
1.3 Hypothesis 39
1.4 Specific Aims and Rationale 4 0
IV


2 MATERIALS AND METHODS
2.1 Specific Aim #1 42
2.1.1 Physical Characteristics and Specifications
of the Model 42
2.1.2 Mass Transport Experiments 46
2.1.3 UV Spectrophotometric Analysis of Vit.B12...47
2.1.4 ELISA of CD25 and (32-M 48
2.2 Specific Aims #2 50
2.2.1 Amination, Pegylation and Protein
Immobilization on PMMA Membranes 50
2.2.2 Amination, Pegylation and Protein
Immobilization on PSF Membranes 51
2.2.3 Stability and Reactivity of Tresylated
Star-PEG Polymers 53
2.2.4 Biospecific Interaction Analysis: Kinetic
and Equilibrilium Analysis Using the
BIACORE 2000 54
2.2.5 Microscopic Analysis: Hollow Fiber Surface
Imaging Using Atomic Force and Scanning
Electron Microscopes 60
2.3 Specific Aims #3 62
2.3.1 Adsorption Capacity Experimental Methods:
AHA ELISA 62
2.3.2 Extracorporeal Immunoadsorption Experimental
Methods 68
3 RESULTS AND DISCUSSION
3.1 Specific Aim #1 71
3.1.1 An In Vitro Model for the Study of
Extracorporeal Therapy: Applications of the
Model 71
3.1.2 The Effects of Dialysate Additives on the
Transport Properties of Donor Solutes Across
Hollow Fiber Dialyzers 71
3.1.3 Summary 83
3.2 Specific Aim #2 83
3.2.1 Estimates of Tresylate Hydrolysis and
Reactivity 83
3.2.2 Immunoreactivity of Immobilized Histones
Characterized Using the BIACORE 2000 86
3.2.3 Surface Analysis of Chemically-Modified
Hollow Fibers 90
3.2.4 Summary 90
3.3 Specific Aim #3 91
3.3.1 Antihistone Adsorption Capacity of PMMA and
PSF Dialysis Membranes 91
3.3.2 Extracorporeal Immunoadsorption of
Circulating Antihistone Antibodies from
Saline Solution 102
v


3.3.3 Extracorporeal Immunoadsorption of
Circulating Antihistone Antibodies from Human
Plasma 108
3.3.4 Extracorporeal Immunoadsorption of
Circulating Antihistone Antibodies from Human
Plasma Unmodified PMMA and PSF Dialyzer...114
3.3.5 Extracorporeal Immunoadsorption of
Circulating Antihistone Antibodies from Human
Plasma Unmodified PMMA and PSF Dialyzer... 114
3.4 Summary 117
4 CONCLUSIONS 118
APPENDICES
I MASS TRANSFER IN HOLLOW FIBER DIALYZERS 122
II TUBING SELECTION TABLE 130
III A LEAST SQUARES LINEAR REGRESSION METHOD 132
IV ANTIHISTONE ANTIBODY HYBRIDOMAS 133
V STAR-PEG POLYMERS 136
VI NON-CONTACT MODE AFM: BURLEIGH ME TRIS NC-2000 145
VII PURIFICATION OF ANTIBODIES BY THIOPHILIC ADSORPTION
CHROMATOGRAPHY: T-GEL FRACTIONATION 152
LIST OF REFERENCES 155
BIOGRAPHICAL SKETCH 163
VI


LIST OF TABLES
Table page
1-1 Criteria for Classification of SLE 2
1-2 Causes of death in a cohort of systemic lupus
patients 3
1-3 Characteristics of SLE Patients Treated in Okinawa,
Japan between 1972 and 1991 4
1-4 Study of the correlation between cause of death, sex,
race and socio-economic status (SES) in patients with
SLE 5
1-5 Principal Characteristics of Calf Thymus Histones 15
1-6 A Summary of the Prevalence of Histone Autoantibodies in
SLE patients 16
1-7 Methods Used to Prepare Microporous Membrane
Materials 21
1-8 Common Surface Analytical Methods 29
1-9 Autoantibodies in SLE 35
2-1 Dialyzer Specifications 45
2-2 Circuit Specifications 45
2-3 Antigen/Antibody Concentrations in Mass Transfer
Experiments 47
3-1 Summary of the Experimental Mass Transfer Results 83
3-2 Data used in the Kinetic and Equilbrium Analysis of the
Histone-Antihistone Interaction 87
3-3 The kinetic and equilibrium constants for AHA binding
to PEG immobilized histones 89
3-4 Total mouse AHA removed from PBS solution 104
vii


3-5 Total human AHA removed from PBS solution 106
3-6 Total mouse AHA eluted from F8 dialyzers: preliminary
study 108
3-7 Total AHA removed from human plasma preliminary
study 113
3-8 Total AHA removed from human plasma by a modified
dialyzer 117
I-1 Mass transfer parameters for different flow
geometries 126
II-1 Tubing Selection and Compatibility 131
V-l Synthetic Methods for Star-PEG Polymers 137
V-2 Efficiency of Dendritic Seed polymer in Star-PEG
Synthesis 144
viii


LIST OF FIGURES
Figure page
1-1 Environmental influences may be related to the
prevalence of SLE 6
1-2 The structure and physical properties of
beta 2-microglobulin 9
1-3 The interleukin-2 receptor complex 10
1-4 Structural models of the "soluble" IL-2R alpha
subunit CD25 11
1-5 Mechanisms of Immune Complex Deposition 14
1-6 Structure of the "Histone octamer-Nucleosome core".... 15
1-7 Proposed Mechanism of Histone-Mediated Glomeruli
Deposition of Immune Complexes 17
1-8 PMMA monomer synthesis 2 0
1-9 Surface-modified PMMA 22
1-10 Formation of Potentially Toxic Hydrazine Derivatives..23
1-11 Condensation polymerization for PSF 25
1-12 Surface-modified PSF 27
1-13 Flowchart for the Surface Characterization of
Biomaterials 30
1-14 Determination of the capacity of an affinity adsorbent
by frontal analysis 33
2-1 In Vitro Model Specifications 4 3
2-2 Structure of commercial hollow fiber dialyzers 44
2-3 Calibration curve and linear regression equation for
the spectrophotometric determination of Vitamin B12....4 8
IX


56
2-4 BIACORE Biosensors
2-5 The BIACORE Sensorgram: surface plasmon angle (SPR)
angle 57
2-6 Activation of a BIACORE sensorchip 61
2-7 Adsorption/Incubation Experiment 63
2-8 Chlorosulfonation of PSF 65
3-1 In Vitro Model: applications 72
3-2 Mass transport curves of vit.Bi2 in a high flux PMMA
dialyzer 74
3-3 Mass transport curves of vit.BX2 in a high flux cellulose
acetate dialyzer 75
3-4 Mass transport curves of vit.Bi2 in a low flux PSF
dialyzer 76
3-5 Mass transport curves of P2-M in a high flux PMMA
dialyzer 77
3-6 Mass transport curves of p2-M in a high flux cellulose
acetate dialyzer 78
3-7 Mass transport curves of [32-M in a low flux PSF
dialyzer 7 9
3-8 Mass transport curves of CD25 in a high flux PMMA
dialyzer 80
3-9 Mass transport curves of CD25 in a high flux cellulose
acetate dialyzer 81
3-10 Mass transport curves of CD25 in a low flux PSF
dialyzer 82
3-11 Tresylated Star-PEG 84
3-12 PEG-Protein coupling reaction 84
3-13 Tresylate Hydrolysis 85
3-14 pH Dependent Hydrolysis of Tresylated Star-PEG
polymers 86
3-15 Antihistone-Histone Interaction Sensorgram 88
3-16 Size Estimation of the Star-PEG 92
x


94
3-17.SEM micrograph of the lumen of a treated PMMA hollow
fiber
3-18 Atomic force micrograph scanned at 2.0pm of a treated
and untreated PMMA hollow fiber 95
3-19 Atomic force micrograph scanned at 1.0pm of a treated
PMMA hollow fiber 97
3-20.Atomic force micrograph scanned at 1.0pm of a untreated
PMMA hollow fiber 98
3-21. Adsorption capacity of surface-modified PMMA
membranes 100
3-22 Adsorption capacity of chlorosulfonated surface-
modified PSF membranes 101
3-23 Adsorption capacity of nitration/reduction surface-
modified PSF membranes 10 3
3-24 Extracorporeal immunoadsorption of murine antihistone
antibodies from saline 105
3-25 Extracorporeal immunoadsorption of human antihistone
antibodies from saline 107
3-26 Extracorporeal immunoadsorption of murine antihistone
antibodies from plasma: preliminary study 109
3-27 Elution of AHA from a modified and a unmodified PSF
dialyzer 110
3-28 T-Gel fractionation of plasma samples from preliminary
study 112
3-29 Extracorporeal immunoadsorption of murine antihistone
antibodies from plasma by unmodified dialyzers 115
3-30 Extracorporeal immunoadsorption of murine antihistone
antibodies from plasma by surface-modified
dialyzers 116
I-l Hemodialysis Configuration 124
V-l Core-First Method using poly(DVB) 137
V-2 A trifunctional initiator: triisobutyl benzene
chloride 138
V-3 A Dendritic PEG star by anionic polymerization 139
V-4 Plurifunctional electrophilic deactivator 140
XI


V-5 Stars by copolymerization with a dialkenyl monomer... 140
V-6 Seed Star Methods 14 2
V-7 Preparation of dendritic polyamidoamine (PAMAM)
polymers 143
VI-1 Atomic Force Microscopy 147
VII-1 Thiophilic Adsorption Chromatography 152
X


LIST OF ABBREVIATIONS
AFM :
AHA :
AP :
B S A :
P2-M :
ds DNA:
DVB :
E DC :
ELISA:
EPO :
ERSD :
EVS :
GFR :
HRP :
IC :
IVS :
mwco :
MM s
NHS :
PAMAM:
PEG :
PEVA s
PMMA :
PNPP s
PSF :
PS :
RID :
RU :
SEM s
SES s
SIL2R:
SLE :
SPM :
SPR :
TEM :
TMB :
Atomic Force Microscope
Antihistone Antibody
Alkaline Phosphatase
Bovine Serum Albumin
beta-2-Microglobulin
double stranded DNA
Divinyl Benzene
1-Ethyl-3-(-Dimethylaminopropyl) Carbodiimide
Enzyme Linked Immunoadsorbent Assay
E rythropoietin
End Stage Renal Disease
Extravascular Space
Glomerular Filtration Rate
Horseradish Peroxidase
Immune Complex
Intravascular Space
Molecular weight cutoff
Middle Molecules
N-Hydroxysuccinimide
Polyamidoamine
Polyethylene glycol
Polyethylene vinyl acetate
polymethylmethacrylate
p-Nitrophenyl phosphate
polysulfone
polystyrene
radioimmunodif fusion
Resonance Units (also response units)
Scanning Electron Microscope
Socio-economic status
Soluble Interleukin-2 Receptor (also CD25)
Systemic lupus erythematosus
Scanning Probe Microscope
Surface Plasmon Resonance
Transmission Electron Microscope
3,3',5,5'Trimethyl Benzidine hydrochloride
xiii


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
THERAPEUTICALLY MODIFIED HOLLOW FIBER DIALYZERS
By
MICHEL LEE BRANHAM
December 1998
Chairman: Ian R. Tebbett, Ph.D.
Major Department: Medicinal Chemistry
Extracorporeal immunoadsorption is a novel strategy for
the removal of disease associated molecules from the
patient's blood stream. This approach possesses unprecedented
molecular-specificity, is potentially more cost effective
than classical drug discovery and is often the only
therapeutic option in medical emergencies or in the
terminally ill. For example, pathogenic candidates for this
method of extraction include (32-microglobulin in dialysis-
related amyloidosis, anti-nuclear autoantibodies in lupus
nephritis, as well as several soluble cytokine receptors
involved in immunodeficiency disorders like AIDS and cancer.
Since the above mentioned toxins are not removed during
conventional hemodialysis therapy, therapeutic membrane
modifications offer new opportunities to treat these and
other diseases.
XIV


Current technology exists for the immunoadsorption of
blood-borne antigens using plasmapheresis, but this method
requires the separation of plasma from the formed elements
prior to immunoadsorption. Some of the disadvantages of
plasma separation include hypervolemic effects, red cell
hemolysis, increased setup time and instrumental complexity.
We have shown that an in vitro model can be constructed which
mimics the extracorporeal circulatory system used in
conventional as well as novel dialysis therapies. The model
was used to evaluate the clearance of a select donor solute
or a group of solutes under both equilibrium and sink
conditions. Passive adsorption of (}2-M to PMMA membranes was
conclusively demonstrated in the model. New dialysate
formulations may be studied in the model as well. Here, we
found that dialysate additives which formed stable complexes
with a specific donor solute (i.e. antibodies) may increase
the clearance of that solute above and beyond steady-state
concentrations. We have also found that surface-modification
of commercial hollow fiber dialyzers is a novel method of
preparation for highly selective bioadsorbent materials. The
effects of these modification were also demonstrated in the
in vitro model. It was found that the use of polyethylene
glycol spacers in the immobilization procedure significantly
increases the surface area of the hollow fibers, it
apparently stabilizes the epitope recognized by the target
molecule (antihistone antibodies) and thereby increases
binding. Polyethylene glycol effectively prevents nonspecific
binding to the surface. The result of the experiments show
that milligram quantities of immunoglobulin can be removed
from human plasma by these surface-modified dialyzers.
XV


CHAPTER 1
INTRODUCTION
1.1 Statement of Problem
Systemic lupus erythematosus (SLE) is an autoimmune
disease of unknown etiology, characterized by the chronic
inflammation of multiple organ systems. Glomerulonephritis is
a frequent complication of SLE; the presence and extent of
kidney involvement greatly influence the outcome of this
disease [Tron and Bach 1977 ]. SLE is characterized by the
production of circulating autoantibodies to constituents of
the cellular nucleus. Although analytical difficulties still
remain, the criterion for diagnosis and classification of SLE
has been established by the American College of Rheumatology
to facilitate the uniform reporting of SLE cases in the
U.S.[Theofilopoulos and Dixon 1982] see table 1-1.
SLE can occur at any age after puberty; with women of
childbearing age being primarily affected. In cases that
begin between ages 15 and 40 yrs, greater than 90% of the
patients are female [Fessel 1974; Hochberg 1993]. These data
strongly suggest that sex hormones influence the probability
of developing or expressing SLE; studies in animal models of
lupus have supported a potential role for estrogen
enhancement and androgen protection against expression of the
disease [Roubinian et al 1978; Hochberg 1993].
1


2
The chance of a white woman in the U.S. developing SLE
in her lifetime is approximately 1 in 700 (0.14%). This
incidence is 3 to 4 times greater for blacks in the U.S., as
well as in certain tribes of Native Americans (e.g. Sioux,
Crow, Arapahoe) [Feldman et al. 1992].
Table 1-1. Criteria for Classification of SLE.
Criterion
Definition
1. Malar rash
2. Discoid rash
3. Photosensitivity
4. Oral ulcers
5. Arthritis
6. Serositis
7 Renal disorder
8. Neurologic disorder
9. Hematologic disorder
10. Immunologic disorders
11. Antinuclear antibody
(ANA)
Fixed erythema, flat or raised over the malar eminences, tending to spare the
nasolabial folds
Erythematous raised patches with adherent keratotic scaling and follicular
plugging; atrophic scarring may occur in older lesions
Skin rash as a result of unusual reaction to sunlight, by patient history or
physician observation
Oral or nasopharyngeal ulceration, usually painless, observed by a physician
Nonerosive arthritis involving two or more peripheral joints, characterized by
tenderness, swelling, or effusion
a. Pleuritis: convincing history of pleuritic pain or rub heard by a physician or
evidence of pleural effusion OR
b. Pericarditis documented by ECG or rub or evidence of pericardial effusion.
a. Persistent proteinuria greater than 0.5 gm/day or >3+ if quantitation not
performed OR
b. Cellular casts (red cell, hemoglobin, granular, tabular, or mixed)
a. Seizures in the absence of offending drugs or known metabolic
derangements, e.g., uremia, ketoacidosis, or electrolyte imbalance OR
b. Psychosis in the absence of offending drugs or known metabolic
derangements, e.g., uremia, ketoacidosis, or electrolyte imbalance
a. Hemolytic anemia with reticulocytosis OR
b. LEUKOPENIA: <4000/pl total on two or more occasions OR
c. Lymphoprnia: <1500/pJ on two or more occasions OR
d. Thrombocytopenia<100,000/pl in the absence of offending drugs
a. Positive LE cell preparation OR
b. Anti-DNA: antibody to native DNA in abnormal titer OR
c. Anti-Sm: presence of antibody to Sm nuclear antigen OR
d. False-positive serologic result for syphilis known to be positive for at least
6 months and confirmed by TPI or FTA-ABS
An abnormal titer of ANA by immunofluorescence or an equivalent assay at
any point in time and in the absence of drugs know to be associated with
drug induced lupus syndrome.
TPI = Treponema pallidum immobilization test; FTA-ABS = fluorescent treponemal antibody absorption test.
'The classification is based on 11 criteria. For the purpose of identifying patients in clinical studies, a
person shall be said to have S:E if any 4 or more of the 11 criteria are present, serially or simultaneously,
during any interval of observation. Source: EM Tan et al. The 1982 revised criteria for the classification of
systemic lupus erythematosus. Arthritis Rheum25:1271; 1982.


3
Causes of death (see tables 1-2 and 1-3) and life
expectancy (table 1-4) of patients with SLE may also be
stratified by socio-economic or environmental parameters (see
fig.1-1) [Fessel 1974; Hochberg 1993; Iseki et al 1994]
Table 1-2. Causes of death in a cohort of systemic
lupus patients.
Cause
No. (%)
Systemic lupus erythrmatosus
49 (34)
Multisystem
16
Nephritis
12
Central nervous system disease
11
Pulmonary
5
Cardiovascular
5
Infection
32 (22)
Cardiovascular disease
23 (16)
Cerebrovascular disease
8(6)
Cancer
8(6)
Iatrogenic
3(2)
Pulmonary embolus
2(1.4)
Gastrointestinal hemorrhage
2(1.4)
Aortic dissection
1 (0.7)
Radiation enteritis
1 (0.7)
Toxic epidermal necrolysis
1 (0.7)
Small bowel obstruction
1 (0.7)
Thrombotic thrombocytopenic purpura
1 (0.7)
Pacemaker malfunction
1 (0.7)
Addisonian crisis
1 (0.7)
Unknown
10(7)


4
Table 1-3. Characteristics of SLE Patients Treated
in Okinawa, Japan, between 1972 and 1991.
No. of Patients
Females
515
(91%)
Males
51
(9%)
Total
566 (100%)
je at diagnosis (yr) (mean SEM)
Females (range)
29.30.6 (5 to 77)
Males (range)
27.01.7 (5 to 63)
Total (range)
29.10.6(5to 77)
Comorbid conditions (%)*
41
(7.2)
Central Nervous System symptoms
108
(19.1)
Aseptic bone necrosis
42
(7.4)
Pericarditis
66
(11.7)
Family history
45
(8.0)
3. of deaths (%)
Females
95
(18.4)
Males
9
(17.6)
Total
104
(18.4)
auses of death (%)
Infection
25
(24.0)
Cerebrovascular disease
16
(15.4)
Uremia
12
(11.5)
Sudden death
6
(5.8)
Cardiac
17
(16.3)
Others
28
(26.9)
jpus nephritis (%)
279
(49.3)
Renal biopsy
174
(30.7)
Nephrotic syndrome
194
(34.3)
Dialysis therapy
78
(13.8)
The occurrence of a comorbid condition at any time in the patients past.


5
Table 1-
4.
Study of
the correlation
between
cause
death, sex,
race and
socio-
economic status (SES) in
patients
with SLE.
Total
SLE Infection
Cardio
Cerebro
Cancer
(n=49)
(n=32)
vascular
vascular
(n = 8)
(n=23)
(n=8)
Female
95
40 (42)
28 (29)
14 (15)
6 (6)
7 (7)
Male
25
9 (36)
4 916)
9 (36)
2 (8)
1 (4)
White
50
22 (44)
7 (14)
15 (300
2 (4)
4 (8)
Black
70
27 (39)
25 (36)
8 (11)
6 (9)
4 (6)
High(SES)
33
15 (45)
5 (15)
7 (21)
2 (6)
4 (12)
Middle(SES)
34
12 (35)
9 (26)
11 (32)
1 (3)
1 (3)
Low(SES)
25
9 (36)
7 (28)
7 (28)
1 (4)
1 94)
Year
1969-1973
43
22 (51)
10 (23)
7 (16)
1 920
3 (7)
1974-1978
52
18 (35)
15 (29)
9 (17)
6 (11)
4 (8)
1979-1983
25
9 (36)
7 (28)
7 (28)
1 (4)
1 (4)
Disease
duration
<2 yrs
35
21 (60)
8 (23)
4 (11)
2 (6)
0 (0)
2-4.9 yrs
28
14 (50)
8 (29)
3 (11)
3 (11)
0 (0)
5-9.9 yrs
31
7 (23)
8 (29)
11 (36)
1 (3)
3 (10)
10-14.9 yrs
14
3 (21)
5 (36)
3 (21)
1 (7)
2 (14)
>15 yrs
12
4 (33)
2 (17)
2 (17)
1 (8)
3 (25)
Of
P
0.16
0.02
0.13
0.54
0.03


6
r400 c
_o
j3
3
'300 o
&
c
o
'200 E
k_
o
Q.
m
O
'100 c
o
CO
>

Figure 1-1. Environmental influences may be related
to the prevalence of SLE. The figure shows the overall
annual incidence (0) and prevalence () of SLE in
Okinawa, Japan between 1972 and 1991. The rates were
expressed as the number of patients per million pop.
of the year of diagnosis.
A number of strategies have been used or proposed to
suppress autoimmune diseases, most notably drugs such as
cyclophosphamide, cyclosporin A, methotrexate and
azathioprine. Steroid compounds, such as prednisone and
methylprednisone, are also employed in many instances. These
drugs have limited long term efficacy against both cell- and
antibody-mediated autoimmunity because of their "global"


7
immunosuppression and their toxic side effects. Prolonged
treatment inhibits the normal protective immune responses to
pathogenic microorganisms, thereby increasing the incidence
of infection. A further drawback is that immune-mediated
elimination of aberrant cells is impaired and there is, thus,
an increased risk of developing malignancies in patients
receiving prolonged immunosuppressive therapy.
The potential to accomplish extracorporeal removal of
certain components from bodily fluids without the above
mentioned risks to the patient has been established using a
variety of bioadsorption devices. A number of patents have
been issued for immunoadsorbent materials which remove
immunoglobulins from plasma [Balint U.S. 4,801,449; Jones
U.S. 5,122,112]. However, there remains a need for improving
efficiency of the existing methods of extracorporeal therapy,
most of which still require plasma separation before the
target molecule can be removed [Schneider 1998]. Therefore,
chemical modifications on the surface of existing devices is
being investigated here and elsewhere. This research is
designed not only to improve the efficiency of the adsorption
process, but also to remove the target molecules from whole
blood.
1.2 Background Review
1.2.1 Clinical Implications of Middle Molecules in
Extracorporeal Therapy.
The normal human kidney is well known for its ability to
excrete water, small-molecular weight end products of
metabolism and to balance electrolytes in the blood. In
addition, it also has a role in the removal of peptides and
low-molecular weight proteins. These "middle molecules" may
be catabolized in the tubular cells after endocytotic


8
absorption from glomerular filtrate and later excreted as
smaller fragments in the urine. Furthermore, middle molecules
(MM's) accumulate in body fluids in the same way as urea and
creatinine when the glomerular filtration rate (GFR) is
decreased [Funck-Brentano et al. 1972; Bergstrom and Wehle
1994]. It is important to note, however, that the
categorization of molecules in this way is somewhat arbitrary
and that MMs in general refer to those solutes that are
poorly dialyzed through conventional cellulosic membranes.
The "large molecules" of dialysis being compounds usually
greater than 7 0kDa and which are not removed to any
significant extent across such membranes. Accordingly, MM
represent compounds having in the broadest sense molecular
weights between 1.5kDa and 55kDa.
There is now sufficient evidence that MMs are elevated
in the serum of patients with chronic renal failure and that
many of them exert toxic effects in vivo [Descamps-Latsca
1994 ]. In the first part of this research project we have
constructed an in vitro model to study the transport of a
test solute, vitamin B12, and two well known pathogenic MM
beta-2 microglobulin (p2-M) and the serum soluble
interleukin-2 receptor (CD25). The result of experiments
undertaken will demonstrate the utility of the model in
evaluating hemodialyzer performance and the study of the
effects of dialysate additives on clearance of middle
molecules. Three test membranes were used in the study; a
low-flux polysulfone hollow fiber, the F8 (Fresenius Corp),
a high-flux cellulose acetate hollow fiber the Altrex 140
(Althin Corp.), and a high-flux polymethylmethacrylate hollow
fiber, the BK2.1P (Toray Ind.).


9
1.2.2 The Effects of Dialysate Additives on the
Transport Properties of Donor Solutes across
Hollow Fiber Membranes.
In order to study the transport properties of MMs in
hollow fiber dialyzers an in vitro model has been constructed
in which equilibrium and sink conditions can be applied to
the three test solutes. Vitamin Bi2 (m.w. 1355) will serve as
a surrogate molecule for dialyzer/circuit performance. Its
high solubility in aqueous solution and its characteristic UV
absorption at 361nm have made it a very useful solute for
clearance studies in hemodialysis [Naitoh 1988].
Beta 2-microglobulin is a 12kDa non-glycosylated protein
present at the surface of nearly all nucleated cells, where
it is non-covalently bound to the heavy chain of the HLA
class I complex. The primary structure and physiochemical
characteristics of (32-M are shown in fig. 1-2.
Figure 1-2. The structure and physical properties
of beta 2-microglobulin. Adapted from Bergstrom and
Wehle (1994).


10
P2-M is markedly retained in ESRD patients, since the
kidney normally accounts for the elimination of about 95.5%
of the normal beta 2-microglobulin generated at 1000-2000
mg/week [Chanard et al. 1993]. Renal retention of P2-M
appears to be the basic requirement for the development of
dialysis-related amyloidosis. p2-M will serve as a mid-range
donor solute in this investigation.
Several cytokine receptors actually consist of two or
more separate polypeptide chains that function as a complex
in the cell membrane. For example, the high affinity IL-2
receptor contains three separate chains (a, p, and y) The
trimeric IL-2RaPy and the IL-2RPy heterodimer both bind IL-2
and mediate signal transduction. See fig. 1-3.
Intermediate lllgli affinity Low affiidty
affinity IL 2R IL2R IL-2R
Subunit
composition: 11.-211(3
IL2lly
Affinity
constant CAj): ltr9Af
IL-2lla lL-2Ra
IL-2R(I
IL2Ry
Itr'Mf l(T8Af
Cells
expressed by: NK cells
Resting T cells
Clow numbers)
Activated CD4*- and
CU8^ cells
Activated D cells
Clow numbers)
Figure 1-3. The interleukin-2 receptor complex. The
receptor is composed of three transmembrane polypeptides
called the a, Pand y subunits. Adapted from Minami (1993).


11
The IL-2Ra serves to increase receptor affinity for the
cytokine but does not contribute at all to signaling. This a-
subunit (CD25 in cluster of differentiation designation) may
be released from a T-cell surface upon activation by a
suitable antigen, mitogen or toxic event (see fig.1-4).
Figure 1-4. Structural models of the "soluble" IL-2R
alpha subunit CD25. The binding regions of the subunit
for IL-2 are shown on the right. Adapted from Miedel
et al. (1988).
Serum soluble CD25 has been shown to retain IL-2
affinity at a level (KD~18nM) sufficient to inhibit IL-2
mediated cell responses. CD25 is elevated in many disease
states, including rheumatoid arthritis, cancer, and AIDS, as
well as in patients with SLE or ongoing maintenance dialysis.
Studies show that serum CD25 levels offer a rapid, and
reliable measure of disease activity [Rubin and Nelson 1990].


12
Therefore, because of its broad clinical significance we have
selected CD25 as the upper-range solute in these mass
transfer experiments.
1.2.3 Molecular Pathogenesis of Lupus Nephritis.
When autoantibody synthesis occurs, it is usually the
result of one of the following: 1) The individual may have an
intrinsic defect in immunity that hinders its ability to
distinguish self from foreign non-self, e.g. T-helper cell
deficiency. 2) Self-antigen may be modified by infection,
injury or drugs. After such modification these self-antigens
are "denatured" or altered in such a way as to be taken as
foreign. 3) If certain self-antigens that are normally
sequestered from the immune system (e.g. cornea) become
exposed to the systemic circulation due to injury, this may
result in stimulation of an immune response and autoantibody
synthesis. A classical example of this mechanism leading to
autoimmunity is found to occur during surgical procedures on
the thyroid gland. When the normally sequestered thyroid
colloid is exposed, thyroglobulin may be released into the
blood. Autoantibodies to plasma thyroglobulin have been found
in patients after thyroid surgery [Theofilopoulos and Dixon
1982].
The development of lupus nephritis depends on the active
deposition of immune complexes onto the glomerular basement
membrane. Immune complexes characteristic of SLE circulate in
a variety of constitutive masses. The largest members of
these circulating immune complexes (ICs) are easily
phagocytized before deposition can occur and are therefore
nonpathogenic. The smallest of these ICs are not well
filtered by kidney glomeruli and thus, are easily excreted in
the urine [Carlson et al. 1988]. However, those ICs of


13
intermediate size (too large for excretion, yet too small
for phagocytosis) have an extented half-life in the
bloodstream and are found predominately in the glomerular
deposits of lupus patients with secondary nephrotic syndrome
[Wener 1986],It has been shown that histones, which are
positively-charged DNA binding proteins, may mediate the
deposition of DNA-anti-DNA immune complexes [Morioka et al.
1994]. Apparently, the histones first affix to the glomerular
basement membrane by electrostatic interactions; then
circulating preformed immune complexes bind to these
subplanted histones to form nephritic lesions between
capillary endothelia [Choi et al. 1995]. Hemodynamic forces
also play an important role in the deposition of circulating
ICs. It has been found that IC deposition occurs
predominantly in those areas of the vasculature under high
shear-stress. This is easy to understand if one considers
that in regions of hemodynamic shear-stress (notably the
kidneys, heart and lung) decreases in the length and luminal
diameter at the capillary bed cause blood flow to become more
turbulent. Blood cells (especially platelets) and endothelial
cells respond to this shear-stress by secreting vasoactive
compounds (i.e. 5-HT and histamine) and changes in morphology
[Kniker and Cochrane 1968; Herbert et al. 1978]. The
pathogenic mechanism involves 1) activation of platelets
followed by release of vasoactive amines and aggregate
formation, 2) retraction of endothelial cells and exposure of
the subendothelial basement membrane, 3) activation of
complement and inflammatory leukocytes followed by, 4) tissue
injury due to the release of proteases and reactive oxygen
species [Mahan et al. 1993; Dixon et al. 1961]; see fig.1-5.


14
I.
Immune complexes
.Proteins
Vessel
lumen
Endothelial
~ cell
Basement
membrane
Basophil
PAF
Vasoactive
amines
increased permeability
Deposition ol immune complexes
Figure 1-5. Mechanisms of Immune Complex Deposition.
Adapted from Kniker and Cochrane (1965).
1.2.4 Anti-(DNA-Histone) Antibodies in Active Lupus
Nephritis.
Histones are the principal structural protein of
eukaryotic chromosomes. They are small polypeptides with 100
to 200 amino acid residues, which do not contain tryptophan
but are rich in lysine and/or arginine. Histones may be
divided into five classes (or types): Hi, H2A, H2B, H3 and
H4. This classification, originally made on the basis of
histone electrophoretic mobility, has been maintained because
histones belonging to each of the classes have different
molecular weights and amino acid sequence. See table 1-5.


15
Table 1-5. Principal Characteristics of Calf Thymus
Histones. Adapted Rubin in Autoantibodies Peter
and Schoenfeld (eds.) Elsevier (1996).
Fraction
Mr'
Number
N-terminal
a-helix
^-pleated
1 mg/mL
Cationic
Molar staining coefficients
of
residue
content
sheet
E
charge
residues
(%)
(%)
275nm
density"
Coomassie
blue
Amido
black
H1
26.500
220
Ac-Ser
55
5
0.079
37.7
1.00
1.00
H2A
14.000
125
Ac-Ser
35
30
0.300
21.4
0.44
0.49
H2B
13.770
129
Pro
40
20
0.510
21.4
0.22
0.59
H3
15.340
135
Ala
39
15
0.300
22.2
0.65
0.78
H4
11.280
102
Ac-Ser
28
31
0.500
23.9
1.67
0.47
The features in the above table distinguish each histone
from those of another class. Two polypeptides of each of the
classes H2A, H2B, H3 and H4 constitute an octamer around
which dsDNA is wound in two turns to form the histone-DNA
complex called a "nucleosome core" (see fig.1-6).
Figure 1-6. Structure of the "Histone octamer-
Nucleosome core." Adapted from Rubin (1996).
A segment of "linker" dsDNA connects one nucleosome to the
next [Sperling and Wachtel 1981; Hnilica 1989].
A recent investigation has confirmed the presence of
anti-(DNA-histone) activity in lupus nephritis [Suenaga and


16
Abdou 1996]. In their study it was found that anti-(DNA-
histone) antibodies specifically recognized DNA-histone
complexes and largely overlapped with antinucleosome
antibodies. While both DNA and histone were required for full
recognition by the antibodies, it is unclear whether DNA
itself is part of the nucleosome specific epitopes or simply
stabilizes the histones and maintains the nucleosomal
structure of the core to form such epitopes. This may be
especially relevant since histones are known to spontaneously
form octamers in physiological pH buffers [Muller 1985;
Burlingame et al. 1985]. Determination of the precise binding
sites, however, must await more sequence information of the
variable regions of both heavy and light chains of fully
characterized antinucleosome antibodies. For a summary of the
prevalence of AHA in SLE, see table 1-6.
Table 1-6. A Summary of the Prevalence of Histone
Autoantibodies in SLE patients. Adapted from Rubin
(1996).
Method
FIAa
ELISA3
ELISA3
IBb
ELISAb
ELISAb
ELISAb
ELISAb
Number of sera
82
39
151
127
32
46
12
40
Histones:
%
%
%
%
%
%
%
%
H1
62
95
52
98
6
59
67
60
H2A
43
54
23
50
34
nd
42
55
H2B
68
79
21
98
63
72
25
55
H3
57
44
47
63
56
59
33
35
H4
54
nd
21
53
22
nd
42
38
ELISA, enzyme-linked immunosorbent assay; FIA. fluorimetrc immonoassay; IB. Immunoblotting; nd, not
determined; a. test of IgG + IgM autoantibodies; b. test of IgG antibodies only.
Anti-(DNA-histone) antibodies are clinically important
because certain subsets of the antibodies are potentially
pathologic for tissue injury such as nephritis.
Antinucleosome activity has been detected in renal eluates


17
from nephritic MRL mice, and a correlation between anti-(H2A-
H2B)-DNA and kidney disease defined by proteinuria has been
described [Burlingame et al. 1994]. In that study it was
found that anti-(DNA-histone) antibodies react with
nucleosomes released from apoptotic mononuclear cells,
forming complexes in patients with circulating nucleosomes.
These complexes may specifically deposit in the renal
glomerular basement membrane by histone-heparan sulphate
mediated mechanisms [Tremaat 1992]. They also may directly
deposit by recognizing DNA-histone/collagen type IV clusters
on the glomerular basement membrane [Morioka et al. 1996];
see fig. 1-7.
blood vessel
blood How
immune
complexes
vessel lumen -f1 -4:
quiescenl
blood vessel under
hemodynamic sheai-stress
blood llow
lammai
immune
complexes -4 -4
deposil
immune complexes deposil onto
exposed basemenl membrane al
ine capillary beds ol Hie glomeruli
activated platelets
5-H1, histamine
Figure 1-7 Proposed Mechanism of Histone-Mediated
Glomeruli Deposition of Immune Complexes.
1.2.5 Surface Modification of PMMA and PSF Dialysis
Membranes.
The preparation of novel affinity adsorption devices has
increased recently due to the great demands of molecular
medicine and biotechnology for efficient yet selective


18
isolation of proteins from complex mixtures. A number of
reports describing various types of microporous membranes
modified with selective-binding ligands continue to appear in
scientific journals.
Conformational factors still play a major role in
binding of the target molecule to the adsorbent. For example,
the orientation of antibody molecules on a surface is often
responsible for differences between their theoretical binding
activity (2:1) and the experimentally determined binding
ratio. This occurs because the antibody conformations are
randomized during the coupling procedure; as a result, the
paratope of some of them becomes inaccessible to the antigen
upon immobilization. This random orientation of immobilized
ligands may be kept to a minimum by the appropriate use of
bifunctional spacers such as gluteraldehyde or polyethylene
glycol [Malmsten et al. 1996]. It has also been shown that
the optimum amount of ligand to be immobilized on an
adsorbent is primarly determined by the molecular size of the
target molecule [Vallar et al. 1995]. For the capture of
large molecules like immunoglobulins, the immobilized ligands
need to have a maximum degree of freedom; this again may be
accomplished by tethering the ligand (antigen) at the end of
a long flexible polymer such as PEG. The ligand is now in a
quasi-fluid state and able to avoid more of the steric
binding impedances between it and the target molecule.
The grafting of (PEG) to solid surfaces is a very useful
technique for obtaining low protein adsorption and low cell
adhesion characteristics. For instance, PEG coating is
reported to give a marked suppression of plasma protein
adsorption and platelet adhesion leading to reduced risk of
thrombus formation. The inert characteristics of PEG surfaces


19
are due to the solution properties of the polymer, its
conformation in aqueous environments and to its electrical
neutrality, its structural similarity with H20 and strong
hydrogen bonding ether linkages provide PEG with very high
water solubility. PEG'S ability to prevent proteins from
approaching the bonded surface is due to both elastic and
osmotic steric stabilization effects. The elastic stabilizing
effect occurs when a protein approaches the PEG-modified
surface and a repulsive force develops. This is due to a loss
in conformational entropy as the polymer chains experience a
reduction in available volume. The osmotic interactions arise
from the increase polymer concentration on compression toward
the surface. When the local concentration of PEG segments
increase osmotic repulsive forces develop. The intensity of
this force depends on compression at high grafting density
but more on interpenetration at lower graft densities.
For a number of applications attachment of proteins or
other biomolecules to PEG-grafted surfaces is of interest.
Implants and artificial organs are rendered more
biocompatible. Strong interactions between the underlying
surface and the immobilized molecule can be avoided as can
nonspecific adsorption of molecules from solution. Provided
that the immobilization procedure used gives minimal
distortion of the biological properties and that the layer of
immobilized molecules is not so dense that crowding phenomena
appear, this method of surface modification can be expected
to lead to stable immunoadsorbents of high affinity. For
details on the synthesis of Star-PEG see appendix V.


20
1.2.6 Surface-Modified Polymethylmethacrylate (PMMA).
PMMA is an amorphous polymer of moderate Tg (105 C),
that has a high transparency consistent with its use in the
formulation of contact lenses. It is highly resistance to
strong acids and environmental deterioration. The solubility
parameter for PMMA is between 11.1-11.3 (cal cm'3)1/2; it is
very soluble in dichloromethane or isopropyl alcohol.
The synthesis of PMMA monomers is based on cyano-hydrin
formation with acetone to prepare the desired methacrylate
ester. See fig. 1-8.
O + HCN
H3C
H3^ /
ch3
HO
O^N
h2so4^
ROH
Figure 1-8. PMMA monomer synthesis.
These esters may then be polymerized by free radical
mechanisms initiated by heat or light.


21
Techniques used to prepare microporous PMMA membranes
are summarized in the table 1-7 below.
Table 1-7. Methods Used to Prepare Microporous Membrane
Materials. Adapted from Fried Polymer Science and
Technology. Prentice Hall p.450 (1995).
Method
Description
Phase Inversion
Phase separation of a ternary mixture of
polymer, solvent, and nonsolvent
Sintering
Melting of a semicrystalline polymer powder
Track etching
Irradiation of polymer films resulting in
the production of fission fragments
followed by caustic etching
Stretching
Combined stretching and annealing of
extruded semicrystalline film
Leaching
Extraction of solid pore formers
Thermally induced
phase separation
Cooling a mixture of a polymer with a
latent solvent to a point of thermal
separation of the mixture followed by
extraction of the latent-solvent phase
The process known as nucleation track-etching is
predominately used for the production of hollow fiber
membranes. Typical pore diameters obtained by this method are
in the range of 0.02 to 20.0|im. Hollow fiber modules for the
separation of complex mixtures typically have outer diameters
(o.d. ) of from 80-200 (im: with wall thicknesses of 20 urn or
greater. Seals for the bundles are usually made of epoxy;
with polycarbonate being the most common housing material.
PMMA is a highly biocompatible membrane material; its
use in hemodialysis began shortly after the discovery if its
ability to bind beta-2 microglobulin (Klinke et al. 1989;


22
Bonomini et al. 1996). For the immobilization of proteins,
PMMA can be activated with hydrazine-monohydrate (NH2NH2) The
resulting aminated surface can then be hydrophilized by
crosslinkage with a tresylated Star-PEG polymer in such a way
that the terminal reactive end-groups are stable enough for
subsequent binding (immobilization) of proteins or other
amines. See fig. 1-9.
slop 1 membrane activation
NH2NH2-H2O
-CONHNH2
COOCH3

50% MeOH-H20
OSO2CH2CF3
step 2 membrane hydrophillzation
SO2CH2CF3
SO2CH2CF3
CONHNH2
tresyl-Star PEG
NH-CH2CH2-O-
50% Me0H-H20
CF3CH2O2SO
OSO2CH2CF3
step 3 antigen immobilization
NH-CH2CH
cf3ch?o?so
CF3CH2C
OS02CH2CF3
S02CH2CF3
histone ,
DOC
PBS
histone
SO2CH2CF3
Figure 1-9. Surface-modified PMMA.


23
Figure 1-10. Formation of Potentially Toxic Hydrazine
Derivatives. Shown above is a) Formation of Hydrazine
Derivatives with Nucleic Acids. Adapted from P. David
Josephy Molecular Toxicology Oxford University
Press p.33 (1997) .
Precautions are warranted in the use of hydrazine or its
derivatives. While hydrazine itself reacts with pyrimidine
bases of nucleic acids to form aminopyrazoles (see fig. l-10a
previous page), most of the free hydrazine reacts with other
molecules extracellularly before reaching the cell interior
or nucleus. The many of these hydrazine derivatives are
single H-substitutions by an aliphatic carbon (see fig.l-
10b).


24
NH2NH2 is a relatively weak nucleophile, but it is still very
reactive with Io and 2 alkyl halides. Acylhydrazines can also
be formed by the reaction of hydrazine and most organic
esters. See fig.l-10b.
R.
R,
"single-H transfer" ACYLHYDRAZINE
ACYLHYDRAZINE N,N' DIACYLHYDRAZINE
Figure 1-10. Formation of Potentially Toxic Hydrazine
Derivatives. Shown above is b) Formation of radical
stabilizing diacylhydrazines.


25
Disubstituted hydrazine derivatives may undergo certain
redox reactions that result in the formation of nitrogen free
radicals. See fig. l-10b. It is these "stabilized" free
radicals that are most likely associated with hydrazine
toxicity [Amdur, Doull, and Klaassen 1991; Purdy 1996].
1.2.7 Surface-Modified Polysulfone (PSF).
The polysulfones are a class of thermoplastics with high
thermal, oxidative and hydrolytic stability. They are usually
synthesized by condensation polymerization involving
substitution of the alkaline salt of bisphenoates with
activated aromatic dihalides. The synthesis of bisphenol-A
PSF is shown in fig. 1-11.
rt
sodium bisphcnolate
dichlorodipbenylsulfone
Figure 1-11. Condensation polymerization for PSF.
PSF has very high resistance to aqueous mineral acids,
salt solutions, oils, and greases. Its high biocompatibility
and ability to be sterilized by several different techniques
make them exceptionally suitable for medical applications.
PSF membranes have both high permeability and permselectivity
and are also fabricated into hollow fibers by nucleation
track-etching.


26
PSF is currently the most successful porous membrane
material for medical applications and the separation of
gases. However, limited use has been made of this polymer as
a substrate for surface-modifications or for the
immobilization of proteins. A few reports of PSF activation
with chlorosulfonic acid followed by crosslinking with a
diamine or diacyl compound for the subsequent binding of
protein have been reported [Klein et al. 1994; Pozniak et al.
1995, Fernando-Salznero et al. 1991], but some of these
methods are carried out in an organic medium, such that the
activation occurs in the slurry rather than on the surface of
an intact membrane or hollow fiber. We, therefore, used an
aqueous phase nitration method to activate intact PSF hollow
fibers which can then be reduced to the corresponding amine
in the presence of a low reduction potential trivalent metal
(i.e. CrCl3). The aminated PSF fashioned in this way can then
be hydrophilized by a tresylated Star-PEG for the subsequent
binding of proteins in a manner similar to the PMMA membranes
described earlier. See fig. 1-12.
1.2.8 Immunoreactivity of Immobilized Histones
Characterized using the BIACORE 2000.
The biospecific interaction between antihistone
antibodies and histones immobilized via PEG Star-polymers was
monitored using the BIACORE 2000. A multistep binding
experiment was conducted where tresylated Star-PEG polymers
were attached to the BIACORE sensorchip in a manner similar
to that used in the modification of PMMA membranes. The
instrument measures in real-time the binding kinetics at the
flow cell surface and then calculates a corresponding
sensorgram which quantifies the recognition process.


27
step 1 membrane activation
hno3\h2oso3
~(OC6H4SC>2C6H4)n
5.0% AcO /H20
CrCI2\HCI
(0C6H4S02C6N02H3)r>
step 2 membrane hydrophilization
tresyl-Star PEG
-(0C6H4S02C6NH2H3)n
50% MeOH-H20
CF3CH2O2S
CF3CH2O2SO
-PSFNH-CH2CH2O-
OSO2CH2CF3
J /OS02CH2CF3
OSO2CH2CF3
step 3 histone immobilization
CF3CH2O2SO OSO2CH2CF3
CF3CH2O2SO \ J >^0S02CH2CF3
-PSFNFFCH2CH2O-.
histone
OSO2CH2CF3
PBS
histon
histone'
histone
histone
histone
histone
Figure 1-12. Surface-modified PSF


28
The BIACORE 2000 operates on the principle of surface
plasmon resonance (SPR) and is designed to detect changes in
the refractive index of the layer of solution in contact with
the sensorchip. Since the refractive index of the medium is
affected by the surface concentration of solutes, monitoring
the SPR angle provides a real-time measurement of changes in
surface concentration and/or coupling.
Units used to express the SPR signal are called response
units (RU). One thousand RU are equivalent to a change of
about 1 ng/mm2 in surface protein concentration or
approximately 6 mg/ml in bulk protein concentration for most
proteins.
1.2.9 Surface Analysis of Surface-Modified Hollow
Fibers.
In order to achieve a more complete understanding of the
molecular-interactions between foreign materials and
biological systems, surface properties and surface structure
must be properly elucidated. The surface structure of a
material is dynamic and often bioreactive, especially in
cases in which components of the surface reconstruct
themselves in response to a local environment.
For these reasons, highly sensitive analytical methods
are required to provide a means to these data. See table 1-8
and fig. 1-13.


29
Table 1-8. Common Surface Analytical Methods.
Adapted from Ratner (ed.) Surface Characterization
of Biomaterials. Elsevier, Amsterdam (1988).
Method
Depth Analyzed
Spatial
resolution
Analytical sensitivity
Contact angles
3-20
1 mm2
Low or high depending on
the chemistry
Scanning force
microscopy (SEM)
5
1
Single atoms
Scanning electron
microscopy (SEM)
5
40 typically
High; not quantitative
Electron spectrocsopy
for chemical analysis
(ESCA)
10-250
8-150 pm
0.1 atomic %
Secondary ion mass
spectrometry (SIMS)
10 to 1 pm6
500
Very high
Attenuated total
reflection infared
(ATR-IR) spectroscopy
1-5 pm
'
10 pm
1 mol %
"The size of a small drop is 1 mm. However, contact angles actually probe the inter-
facial line at the edge of the drop. The spatial resolution of this zone might be approx
imately 0.1 pm
Static SIMS -10 ; dynamic SIMS to 1 pm


30
useful for examining the solid-aqueous interface
Figure 1-13. Flowchart for the Surface
Characterization of Biomaterials. Adapted from Ratner
(ed.) Surface Characterization of Biomaterials.
Elsevier, Amsterdam (1988).


31
The simplest picture of a sample surface is provided by
the optical microscope. The optical microscopy may be
extended by the use of specialized methods which include
polarized microscopy, fluorescence microscopy or confocal
microscopy. A more detailed picture of the surface is
provided by the scanning electron microscope (SEM) which
produces surface images of greater resolution and depth of
field than the above mentioned techniques. However, the major
disadvantage of the SEM is that nonconductive materials have
to be "sputter-coated" before images can be made.
Advanced imaging of surfaces at the molecular level is
now obtained with the so called scanning probe microscopes
(SPM). The atomic force microscope (AFM) is such a scanning
probe that does not require the subject surface to be
electrically conductive, since it merely measures the
interaction between a microfabricated tip mounted on a small
cantilever spring and the atoms on the surface of the sample.
Scanning the probe tip over soft (compliant) surfaces may
induce deformations in the image. Therefore, to avoid damage
to such samples, the interaction force may be kept below 8-
10N by operating the AFM in "non-contact mode" [Porter, TL.,
Sykes, AG and Caple, G 1994]; see appendix VI.
1.2.10 Extracorporeal Immunoadsorption Therapies in
SLE.
The selective extraction of constituents from plasma is
based on the principle of chromatographic affinity. A
substance (ligand or sorbent) with specific affinity may be
fixed to an insoluble, inert or "passive" matrix with the aim
of selecting and binding its complementary substance. The
ligand (the active part of the column) may vary broadly; it
can be a chemical compound like heparin, carbon, dextran


32
sulfate, or a protein i.e. an enzyme, antigen, antibody and
so on. Selective or semi-selective adsorption may occur
through physiochemical interactions (i.e. hydrophobic or
ionic bonding) or by biological interactions (e.g. the
interactions between enzyme-substrate, Ag-Ab molecular
recognition, complement fixation, or immunoglobulin Fc-lectin
interactions).
The binding capacity of affinity adsorbents is perhaps
the most important parameter in immunoadsorption therapy. The
capacity of a selective adsorbent is determined principally
by two interdependent sets of conditions: (1) the correct
choice of the matrix, spacer molecule and ligand, in order to
optimize the enzyme-ligand interaction and (2) the way in
which the capacity is affected by such dynamic factors as
flow rate, equilibration time and adsorption technique.
Assuming the design of the adsorbents has been
optimized, the operational adsorption capacity of an affinity
gel is best determined by frontal analysis. Here, a given
concentration of the complementary protein (C0) is applied to
the adsorbent continuously and then its emergence monitored.
As the adsorbent material becomes saturated with the
adsorbate, the solution breaks through at the same
concentration it had on entering the column (see fig. 1-14).
The volume of eluant that appears up to the "step", where the
concentration of the complemnetary protein increases rapidly
to Co over a small volume, is called the retention volume
(V.). It comprises the interstitial volume (VD) and the volume
solution from which the adsorbate was removed (V), i.e.
Ve
Vo + V


33
C0
Cone. in
eluate (CQ)
Vol.(mL)
Figure 1-14. Determination of the capacity of an
affinity adsorbent by frontal analysis. Adopted from
Lowe and Dean. Affinity Chromatography. John Wiley
and Sons Ltd., London (1974).
If m is the total weight of the affinity adsorbent in
grams, then the capacity of the gel, i.e. the amount of
adsorbate specifically adsorded per gram, Q, is
Q = (V/m)C0 = (V.- V/m) C0
and the total amount adsorded by the bed is q = Q xm, i.e.
VCo.
The capacity, consistent with the emergence point of the
monitored species, is dependent on the rate of application of


34
the original sample. At relatively high sample flow rates,
affinity equilibrium is not attained and premature emergence
of adsorbate may be observed. This will lead to a under
estimation of the operational capacity of the adsorbent. The
effective capacity of an adsorbent, however, may be deduced
by incubating a known weight of adsorbent (m) with a given
volume of concentration CQ and subsequently, after equilibrium
has been established, measuring the new lower concentration
C. The capacity, Q, is then calculated from Q. = (C-C)/m.
Currently, few data is available to estimate the
capacity theoretically from a known immobilized ligand
concentration and other parametrers of the system. It appears
that the effective capacity of an adsorbent is considerably
lower, in fact often <1%, of the theoretical capacity based
on the ligand concentration [Lowe, CR et al., 1974; Harvey,
MJ 1974]. Presumably, the effective capacity of a specific
adsorbent is determined by the concentration of immobilized
ligand that is freely available for interaction with the
complementary protein. We have used a similar incubation
experiment to determine the "effective adsorption capacity"
of surface-modified membranes (PSF and PMMA) as will be
described in section 2.31. The results will help us determine
if therapeutically significant amounts of autoantibodies can
be removed by surface-modified commercial dialyzers.
SLE is fundamentally characterized by the synthesis of
numerous antibodies; at least 10 antigen/antibody systems
have been detected and are present in at least 10% of all
patients studied see table 1-9.


35
Table 1-9. Autoantibodies in SLE. Adapted from
Tan et al. (1982) .
Target
Clinical Associations
dsDNA
High diagnostic specificity
Correlation with disease activity (especially renal
disease)
ssDNA
Low diagnostic specificity
Histones (H1, H2A, H2B, H3, H4)
SLE and drug-induced lupus
Sm (SnRNP core proteins B, B1,
D. E)
High diagnostic specificity
No correlation with disease activity
U1-RNP (snRNP-specific
proteins A, C, 70 kd)
Mixed connective tissue disease or overlap
syndrome (when not accompanied by anti-Sm
antibodies)
Ro/SS-A (60-kd and 52-kd
proteins)
Neonatal lupus (with anti-SS-B/La)
Photosensitivity
Subacute cutaneous lupus
La/SS-B (48-kd protein)
Neonatal lupus (with anti-SS-A/Ro)
Associated Sjogren's syndrome
Ku
Diagnostic specificity for SLE and related overlap
syndromes
Proliferating cell nuclear
antigen (PCNAO/cyclin
High diagnostic specificity
Ribosomal P
High diagnostic specificity
Cytoplasmic staining
Psychiatric disease
Phospholipids
Inhibition of In vitro coagulation tests (lupus
anticoagulant")
Thrombosis
Recurrent abortions/fetal wasting
Neurologic disease (focal presentations)
Thrombocytopenia
Cell surface antigens
Red blood cells
Platelets
Neuronal cells
Hemolytic anemia
Thrombocytopenia
Neurologic disease (diffuse presentations)
Tissue and organ damage mostly derives from the
formation and deposition of immune complexes (ICs),
especially in the vascular endothelium. The logic behind the
use of plasma exchange or adsorption as a support to drug


36
therapy is obvious; removal of the autoantibodies before they
bind with antigen and are deposited avoids tissue damage.
This apparently logical consideration is not aways evaluated
correctly and is often not withstanding in even the most
painstaking therapies. Many arguments have been presented
both for and against plasmapheresis in SLE since its first
application in 1976 by the Lockwood and Jones Groups. In
their groundbreaking study [Jones et al 1976] they showed
that therapeutic levels of anti-DNA antibodies could be
removed by plasmapheresis. Even though no adsorbent mechanism
was technically described the claims that immune complexes
were removed were based on reductions in serum anti-DNA
antibodies determined by a Clq precipitation assay. In
addition, an apparent reduction in complement fixation was
measured by RID of C3 and C4 levels in each of the 8 patients
of the study.
Early in the practice of therapeutic apheresis it was
found that the equilibrium between the cell product
(antibody) and the producing cell itself needs to be taken
into account. The term "rebound" is often used to express the
resynthesis of the antibody after its removal by
plasmapheresis. This resynthesis sometimes exceeds basal
levels and is cause for clinical discretion when apheresis
may be applied. The effects of extracorporeal removal of IgG
from rabbits was studied [Carlton et al 1983 ] and the
findings infer that 1) selective removal of identified
pathologic factors is more efficient and potentially more
cost effective than non-specific plasma exchange and 2)
selective removal from whole blood has no required
replacement fluid or associated dangers and inadequacies.
Apart from the limitations imposed by non-selective plasma


37
exchange therapy it is also hindered by the lack of accurate
kinetic models which are required to enable an objective
approach to the treatment. In their model they used a hollow
fiber plasma separator in-line with a protein A adsorbent
column. What they found was that the system was effective in
lowering circulating IgGs in rabbitts by 60% in 30-40 mins.
The adsorption was also adequately modeled using first-order
kinetics, which indicates that intrabody transfer and
endogenous regeneration of the IgG were negligible during the
treatment period. This may not be true in non-healthy rabbits
or in patients with SLE. IgG transfer rate from the EVS to
the IVS is normally three times the rate in the opposite
direction. This is because the EVS to IVS transfer occurs
predominately by convection transport through the lymphatic
system. Following removal of substantial IgG from the IVS,
mobilization of the EV IgG occurs, resulting in a rebound
increase in plasma levels that usually reaches a peak within
24 hrs before a substantial decline.
The rise in plasma IgG level after plasmapheresis may be
as much as 50% due to IgG resynthesis, thus immunosuppressant
therapy is warranted in most cases [Higgins 1995]. In
addition, IgG catabolic rates may fall by 30% as a result of
IgG extraction, thus even in the presence of suppressant
therapy, increases of up to 50% post-apheresis may occur in
an individual.
Even with these forseen difficulties in the early
1980's, investigators continued to design plasma adsorption
methods for the treatment of SLE and other immune complex
(IC) diseases. Most of the earlier immunoadsorbent devices
suffered from two major limitations. Firstly, the plasma
exchange systems were technically not suitable for routine


38
clinical application due to their mechanical complexity
and/or the poor histocompatibility of the sorbent-materials.
This would often lead to unwanted clotting of the blood or
leakage. Secondly, the adsorption capacity of the earlier
designs were either insufficient or fixed. Therefore,
increased surface areas on highly biocompatible, yet specific
adsorbents were being proposed. This objective remains the
most important goal behind current blood purification
research. One group (El-Habib et al. 1984) has successfully
fabricated flat-sheets of chemically-modified collagen onto
which proteins or DNA can be attached under physiologic
conditions. These antigen-coated sheets were then set up in
hemodialyzer frames (RPS Hospal) and sterilized by gamma-
irradiation at 2-5 Mrad. The devices were used to treat 5
SLE-patients after plasma separation which resulted in a 3.7
g/L reduction in IgG after 180 mins, of treatment. Again
differences in the amount of anti-DNA removed were estimated
when different assay methods were used. Bratt and Ohlson
(1988) have developed an immune complex adsorbent material
based on covalently immobilized anti-Clg on Sepharose 4B.
Onto the anti-Clq column, Clq was non-covalently attached and
plasma perfused through such a device was cleared of 1.0 mg
IgG/ml gel. It was estimated that in order to remove all the
Clq-reactive material in the blood of one patient a column
containing only 100-180 ml of the Clq-anti-Clq gel would be
needed. It has been demonstrated in yet another study that
Clq immobilized onto Sepharose 4B typically extracts lmg of
immune complex per ml of the treated gel [Hiepe et al. 1990].
However, this impressive adsorption capacity is not
independent of immune complex concentration and the
efficiency of the separation rapidly decreases with serum


39
content. Furthermore, platelets expressing the Clq receptor
may be activated by the immobilized Clq or soluble Clq
complexes [Casali 1979]. Anti-DNA antibodies have been shown
to adsorb to other materials as well, these include
polyanionic dextran sulfate using both single [Aotsuka et al.
1990] and tandem columns [Suzuki, K. et al 1991]; protein A
[Bygren 1985; Palmer et al. 1991], tryptophan or
phenylalanine adsorbent materials [Yamazaki, Z et al 1989],
DNA on carbonized resin beads [Gao et al. 1995] and DNA
cellulose [Susuki et al. 1994].
A report of a surface-modified hollow fiber for the
removal of human igG from human plasma or serum has been
recently published [Bueno et al. 1995]. The authors studied a
pseudobiospecific affinity ligand, L-histidine, immobilized
through an ether linkage onto a PEVA hollow fiber cartridge.
The membranes effectively adsorbed igGi, IgG2, and IgG3 if MOPS
buffer is used, but were more selective toward IgGi and IgG3
in Tris-HCl buffer. The ligand also showed a higher capacity
than protein A-membranes and may offer a low cost alternative
to protein A based devices. While a K of (10' 5) was reported
for the igG-histidine complex on PEVA a lower KD of(5.0*10"7)
has been reported for protein A immobilized onto a PSF
modified hollow fiber [Klein et al. 1994].
1.3 Hypothesis
Therapeutically modified hollow fiber dialyzers can be
used to remove undesirable antibodies from plasma in the
treatment of various disease states. My hypothesis is that
the extraction of immune complexes and other blood borne
molecules by immobilized ligands against them, can be
demonstrated in an in vitro model using therapeutically


40
modified hollow-fiber dialyzers. The findings of these
studies may be used to optimize extracorporeal therapy of
patients with a variety of immune or metabolic diseases.
1.4 Specific Aims and Rationale
The purpose of this research is to demonstrate the
potential use of therapeutically-modified hollow fiber
dialyzers for the ex vivo extraction of autoantibodies. We
have selected the extraction of antihistone antibodies of the
type found in SLE as one of many possible experimental
models. Our approach will be to immobilize whole histones on
the surface of commercially available dialyzers in such a way
that solutions containing antihistone antibodies can be
cleared of substantial amounts of it by perfusion of the
solution through the device within 4hr hours. To demonstrate
the efficacy of these modified dialyzers, we defined three
specific aims which must be met to support our hypothesis.
In specific aim #1 an in vitro model of an
extracorporeal circuit was built to evaluate the ability of
dialysate additives or dialyzer surface ligands to
selectively remove donor solutes from a series of complex
mixtures. The model used is a simple two compartment model
constructed from two peristaltic pumps, hemodialysis blood
line tubing and dialyzer housing frames. It will allow us to
study transport properties of a given donor solute onto or
across the hollow fiber surface under conditions which mimic
those during a patient extracorporeal therapy session.
Specific aim #2 of the research will be to develop
coupling reactions that immobilize ligands (histones) onto
the hollow fiber surface, to characterize these surfaces, and
to determine that the immobilized ligand is still recognized


41
by the antibody. The characterization of the treated surfaces
will be carried out by scanning electron microscopy of both
chemically modified and unmodified surfaces. The atomic force
microscope (AFM) has the ability to image these surfaces
without application of a conductive metal-coating to the
sample; AFM images were obtained as well. Binding of AHA to
the modified surfaces will be tested in a series of
adsorption capacity experiments. From these studies the
optimal activation site density and AHA extraction per area
of treated membrane will be determined.
In specific aim #3 AHA immunoadsorption was tested.
Therapeutically-modified dialyzers (PMMA and PSF) were placed
in the in vitro model and perfusion with saline or plasma
containing antihistone antibodies. The extracorporeal circuit
was sampled for 4hrs, then assayed for AHA concentration in
the donor compartment. The rate of and total AHA clearance
from the donor reservoir was then calculated and averaged for
each dialyzer in the study.


CHAPTER 2
MATERIALS AND METHODS
2.1 Specific Aim #1
2.1.1 Physical characteristics and specifications of
the model
An extracorporeal circuit was built from conventional
dialysis blood line tubing and two peristaltic (Masterflex
Barnant Co. Barrington, Ill.) pumps. Two CF23 (Baxter Corp.)
dialyzers were fitted so that the extracapillary space now
serves as a 230ml reservoir for each of the donor and
receiver compartments in the mass transfer experiments. Two
F8 dialyzers were opened and the hollow fibers removed so
that the housing capsule now serves as a 475mL reservoir for
each of the donor and receiver compartments in the
immunoadsorption experiments. The physical characteristics
and specifications of the model are shown in fig. 2-1. Hollow
fiber devices have been well developed and produced
commercially due to their application in hemodialysis for the
past 40 years. Most hollow fiber dialyzers currently being
used have structure and composition similar to that shown in
fig. 2-2.
42


43
DONOR RESERVOIR
RECEIVER
RESERVOIR
bypass
oulle! sampling
port
1
inlet sampling
a
'
pon
-
1r
inlet sampling
a
C port
outlet sampling
port
bypass
**4
BLOOD PUMP
DIALYSATE PUMP
VOLUME OF EACH TUBING SEGMENT
DONORCIRCUIT
RECEIVER CIRCUIT
LENGTH(IN) DIAMETER (IN)
VOL (IN3)
LENGTH (IN)
DIAMETER(IN)
VOL (IN3)
ab=33.5 0.125
0 41
aV=21 0
0.25
1 00
bc= 74.5 0.25
3.65
b c'=27 0
0 125
0.33
de= 18 5 0.25
0.91
d'e =5 l 0
0.25
2 49
eb= 18 0 0.25
0.88
e b'= 16 0
0 25
0 78
el= 18 0 0 125
0 22
e r= 16 0
0 25
0 78
donor reservoir
475ml
receiver reservoir
475ml
DONOR BLOOD LINES
= 99 6 ml
donor bypass
= 14 4 ml
RECEIVER BLOOD LINES
= 88.2 ml
receiver bypass
= 12.8 ml
187 8 ml
with 60ml dialyzer
60 ml
two resevoirs
950 ml
total circuit volume
1197 2 ml
MATERIAL COMPOSITION
dialyzer housing
polycarbonate
blood line tubing
polyvinyl
Figure 2-1. In Vitro Model Specifications


44
Blood Out
Figure 2-2. Structure of commercial hollow fiber
dialyzers. Adapted from Colton and Lowrie (1981).


45
Three test membranes were used in the study; a low-flux
polysulfone hollow fiber, the F8 (Fresenius Corp), a high-
flux cellulose acetate hollow fiber the Altrex 140 (Althin
Corp.)/ and a high-flux polymethylmethacrylate hollow fiber,
the BK2.1P (Toray Co.). Dialyzer specifications provided by
the manufacturer are listed in table 2-1.
Table 2-1. Dialyzer Specifications.
F8
Altrex
BK2.1P/B2-1.0H
Surface Area m2
1.8
1.4
2.1/1.0
Material
PSF
CDA
PMMA/PMMA
Ku£ (mL/hr)
4.17
17.0
11.3
Blood Vol.(mL)
120
76
126
Clcr( mL/min)
175
153
181
ClUr (mL/min)
192
182
195
C1bi2 (mL/min)
76
103
127
Flow rates
and flow volumes typically
used in the :
vitro model are
listed in
table 2-2.
Table 2-2.
Circuit Specifications.
Flow rates
Qb =
65mL/min
Qd = 55mL/min
Volumes
Vdon
230mL
Vrec = 200mL
The reservoirs were connected by plastic (polyvinyl)
tubing with Hanson type connectors at one end and Luer type
connectors at the other. For tubing selection and
compatibility tables see appendix II. The H-shaped circuit
allows the intracapillary space to be perfused with a donor
compartment fluid (e.g. blood) while the extracapillary space
is perfused by a receiver compartment fluid (e.g. dialysate).


46
The center section of tubing in each compartment serves
as a circuit bypass so that the reservoir fluids may
equilibrate prior to exposure to the test hollow fiber
dialyzer, or for dialyzer exchange.
2.1.2 Mass Transport Experiments.
Materials: Vitamin Bn was obtained from (Sigma Chemical Co.)
anti-Vitamin B12 monoclonal antibodies (Sigma-Aldrich Co.).
The (32-microglobulin and anti-P2-microglobulin polyclonal
antibodies were also obtained from Sigma Chemical Co. SIL2R
(CD25) and anti-SIL2R monoclonal antibodies were obtained
from R&D Systems
Methods: The clearance of each of three solutes was tested
in three different hollow fiber dialyzers in the presence and
absence of dialysate additives (antibodies) which form stable
complexes with each of them. At the beginning of each
experiment the circuit was purged with 0.1% BSA in PBS and
the donor compartment was put into bypass.
A known amount of donor solute was then injected into
the donor circuit and allowed to mix (in bypass) for 5 mins.,
after which time the circuit was taken out of bypass and
opened to one of the test dialyzers. Donor and receiver
compartment samples were then taken at 5, 10, 30, 60, 120
minute intervals from the time the donor circuit was opened.
After the 120 min. sample the donor circuit was again put
into bypass and antibodies against the donor solute were
injected into the receiver compartment. After 5 minutes of
mixing the donor circuit was again taken out of bypass and
donor compartment samples were taken at 5, 10, 30, 60, 120,
240 min. after the donor circuit was reopened.
The vitamin B12 samples were analyzed via
spectrophotometric assay at 361nm. The (32-microglobulin


47
samples were analyzed via ELISA (Elias Co); the CD25 samples
were also analyzed via ELISA (R&D Systems).
For a summary of the experimental antigen-antibody
concentrations used see table 2-3 below.
Table 2-3. Antigen/Antibody Concentrations in
the Mass Transfer Experiments.
Vitamin B (32-M CD25
antibody 400pg 1060¡j.g 333(ig
antigen 504mg 16.7[ig 1.67|ig
2.1.3 UV Spectrophotometric Analysis of Vit.B12.
The transport of Vit.Bi2 across hollow fiber dialyzers
was studied in the presence and absence of anti-Vit.B12
antibodies as dialysate additives. The study was performed in
an in vitro model (see fig.2-1), both donor and receiver
compartment Vit.B12 concentrations were monitored
spectrophotometrically at 361nm.
Method:
1.Prepare lOOmL stock solution of vitamin B12 in DI H20
1.000g/L
2.Prepare nine standards by serial dilution.
5.0 mg/lOOmL 2.8 mg/lOOmL 2.5 mg/lOOmL
1.00 mg/lOOmL 0.5 mg/lOOmL 0.1 mg/lOOmL
0.05 mg/lOOmL 0.01 mg/lOOmL 0.005 mg/lOOmL
3. Obtain an absorbance spectrum for each of the standards
and plot the calibration curve (see fig. 2-3). For the least
squares regression method used see appendix III.


48
Standard Calibration Curve
Vitamin Absorbance BI2 at 361nm
Effluent Concentration Vitamin B12 mg/lOOmL
Figure 2-3 Calibration curve and linear regression
equation for the spectrophotometric determination of
Vitamin Bi2.
2.1.4 ELISA of CD2 5, f}2-M.
The enzyme-linked immunosorbent assays are solid phase
biospecific binding-assays which have been taking on an
increasingly more prominent role in laboratory medicine and
biotechnology. In comparison to most compound identification
techniques, ELISA's now provide greater objectivity and
potential for automation. Since a single unit of enzyme label
can amplify a reaction product several fold, many ELISA's can
be optimized for the detection at the picomole or attomole
level [Tijssen 1985., Gosling 1990].
Materials: Quantikine human IL-2sRa (R&D Systems, Inc.,
Minneapolis USA); beta 2-Microglobulin Immunoassay (Elias
Co. Madison, Wisconsin USA). For the in-house assays we used
the Reacti-Bind a maleic anhydride activated polystyrene
plate (Pierce Chemical Company, Rockford, Illinois USA) and
performed the selected assays according to the following
methods.


49
Method: For the commercial kits the protocol followed was
included in the package. For the "in-house" assays the
following protocol was used.
1. Add lOOmL of the ligand (antigen or antibody) solution to
each well. For optimal results the ligand immobilization
should be made from solutions containing 10-30|ig/mL in PBS or
Hepes buffer. Incubate at room temperature for l-8hrs or
overnight if refrigerated.
2. Rinse three times with 200(J,L blocking agent to prevent
non-specific binding. Several different blocking agents can
be used, our best results were obtained with 5% BSA-PBS and
1% gelatin-PBS. Thirty minute incubations of the blocking
agent improve selectivity of the assay. These blocked
microtiter plates may now be stored a 4 C until samples are
to be assayed.
3. Add lOOpL of the analyte solution or standards (serum,
plasma or supernatant) to each well. Incubate the plates for
lhr at room temperature. Rinse three times with 20 0|iL wash
buffer. The wash buffer used was TBS-0.05% Tween for alkaline
phosphatase (AP) enzyme-conjugates or PBS-0.05% Tween for
horseradish peroxidase (HRP) enzyme-conjugates.
4. Add 100(iL of the enzyme-labelled secondary antibody to
each well, Incubate the plates for lhr at room temperature.
Rinse three times with 200(iL wash buffer.
5. Add 100(iL of the appropriate substrate (fresh TMB or ABTS
for HRP; and PNPP for AP enzyme-conjugates). Allow the color
to develop for 2 0 mins, then add 50[J.L of stop solution (1%
SDS, 3M HCL or 3M H2S04).
6. Measure the absorbance of the plate at 405nm. The
correlation between analyte concentration and 405nm


50
absorbance was made each time using the DeltaSoft data
analysis software from BIOTEK or by the least squares
regression method shown in appendix III.
2.2 Specific Aims #2
2.2.1 Amination, Pegylation and Protein Immobilization
on PMMA Membranes.
Histones were immobilized on the surface of intact PMMA
hollow fibers via surface modifications that result in stable
primary amine groups which can undergoe subsequent reactions
with a tresylated star polymer. When the Star-PEG polymers
are coupled to the surface, histones or other proteins can be
immobilized onto the hollow fibers. We describe below the
procedure for aminating and then crosslinking whole histones
to the PMMA hollow fiber via a functionalized 64-arm
polyethylene glycol star. The histone-modified hollow fibers
were then characterized by SEM and AFM or else used in the
adsorption capacity experiments of section 2.3.1.
Materials: PMMA hollow fiber dialyzers were obtained from
TORAY Industries, Inc. (Tokyo). Hydrazine-monohydrate (Fluka
Chemical Corp. New York USA). Tresyl Star-PEG was obtained
from Shearwater Polymers (Huntsville, Alabama USA). Whole
histones (derived from calf thymus) were obtained from Sigma
Chemical Company.
Methods: Histone immobilization on PMMA hollow fibers for
microscopic analysis.
1.Weight out four 75mg samples of PMMA hollow fibers. Rinse
twice with 50ml dd H20 then soak in 50% MeOH for 30min
discard.
2. Add 4ml of 50% MeOH to the hollow fibers in a test tube
then add 370.0|iL of N2H4-H20 activator solution (10.32mg/mL) .
Allow the amination reaction to proceed for 3 hrs.


51
3. After the hollow fibers have been aminated they are rinsed
with 50mL of dd H20, then soaked in 50% MeOH for 30min. This
should result in a maximum of 7 6[imol of activation sites on
the fibers.
4. The aminated PMMA hollow fibers are then incubated in test
tubes containing 4mL of 0.7|iM (in 50% MeOH pH 8.0)
tresylated-Star-PEG for 3hrs.
5. These hydrophilized PMMA hollow fibers are then incubated
as before in 4mL of 3.7mg/mL whole histones in PBS overnight.
Histones are immobilized onto the Star-PEG by covalent
binding to the tresylated end groups. Histone octamers may
spontaneously form at physiologic pH.
6. The histone-modified PMMA hollow fibers are then stored in
PBS-0.1% BSA (to block non-specific binding) until
characterized by AFM or used in adsorption capacity
experiments.
2.2.2 Amination, Pegylation and Protein Immobilization
on PSF Membranes.
Histones were immobilized on the surface of intact PSF
hollow fibers via surface modifications that result in stable
primary amine groups which can undergoe subsequent reactions
with a tresylated Star polymer. When the Star-PEG polymers
are coupled to the surface, histones or other proteins can be
immobilized onto the hollow fibers. We describe below the
procedure for aminating and then crosslinking whole histone
to the PSF hollow fiber via a functionalized 64-arm
polyethylene glycol star.
Materials: PSF hollow fiber dialyzers were obtained from
Fresenius AG, (Germany). Hydrazine-monohydrate (Fluka
Chemical Corp. New York USA). CrCL2 was obtained from Fluka
Chemical Corp. (New York USA). Tresyl Star-PEG was obtained
from Shearwater Polymers (Huntsville, Alabama USA)


52
Methods: Histone immobilization on PSF hollow fibers or
films. A new method for aminating intact PSF hollow fibers in
aqueous solution is presented. Nitration in aqueous solution
followed by reduction of the meta-nitro groups to the
corresponding amine was used to aminate PSF membranes [Fried,
JR 1995]. The aminated films are then surface-modified by
attaching whole histones to the surface via a tresylated-
Star-PEG.
1. The nitration reagent contained
lOmL cone. HN03
20mL conc.H2S04
30mL cone. Acetic acid
30mL distilled H,0
90mL nitration reagent
The acetate is used to stabilize the nitronium ion [N02]~ in
the form of the mixed anhydride.
The PSF hollow fibers or films were then incubated in
2mL of the nitration reagent for 1, 3, 12, and 24 hrs to
acheive four levels of nitration.
2. The nitrated PSF membranes were then incubated for lhr in
4mL of reducing agent (5g CrCl2 in 250mL of 1M HC1). The
reduction was monitored by the change in color of the reagent
from green Cr2+ to Cr3+ which is blue.
3. The aminated films were then hydrophilized by incubation
for 3 hrs with 2mL of 14.2|iM tresyl Star-PEG.
4. Histones were immobilized by incubation with 2mL of 8.6
mg/lOOmL whole histone overnight. The histone-modified films
were blocked against non-specific binding by storage in 0.1%
BSA PBS.


53
2.2.3 Stability and Reactivity of Tresylated Star-PEG
Polymers.
Polyethylene glycol can be coupled to other molecules
through its hydroxyl end-groups. One commonly used reagent is
tresyl chloride which activates the primary-OH into a good
leaving group i.e. tresylate. The activated function can then
be displaced by a protein nucleophile, for example, the
amino-terminal side chain of lysine. Loss of activity by
hydrolysis, however, does occur but at a slower rate than
aminolysis. Hydrolysis is most successful where the ester is
that of a stronger acid, thus the stability of tresylate
esters is intermediate between esters of sulfonic acid and
the more stable tosylates [Mosbach and Nilsson 1984].
Materials: TMB ( 3,3',5,5'-tetramethyl benzidine HCL) and
Reacti-Bind a maleic anhydride activated polystyrene plate;
were both obtained from Pierce Chemical Company (Rockford,
Illinois USA); Tresyl Star-PEG was obtained from Shearwater
Polymers (Huntsville, Alabama USA); Diaminobutane
(putrescine) was obtained from Fluka Chemical Corp. (New York
USA); Tresyl chloride Fluka Chemical Corp. (New York USA);
HRP-antibody conjugate (horseradish peroxidase-IgG) was
obtain from ALPCO, Ltd. (New Hampshire USA)
Method: 1. 300|iL of 1M diaminobutane was incubated in maleic
anhydride activated polystyrene microtiter plates for lhr.
Remove the solution, do not wash.
2. 300pL of the tresyl Star-PEG reagent (0.14|iM) was then
coupled to the aminated wells for 2hr. Remove the solution,
do not wash.
3. Thereafter wells 1-16 were incubate under nitrogen, while
the remaining wells were incubated with 300|iL of 50% MeOH pH
(2.4-11.3) buffers for 30 mins. Remove the solution do not
wash.


54
4. Add 300|iL of HRP-antibody conjugate and incubate for lhr.
Remove the solution then wash three times with wash buffer
(PBS 0.05% Tween).
5. Add 300fiL of the HRP substrate TMB, allow the color to
develop for 30 mins, then read the absorbance at 405nm.
For the control wells the average 405nm absorbance was
0.465 as determined by HRP-conjugate substrate activity.
Taking this value to represent 100% protein binding capacity
of the immobilized Star's, the percent hydrolysis was
estimated iron the equation below
%protein binding + % hydrolysis = 100%
and therefore
% hydrolysis = 1- A4OS/0.465.
We determined the extent of tresylate hydrolysis of
these Star-PEG derivatives over the pH range 2.4-11.3.
2.2.4 Biospecific Interaction Analysis: Kinetic and
Equilibrilium Analysis using the BIACORE 2000
The biospecific interaction between antihistone
antibodies and histones immobilized via PEG Star-polymers was
monitored using the BIACORE 2000. A multistep binding
experiment was conducted where tresylated Star-PEG polymers
were attached to the BIACORE sensorchip in a manner similar
to that in the modification of PMMA membranes. The instrument
measures in real-time the binding kinetics at the flow cell
surface and then displays a corresponding sensorgram which
quantifies the recognition process, see fig.2-4a. The BIACORE
2000 operates on the principle of surface plasmon resonance
(SPR) and is designed to detect changes in the refractive
index of the layer of solution in contact with the sensorchip
[Malmqvist et al. 1993].


55
Since the refractive index of the medium is affected by
the surface concentration of solutes, monitoring the SPR
angle provides a real-time measurement of changes in surface
concentration and/or coupling, see fig.2-5. Units used to
express the SPR signal are called response units (RU). One
thousand RU are equivalent to a change of about lng/mm2 in
surface protein concentration.
We confirmed the immobilization of immunoreactive
histones onto tresylated Star-PEG's and their subsequent
binding to circulating antihistone antibodies using the
BIACORE biosensor. The magnitude of the shift in SPR angle is
proportional to the concentration of proteins bound to the
surface; see fig. 2-4b.
A real-time BIA sensorgram provides essentially two
kinds of information; 1) the rate of a molecular interaction
and 2) the binding level or concentration of the analyte and
the immobilized ligand see fig.2-4c. For all applications the
SPR response is related to the change in surface mass
concentration of the analyte. It also depends on the
molecular weight of the analyte in relation to the number of
ligands on the surface. For any given number of binding
sites, a higher molecular weight analyte will give a
proportionally larger response in RU's, conversely, a higher
molecular weight ligand will provide few binding sites per
unit area (unless the valency is much greater than one) see
fig.2-4d.


56
Figure 2-4. BIACORE Biosensors, a) Schematic of the
optical system and flow cell used in BIACORE
biosensors, b) The surface of a sensor chip consists
of three layers; glass, a thin gold film layer and
a carboxymethylated dextran layer, c) The ligand
or capturing molecule is covalently immobilized
on the sensor chip whereas the analyte is injected
in the flow cell over the surface, d) The response
in resonance units (RU) is recorded as a function of
time; it consists of an association phase (A), an
equilibrium phase (E) and a dissociation phase (D).


57
nnetd tnt*ntty
c)
Retoarte* signal |RU|
Figure 2-5. The BIACORE Sensorgram: surface plasmon
resonance angle (SPR). The sensorgram is a plot of the
(SPR) angle against time; it displays the progress of
the molecular interaction at the sensor surface.
Binding of analyte shifts the SPR angle from
position 1 to position 2. in the figures a) SPR is
observed as a dip in reflection intensity at an angle
which depends on the refractive index of the medium on
non-illuminated side of the surface; b) the SPR angle
shifts when biomolecules bind to the surface and change
the refractive index of the surface layer.c) the
sensorgram shows in real-time a quantitative measurement
of changes in surface concentration of the analyte.


58
Assuming that the relationship between response and mass
is the same for ligand and analyte (i.e. 1000RU-lng/mm2 for
proteins), the following equation may be applied.
#binding sites = (ligand response/ligandMW) (valence of the ligand)
The analytic response = analyte MW(#analyte molecules) at a
fully saturated ligand surface the #binding sites = #analyte
molecules. The response here would represent the maximum
possible resonance signal obtain for this system. Therefore,
we may substitute the expression for the #binding sites for
the #analyte molecules in the equation for analyte response
and obtain an estimate of Rmax
R, = (valence of the ligand) (analyte MW) (ligand response/ligand MW)
R. = nM.RL/ML
Where R* is the maximum binding capacity of the surface
ligands for this particular analyte in RU's.
Materials: BIACORE 2000 (Pharmacia Biosensor., Uppsala,
Sweden). Whole histones (Sigma Chemical Co.); Monoclonal
anti-histone antibodies were derived from WEHI hybridomas
obtained courtesy of Dr. Joel Schiffenbauer of the University
of Florida. Tresyl Star-PEG (Shearwater Polymers.,
Huntsville, Alabama USA)
Methods: Biospecific interaction analysis using either of
the BIACORE biosensors has many methodological similarities
with solid-phase immunoassays. The sensor chip is, however,
the signal transducer instead of the enzyme-antibody
conjugate that yields a signal in ELISA's after exposure to
substrate. The chip is a glass slide with a thin layer of
gold deposited on one side. The gold film is in turn covered
with a covalently bound matrix (e.g. carboxymethyl dextran)
onto which biomolecules can be immobilized.


59
The matrix serves four purposes.
It allows covalent immobilization of biomolecules using
well-characterized coupling chemistry.
It increases sensitivity by increasing the binding
capacity of the surface.
It provides a hydrophilic environment suitable for most
interactions of biological interest.
It provides a very low degree of non-specific binding to
the surface.
Multi-step methods rely on a series of binding steps to
analyze the interaction of interest. For multi-step methods,
real-time BIA has a marked advantage over many other surface
interaction techniques, in that surface binding is registered
on the sensorgram at each stage in the process. This gives
unique facilities for controlling the intermediate binding
events. Multi-step methods are typically used to study
multimolecular complex formation in sequential binding
experiments. This can give valuable information concerning
steric and allosteric effects, since the order of the
interaction and binding can be varied with ease. For an
example see the multi-binding experiment sensorgram shown in
fig. 2-6a and b.
Method: Start the immobilization sequence using an amine
coupling to bind proteins to the surface. The numbered steps
refer to the following stages in the immobilization
experiment.
1. Baseline for the unmodified sensor chip surface with
continuous flow buffer.
2. Injection of NHS/EDC to activate the surface gives an
increase in the SPR signal due to the change in bulk


60
refractive index. Resulting in a NHS ester on the surface
which readily reacts with primary amines.
3. Next we aminated the surface with 1M ethylenediaamine at
pH 8.3. Deactivation of unreacted NHS-esters using 1M
ethanolamine HCL, adjusted to pH 8.5 with NaOH. The
deactivation process also removes any remaining
electrostatically bound ligand.
4. 55 (IM tresyl-Star-PEG is immobilized on the surface (with
64-1 = 63 free arms for subsequent histone binding). Binding
curve was observed.
5. 1:10 whole histones in Hepes buffer was added. Binding
curve was observed.
6. Deactivation of unreacted tresyl groups using 1M
ethanolamine HCL, adjusted to pH 8.5 with NaOH. The increased
SPR signal is due to a change in the bulk refractive index.
The deactivation process also removes any remaining
electrostatically bound ligand.
7. Record the response (in RU's) of the immobilized ligand
after deactivation.
From the resulting sensorgram (see section 3.22) the
response and slope at each time point can be observed. These
parameters are then used to calculate the rate and
equilibrium constants of the histone-antihistone interaction.
2.2.5 Microscopic Analysis: Hollow Fiber Surface
Imaging using Atomic Force and Scanning Electron
Microscopies.
Materials and Method: We used the transmission electron
microscope to estimate the size range of the Star-PEG
polymers. A Star-PEG solution (2mg/mL Star-PEG in 50% MeOH) a
5(iL drop was evaporated on the grid then imaged by TEM. The
TEM images were performed courtesy of Mr. Scott Whitaker of
the University of Florida ICBR Microscopy Core Lab on a


61
Phillips EM-400T transmission electron microscope
(Hassborough, Oregon) at 6kV accelerating voltage.
Figure 2-6. Activation of a BIACORE sensorchip.
a) The sensor chip is activated with a pulse of EDC/NHS.
The resulting response curve displays (streptavidin)
immobilization to the surface. Residual activation sites
are quenched with ethanolamine b) Compares the response
curves of equal concentrations (4 0(ig/mL) of two ligands,
avidin and streptavidin.


62
We used both the SEM and "non-contact mode" AFM to
compare the chemically-modified PMMA with control hollow-
fibers and found organized structures in the size range of a
histone-Star i.e. between 150 and 250 nm. The SEM images were
performed courtesy of Mr. Scott Whitaker of the University of
Florida ICBR Microscopy Core Lab on a Jeol JSM 35CF scanning
electron microscope (Soquelec, Montreal) at 15kV accelerating
voltage. The AFM images were performed courtesy of Burleigh
Instruments, Inc. on a METRIS-NC-2000 scanning probe
microscope (Fishers, New York).
2.3 Specific Aims #3
2.3.1 Adsorption Capacity Experimental Methods: AHA
ELISA
An adsorption capacity experiment was performed to
assess the binding of antihistone antibodies (AHA) from
hybridoma supernatant solution (see fig. 2-7).
Materials Anti-Histone Immunoassay (ALPCO, Ltd. New
Hampshire USA); PMMA hollow fiber dialyzers were obtained
from TORAY Industries Inc. (Tokyo). PSF hollow fiber
dialyzers were obtained from Fresenius AG, Germany.
Hydrazine-monohydrate (Fluka Chemical Corp. New York USA).
CrCL2 was obtained from Fluka Chemical Corp. New York USA.
Tresyl Star-PEG was obtained from Shearwater Polymers
(Huntsville, Alabama USA)
Method: AHA ELISA
For the commercial kits the protocol followed was that
included in the package with one exception. Since the ALPCO
Anti-Histone Immunoassay was designed for the detection of
human AHA and we have designed experiments for the
immunoadsorption of murine (hybridoma) AHA, a anti-mouse IgG-


63
HRP conjugate was used instead in the detection step. In
house ELISA's were made according to method outlined in
section 2.14.
Figure 2-7. Adsorption/Incubation Experiment. In the
adsorption capacity experiments a thin film of the membrane
material (PMMA or PSF) was sheet cast onto the base of test
tubes at a well-defined surface area. The films were then
incubated for 15 mins.in hybridoma cell culture supernatant
containing a known concentration of AHA. The solutions were
sampled before and after each incubation then assay by
ELISA to determine AHA adsorption. The delta sub (i) values
represent the changes in AHA concentration during each
incubation and the sum of these changes was used to
estimate the effective adsorption capacity of these
chemically-modified membranes.


64
Method: Adsorption capacity of surface-modified PMMA films.
PMMA (0.8g) was dissolved in warm CHC13 (lOOmL). After
the slurry cooled to room temperature a portion of it was
covered with 50% Me0H/H20. Plastic cuvettes were then dipped
into the slurry layer and carefully raised up through the
aqueous layer repeatedly until a thin film of PMMA coated the
lower 13 cm2 of the cuvette. The cuvettes were then air dryed
for 4hrs. before being incubated in the hydrazine activator
solution as follows.
Ll = 69.0mg hr/cm2 polymer
L2 = 208.Omg hr/cm2 polymer
L3 = 690.Omg hr/cm2 polymer
L4 = 2081.Omg hr/cm2polymer
After each of the films were activated they were placed
with 250]iL of 12.5^iM tresylated Star-PEG in 2mL 50% MeOH (pH
8.0) for 12hrs. These "hydrophilized" PMMA films were then
incubated in 2mL of 2.5 mg/mL whole histones (PBS buffer)
overnight. The histone modified PMMA films are then placed in
PBS 0.1% BSA to block non-spacific binding sites then stored
at 4 C until used in the adsorption capacity experiments.
The AHA adsorption of a group of modified PMMA films
(L1-L4) was determined by incubating each of the films (13
cm2) in filtered hybridoma cell culture supernatant containing
approximately 1.3mg/3mL of antihistone antibodies as follows:
1. 3mL of the supernatant (1.3mg) was added to each of
twenty., 4mL reacti-vials (previously washed in buffer).
2. Then each of the four films (test tubes) were incubated
sequentially in each of five of these vials for 15 mins. The
initial concentration in each vial being approximately 1.32mg
the change in AHA concentration ACaha in each vial was
measured and plotted against incubation number (see fig.2-7).


65
Extrapolation of the curve yields an approximate adsorption
capacity of the films.
3. Summation of the ACaha values equals the AHA adsorption
capacity of the modified surface per unit area. Half (0.5)
milliliter samples of the supernatant were taken before and
after incubation and assayed for AHA concentration by ELISA
(see methods section 2.14).
4. The total adsorption for the four levels of PMMA surface
modification were calculated and the adsorption capacity per
unit area of each of the treated membranes was determined as
follows, (also see Results abd Discussion specific aim #3)
Q1 = 0.84mg/m2 polymer Q3 = 1.80mg/m2 polymer
Q2 = 0.335mg/m2 polymer Q4 = 0.066mg/m2 polymer
Method: Adsorption capacity of chlorosulfonated surface-
modified PSF film. The mechanism of polysulfone
chlorosulfonation theoretically takes place according to the
following reactions; see fig. 2-8 below [Gilbert 1965;
Pozniak et al. 1995].
1. ArH + C1S03H > ArS03H + HC1
2. ArS03H + CISChH <===> ArS02Cl + H2S04
(basic medium) (acidic medium)
3. ArS02Cl + H20 > AXSO3H + HC1
Figure 2-8. Chlorosulfonation of PSF.
Note: It is clear from the reaction scheme that polysulfone
amination should take place in acid medium free of intact H20.
The viscosity of the polysulfone solution also decreases with
increasing sulfonation rate, and the solvent affinity of
chlorosulfonated PSF is less than that of PSF resulting in


66
decreased solubility as the reaction proceeds. However,
chlorosulfonated PSF film may still be sheet cast onto the
base of glass cuvettes according to the following protocol.
Four levels of films were activated (one control)
Ll = control
L2 = 281 mg C1S03H /gm polymer
L3 = 562 mg C1S03H /gm polymer
L4 = 1124 mg C1S03H /gm polymer
After each of the activated films were sheet casted on
to the base of 11cm2 glass test tubes then incubated in a
solution of 1M diaminobutane for 2hrs. The aminated films
were then allowed to react with 250|iL of 12.5(iM tresylated
Star-PEG in 2mL 50% MeOH for 12hrs. These "hydrophilized" PSF
films were then incubated in 2mL of 2.5 mg/mL whole histones
overnight. The histone-modified PSF films were then placed in
PBS 0.1% BSA to block non-spacific binding sites then stored
at 4 C until used in the adsorption capacity experiments.
The aha adsorption of a group of modified PSF film (Ll-
L4) was determined by incubating each of the films (11 cm2) in
filtered hybridoma cell culture supernatant containing
approximately 88.32|ig/3mL of antihistone antibodies as
follows:
1. 3mL of the supernatant (88.32|ig) was added to each of
twenty, 4mL reacti-vials (previously washed in buffer).
2. Then each of the 4 films (test tubes) were incubated
sequentially in each of five of these vials for 15 mins. The
initial concentration in each vial being 88.32fig the change
in AHA concentration ACaha in each vial was measured and
plotted against incubation number (see fig. 2-7).
Extrapolation of the curve yields an approximate adsorption
capacity of the films.


67
3. Summation of the ACaha values equals the AHA adsorption
capacity of the modified surface per unit area. Half (0.5)
milliliter samples of the supernatant were taken before and
after incubation and assayed for AHA concentration by ELISA
(see methods section 2.14).
4. The total adsorption for the four levels of PMMA surface
modification were calculated and the adsorption capacity per
unit area of each of the treated membranes was determined,
(also see Results and Discussion specific aim #3)
Q1 = 0.0mg/m2 polymer Q3 = 1.06mg/m2 polymer
Q2 = 0.67mg/m2 polymer Q4 = 3.93mg/m2 polyme
Method: Adsorption capacity of nitration surface-modified
PSF film. Since chlorosulfonation of PSF requires the polymer
to be activated in organic solvents which dissolve the PSF
hollow fiber, methods of aminating the intact PSF hollow
fibers in aqueous were attempted. Nitration in aqueous
solution followed by reduction of the nitro-groups to the
corresponding amine was used to aminate PSF films sheet cast
at the base of plastic cuvettes. The aminated films were then
surface-modified by attaching whole histones to the surface
via a tresylated Star-PEG.
1. The nitration reagent contained
lOmL cone. HN03
20mL cone. H2SCh
30mL cone. Acetic acid
30mL distilled H,0
90mL nitration reagent.
The acetate is used to stabilize the nitronium ion [NO2]- in
the form of the mixed anhydride.


68
2. 13cm2 PSF films were sheet cast onto the base of plastic
cuvettes from the slurry and allowed to air dry for lhr, then
stored in distilled H20 before activation. The films were then
incubated in 2mL of the nitration reagent for 1, 3, 12, and
24 hrs to acheive four levels of nitration i.e. 1, 3, 12, and
24 hrs.
3. The nitrated PSF films were then each incubated for lhr in
4mL of reducting agent (5g CrCl2 in 250ml of 1M HC1). The
reduction was monitored by the change in color of the reagent
from green Cr2+ to Cr3+ which is blue.
4. The aminated films were then surface-modified by reacting
with 2mL of 14.2|iM tresyl Star-PEG and then 2mL of
8.6mg/100mL whole histone. The histone-modified films were
blocked against non-specific binding by storage in 0.1% BSA
PBS.
5. AHA adsorption capacity of the modified PSF films was then
estimated by incubating each of the treated films in
antihistone antibody hybridoma supernatant (1400U/2mL). The
concentration of the AHA in each reacti-vial was assayed
before and after the 15 min. incubation with the treated film
and the change in AHA concentration ACaha determined by ELISA
(see fig.2-7). The results of the experiment showed that
films activated in the nitration reagent for 1, 3, 12, and 24
hrs. respectively adsorbed the following amounts.
Q1 = 1.844 mg/m2 polymer Q3 = 2.29 mg/m2 polymer
Q2 = 1.190 mg/m2 polymer Q4 = 2.18 mg/m2 polymer
2.3.2 Extracorporeal Immunoadsorption Experimental
Methods.
We have used the in vitro model (described earlier) to
produce and evaluate chemically-modified hollow fiber
dialyzers as potential blood purification devices. Binding
agents were covalently immobilized onto the inside or outside


69
surface of polymethylmethacrylate (PMMA) and polysulfone
(PSF) hollow fibers by coupling reactions designed to remove
antihistone antibodies from donor compartment solutions over
the course of 4hrs.
Materials: Monoclonal anti-histone antibodies were derived
from WEHI hybridomas obtained courtesy of Dr. Joel
Schiffenbauer of the University of Florida. Monoclonal anti
histone antibody standards were obtained from Boehringer
Mannheim GmbH, (Indianapolis, USA). Anti-Histone Immunoassay
ALPCO, Ltd. (New Hampshire USA); PMMA hollow fiber dialyzers
were obtained from TORAY Industries, Inc. (Tokyo). PSF hollow
fiber dialyzers were obtained from Fresenius AG, Germany.
Hydrazine-monohydrate Fluka Chemical Corp. (New York USA).
CrCL2 was obtained from Fluka Chemical Corp. (New York USA).
Tresyl Star-PEG was obtained from Shearwater Polymers
(Huntsville, Alabama USA)
Methods:
1. The model circuit was rinsed throughly and then primed
with 50% MeOH/HjO. PMMA dialyzers were activated with 5.2mM
hydrazine-monohydrate (in 50% MeOH) for 4 hr. Polysulfone is
activated instead with "strong nitration reagent" then
reduced with "strong reducing agent" (5g CrCL2 in 250mL in
1M HCL).
2. The aminated hollow-fibers are then rinsed thoroughly with
50% Me0H/H20 then allowed to react with 14.2(iM tresyl-Star-PEG
for 8 hrs. Since each immobilized Star-PEG may bind to the
activated polymer surface by a single tresylate end group,
this would leave 64 1 = 63 PEG arm's free to bind proteins.
The immobilized star polymers dramatically increase the
active surface area of the hollow fibers.


70
3. After the Star-PEG coupling the circuit was washed
throughly with PBS. Then 0.4g (3.67|imol octamers) of histones
in PBS was injected into the circuit and allowed to bind onto
the star polymers. Hypothetically each arm of the immobilized
Star-PEG is available for each protein molecule.
4. After these modified dialyzers are rinsed with PBS 0.1%
albumin they were refrigerated at 8 C until ready for use in
an extracorporeal immunoadsorption experiment.
The clearance of antihistone antibody was tested in PMMA
hollow fiber dialyzers with whole histones immobilized on the
outside and inside lumen of the membrane. The results were
then compared against controls. The dialyzers were modified
as described earlier.
5. A known amount of antihistone antibody (0.289mg/mL) was
then injected into the donor circuit and allowed to mix (in
bypass) for 5 mins., after which time the circuit was taken
out of bypass and opened to one of the test dialyzers. Donor
compartment inlet and outlet samples (0.5mL) were then taken
at 5, 10, 30, 60, 120 and 240 minute intervals from the time
the donor circuit was opened. The samples were analyzed by
Anti-Histone Immunoassay for AHA content.
In the plasma studies the donor reservoir (475mL) was
filled with plasma and spiked with approximately 16.75mg of
murine AHA, then mixed in bypass at room temperature for
0.5hr. Modified or control dialyzers were then placed in the
extracorporeal circuit and perfused with the spiked plasma
for 4hrs. Plasma samples were taken from the dialyzer inlet
and outlet sampling port and assayed for AHA concentration.


CHAPTER 3
RESULTS AND DISCUSSION
3.1Specific Aim #1
3.1.1 An In Vitro Model for the Study of
Extracorporeal Therapy: Applications of the Model.
The results of the experiments that follow will
demonstrated the utility of the model in evaluating
hemodialyzer performance and in studying the effect of
dialysate additives on the clearance rate of "middle
molecules". Other potential applications of the model are
also shown, see fig.3-1.
3.1.2 The Effects of Dialysate Additives on the
Transport Properties of Donor Solutes across Hollow
Fiber Dialyzers.
It was found that the presence of antibodies in the
dialysate compartment potentiates the clearance of specific
antigens beyond steady-state concentrations only when the
permeability of the dialyzer membrane is sufficiently great
for antigen diffusion. In fig.3-2 the transport
characteristics of vit.B12 in a high flux PMMA dialyzer are
shown. The donor compartment concentration decreases from
2.75mg/100mL to a steady-state concentration of 1.85mg/100mL
in the first 60 minutes. After anti-vit.B12 antibodies were
added to the receiver compartment the mass transport was
increased again and donor compartment vit.Bi2 continued to
decrease from 1.85mg/100mL to 1.35mg/100mL within the next
60mins. The experiment shows that sink conditions for a
specific donor solute can be re-established by the presence
of complexing additives in the receiver compartment.
71


72
APPLICATIONS
DRUG METABOLISM: Immobilized microsomal enzymes may be used to study the
metabolism of new lead compounds in a physiologically controlled environment.
EXTRACORPOREAL IMMUNOADSORPTION: Immobilized self-antigens (i.e. histones) may
be use to study the adsorption of autoantibodies (i.e. anti-histone antibodies) or other
disease associated antigens.
MEMBRANE TRANSPORT: The transport properties of blood-borne solutes across hollow-
fibers may studied and/or manipulated by dialysate additives (i.e. antibodies,
cyclodextrins)
CELL CULTURE SUPERNATANT FRACTIONATION: By-products of cellular metabolism and
gene expression can be monitored and separated on-line in the in vitro model.
RADIOIMMUNOTARGET1NG : Targeting of radioimmunoconiugates to selected tissue-types
can be studied using a radiolabeled biotinylated-monoclonal antibody to tissue cell
antigens. After the radioimmunoconjugate has accumulated in the target tissue, the
circuit may be taken out of bypass so that the unbound radioimmunoconjugate can
adsorb to the surface modified (e g. avidin) hollow-fiber dialyser.
BLOOD COMPARTMENT DIALYSATE
COMPARTMENT
BLOOD PUMP DIALYSATE PUMP
Figure 3-1
In Vitro Model: applications


73
The mass clearance under equilibrium conditions was
(2.75-1.85)mg/100mL hr
=> (0.9mg/100mL hr)*(230mL)
= 2.07mg/hr
and that under "sink" conditions was
(1.85-1.35)mg/100mL hr
=>(0.5mg/100mL hr)*(230mL)
- 1.15mg/hr
This should be considered as a rate enhancement effect
against an equilibrium concentration gradient. Thus a
"clearance enhancement factor" can then be approximated for
anti-vit.Bi2 in this system which is equal to
1.15 + 2.07 = 3.22mg/hr
(3.22/2.07)-1 = 55.6%
A similar experiment was conducted using a high flux
cellulose acetate dialyzer see fig.3-3. The equilibrium and
sink condition clearances were equal to 1.84mg/hr and
1.27mg/hr.
The a "clearance enhancement factor" for anti-vit. B12 in
this system is equal to
(3.105/1.84)-l = 68.8%
In fig. 3-4, the transport of vit.Bi2 in a low flux
polysulfone dialyzer was studied. Equilibrium and sink
condition clearances were again determined and found equal to
0.644mg/hr and 0.575mg/hr, respectively. A clearance
enhancement factor of 89.3% was calculated here.


(IUJOOI/Buj) ooq
74
Transport ot vitamin B12 in a BK2.1P dialyzer
Figure 3-2. Mass transport curves of vit.Bi2 in a high
flux PMMA dialyzer. The solid () and hollow (o) circles
represent donor vit.B12 cone, before anti-vit.B12 antibody
administration to the receiver compartment at 60 mins.
The hollow triangles (A) represent receiver vit.B12
during the same time period. The hollow squares() and
solid triangles() represent donor vit.B22 inlet and
outlet concentrations respectively^, after anti-vit.B12
antibodies were injected into the receiver compartment.


Cone, (mg/100ml)
75
Transport ol vitamin B12 In a Altrex dlalyzer
Figure 3-3. Mass transport curves of vit.B12 in a high
flux cellulose acetate dialyzer. The hollow (o) circles
represent donor vit.Bu cone, before anti-vit.B22 antibody
administration to the receiver compartment at 120 mins.
The hollow triangles (A) represent receiver vit.B12
during the same time period. The hollow squares() inlet
and concentrations respectively, after anti-vit.B12
antibodies were injected into the receiver compartment.


76
Transport of vitamin B12 In a F8 dialyzer
Figure 3-4. Mass transport curves of vit.Bi2 in a low
flux PSF dialyzer. The solid () and hollow (o) circles
represent donor vit.B12 cone, before anti-vit.Bl2 antibody
administration to the receiver compartment at 60 mins.
The hollow triangles (A) represent receiver vit.B12
during the same time period. The hollow squares ()
represent donor vit.Bi2 inlet concentrations after anti-
vit.Bij antibodies were injected into the receiver
compartment.


P2-M Cone.(pg/100ml)
77
Similar transport studies were performed with the
polypeptides (32-M and CD25. Fig. 3-5 shows that for 02-M in
the high flux PMMA dialyzer, hydrophobic interactions were
responsible for the rapid decrease in donor concentration
from 10 to 0.7|ig/mL in the first 2hrs of the experiment. No
significant amounts of P2-M were measured in the receiver
compartment. Therefore, no clearance enhancement due to
dialysate additives (antibodies) was observed.
Transport of P2-M in a BK2.1 P dialyzer
Figure 3-5. Mass transport curves of p2-M in a high
flux PMMA dialyzer. The hollow (o) circles represent
donor vit.Bi2 cone, before anti-vit.B12 antibody
administration to the receiver compartment at 120 mins.
The hollow triangles (A) represent receiver vit.B12
during the same time period. The hollow squares()
represent donor vit.Bi2 inlet concentrations
respectively, after anti-vit.Bu antibodies were injected
into the receiver compartment.


P2-M Cone. (pg/100ml)
78
In fig. 3-6, the clearances of p2-M were studied in a
high flux cellulose acetate dialyzer. The equilibrium and
sink condition clearances were determined as 3.62(ig/mL and
3.22(ig/mL with a clearance enhancement factor of 8 8.9%.
Transport of P2-M in a Altrex dialyzer
Figure 3-6. Mass transport curves of P2-M in a high
flux cellulose acetate dialyzer. The hollow () circles
represent donor P2-M cone, before anti-P2-M antibody
administration to the receiver compartment at 120 mins.
The hollow triangles (A) represent receiver vit.Bl2
during the same time period. The hollow squares() inlet
and concentrations respectively, after anti-P2-M
antibodies were injected into the receiver compartment.


79
In fig. 3-7 (32-M transport in the low flux polysulfone
dialyzer was impeded due to the smaller pore size of the
hollow fibers and the larger molecular diameter of the
polypeptide. No clearance into the receiver compartment was
observed.
Transport of (32-M in a F8 dialyzer
E
o
o
O)
=L
d
c
o
O
C\J
CD_
Figure 3-7. Mass transport curves of (32-M in a low flux
PSF dialyzer^The solid () represent donor (32-M cone,
before anti-(32-M antibody administration to the receiver
compartment at 60 mins. The hollow triangles (A)
represent receiver (32-M during the same time period. The
hollow squares(Q) represent donor (32-M inlet and
outlet concentrations respectively, after anti-(32-M
antibodies were injected into the receiver compartment.


80
In fig. 3-8 CD25 in the high flux PMMA dialyzer, the
equilibrium and sink condition clearances were determined as
0.19ng/mL and 0.11ng/mL with a clearance enhancement factor
of 57.6%.
Transport of CD25 in a BK2.1P dialyzer
time mins.
Figure 3-8. Mass transport curves of CD25 in a high
flux PMMA dialyzer. The hollow () circles represent
donor CD25 cone, before anti-CD25 antibody
administration to the receiver compartment at 120 mins.
The hollow triangles (A) represent receiver CD25 during
the same time period. The hollow squares() represent
donor CD25 inlet concentrations respectively, after
anti-CD25 antibodies were injected into the receiver
compartment.


81
In fig.3-9 the clearances of CD25 were also studied in a
high flux cellulose acetate dialyzer.jlhe equilibrium and sink
condition clearances were determined as 0.172ng/mL and
0.422ng/mL with a clearance enhancement factor of 245%.
Transport of CD25 in a Altrex dialyzer
time mins.
Figure 3-9. Mass transport curves of CD25 in a high
flux cellulose acetate dialyzer. The solid () circles
represent donor CD25 cone, before anti-CD25 antibody
administration to the receiver compartment at 120 mins.
The hollow triangles (A) represent receiver CD25 during
the same time period. The hollow squares() represent
donor CD25 inlet concentrations respectively, after
anti-CD25 antibodies were injected into the receiver
compartment.


82
In fig.3-10 CD25 transport in the low flux polysulfone
dialyzer was also impeded due to the smaller pore size of the
hollow fiber membrane and the larger molecular diameter of
the polypeptide. No clearance into the receiver compartment
was observed.
Transport of CD25 in a F8 dialyzer
time mins.
Figure 3-10. Mass transport curves of CD25 in a low
flux PSF dialyzer. The hollow () circles represent
donor CD25 cone, before anti-CD25 antibody
administration to the receiver compartment at 120 mins.
The hollow triangles (A) represent receiver CD25
during the same time period.The hollow squares ()
represent donor CD25 inlet concentrations respectively
after anti-CD25 antibodies were injected into the
receiver compartment.


83
3.1.3 Summary.
We have shown that an in vitro model can be constructed
which mimics the extracorporeal circulatory system used in
conventional as well as novel dialysis therapies. The model
can be used to evaluate the clearance of a select donor
solute or a group of solutes under both equilibrium or sink
conditions. Passive adsorption of (32-M to PMMA membranes was
conclusively demonstrated in the model. New dialysate
formulations may be studied in the model as well. Here, we
found that dialysate additives which formed stable complexes
with a specific donor solute (i.e. antibodies) may increase
the clearance of that solute above and beyond steady-state
concentrations. For a summary of the experimental mass
transfer results see table 3-1.
Table 3-1. Summary of the experimental mass transfer
results. Percentages are clearance enhancement factors
described in the text.
Vitamin B,,
beta 2-M
CD25
BK2.1P
78.0%
22.06pg
57.6%
adsorbed
Altrex
68.8%
88.9%
245%
F8
89.3%
no clearance
no clearance
>mwco
>mwco
3.2 SPECIFIC AIM #2
3.2.1 Estimates of Tresylate Hydrolysis and
Reactivity.
Polyethylene glycol (PEG) can be coupled to other
molecules through its terminal hydroxide. One commonly used
coupling agent, tresyl chloride, activates these primary-OH's


84
into good leaving groups. The activated function can then be
displaced by a protein nucleophile, such as the amino-side
chain of lysine. See fig.3-11.
RO
+ ciso2-ch2-cf3
RO
n 0S02CH2CF
tresyl chloride
tresylated MPEG
Figure 3-11. Tresylated Star-PEG.
The (electrophilically) activated PEG-tresylate is a
very efficient derivative for coupling primary amines. The
product secondary amine is very stable toward hydrolysis and
the conjugate formed has similar charge (pi) to the native
protein, so that its bioreactivity is preserved; see fig.3-
12.
pH 7.5 8.5
PEG-0S02CH2CF3 + Protein-NH2 PEG-NH-Protein
Figure 3-12. PEG-protein coupling reaction.


85
Loss of activity by hydrolysis however, does occur but
at a slower rate than aminolysis. See fig. 3-13.
O
II
HOsch2cf3
o
Figure 3-13. Tresylate Hydrolysis.
Hydrolysis is most successful where the ester is that of a
strong acid i.e. -S030H > -S03CH2CF3 > -S03-ArR
Using a solid phase protein binding assay, the percent
hydrolysis of tresylated Star-PEG was determined 1) to be pH
dependent and 2) to be minimal at pH values 4.0 and 8.0; at
which tresylate hydrolysis was less than 35% in 0.5 hrs. See
fig. 3-14.


Full Text
THERAPEUTICALLY MODIFIED HOLLOW FIBER DIALYZERS
By
MICHEL LEE BRANHAM
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
1998

This work is dedicated to my parents Landon Frank Branham Jr.
and Norma Elizabeth Branham, who through love and guidance
have inspired in me the joy of learning.

ACKNOWLEDGEMENTS
I would like to express my gratitude and admiration to
Professor Ian R. Tebbett for his guidance, patience and
encouragement throughout the course of this research program.
Also, I thank the very talented members of my supervisory
committee, Drs. Christopher Batich Ph.D., Edward Ross M.D.,
Kenneth Sloan Ph.D., Roger Tran-Son-Tay Ph.D. and Donna
Wielbo Ph.D. I do greatly appreciate their advice and support
during my graduate study. This project would not have been
successful without the financial support of the Florida
Education Fund and RW Johnson Graduate Research Fellowships
for which I am both honored and grateful.
For their endearing kinship and personal support I owe
special thanks to Dr. Betty Parker-Smith (Florida Education
Fund) and Mr. Robert L. Woods (University of Florida Graduate
Minority Programs). Finally, since much of the laboratory
work was done in the immunology laboratory of the Department
of Pharmacy Practice; I thank Dr. Janet Karlix for all her
assistance and the use of those facilities.

TABLE OF CONTENTS
page
ACKNOWLEDGEMENTS iii
LIST OF TABLES vii
LIST OF FIGURES ix
LIST OF ABBREVIATIONS xiii
ABSTRACT xiv
CHAPTERS
1 INTRODUCTION
1.1 Statement of Problem 1
1.2 Background Review 7
1.2.1 Clinical Implications of Middle Molecules in
Extracorporeal Therapy 7
1.2.2 The Effects of Dialysate Additives on the
Transport Properties of Donor Solutes across
Hollow Fiber Membranes 9
1.2.3 Molecular Pathogensis of Lupus Nephritis... 12
1.2.4 Anti-(DNA-Histone) Antibodies in Active Lupus
Nephritis 14
1.2.5 Surface Modification of PMMA and PSF Dialysis
Membranes 17
1.2.6 Surface-modified polymethylmethacrylate
(PMMA) 20
1.2.7 Surface-modified polysulfone (PSF) 25
1.2.8 Immunoreactivity of Immobilized Histones
Characterized using the BIACORE 2000 26
1.2.9 Surface Analysis of Surface-Modified Hollow
Fibers 28
1.2.10 Extracorporeal Immunoadsorption Therapies
in SLE 31
1.3 Hypothesis 39
1.4 Specific Aims and Rationale 4 0
IV

2 MATERIALS AND METHODS
2.1 Specific Aim #1 42
2.1.1 Physical Characteristics and Specifications
of the Model 42
2.1.2 Mass Transport Experiments 46
2.1.3 UV Spectrophotometric Analysis of Vit.B12...47
2.1.4 ELISA of CD25 and (32-M 48
2.2 Specific Aims #2 50
2.2.1 Amination, Pegylation and Protein
Immobilization on PMMA Membranes 50
2.2.2 Amination, Pegylation and Protein
Immobilization on PSF Membranes 51
2.2.3 Stability and Reactivity of Tresylated
Star-PEG Polymers 53
2.2.4 Biospecific Interaction Analysis: Kinetic
and Equilibrilium Analysis Using the
BIACORE 2000 54
2.2.5 Microscopic Analysis: Hollow Fiber Surface
Imaging Using Atomic Force and Scanning
Electron Microscopes 60
2.3 Specific Aims #3 62
2.3.1 Adsorption Capacity Experimental Methods:
AHA ELISA 62
2.3.2 Extracorporeal Immunoadsorption Experimental
Methods 68
3 RESULTS AND DISCUSSION
3.1 Specific Aim #1 71
3.1.1 An In Vitro Model for the Study of
Extracorporeal Therapy: Applications of the
Model 71
3.1.2 The Effects of Dialysate Additives on the
Transport Properties of Donor Solutes Across
Hollow Fiber Dialyzers 71
3.1.3 Summary 83
3.2 Specific Aim #2 83
3.2.1 Estimates of Tresylate Hydrolysis and
Reactivity 83
3.2.2 Immunoreactivity of Immobilized Histones
Characterized Using the BIACORE 2000 86
3.2.3 Surface Analysis of Chemically-Modified
Hollow Fibers 90
3.2.4 Summary 90
3.3 Specific Aim #3 91
3.3.1 Antihistone Adsorption Capacity of PMMA and
PSF Dialysis Membranes 91
3.3.2 Extracorporeal Immunoadsorption of
Circulating Antihistone Antibodies from
Saline Solution 102
v

3.3.3 Extracorporeal Immunoadsorption of
Circulating Antihistone Antibodies from Human
Plasma 108
3.3.4 Extracorporeal Immunoadsorption of
Circulating Antihistone Antibodies from Human
Plasma Unmodified PMMA and PSF Dialyzer...114
3.3.5 Extracorporeal Immunoadsorption of
Circulating Antihistone Antibodies from Human
Plasma Unmodified PMMA and PSF Dialyzer... 114
3.4 Summary 117
4 CONCLUSIONS 118
APPENDICES
I MASS TRANSFER IN HOLLOW FIBER DIALYZERS 122
II TUBING SELECTION TABLE 130
III A LEAST SQUARES LINEAR REGRESSION METHOD 132
IV ANTIHISTONE ANTIBODY HYBRIDOMAS 133
V STAR-PEG POLYMERS 136
VI NON-CONTACT MODE AFM: BURLEIGH ME TRIS NC-2000 145
VII PURIFICATION OF ANTIBODIES BY THIOPHILIC ADSORPTION
CHROMATOGRAPHY: T-GEL FRACTIONATION 152
LIST OF REFERENCES 155
BIOGRAPHICAL SKETCH 163
VI

LIST OF TABLES
Table page
1-1 Criteria for Classification of SLE 2
1-2 Causes of death in a cohort of systemic lupus
patients 3
1-3 Characteristics of SLE Patients Treated in Okinawa,
Japan between 1972 and 1991 4
1-4 Study of the correlation between cause of death, sex,
race and socio-economic status (SES) in patients with
SLE 5
1-5 Principal Characteristics of Calf Thymus Histones 15
1-6 A Summary of the Prevalence of Histone Autoantibodies in
SLE patients 16
1-7 Methods Used to Prepare Microporous Membrane
Materials 21
1-8 Common Surface Analytical Methods 29
1-9 Autoantibodies in SLE 35
2-1 Dialyzer Specifications 45
2-2 Circuit Specifications 45
2-3 Antigen/Antibody Concentrations in Mass Transfer
Experiments 47
3-1 Summary of the Experimental Mass Transfer Results 83
3-2 Data used in the Kinetic and Equilbrium Analysis of the
Histone-Antihistone Interaction 87
3-3 The kinetic and equilibrium constants for AHA binding
to PEG immobilized histones 89
3-4 Total mouse AHA removed from PBS solution 104
vii

3-5 Total human AHA removed from PBS solution 106
3-6 Total mouse AHA eluted from F8 dialyzers: preliminary
study 108
3-7 Total AHA removed from human plasma preliminary
study 113
3-8 Total AHA removed from human plasma by a modified
dialyzer 117
I-1 Mass transfer parameters for different flow
geometries 126
II-1 Tubing Selection and Compatibility 131
V-l Synthetic Methods for Star-PEG Polymers 137
V-2 Efficiency of Dendritic Seed polymer in Star-PEG
Synthesis 144
viii

LIST OF FIGURES
Figure page
1-1 Environmental influences may be related to the
prevalence of SLE 6
1-2 The structure and physical properties of
beta 2-microglobulin 9
1-3 The interleukin-2 receptor complex 10
1-4 Structural models of the "soluble" IL-2R alpha
subunit CD25 11
1-5 Mechanisms of Immune Complex Deposition 14
1-6 Structure of the "Histone octamer-Nucleosome core".... 15
1-7 Proposed Mechanism of Histone-Mediated Glomeruli
Deposition of Immune Complexes 17
1-8 PMMA monomer synthesis 2 0
1-9 Surface-modified PMMA 22
1-10 Formation of Potentially Toxic Hydrazine Derivatives..23
1-11 Condensation polymerization for PSF 25
1-12 Surface-modified PSF 27
1-13 Flowchart for the Surface Characterization of
Biomaterials 30
1-14 Determination of the capacity of an affinity adsorbent
by frontal analysis 33
2-1 In Vitro Model Specifications 4 3
2-2 Structure of commercial hollow fiber dialyzers 44
2-3 Calibration curve and linear regression equation for
the spectrophotometric determination of Vitamin B12....4 8
IX

56
2-4 BIACORE Biosensors
2-5 The BIACORE Sensorgram: surface plasmon angle (SPR)
angle 57
2-6 Activation of a BIACORE sensorchip 61
2-7 Adsorption/Incubation Experiment 63
2-8 Chlorosulfonation of PSF 65
3-1 In Vitro Model: applications 72
3-2 Mass transport curves of vit.Bi2 in a high flux PMMA
dialyzer 74
3-3 Mass transport curves of vit.BX2 in a high flux cellulose
acetate dialyzer 75
3-4 Mass transport curves of vit.Bi2 in a low flux PSF
dialyzer 76
3-5 Mass transport curves of P2-M in a high flux PMMA
dialyzer 77
3-6 Mass transport curves of p2-M in a high flux cellulose
acetate dialyzer 78
3-7 Mass transport curves of [32-M in a low flux PSF
dialyzer 7 9
3-8 Mass transport curves of CD25 in a high flux PMMA
dialyzer 80
3-9 Mass transport curves of CD25 in a high flux cellulose
acetate dialyzer 81
3-10 Mass transport curves of CD25 in a low flux PSF
dialyzer 82
3-11 Tresylated Star-PEG 84
3-12 PEG-Protein coupling reaction 84
3-13 Tresylate Hydrolysis 85
3-14 pH Dependent Hydrolysis of Tresylated Star-PEG
polymers 86
3-15 Antihistone-Histone Interaction Sensorgram 88
3-16 Size Estimation of the Star-PEG 92
x

94
3-17.SEM micrograph of the lumen of a treated PMMA hollow
fiber
3-18 Atomic force micrograph scanned at 2.0pm of a treated
and untreated PMMA hollow fiber 95
3-19 Atomic force micrograph scanned at 1.0pm of a treated
PMMA hollow fiber 97
3-20.Atomic force micrograph scanned at 1.0pm of a untreated
PMMA hollow fiber 98
3-21. Adsorption capacity of surface-modified PMMA
membranes 100
3-22 Adsorption capacity of chlorosulfonated surface-
modified PSF membranes 101
3-23 Adsorption capacity of nitration/reduction surface-
modified PSF membranes 10 3
3-24 Extracorporeal immunoadsorption of murine antihistone
antibodies from saline 105
3-25 Extracorporeal immunoadsorption of human antihistone
antibodies from saline 107
3-26 Extracorporeal immunoadsorption of murine antihistone
antibodies from plasma: preliminary study 109
3-27 Elution of AHA from a modified and a unmodified PSF
dialyzer 110
3-28 T-Gel fractionation of plasma samples from preliminary
study 112
3-29 Extracorporeal immunoadsorption of murine antihistone
antibodies from plasma by unmodified dialyzers 115
3-30 Extracorporeal immunoadsorption of murine antihistone
antibodies from plasma by surface-modified
dialyzers 116
I-l Hemodialysis Configuration 124
V-l Core-First Method using poly(DVB) 137
V-2 A trifunctional initiator: triisobutyl benzene
chloride 138
V-3 A Dendritic PEG star by anionic polymerization 139
V-4 Plurifunctional electrophilic deactivator 140
XI

V-5 Stars by copolymerization with a dialkenyl monomer... 140
V-6 Seed Star Methods 14 2
V-7 Preparation of dendritic polyamidoamine (PAMAM)
polymers 143
VI-1 Atomic Force Microscopy 147
VII-1 Thiophilic Adsorption Chromatography 152
X

LIST OF ABBREVIATIONS
AFM :
AHA :
AP :
B S A :
P2-M :
ds DNA:
DVB :
E DC :
ELISA:
EPO :
ERSD :
EVS :
GFR :
HRP :
IC :
IVS :
mwco :
MM s
NHS :
PAMAM:
PEG :
PEVA s
PMMA :
PNPP s
PSF :
PS :
RID :
RU :
SEM s
SES s
SIL2R:
SLE :
SPM :
SPR :
TEM :
TMB :
Atomic Force Microscope
Antihistone Antibody
Alkaline Phosphatase
Bovine Serum Albumin
beta-2-Microglobulin
double stranded DNA
Divinyl Benzene
1-Ethyl-3-(-Dimethylaminopropyl) Carbodiimide
Enzyme Linked Immunoadsorbent Assay
E rythropoietin
End Stage Renal Disease
Extravascular Space
Glomerular Filtration Rate
Horseradish Peroxidase
Immune Complex
Intravascular Space
Molecular weight cutoff
Middle Molecules
N-Hydroxysuccinimide
Polyamidoamine
Polyethylene glycol
Polyethylene vinyl acetate
polymethylmethacrylate
p-Nitrophenyl phosphate
polysulfone
polystyrene
radioimmunodif fusion
Resonance Units (also response units)
Scanning Electron Microscope
Socio-economic status
Soluble Interleukin-2 Receptor (also CD25)
Systemic lupus erythematosus
Scanning Probe Microscope
Surface Plasmon Resonance
Transmission Electron Microscope
3,3',5,5'Trimethyl Benzidine hydrochloride
xiii

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
THERAPEUTICALLY MODIFIED HOLLOW FIBER DIALYZERS
By
MICHEL LEE BRANHAM
December 1998
Chairman: Ian R. Tebbett, Ph.D.
Major Department: Medicinal Chemistry
Extracorporeal immunoadsorption is a novel strategy for
the removal of disease associated molecules from the
patient's blood stream. This approach possesses unprecedented
molecular-specificity, is potentially more cost effective
than classical drug discovery and is often the only
therapeutic option in medical emergencies or in the
terminally ill. For example, pathogenic candidates for this
method of extraction include (32-microglobulin in dialysis-
related amyloidosis, anti-nuclear autoantibodies in lupus
nephritis, as well as several soluble cytokine receptors
involved in immunodeficiency disorders like AIDS and cancer.
Since the above mentioned toxins are not removed during
conventional hemodialysis therapy, therapeutic membrane
modifications offer new opportunities to treat these and
other diseases.
XIV

Current technology exists for the immunoadsorption of
blood-borne antigens using plasmapheresis, but this method
requires the separation of plasma from the formed elements
prior to immunoadsorption. Some of the disadvantages of
plasma separation include hypervolemic effects, red cell
hemolysis, increased setup time and instrumental complexity.
We have shown that an in vitro model can be constructed which
mimics the extracorporeal circulatory system used in
conventional as well as novel dialysis therapies. The model
was used to evaluate the clearance of a select donor solute
or a group of solutes under both equilibrium and sink
conditions. Passive adsorption of (}2-M to PMMA membranes was
conclusively demonstrated in the model. New dialysate
formulations may be studied in the model as well. Here, we
found that dialysate additives which formed stable complexes
with a specific donor solute (i.e. antibodies) may increase
the clearance of that solute above and beyond steady-state
concentrations. We have also found that surface-modification
of commercial hollow fiber dialyzers is a novel method of
preparation for highly selective bioadsorbent materials. The
effects of these modification were also demonstrated in the
in vitro model. It was found that the use of polyethylene
glycol spacers in the immobilization procedure significantly
increases the surface area of the hollow fibers, it
apparently stabilizes the epitope recognized by the target
molecule (antihistone antibodies) and thereby increases
binding. Polyethylene glycol effectively prevents nonspecific
binding to the surface. The result of the experiments show
that milligram quantities of immunoglobulin can be removed
from human plasma by these surface-modified dialyzers.
XV

CHAPTER 1
INTRODUCTION
1.1 Statement of Problem
Systemic lupus erythematosus (SLE) is an autoimmune
disease of unknown etiology, characterized by the chronic
inflammation of multiple organ systems. Glomerulonephritis is
a frequent complication of SLE; the presence and extent of
kidney involvement greatly influence the outcome of this
disease [Tron and Bach 1977 ]. SLE is characterized by the
production of circulating autoantibodies to constituents of
the cellular nucleus. Although analytical difficulties still
remain, the criterion for diagnosis and classification of SLE
has been established by the American College of Rheumatology
to facilitate the uniform reporting of SLE cases in the
U.S.[Theofilopoulos and Dixon 1982] see table 1-1.
SLE can occur at any age after puberty; with women of
childbearing age being primarily affected. In cases that
begin between ages 15 and 40 yrs, greater than 90% of the
patients are female [Fessel 1974; Hochberg 1993]. These data
strongly suggest that sex hormones influence the probability
of developing or expressing SLE; studies in animal models of
lupus have supported a potential role for estrogen
enhancement and androgen protection against expression of the
disease [Roubinian et al 1978; Hochberg 1993].
1

2
The chance of a white woman in the U.S. developing SLE
in her lifetime is approximately 1 in 700 (0.14%). This
incidence is 3 to 4 times greater for blacks in the U.S., as
well as in certain tribes of Native Americans (e.g. Sioux,
Crow, Arapahoe) [Feldman et al. 1992].
Table 1-1. Criteria for Classification of SLE.
Criterion
Definition
1. Malar rash
2. Discoid rash
3. Photosensitivity
4. Oral ulcers
5. Arthritis
6. Serositis
7 Renal disorder
8. Neurologic disorder
9. Hematologic disorder
10. Immunologic disorders
11. Antinuclear antibody
(ANA)
Fixed erythema, flat or raised over the malar eminences, tending to spare the
nasolabial folds
Erythematous raised patches with adherent keratotic scaling and follicular
plugging; atrophic scarring may occur in older lesions
Skin rash as a result of unusual reaction to sunlight, by patient history or
physician observation
Oral or nasopharyngeal ulceration, usually painless, observed by a physician
Nonerosive arthritis involving two or more peripheral joints, characterized by
tenderness, swelling, or effusion
a. Pleuritis: convincing history of pleuritic pain or rub heard by a physician or
evidence of pleural effusion OR
b. Pericarditis documented by ECG or rub or evidence of pericardial effusion.
a. Persistent proteinuria greater than 0.5 gm/day or >3+ if quantitation not
performed OR
b. Cellular casts (red cell, hemoglobin, granular, tabular, or mixed)
a. Seizures in the absence of offending drugs or known metabolic
derangements, e.g., uremia, ketoacidosis, or electrolyte imbalance OR
b. Psychosis in the absence of offending drugs or known metabolic
derangements, e.g., uremia, ketoacidosis, or electrolyte imbalance
a. Hemolytic anemia with reticulocytosis OR
b. LEUKOPENIA: <4000/pl total on two or more occasions OR
c. Lymphoprnia: <1500/pJ on two or more occasions OR
d. Thrombocytopenia<100,000/pl in the absence of offending drugs
a. Positive LE cell preparation OR
b. Anti-DNA: antibody to native DNA in abnormal titer OR
c. Anti-Sm: presence of antibody to Sm nuclear antigen OR
d. False-positive serologic result for syphilis known to be positive for at least
6 months and confirmed by TPI or FTA-ABS
An abnormal titer of ANA by immunofluorescence or an equivalent assay at
any point in time and in the absence of drugs know to be associated with
drug induced lupus syndrome.
TPI = Treponema pallidum immobilization test; FTA-ABS = fluorescent treponemal antibody absorption test.
'The classification is based on 11 criteria. For the purpose of identifying patients in clinical studies, a
person shall be said to have S:E if any 4 or more of the 11 criteria are present, serially or simultaneously,
during any interval of observation. Source: EM Tan et al. The 1982 revised criteria for the classification of
systemic lupus erythematosus. Arthritis Rheum25:1271; 1982.

3
Causes of death (see tables 1-2 and 1-3) and life
expectancy (table 1-4) of patients with SLE may also be
stratified by socio-economic or environmental parameters (see
fig.1-1) [Fessel 1974; Hochberg 1993; Iseki et al 1994]
Table 1-2. Causes of death in a cohort of systemic
lupus patients.
Cause
No. (%)
Systemic lupus erythrmatosus
49 (34)
Multisystem
16
Nephritis
12
Central nervous system disease
11
Pulmonary
5
Cardiovascular
5
Infection
32 (22)
Cardiovascular disease
23 (16)
Cerebrovascular disease
8(6)
Cancer
8(6)
Iatrogenic
3(2)
Pulmonary embolus
2(1.4)
Gastrointestinal hemorrhage
2(1.4)
Aortic dissection
1 (0.7)
Radiation enteritis
1 (0.7)
Toxic epidermal necrolysis
1 (0.7)
Small bowel obstruction
1 (0.7)
Thrombotic thrombocytopenic purpura
1 (0.7)
Pacemaker malfunction
1 (0.7)
Addisonian crisis
1 (0.7)
Unknown
10(7)

4
Table 1-3. Characteristics of SLE Patients Treated
in Okinawa, Japan, between 1972 and 1991.
No. of Patients
Females
515
(91%)
Males
51
(9%)
Total
566 (100%)
je at diagnosis (yr) (mean SEM)
Females (range)
29.30.6 (5 to 77)
Males (range)
27.01.7 (5 to 63)
Total (range)
29.10.6(5to 77)
Comorbid conditions (%)*
41
(7.2)
Central Nervous System symptoms
108
(19.1)
Aseptic bone necrosis
42
(7.4)
Pericarditis
66
(11.7)
Family history
45
(8.0)
3. of deaths (%)
Females
95
(18.4)
Males
9
(17.6)
Total
104
(18.4)
auses of death (%)
Infection
25
(24.0)
Cerebrovascular disease
16
(15.4)
Uremia
12
(11.5)
Sudden death
6
(5.8)
Cardiac
17
(16.3)
Others
28
(26.9)
jpus nephritis (%)
279
(49.3)
Renal biopsy
174
(30.7)
Nephrotic syndrome
194
(34.3)
Dialysis therapy
78
(13.8)
The occurrence of a comorbid condition at any time in the patients past.

5
Table 1-
4.
Study of
the correlation
between
cause
death, sex,
race and
socio-
economic status (SES) in
patients
with SLE.
Total
SLE Infection
Cardio
Cerebro
Cancer
(n=49)
(n=32)
vascular
vascular
(n = 8)
(n=23)
(n=8)
Female
95
40 (42)
28 (29)
14 (15)
6 (6)
7 (7)
Male
25
9 (36)
4 916)
9 (36)
2 (8)
1 (4)
White
50
22 (44)
7 (14)
15 (300
2 (4)
4 (8)
Black
70
27 (39)
25 (36)
8 (11)
6 (9)
4 (6)
High(SES)
33
15 (45)
5 (15)
7 (21)
2 (6)
4 (12)
Middle(SES)
34
12 (35)
9 (26)
11 (32)
1 (3)
1 (3)
Low(SES)
25
9 (36)
7 (28)
7 (28)
1 (4)
1 94)
Year
1969-1973
43
22 (51)
10 (23)
7 (16)
1 920
3 (7)
1974-1978
52
18 (35)
15 (29)
9 (17)
6 (11)
4 (8)
1979-1983
25
9 (36)
7 (28)
7 (28)
1 (4)
1 (4)
Disease
duration
<2 yrs
35
21 (60)
8 (23)
4 (11)
2 (6)
0 (0)
2-4.9 yrs
28
14 (50)
8 (29)
3 (11)
3 (11)
0 (0)
5-9.9 yrs
31
7 (23)
8 (29)
11 (36)
1 (3)
3 (10)
10-14.9 yrs
14
3 (21)
5 (36)
3 (21)
1 (7)
2 (14)
>15 yrs
12
4 (33)
2 (17)
2 (17)
1 (8)
3 (25)
Of
P
0.16
0.02
0.13
0.54
0.03

6
r400 c
_o
j3
3
'300 o
&
c
o
'200 E
k_
o
Q.
m
O
'100 c
o
CO
>

Figure 1-1. Environmental influences may be related
to the prevalence of SLE. The figure shows the overall
annual incidence (0) and prevalence () of SLE in
Okinawa, Japan between 1972 and 1991. The rates were
expressed as the number of patients per million pop.
of the year of diagnosis.
A number of strategies have been used or proposed to
suppress autoimmune diseases, most notably drugs such as
cyclophosphamide, cyclosporin A, methotrexate and
azathioprine. Steroid compounds, such as prednisone and
methylprednisone, are also employed in many instances. These
drugs have limited long term efficacy against both cell- and
antibody-mediated autoimmunity because of their "global"

7
immunosuppression and their toxic side effects. Prolonged
treatment inhibits the normal protective immune responses to
pathogenic microorganisms, thereby increasing the incidence
of infection. A further drawback is that immune-mediated
elimination of aberrant cells is impaired and there is, thus,
an increased risk of developing malignancies in patients
receiving prolonged immunosuppressive therapy.
The potential to accomplish extracorporeal removal of
certain components from bodily fluids without the above
mentioned risks to the patient has been established using a
variety of bioadsorption devices. A number of patents have
been issued for immunoadsorbent materials which remove
immunoglobulins from plasma [Balint U.S. 4,801,449; Jones
U.S. 5,122,112]. However, there remains a need for improving
efficiency of the existing methods of extracorporeal therapy,
most of which still require plasma separation before the
target molecule can be removed [Schneider 1998]. Therefore,
chemical modifications on the surface of existing devices is
being investigated here and elsewhere. This research is
designed not only to improve the efficiency of the adsorption
process, but also to remove the target molecules from whole
blood.
1.2 Background Review
1.2.1 Clinical Implications of Middle Molecules in
Extracorporeal Therapy.
The normal human kidney is well known for its ability to
excrete water, small-molecular weight end products of
metabolism and to balance electrolytes in the blood. In
addition, it also has a role in the removal of peptides and
low-molecular weight proteins. These "middle molecules" may
be catabolized in the tubular cells after endocytotic

8
absorption from glomerular filtrate and later excreted as
smaller fragments in the urine. Furthermore, middle molecules
(MM's) accumulate in body fluids in the same way as urea and
creatinine when the glomerular filtration rate (GFR) is
decreased [Funck-Brentano et al. 1972; Bergstrom and Wehle
1994]. It is important to note, however, that the
categorization of molecules in this way is somewhat arbitrary
and that MMs in general refer to those solutes that are
poorly dialyzed through conventional cellulosic membranes.
The "large molecules" of dialysis being compounds usually
greater than 7 0kDa and which are not removed to any
significant extent across such membranes. Accordingly, MM
represent compounds having in the broadest sense molecular
weights between 1.5kDa and 55kDa.
There is now sufficient evidence that MMs are elevated
in the serum of patients with chronic renal failure and that
many of them exert toxic effects in vivo [Descamps-Latsca
1994 ]. In the first part of this research project we have
constructed an in vitro model to study the transport of a
test solute, vitamin B12, and two well known pathogenic MM
beta-2 microglobulin (p2-M) and the serum soluble
interleukin-2 receptor (CD25). The result of experiments
undertaken will demonstrate the utility of the model in
evaluating hemodialyzer performance and the study of the
effects of dialysate additives on clearance of middle
molecules. Three test membranes were used in the study; a
low-flux polysulfone hollow fiber, the F8 (Fresenius Corp),
a high-flux cellulose acetate hollow fiber the Altrex 140
(Althin Corp.), and a high-flux polymethylmethacrylate hollow
fiber, the BK2.1P (Toray Ind.).

9
1.2.2 The Effects of Dialysate Additives on the
Transport Properties of Donor Solutes across
Hollow Fiber Membranes.
In order to study the transport properties of MMs in
hollow fiber dialyzers an in vitro model has been constructed
in which equilibrium and sink conditions can be applied to
the three test solutes. Vitamin Bi2 (m.w. 1355) will serve as
a surrogate molecule for dialyzer/circuit performance. Its
high solubility in aqueous solution and its characteristic UV
absorption at 361nm have made it a very useful solute for
clearance studies in hemodialysis [Naitoh 1988].
Beta 2-microglobulin is a 12kDa non-glycosylated protein
present at the surface of nearly all nucleated cells, where
it is non-covalently bound to the heavy chain of the HLA
class I complex. The primary structure and physiochemical
characteristics of (32-M are shown in fig. 1-2.
Figure 1-2. The structure and physical properties
of beta 2-microglobulin. Adapted from Bergstrom and
Wehle (1994).

10
P2-M is markedly retained in ESRD patients, since the
kidney normally accounts for the elimination of about 95.5%
of the normal beta 2-microglobulin generated at 1000-2000
mg/week [Chanard et al. 1993]. Renal retention of P2-M
appears to be the basic requirement for the development of
dialysis-related amyloidosis. p2-M will serve as a mid-range
donor solute in this investigation.
Several cytokine receptors actually consist of two or
more separate polypeptide chains that function as a complex
in the cell membrane. For example, the high affinity IL-2
receptor contains three separate chains (a, p, and y) The
trimeric IL-2RaPy and the IL-2RPy heterodimer both bind IL-2
and mediate signal transduction. See fig. 1-3.
Intermediate lllgli affinity Low affiidty
affinity IL 2R IL2R IL-2R
Subunit
composition: 11.-211(3
IL2lly
Affinity
constant CAj): ltr9Af
IL-2lla lL-2Ra
IL-2R(I
IL2Ry
Itr'Mf l(T8Af
Cells
expressed by: NK cells
Resting T cells
Clow numbers)
Activated CD4*- and
CU8^ cells
Activated D cells
Clow numbers)
Figure 1-3. The interleukin-2 receptor complex. The
receptor is composed of three transmembrane polypeptides
called the a, Pand y subunits. Adapted from Minami (1993).

11
The IL-2Ra serves to increase receptor affinity for the
cytokine but does not contribute at all to signaling. This a-
subunit (CD25 in cluster of differentiation designation) may
be released from a T-cell surface upon activation by a
suitable antigen, mitogen or toxic event (see fig.1-4).
Figure 1-4. Structural models of the "soluble" IL-2R
alpha subunit CD25. The binding regions of the subunit
for IL-2 are shown on the right. Adapted from Miedel
et al. (1988).
Serum soluble CD25 has been shown to retain IL-2
affinity at a level (KD~18nM) sufficient to inhibit IL-2
mediated cell responses. CD25 is elevated in many disease
states, including rheumatoid arthritis, cancer, and AIDS, as
well as in patients with SLE or ongoing maintenance dialysis.
Studies show that serum CD25 levels offer a rapid, and
reliable measure of disease activity [Rubin and Nelson 1990].

12
Therefore, because of its broad clinical significance we have
selected CD25 as the upper-range solute in these mass
transfer experiments.
1.2.3 Molecular Pathogenesis of Lupus Nephritis.
When autoantibody synthesis occurs, it is usually the
result of one of the following: 1) The individual may have an
intrinsic defect in immunity that hinders its ability to
distinguish self from foreign non-self, e.g. T-helper cell
deficiency. 2) Self-antigen may be modified by infection,
injury or drugs. After such modification these self-antigens
are "denatured" or altered in such a way as to be taken as
foreign. 3) If certain self-antigens that are normally
sequestered from the immune system (e.g. cornea) become
exposed to the systemic circulation due to injury, this may
result in stimulation of an immune response and autoantibody
synthesis. A classical example of this mechanism leading to
autoimmunity is found to occur during surgical procedures on
the thyroid gland. When the normally sequestered thyroid
colloid is exposed, thyroglobulin may be released into the
blood. Autoantibodies to plasma thyroglobulin have been found
in patients after thyroid surgery [Theofilopoulos and Dixon
1982].
The development of lupus nephritis depends on the active
deposition of immune complexes onto the glomerular basement
membrane. Immune complexes characteristic of SLE circulate in
a variety of constitutive masses. The largest members of
these circulating immune complexes (ICs) are easily
phagocytized before deposition can occur and are therefore
nonpathogenic. The smallest of these ICs are not well
filtered by kidney glomeruli and thus, are easily excreted in
the urine [Carlson et al. 1988]. However, those ICs of

13
intermediate size (too large for excretion, yet too small
for phagocytosis) have an extented half-life in the
bloodstream and are found predominately in the glomerular
deposits of lupus patients with secondary nephrotic syndrome
[Wener 1986],It has been shown that histones, which are
positively-charged DNA binding proteins, may mediate the
deposition of DNA-anti-DNA immune complexes [Morioka et al.
1994]. Apparently, the histones first affix to the glomerular
basement membrane by electrostatic interactions; then
circulating preformed immune complexes bind to these
subplanted histones to form nephritic lesions between
capillary endothelia [Choi et al. 1995]. Hemodynamic forces
also play an important role in the deposition of circulating
ICs. It has been found that IC deposition occurs
predominantly in those areas of the vasculature under high
shear-stress. This is easy to understand if one considers
that in regions of hemodynamic shear-stress (notably the
kidneys, heart and lung) decreases in the length and luminal
diameter at the capillary bed cause blood flow to become more
turbulent. Blood cells (especially platelets) and endothelial
cells respond to this shear-stress by secreting vasoactive
compounds (i.e. 5-HT and histamine) and changes in morphology
[Kniker and Cochrane 1968; Herbert et al. 1978]. The
pathogenic mechanism involves 1) activation of platelets
followed by release of vasoactive amines and aggregate
formation, 2) retraction of endothelial cells and exposure of
the subendothelial basement membrane, 3) activation of
complement and inflammatory leukocytes followed by, 4) tissue
injury due to the release of proteases and reactive oxygen
species [Mahan et al. 1993; Dixon et al. 1961]; see fig.1-5.

14
I.
Immune complexes
.Proteins
Vessel
lumen
Endothelial
~ cell
Basement
membrane
Basophil
PAF
Vasoactive
amines
increased permeability
Deposition ol immune complexes
Figure 1-5. Mechanisms of Immune Complex Deposition.
Adapted from Kniker and Cochrane (1965).
1.2.4 Anti-(DNA-Histone) Antibodies in Active Lupus
Nephritis.
Histones are the principal structural protein of
eukaryotic chromosomes. They are small polypeptides with 100
to 200 amino acid residues, which do not contain tryptophan
but are rich in lysine and/or arginine. Histones may be
divided into five classes (or types): Hi, H2A, H2B, H3 and
H4. This classification, originally made on the basis of
histone electrophoretic mobility, has been maintained because
histones belonging to each of the classes have different
molecular weights and amino acid sequence. See table 1-5.

15
Table 1-5. Principal Characteristics of Calf Thymus
Histones. Adapted Rubin in Autoantibodies Peter
and Schoenfeld (eds.) Elsevier (1996).
Fraction
Mr'
Number
N-terminal
a-helix
^-pleated
1 mg/mL
Cationic
Molar staining coefficients
of
residue
content
sheet
E
charge
residues
(%)
(%)
275nm
density"
Coomassie
blue
Amido
black
H1
26.500
220
Ac-Ser
55
5
0.079
37.7
1.00
1.00
H2A
14.000
125
Ac-Ser
35
30
0.300
21.4
0.44
0.49
H2B
13.770
129
Pro
40
20
0.510
21.4
0.22
0.59
H3
15.340
135
Ala
39
15
0.300
22.2
0.65
0.78
H4
11.280
102
Ac-Ser
28
31
0.500
23.9
1.67
0.47
The features in the above table distinguish each histone
from those of another class. Two polypeptides of each of the
classes H2A, H2B, H3 and H4 constitute an octamer around
which dsDNA is wound in two turns to form the histone-DNA
complex called a "nucleosome core" (see fig.1-6).
Figure 1-6. Structure of the "Histone octamer-
Nucleosome core." Adapted from Rubin (1996).
A segment of "linker" dsDNA connects one nucleosome to the
next [Sperling and Wachtel 1981; Hnilica 1989].
A recent investigation has confirmed the presence of
anti-(DNA-histone) activity in lupus nephritis [Suenaga and

16
Abdou 1996]. In their study it was found that anti-(DNA-
histone) antibodies specifically recognized DNA-histone
complexes and largely overlapped with antinucleosome
antibodies. While both DNA and histone were required for full
recognition by the antibodies, it is unclear whether DNA
itself is part of the nucleosome specific epitopes or simply
stabilizes the histones and maintains the nucleosomal
structure of the core to form such epitopes. This may be
especially relevant since histones are known to spontaneously
form octamers in physiological pH buffers [Muller 1985;
Burlingame et al. 1985]. Determination of the precise binding
sites, however, must await more sequence information of the
variable regions of both heavy and light chains of fully
characterized antinucleosome antibodies. For a summary of the
prevalence of AHA in SLE, see table 1-6.
Table 1-6. A Summary of the Prevalence of Histone
Autoantibodies in SLE patients. Adapted from Rubin
(1996).
Method
FIAa
ELISA3
ELISA3
IBb
ELISAb
ELISAb
ELISAb
ELISAb
Number of sera
82
39
151
127
32
46
12
40
Histones:
%
%
%
%
%
%
%
%
H1
62
95
52
98
6
59
67
60
H2A
43
54
23
50
34
nd
42
55
H2B
68
79
21
98
63
72
25
55
H3
57
44
47
63
56
59
33
35
H4
54
nd
21
53
22
nd
42
38
ELISA, enzyme-linked immunosorbent assay; FIA. fluorimetrc immonoassay; IB. Immunoblotting; nd, not
determined; a. test of IgG + IgM autoantibodies; b. test of IgG antibodies only.
Anti-(DNA-histone) antibodies are clinically important
because certain subsets of the antibodies are potentially
pathologic for tissue injury such as nephritis.
Antinucleosome activity has been detected in renal eluates

17
from nephritic MRL mice, and a correlation between anti-(H2A-
H2B)-DNA and kidney disease defined by proteinuria has been
described [Burlingame et al. 1994]. In that study it was
found that anti-(DNA-histone) antibodies react with
nucleosomes released from apoptotic mononuclear cells,
forming complexes in patients with circulating nucleosomes.
These complexes may specifically deposit in the renal
glomerular basement membrane by histone-heparan sulphate
mediated mechanisms [Tremaat 1992]. They also may directly
deposit by recognizing DNA-histone/collagen type IV clusters
on the glomerular basement membrane [Morioka et al. 1996];
see fig. 1-7.
blood vessel
blood How
immune
complexes
vessel lumen -f1 -4:
quiescenl
blood vessel under
hemodynamic sheai-stress
blood llow
lammai
immune
complexes -4 -4
deposil
immune complexes deposil onto
exposed basemenl membrane al
ine capillary beds ol Hie glomeruli
activated platelets
5-H1, histamine
Figure 1-7 Proposed Mechanism of Histone-Mediated
Glomeruli Deposition of Immune Complexes.
1.2.5 Surface Modification of PMMA and PSF Dialysis
Membranes.
The preparation of novel affinity adsorption devices has
increased recently due to the great demands of molecular
medicine and biotechnology for efficient yet selective

18
isolation of proteins from complex mixtures. A number of
reports describing various types of microporous membranes
modified with selective-binding ligands continue to appear in
scientific journals.
Conformational factors still play a major role in
binding of the target molecule to the adsorbent. For example,
the orientation of antibody molecules on a surface is often
responsible for differences between their theoretical binding
activity (2:1) and the experimentally determined binding
ratio. This occurs because the antibody conformations are
randomized during the coupling procedure; as a result, the
paratope of some of them becomes inaccessible to the antigen
upon immobilization. This random orientation of immobilized
ligands may be kept to a minimum by the appropriate use of
bifunctional spacers such as gluteraldehyde or polyethylene
glycol [Malmsten et al. 1996]. It has also been shown that
the optimum amount of ligand to be immobilized on an
adsorbent is primarly determined by the molecular size of the
target molecule [Vallar et al. 1995]. For the capture of
large molecules like immunoglobulins, the immobilized ligands
need to have a maximum degree of freedom; this again may be
accomplished by tethering the ligand (antigen) at the end of
a long flexible polymer such as PEG. The ligand is now in a
quasi-fluid state and able to avoid more of the steric
binding impedances between it and the target molecule.
The grafting of (PEG) to solid surfaces is a very useful
technique for obtaining low protein adsorption and low cell
adhesion characteristics. For instance, PEG coating is
reported to give a marked suppression of plasma protein
adsorption and platelet adhesion leading to reduced risk of
thrombus formation. The inert characteristics of PEG surfaces

19
are due to the solution properties of the polymer, its
conformation in aqueous environments and to its electrical
neutrality, its structural similarity with H20 and strong
hydrogen bonding ether linkages provide PEG with very high
water solubility. PEG'S ability to prevent proteins from
approaching the bonded surface is due to both elastic and
osmotic steric stabilization effects. The elastic stabilizing
effect occurs when a protein approaches the PEG-modified
surface and a repulsive force develops. This is due to a loss
in conformational entropy as the polymer chains experience a
reduction in available volume. The osmotic interactions arise
from the increase polymer concentration on compression toward
the surface. When the local concentration of PEG segments
increase osmotic repulsive forces develop. The intensity of
this force depends on compression at high grafting density
but more on interpenetration at lower graft densities.
For a number of applications attachment of proteins or
other biomolecules to PEG-grafted surfaces is of interest.
Implants and artificial organs are rendered more
biocompatible. Strong interactions between the underlying
surface and the immobilized molecule can be avoided as can
nonspecific adsorption of molecules from solution. Provided
that the immobilization procedure used gives minimal
distortion of the biological properties and that the layer of
immobilized molecules is not so dense that crowding phenomena
appear, this method of surface modification can be expected
to lead to stable immunoadsorbents of high affinity. For
details on the synthesis of Star-PEG see appendix V.

20
1.2.6 Surface-Modified Polymethylmethacrylate (PMMA).
PMMA is an amorphous polymer of moderate Tg (105 C),
that has a high transparency consistent with its use in the
formulation of contact lenses. It is highly resistance to
strong acids and environmental deterioration. The solubility
parameter for PMMA is between 11.1-11.3 (cal cm'3)1/2; it is
very soluble in dichloromethane or isopropyl alcohol.
The synthesis of PMMA monomers is based on cyano-hydrin
formation with acetone to prepare the desired methacrylate
ester. See fig. 1-8.
O + HCN
H3C
H3^ /
ch3
HO
O^N
h2so4^
ROH
Figure 1-8. PMMA monomer synthesis.
These esters may then be polymerized by free radical
mechanisms initiated by heat or light.

21
Techniques used to prepare microporous PMMA membranes
are summarized in the table 1-7 below.
Table 1-7. Methods Used to Prepare Microporous Membrane
Materials. Adapted from Fried Polymer Science and
Technology. Prentice Hall p.450 (1995).
Method
Description
Phase Inversion
Phase separation of a ternary mixture of
polymer, solvent, and nonsolvent
Sintering
Melting of a semicrystalline polymer powder
Track etching
Irradiation of polymer films resulting in
the production of fission fragments
followed by caustic etching
Stretching
Combined stretching and annealing of
extruded semicrystalline film
Leaching
Extraction of solid pore formers
Thermally induced
phase separation
Cooling a mixture of a polymer with a
latent solvent to a point of thermal
separation of the mixture followed by
extraction of the latent-solvent phase
The process known as nucleation track-etching is
predominately used for the production of hollow fiber
membranes. Typical pore diameters obtained by this method are
in the range of 0.02 to 20.0|im. Hollow fiber modules for the
separation of complex mixtures typically have outer diameters
(o.d. ) of from 80-200 (im: with wall thicknesses of 20 urn or
greater. Seals for the bundles are usually made of epoxy;
with polycarbonate being the most common housing material.
PMMA is a highly biocompatible membrane material; its
use in hemodialysis began shortly after the discovery if its
ability to bind beta-2 microglobulin (Klinke et al. 1989;

22
Bonomini et al. 1996). For the immobilization of proteins,
PMMA can be activated with hydrazine-monohydrate (NH2NH2) The
resulting aminated surface can then be hydrophilized by
crosslinkage with a tresylated Star-PEG polymer in such a way
that the terminal reactive end-groups are stable enough for
subsequent binding (immobilization) of proteins or other
amines. See fig. 1-9.
slop 1 membrane activation
NH2NH2-H2O
-CONHNH2
COOCH3

50% MeOH-H20
OSO2CH2CF3
step 2 membrane hydrophillzation
SO2CH2CF3
SO2CH2CF3
CONHNH2
tresyl-Star PEG
NH-CH2CH2-O-
50% Me0H-H20
CF3CH2O2SO
OSO2CH2CF3
step 3 antigen immobilization
NH-CH2CH
cf3ch?o?so
CF3CH2C
OS02CH2CF3
S02CH2CF3
histone ,
DOC
PBS
histone
SO2CH2CF3
Figure 1-9. Surface-modified PMMA.

23
Figure 1-10. Formation of Potentially Toxic Hydrazine
Derivatives. Shown above is a) Formation of Hydrazine
Derivatives with Nucleic Acids. Adapted from P. David
Josephy Molecular Toxicology Oxford University
Press p.33 (1997) .
Precautions are warranted in the use of hydrazine or its
derivatives. While hydrazine itself reacts with pyrimidine
bases of nucleic acids to form aminopyrazoles (see fig. l-10a
previous page), most of the free hydrazine reacts with other
molecules extracellularly before reaching the cell interior
or nucleus. The many of these hydrazine derivatives are
single H-substitutions by an aliphatic carbon (see fig.l-
10b).

24
NH2NH2 is a relatively weak nucleophile, but it is still very
reactive with Io and 2 alkyl halides. Acylhydrazines can also
be formed by the reaction of hydrazine and most organic
esters. See fig.l-10b.
R.
R,
"single-H transfer" ACYLHYDRAZINE
ACYLHYDRAZINE N,N' DIACYLHYDRAZINE
Figure 1-10. Formation of Potentially Toxic Hydrazine
Derivatives. Shown above is b) Formation of radical
stabilizing diacylhydrazines.

25
Disubstituted hydrazine derivatives may undergo certain
redox reactions that result in the formation of nitrogen free
radicals. See fig. l-10b. It is these "stabilized" free
radicals that are most likely associated with hydrazine
toxicity [Amdur, Doull, and Klaassen 1991; Purdy 1996].
1.2.7 Surface-Modified Polysulfone (PSF).
The polysulfones are a class of thermoplastics with high
thermal, oxidative and hydrolytic stability. They are usually
synthesized by condensation polymerization involving
substitution of the alkaline salt of bisphenoates with
activated aromatic dihalides. The synthesis of bisphenol-A
PSF is shown in fig. 1-11.
rt
sodium bisphcnolate
dichlorodipbenylsulfone
Figure 1-11. Condensation polymerization for PSF.
PSF has very high resistance to aqueous mineral acids,
salt solutions, oils, and greases. Its high biocompatibility
and ability to be sterilized by several different techniques
make them exceptionally suitable for medical applications.
PSF membranes have both high permeability and permselectivity
and are also fabricated into hollow fibers by nucleation
track-etching.

26
PSF is currently the most successful porous membrane
material for medical applications and the separation of
gases. However, limited use has been made of this polymer as
a substrate for surface-modifications or for the
immobilization of proteins. A few reports of PSF activation
with chlorosulfonic acid followed by crosslinking with a
diamine or diacyl compound for the subsequent binding of
protein have been reported [Klein et al. 1994; Pozniak et al.
1995, Fernando-Salznero et al. 1991], but some of these
methods are carried out in an organic medium, such that the
activation occurs in the slurry rather than on the surface of
an intact membrane or hollow fiber. We, therefore, used an
aqueous phase nitration method to activate intact PSF hollow
fibers which can then be reduced to the corresponding amine
in the presence of a low reduction potential trivalent metal
(i.e. CrCl3). The aminated PSF fashioned in this way can then
be hydrophilized by a tresylated Star-PEG for the subsequent
binding of proteins in a manner similar to the PMMA membranes
described earlier. See fig. 1-12.
1.2.8 Immunoreactivity of Immobilized Histones
Characterized using the BIACORE 2000.
The biospecific interaction between antihistone
antibodies and histones immobilized via PEG Star-polymers was
monitored using the BIACORE 2000. A multistep binding
experiment was conducted where tresylated Star-PEG polymers
were attached to the BIACORE sensorchip in a manner similar
to that used in the modification of PMMA membranes. The
instrument measures in real-time the binding kinetics at the
flow cell surface and then calculates a corresponding
sensorgram which quantifies the recognition process.

27
step 1 membrane activation
hno3\h2oso3
~(OC6H4SC>2C6H4)n
5.0% AcO /H20
CrCI2\HCI
(0C6H4S02C6N02H3)r>
step 2 membrane hydrophilization
tresyl-Star PEG
-(0C6H4S02C6NH2H3)n
50% MeOH-H20
CF3CH2O2S
CF3CH2O2SO
-PSFNH-CH2CH2O-
OSO2CH2CF3
J /OS02CH2CF3
OSO2CH2CF3
step 3 histone immobilization
CF3CH2O2SO OSO2CH2CF3
CF3CH2O2SO \ J >^0S02CH2CF3
-PSFNFFCH2CH2O-.
histone
OSO2CH2CF3
PBS
histon
histone'
histone
histone
histone
histone
Figure 1-12. Surface-modified PSF

28
The BIACORE 2000 operates on the principle of surface
plasmon resonance (SPR) and is designed to detect changes in
the refractive index of the layer of solution in contact with
the sensorchip. Since the refractive index of the medium is
affected by the surface concentration of solutes, monitoring
the SPR angle provides a real-time measurement of changes in
surface concentration and/or coupling.
Units used to express the SPR signal are called response
units (RU). One thousand RU are equivalent to a change of
about 1 ng/mm2 in surface protein concentration or
approximately 6 mg/ml in bulk protein concentration for most
proteins.
1.2.9 Surface Analysis of Surface-Modified Hollow
Fibers.
In order to achieve a more complete understanding of the
molecular-interactions between foreign materials and
biological systems, surface properties and surface structure
must be properly elucidated. The surface structure of a
material is dynamic and often bioreactive, especially in
cases in which components of the surface reconstruct
themselves in response to a local environment.
For these reasons, highly sensitive analytical methods
are required to provide a means to these data. See table 1-8
and fig. 1-13.

29
Table 1-8. Common Surface Analytical Methods.
Adapted from Ratner (ed.) Surface Characterization
of Biomaterials. Elsevier, Amsterdam (1988).
Method
Depth Analyzed
Spatial
resolution
Analytical sensitivity
Contact angles
3-20
1 mm2
Low or high depending on
the chemistry
Scanning force
microscopy (SEM)
5
1
Single atoms
Scanning electron
microscopy (SEM)
5
40 typically
High; not quantitative
Electron spectrocsopy
for chemical analysis
(ESCA)
10-250
8-150 pm
0.1 atomic %
Secondary ion mass
spectrometry (SIMS)
10 to 1 pm6
500
Very high
Attenuated total
reflection infared
(ATR-IR) spectroscopy
1-5 pm
'
10 pm
1 mol %
"The size of a small drop is 1 mm. However, contact angles actually probe the inter-
facial line at the edge of the drop. The spatial resolution of this zone might be approx
imately 0.1 pm
Static SIMS -10 ; dynamic SIMS to 1 pm

30
useful for examining the solid-aqueous interface
Figure 1-13. Flowchart for the Surface
Characterization of Biomaterials. Adapted from Ratner
(ed.) Surface Characterization of Biomaterials.
Elsevier, Amsterdam (1988).

31
The simplest picture of a sample surface is provided by
the optical microscope. The optical microscopy may be
extended by the use of specialized methods which include
polarized microscopy, fluorescence microscopy or confocal
microscopy. A more detailed picture of the surface is
provided by the scanning electron microscope (SEM) which
produces surface images of greater resolution and depth of
field than the above mentioned techniques. However, the major
disadvantage of the SEM is that nonconductive materials have
to be "sputter-coated" before images can be made.
Advanced imaging of surfaces at the molecular level is
now obtained with the so called scanning probe microscopes
(SPM). The atomic force microscope (AFM) is such a scanning
probe that does not require the subject surface to be
electrically conductive, since it merely measures the
interaction between a microfabricated tip mounted on a small
cantilever spring and the atoms on the surface of the sample.
Scanning the probe tip over soft (compliant) surfaces may
induce deformations in the image. Therefore, to avoid damage
to such samples, the interaction force may be kept below 8-
10N by operating the AFM in "non-contact mode" [Porter, TL.,
Sykes, AG and Caple, G 1994]; see appendix VI.
1.2.10 Extracorporeal Immunoadsorption Therapies in
SLE.
The selective extraction of constituents from plasma is
based on the principle of chromatographic affinity. A
substance (ligand or sorbent) with specific affinity may be
fixed to an insoluble, inert or "passive" matrix with the aim
of selecting and binding its complementary substance. The
ligand (the active part of the column) may vary broadly; it
can be a chemical compound like heparin, carbon, dextran

32
sulfate, or a protein i.e. an enzyme, antigen, antibody and
so on. Selective or semi-selective adsorption may occur
through physiochemical interactions (i.e. hydrophobic or
ionic bonding) or by biological interactions (e.g. the
interactions between enzyme-substrate, Ag-Ab molecular
recognition, complement fixation, or immunoglobulin Fc-lectin
interactions).
The binding capacity of affinity adsorbents is perhaps
the most important parameter in immunoadsorption therapy. The
capacity of a selective adsorbent is determined principally
by two interdependent sets of conditions: (1) the correct
choice of the matrix, spacer molecule and ligand, in order to
optimize the enzyme-ligand interaction and (2) the way in
which the capacity is affected by such dynamic factors as
flow rate, equilibration time and adsorption technique.
Assuming the design of the adsorbents has been
optimized, the operational adsorption capacity of an affinity
gel is best determined by frontal analysis. Here, a given
concentration of the complementary protein (C0) is applied to
the adsorbent continuously and then its emergence monitored.
As the adsorbent material becomes saturated with the
adsorbate, the solution breaks through at the same
concentration it had on entering the column (see fig. 1-14).
The volume of eluant that appears up to the "step", where the
concentration of the complemnetary protein increases rapidly
to Co over a small volume, is called the retention volume
(V.). It comprises the interstitial volume (VD) and the volume
solution from which the adsorbate was removed (V), i.e.
Ve
Vo + V

33
C0
Cone. in
eluate (CQ)
Vol.(mL)
Figure 1-14. Determination of the capacity of an
affinity adsorbent by frontal analysis. Adopted from
Lowe and Dean. Affinity Chromatography. John Wiley
and Sons Ltd., London (1974).
If m is the total weight of the affinity adsorbent in
grams, then the capacity of the gel, i.e. the amount of
adsorbate specifically adsorded per gram, Q, is
Q = (V/m)C0 = (V.- V/m) C0
and the total amount adsorded by the bed is q = Q xm, i.e.
VCo.
The capacity, consistent with the emergence point of the
monitored species, is dependent on the rate of application of

34
the original sample. At relatively high sample flow rates,
affinity equilibrium is not attained and premature emergence
of adsorbate may be observed. This will lead to a under
estimation of the operational capacity of the adsorbent. The
effective capacity of an adsorbent, however, may be deduced
by incubating a known weight of adsorbent (m) with a given
volume of concentration CQ and subsequently, after equilibrium
has been established, measuring the new lower concentration
C. The capacity, Q, is then calculated from Q. = (C-C)/m.
Currently, few data is available to estimate the
capacity theoretically from a known immobilized ligand
concentration and other parametrers of the system. It appears
that the effective capacity of an adsorbent is considerably
lower, in fact often <1%, of the theoretical capacity based
on the ligand concentration [Lowe, CR et al., 1974; Harvey,
MJ 1974]. Presumably, the effective capacity of a specific
adsorbent is determined by the concentration of immobilized
ligand that is freely available for interaction with the
complementary protein. We have used a similar incubation
experiment to determine the "effective adsorption capacity"
of surface-modified membranes (PSF and PMMA) as will be
described in section 2.31. The results will help us determine
if therapeutically significant amounts of autoantibodies can
be removed by surface-modified commercial dialyzers.
SLE is fundamentally characterized by the synthesis of
numerous antibodies; at least 10 antigen/antibody systems
have been detected and are present in at least 10% of all
patients studied see table 1-9.

35
Table 1-9. Autoantibodies in SLE. Adapted from
Tan et al. (1982) .
Target
Clinical Associations
dsDNA
High diagnostic specificity
Correlation with disease activity (especially renal
disease)
ssDNA
Low diagnostic specificity
Histones (H1, H2A, H2B, H3, H4)
SLE and drug-induced lupus
Sm (SnRNP core proteins B, B1,
D. E)
High diagnostic specificity
No correlation with disease activity
U1-RNP (snRNP-specific
proteins A, C, 70 kd)
Mixed connective tissue disease or overlap
syndrome (when not accompanied by anti-Sm
antibodies)
Ro/SS-A (60-kd and 52-kd
proteins)
Neonatal lupus (with anti-SS-B/La)
Photosensitivity
Subacute cutaneous lupus
La/SS-B (48-kd protein)
Neonatal lupus (with anti-SS-A/Ro)
Associated Sjogren's syndrome
Ku
Diagnostic specificity for SLE and related overlap
syndromes
Proliferating cell nuclear
antigen (PCNAO/cyclin
High diagnostic specificity
Ribosomal P
High diagnostic specificity
Cytoplasmic staining
Psychiatric disease
Phospholipids
Inhibition of In vitro coagulation tests (lupus
anticoagulant")
Thrombosis
Recurrent abortions/fetal wasting
Neurologic disease (focal presentations)
Thrombocytopenia
Cell surface antigens
Red blood cells
Platelets
Neuronal cells
Hemolytic anemia
Thrombocytopenia
Neurologic disease (diffuse presentations)
Tissue and organ damage mostly derives from the
formation and deposition of immune complexes (ICs),
especially in the vascular endothelium. The logic behind the
use of plasma exchange or adsorption as a support to drug

36
therapy is obvious; removal of the autoantibodies before they
bind with antigen and are deposited avoids tissue damage.
This apparently logical consideration is not aways evaluated
correctly and is often not withstanding in even the most
painstaking therapies. Many arguments have been presented
both for and against plasmapheresis in SLE since its first
application in 1976 by the Lockwood and Jones Groups. In
their groundbreaking study [Jones et al 1976] they showed
that therapeutic levels of anti-DNA antibodies could be
removed by plasmapheresis. Even though no adsorbent mechanism
was technically described the claims that immune complexes
were removed were based on reductions in serum anti-DNA
antibodies determined by a Clq precipitation assay. In
addition, an apparent reduction in complement fixation was
measured by RID of C3 and C4 levels in each of the 8 patients
of the study.
Early in the practice of therapeutic apheresis it was
found that the equilibrium between the cell product
(antibody) and the producing cell itself needs to be taken
into account. The term "rebound" is often used to express the
resynthesis of the antibody after its removal by
plasmapheresis. This resynthesis sometimes exceeds basal
levels and is cause for clinical discretion when apheresis
may be applied. The effects of extracorporeal removal of IgG
from rabbits was studied [Carlton et al 1983 ] and the
findings infer that 1) selective removal of identified
pathologic factors is more efficient and potentially more
cost effective than non-specific plasma exchange and 2)
selective removal from whole blood has no required
replacement fluid or associated dangers and inadequacies.
Apart from the limitations imposed by non-selective plasma

37
exchange therapy it is also hindered by the lack of accurate
kinetic models which are required to enable an objective
approach to the treatment. In their model they used a hollow
fiber plasma separator in-line with a protein A adsorbent
column. What they found was that the system was effective in
lowering circulating IgGs in rabbitts by 60% in 30-40 mins.
The adsorption was also adequately modeled using first-order
kinetics, which indicates that intrabody transfer and
endogenous regeneration of the IgG were negligible during the
treatment period. This may not be true in non-healthy rabbits
or in patients with SLE. IgG transfer rate from the EVS to
the IVS is normally three times the rate in the opposite
direction. This is because the EVS to IVS transfer occurs
predominately by convection transport through the lymphatic
system. Following removal of substantial IgG from the IVS,
mobilization of the EV IgG occurs, resulting in a rebound
increase in plasma levels that usually reaches a peak within
24 hrs before a substantial decline.
The rise in plasma IgG level after plasmapheresis may be
as much as 50% due to IgG resynthesis, thus immunosuppressant
therapy is warranted in most cases [Higgins 1995]. In
addition, IgG catabolic rates may fall by 30% as a result of
IgG extraction, thus even in the presence of suppressant
therapy, increases of up to 50% post-apheresis may occur in
an individual.
Even with these forseen difficulties in the early
1980's, investigators continued to design plasma adsorption
methods for the treatment of SLE and other immune complex
(IC) diseases. Most of the earlier immunoadsorbent devices
suffered from two major limitations. Firstly, the plasma
exchange systems were technically not suitable for routine

38
clinical application due to their mechanical complexity
and/or the poor histocompatibility of the sorbent-materials.
This would often lead to unwanted clotting of the blood or
leakage. Secondly, the adsorption capacity of the earlier
designs were either insufficient or fixed. Therefore,
increased surface areas on highly biocompatible, yet specific
adsorbents were being proposed. This objective remains the
most important goal behind current blood purification
research. One group (El-Habib et al. 1984) has successfully
fabricated flat-sheets of chemically-modified collagen onto
which proteins or DNA can be attached under physiologic
conditions. These antigen-coated sheets were then set up in
hemodialyzer frames (RPS Hospal) and sterilized by gamma-
irradiation at 2-5 Mrad. The devices were used to treat 5
SLE-patients after plasma separation which resulted in a 3.7
g/L reduction in IgG after 180 mins, of treatment. Again
differences in the amount of anti-DNA removed were estimated
when different assay methods were used. Bratt and Ohlson
(1988) have developed an immune complex adsorbent material
based on covalently immobilized anti-Clg on Sepharose 4B.
Onto the anti-Clq column, Clq was non-covalently attached and
plasma perfused through such a device was cleared of 1.0 mg
IgG/ml gel. It was estimated that in order to remove all the
Clq-reactive material in the blood of one patient a column
containing only 100-180 ml of the Clq-anti-Clq gel would be
needed. It has been demonstrated in yet another study that
Clq immobilized onto Sepharose 4B typically extracts lmg of
immune complex per ml of the treated gel [Hiepe et al. 1990].
However, this impressive adsorption capacity is not
independent of immune complex concentration and the
efficiency of the separation rapidly decreases with serum

39
content. Furthermore, platelets expressing the Clq receptor
may be activated by the immobilized Clq or soluble Clq
complexes [Casali 1979]. Anti-DNA antibodies have been shown
to adsorb to other materials as well, these include
polyanionic dextran sulfate using both single [Aotsuka et al.
1990] and tandem columns [Suzuki, K. et al 1991]; protein A
[Bygren 1985; Palmer et al. 1991], tryptophan or
phenylalanine adsorbent materials [Yamazaki, Z et al 1989],
DNA on carbonized resin beads [Gao et al. 1995] and DNA
cellulose [Susuki et al. 1994].
A report of a surface-modified hollow fiber for the
removal of human igG from human plasma or serum has been
recently published [Bueno et al. 1995]. The authors studied a
pseudobiospecific affinity ligand, L-histidine, immobilized
through an ether linkage onto a PEVA hollow fiber cartridge.
The membranes effectively adsorbed igGi, IgG2, and IgG3 if MOPS
buffer is used, but were more selective toward IgGi and IgG3
in Tris-HCl buffer. The ligand also showed a higher capacity
than protein A-membranes and may offer a low cost alternative
to protein A based devices. While a K of (10' 5) was reported
for the igG-histidine complex on PEVA a lower KD of(5.0*10"7)
has been reported for protein A immobilized onto a PSF
modified hollow fiber [Klein et al. 1994].
1.3 Hypothesis
Therapeutically modified hollow fiber dialyzers can be
used to remove undesirable antibodies from plasma in the
treatment of various disease states. My hypothesis is that
the extraction of immune complexes and other blood borne
molecules by immobilized ligands against them, can be
demonstrated in an in vitro model using therapeutically

40
modified hollow-fiber dialyzers. The findings of these
studies may be used to optimize extracorporeal therapy of
patients with a variety of immune or metabolic diseases.
1.4 Specific Aims and Rationale
The purpose of this research is to demonstrate the
potential use of therapeutically-modified hollow fiber
dialyzers for the ex vivo extraction of autoantibodies. We
have selected the extraction of antihistone antibodies of the
type found in SLE as one of many possible experimental
models. Our approach will be to immobilize whole histones on
the surface of commercially available dialyzers in such a way
that solutions containing antihistone antibodies can be
cleared of substantial amounts of it by perfusion of the
solution through the device within 4hr hours. To demonstrate
the efficacy of these modified dialyzers, we defined three
specific aims which must be met to support our hypothesis.
In specific aim #1 an in vitro model of an
extracorporeal circuit was built to evaluate the ability of
dialysate additives or dialyzer surface ligands to
selectively remove donor solutes from a series of complex
mixtures. The model used is a simple two compartment model
constructed from two peristaltic pumps, hemodialysis blood
line tubing and dialyzer housing frames. It will allow us to
study transport properties of a given donor solute onto or
across the hollow fiber surface under conditions which mimic
those during a patient extracorporeal therapy session.
Specific aim #2 of the research will be to develop
coupling reactions that immobilize ligands (histones) onto
the hollow fiber surface, to characterize these surfaces, and
to determine that the immobilized ligand is still recognized

41
by the antibody. The characterization of the treated surfaces
will be carried out by scanning electron microscopy of both
chemically modified and unmodified surfaces. The atomic force
microscope (AFM) has the ability to image these surfaces
without application of a conductive metal-coating to the
sample; AFM images were obtained as well. Binding of AHA to
the modified surfaces will be tested in a series of
adsorption capacity experiments. From these studies the
optimal activation site density and AHA extraction per area
of treated membrane will be determined.
In specific aim #3 AHA immunoadsorption was tested.
Therapeutically-modified dialyzers (PMMA and PSF) were placed
in the in vitro model and perfusion with saline or plasma
containing antihistone antibodies. The extracorporeal circuit
was sampled for 4hrs, then assayed for AHA concentration in
the donor compartment. The rate of and total AHA clearance
from the donor reservoir was then calculated and averaged for
each dialyzer in the study.

CHAPTER 2
MATERIALS AND METHODS
2.1 Specific Aim #1
2.1.1 Physical characteristics and specifications of
the model
An extracorporeal circuit was built from conventional
dialysis blood line tubing and two peristaltic (Masterflex
Barnant Co. Barrington, Ill.) pumps. Two CF23 (Baxter Corp.)
dialyzers were fitted so that the extracapillary space now
serves as a 230ml reservoir for each of the donor and
receiver compartments in the mass transfer experiments. Two
F8 dialyzers were opened and the hollow fibers removed so
that the housing capsule now serves as a 475mL reservoir for
each of the donor and receiver compartments in the
immunoadsorption experiments. The physical characteristics
and specifications of the model are shown in fig. 2-1. Hollow
fiber devices have been well developed and produced
commercially due to their application in hemodialysis for the
past 40 years. Most hollow fiber dialyzers currently being
used have structure and composition similar to that shown in
fig. 2-2.
42

43
DONOR RESERVOIR
RECEIVER
RESERVOIR
bypass
oulle! sampling
port
1
inlet sampling
a
'
pon
-
1r
inlet sampling
a
C port
outlet sampling
port
bypass
**4
BLOOD PUMP
DIALYSATE PUMP
VOLUME OF EACH TUBING SEGMENT
DONORCIRCUIT
RECEIVER CIRCUIT
LENGTH(IN) DIAMETER (IN)
VOL (IN3)
LENGTH (IN)
DIAMETER(IN)
VOL (IN3)
ab=33.5 0.125
0 41
aV=21 0
0.25
1 00
bc= 74.5 0.25
3.65
b c'=27 0
0 125
0.33
de= 18 5 0.25
0.91
d'e =5 l 0
0.25
2 49
eb= 18 0 0.25
0.88
e b'= 16 0
0 25
0 78
el= 18 0 0 125
0 22
e r= 16 0
0 25
0 78
donor reservoir
475ml
receiver reservoir
475ml
DONOR BLOOD LINES
= 99 6 ml
donor bypass
= 14 4 ml
RECEIVER BLOOD LINES
= 88.2 ml
receiver bypass
= 12.8 ml
187 8 ml
with 60ml dialyzer
60 ml
two resevoirs
950 ml
total circuit volume
1197 2 ml
MATERIAL COMPOSITION
dialyzer housing
polycarbonate
blood line tubing
polyvinyl
Figure 2-1. In Vitro Model Specifications

44
Blood Out
Figure 2-2. Structure of commercial hollow fiber
dialyzers. Adapted from Colton and Lowrie (1981).

45
Three test membranes were used in the study; a low-flux
polysulfone hollow fiber, the F8 (Fresenius Corp), a high-
flux cellulose acetate hollow fiber the Altrex 140 (Althin
Corp.)/ and a high-flux polymethylmethacrylate hollow fiber,
the BK2.1P (Toray Co.). Dialyzer specifications provided by
the manufacturer are listed in table 2-1.
Table 2-1. Dialyzer Specifications.
F8
Altrex
BK2.1P/B2-1.0H
Surface Area m2
1.8
1.4
2.1/1.0
Material
PSF
CDA
PMMA/PMMA
Ku£ (mL/hr)
4.17
17.0
11.3
Blood Vol.(mL)
120
76
126
Clcr( mL/min)
175
153
181
ClUr (mL/min)
192
182
195
C1bi2 (mL/min)
76
103
127
Flow rates
and flow volumes typically
used in the :
vitro model are
listed in
table 2-2.
Table 2-2.
Circuit Specifications.
Flow rates
Qb =
65mL/min
Qd = 55mL/min
Volumes
Vdon
230mL
Vrec = 200mL
The reservoirs were connected by plastic (polyvinyl)
tubing with Hanson type connectors at one end and Luer type
connectors at the other. For tubing selection and
compatibility tables see appendix II. The H-shaped circuit
allows the intracapillary space to be perfused with a donor
compartment fluid (e.g. blood) while the extracapillary space
is perfused by a receiver compartment fluid (e.g. dialysate).

46
The center section of tubing in each compartment serves
as a circuit bypass so that the reservoir fluids may
equilibrate prior to exposure to the test hollow fiber
dialyzer, or for dialyzer exchange.
2.1.2 Mass Transport Experiments.
Materials: Vitamin Bn was obtained from (Sigma Chemical Co.)
anti-Vitamin B12 monoclonal antibodies (Sigma-Aldrich Co.).
The (32-microglobulin and anti-P2-microglobulin polyclonal
antibodies were also obtained from Sigma Chemical Co. SIL2R
(CD25) and anti-SIL2R monoclonal antibodies were obtained
from R&D Systems
Methods: The clearance of each of three solutes was tested
in three different hollow fiber dialyzers in the presence and
absence of dialysate additives (antibodies) which form stable
complexes with each of them. At the beginning of each
experiment the circuit was purged with 0.1% BSA in PBS and
the donor compartment was put into bypass.
A known amount of donor solute was then injected into
the donor circuit and allowed to mix (in bypass) for 5 mins.,
after which time the circuit was taken out of bypass and
opened to one of the test dialyzers. Donor and receiver
compartment samples were then taken at 5, 10, 30, 60, 120
minute intervals from the time the donor circuit was opened.
After the 120 min. sample the donor circuit was again put
into bypass and antibodies against the donor solute were
injected into the receiver compartment. After 5 minutes of
mixing the donor circuit was again taken out of bypass and
donor compartment samples were taken at 5, 10, 30, 60, 120,
240 min. after the donor circuit was reopened.
The vitamin B12 samples were analyzed via
spectrophotometric assay at 361nm. The (32-microglobulin

47
samples were analyzed via ELISA (Elias Co); the CD25 samples
were also analyzed via ELISA (R&D Systems).
For a summary of the experimental antigen-antibody
concentrations used see table 2-3 below.
Table 2-3. Antigen/Antibody Concentrations in
the Mass Transfer Experiments.
Vitamin B (32-M CD25
antibody 400pg 1060¡j.g 333(ig
antigen 504mg 16.7[ig 1.67|ig
2.1.3 UV Spectrophotometric Analysis of Vit.B12.
The transport of Vit.Bi2 across hollow fiber dialyzers
was studied in the presence and absence of anti-Vit.B12
antibodies as dialysate additives. The study was performed in
an in vitro model (see fig.2-1), both donor and receiver
compartment Vit.B12 concentrations were monitored
spectrophotometrically at 361nm.
Method:
1.Prepare lOOmL stock solution of vitamin B12 in DI H20
1.000g/L
2.Prepare nine standards by serial dilution.
5.0 mg/lOOmL 2.8 mg/lOOmL 2.5 mg/lOOmL
1.00 mg/lOOmL 0.5 mg/lOOmL 0.1 mg/lOOmL
0.05 mg/lOOmL 0.01 mg/lOOmL 0.005 mg/lOOmL
3. Obtain an absorbance spectrum for each of the standards
and plot the calibration curve (see fig. 2-3). For the least
squares regression method used see appendix III.

48
Standard Calibration Curve
Vitamin Absorbance BI2 at 361nm
Effluent Concentration Vitamin B12 mg/lOOmL
Figure 2-3 Calibration curve and linear regression
equation for the spectrophotometric determination of
Vitamin Bi2.
2.1.4 ELISA of CD2 5, f}2-M.
The enzyme-linked immunosorbent assays are solid phase
biospecific binding-assays which have been taking on an
increasingly more prominent role in laboratory medicine and
biotechnology. In comparison to most compound identification
techniques, ELISA's now provide greater objectivity and
potential for automation. Since a single unit of enzyme label
can amplify a reaction product several fold, many ELISA's can
be optimized for the detection at the picomole or attomole
level [Tijssen 1985., Gosling 1990].
Materials: Quantikine human IL-2sRa (R&D Systems, Inc.,
Minneapolis USA); beta 2-Microglobulin Immunoassay (Elias
Co. Madison, Wisconsin USA). For the in-house assays we used
the Reacti-Bind a maleic anhydride activated polystyrene
plate (Pierce Chemical Company, Rockford, Illinois USA) and
performed the selected assays according to the following
methods.

49
Method: For the commercial kits the protocol followed was
included in the package. For the "in-house" assays the
following protocol was used.
1. Add lOOmL of the ligand (antigen or antibody) solution to
each well. For optimal results the ligand immobilization
should be made from solutions containing 10-30|ig/mL in PBS or
Hepes buffer. Incubate at room temperature for l-8hrs or
overnight if refrigerated.
2. Rinse three times with 200(J,L blocking agent to prevent
non-specific binding. Several different blocking agents can
be used, our best results were obtained with 5% BSA-PBS and
1% gelatin-PBS. Thirty minute incubations of the blocking
agent improve selectivity of the assay. These blocked
microtiter plates may now be stored a 4 C until samples are
to be assayed.
3. Add lOOpL of the analyte solution or standards (serum,
plasma or supernatant) to each well. Incubate the plates for
lhr at room temperature. Rinse three times with 20 0|iL wash
buffer. The wash buffer used was TBS-0.05% Tween for alkaline
phosphatase (AP) enzyme-conjugates or PBS-0.05% Tween for
horseradish peroxidase (HRP) enzyme-conjugates.
4. Add 100(iL of the enzyme-labelled secondary antibody to
each well, Incubate the plates for lhr at room temperature.
Rinse three times with 200(iL wash buffer.
5. Add 100(iL of the appropriate substrate (fresh TMB or ABTS
for HRP; and PNPP for AP enzyme-conjugates). Allow the color
to develop for 2 0 mins, then add 50[J.L of stop solution (1%
SDS, 3M HCL or 3M H2S04).
6. Measure the absorbance of the plate at 405nm. The
correlation between analyte concentration and 405nm

50
absorbance was made each time using the DeltaSoft data
analysis software from BIOTEK or by the least squares
regression method shown in appendix III.
2.2 Specific Aims #2
2.2.1 Amination, Pegylation and Protein Immobilization
on PMMA Membranes.
Histones were immobilized on the surface of intact PMMA
hollow fibers via surface modifications that result in stable
primary amine groups which can undergoe subsequent reactions
with a tresylated star polymer. When the Star-PEG polymers
are coupled to the surface, histones or other proteins can be
immobilized onto the hollow fibers. We describe below the
procedure for aminating and then crosslinking whole histones
to the PMMA hollow fiber via a functionalized 64-arm
polyethylene glycol star. The histone-modified hollow fibers
were then characterized by SEM and AFM or else used in the
adsorption capacity experiments of section 2.3.1.
Materials: PMMA hollow fiber dialyzers were obtained from
TORAY Industries, Inc. (Tokyo). Hydrazine-monohydrate (Fluka
Chemical Corp. New York USA). Tresyl Star-PEG was obtained
from Shearwater Polymers (Huntsville, Alabama USA). Whole
histones (derived from calf thymus) were obtained from Sigma
Chemical Company.
Methods: Histone immobilization on PMMA hollow fibers for
microscopic analysis.
1.Weight out four 75mg samples of PMMA hollow fibers. Rinse
twice with 50ml dd H20 then soak in 50% MeOH for 30min
discard.
2. Add 4ml of 50% MeOH to the hollow fibers in a test tube
then add 370.0|iL of N2H4-H20 activator solution (10.32mg/mL) .
Allow the amination reaction to proceed for 3 hrs.

51
3. After the hollow fibers have been aminated they are rinsed
with 50mL of dd H20, then soaked in 50% MeOH for 30min. This
should result in a maximum of 7 6[imol of activation sites on
the fibers.
4. The aminated PMMA hollow fibers are then incubated in test
tubes containing 4mL of 0.7|iM (in 50% MeOH pH 8.0)
tresylated-Star-PEG for 3hrs.
5. These hydrophilized PMMA hollow fibers are then incubated
as before in 4mL of 3.7mg/mL whole histones in PBS overnight.
Histones are immobilized onto the Star-PEG by covalent
binding to the tresylated end groups. Histone octamers may
spontaneously form at physiologic pH.
6. The histone-modified PMMA hollow fibers are then stored in
PBS-0.1% BSA (to block non-specific binding) until
characterized by AFM or used in adsorption capacity
experiments.
2.2.2 Amination, Pegylation and Protein Immobilization
on PSF Membranes.
Histones were immobilized on the surface of intact PSF
hollow fibers via surface modifications that result in stable
primary amine groups which can undergoe subsequent reactions
with a tresylated Star polymer. When the Star-PEG polymers
are coupled to the surface, histones or other proteins can be
immobilized onto the hollow fibers. We describe below the
procedure for aminating and then crosslinking whole histone
to the PSF hollow fiber via a functionalized 64-arm
polyethylene glycol star.
Materials: PSF hollow fiber dialyzers were obtained from
Fresenius AG, (Germany). Hydrazine-monohydrate (Fluka
Chemical Corp. New York USA). CrCL2 was obtained from Fluka
Chemical Corp. (New York USA). Tresyl Star-PEG was obtained
from Shearwater Polymers (Huntsville, Alabama USA)

52
Methods: Histone immobilization on PSF hollow fibers or
films. A new method for aminating intact PSF hollow fibers in
aqueous solution is presented. Nitration in aqueous solution
followed by reduction of the meta-nitro groups to the
corresponding amine was used to aminate PSF membranes [Fried,
JR 1995]. The aminated films are then surface-modified by
attaching whole histones to the surface via a tresylated-
Star-PEG.
1. The nitration reagent contained
lOmL cone. HN03
20mL conc.H2S04
30mL cone. Acetic acid
30mL distilled H,0
90mL nitration reagent
The acetate is used to stabilize the nitronium ion [N02]~ in
the form of the mixed anhydride.
The PSF hollow fibers or films were then incubated in
2mL of the nitration reagent for 1, 3, 12, and 24 hrs to
acheive four levels of nitration.
2. The nitrated PSF membranes were then incubated for lhr in
4mL of reducing agent (5g CrCl2 in 250mL of 1M HC1). The
reduction was monitored by the change in color of the reagent
from green Cr2+ to Cr3+ which is blue.
3. The aminated films were then hydrophilized by incubation
for 3 hrs with 2mL of 14.2|iM tresyl Star-PEG.
4. Histones were immobilized by incubation with 2mL of 8.6
mg/lOOmL whole histone overnight. The histone-modified films
were blocked against non-specific binding by storage in 0.1%
BSA PBS.

53
2.2.3 Stability and Reactivity of Tresylated Star-PEG
Polymers.
Polyethylene glycol can be coupled to other molecules
through its hydroxyl end-groups. One commonly used reagent is
tresyl chloride which activates the primary-OH into a good
leaving group i.e. tresylate. The activated function can then
be displaced by a protein nucleophile, for example, the
amino-terminal side chain of lysine. Loss of activity by
hydrolysis, however, does occur but at a slower rate than
aminolysis. Hydrolysis is most successful where the ester is
that of a stronger acid, thus the stability of tresylate
esters is intermediate between esters of sulfonic acid and
the more stable tosylates [Mosbach and Nilsson 1984].
Materials: TMB ( 3,3',5,5'-tetramethyl benzidine HCL) and
Reacti-Bind a maleic anhydride activated polystyrene plate;
were both obtained from Pierce Chemical Company (Rockford,
Illinois USA); Tresyl Star-PEG was obtained from Shearwater
Polymers (Huntsville, Alabama USA); Diaminobutane
(putrescine) was obtained from Fluka Chemical Corp. (New York
USA); Tresyl chloride Fluka Chemical Corp. (New York USA);
HRP-antibody conjugate (horseradish peroxidase-IgG) was
obtain from ALPCO, Ltd. (New Hampshire USA)
Method: 1. 300|iL of 1M diaminobutane was incubated in maleic
anhydride activated polystyrene microtiter plates for lhr.
Remove the solution, do not wash.
2. 300pL of the tresyl Star-PEG reagent (0.14|iM) was then
coupled to the aminated wells for 2hr. Remove the solution,
do not wash.
3. Thereafter wells 1-16 were incubate under nitrogen, while
the remaining wells were incubated with 300|iL of 50% MeOH pH
(2.4-11.3) buffers for 30 mins. Remove the solution do not
wash.

54
4. Add 300|iL of HRP-antibody conjugate and incubate for lhr.
Remove the solution then wash three times with wash buffer
(PBS 0.05% Tween).
5. Add 300fiL of the HRP substrate TMB, allow the color to
develop for 30 mins, then read the absorbance at 405nm.
For the control wells the average 405nm absorbance was
0.465 as determined by HRP-conjugate substrate activity.
Taking this value to represent 100% protein binding capacity
of the immobilized Star's, the percent hydrolysis was
estimated iron the equation below
%protein binding + % hydrolysis = 100%
and therefore
% hydrolysis = 1- A4OS/0.465.
We determined the extent of tresylate hydrolysis of
these Star-PEG derivatives over the pH range 2.4-11.3.
2.2.4 Biospecific Interaction Analysis: Kinetic and
Equilibrilium Analysis using the BIACORE 2000
The biospecific interaction between antihistone
antibodies and histones immobilized via PEG Star-polymers was
monitored using the BIACORE 2000. A multistep binding
experiment was conducted where tresylated Star-PEG polymers
were attached to the BIACORE sensorchip in a manner similar
to that in the modification of PMMA membranes. The instrument
measures in real-time the binding kinetics at the flow cell
surface and then displays a corresponding sensorgram which
quantifies the recognition process, see fig.2-4a. The BIACORE
2000 operates on the principle of surface plasmon resonance
(SPR) and is designed to detect changes in the refractive
index of the layer of solution in contact with the sensorchip
[Malmqvist et al. 1993].

55
Since the refractive index of the medium is affected by
the surface concentration of solutes, monitoring the SPR
angle provides a real-time measurement of changes in surface
concentration and/or coupling, see fig.2-5. Units used to
express the SPR signal are called response units (RU). One
thousand RU are equivalent to a change of about lng/mm2 in
surface protein concentration.
We confirmed the immobilization of immunoreactive
histones onto tresylated Star-PEG's and their subsequent
binding to circulating antihistone antibodies using the
BIACORE biosensor. The magnitude of the shift in SPR angle is
proportional to the concentration of proteins bound to the
surface; see fig. 2-4b.
A real-time BIA sensorgram provides essentially two
kinds of information; 1) the rate of a molecular interaction
and 2) the binding level or concentration of the analyte and
the immobilized ligand see fig.2-4c. For all applications the
SPR response is related to the change in surface mass
concentration of the analyte. It also depends on the
molecular weight of the analyte in relation to the number of
ligands on the surface. For any given number of binding
sites, a higher molecular weight analyte will give a
proportionally larger response in RU's, conversely, a higher
molecular weight ligand will provide few binding sites per
unit area (unless the valency is much greater than one) see
fig.2-4d.

56
Figure 2-4. BIACORE Biosensors, a) Schematic of the
optical system and flow cell used in BIACORE
biosensors, b) The surface of a sensor chip consists
of three layers; glass, a thin gold film layer and
a carboxymethylated dextran layer, c) The ligand
or capturing molecule is covalently immobilized
on the sensor chip whereas the analyte is injected
in the flow cell over the surface, d) The response
in resonance units (RU) is recorded as a function of
time; it consists of an association phase (A), an
equilibrium phase (E) and a dissociation phase (D).

57
nnetd tnt*ntty
c)
Retoarte* signal |RU|
Figure 2-5. The BIACORE Sensorgram: surface plasmon
resonance angle (SPR). The sensorgram is a plot of the
(SPR) angle against time; it displays the progress of
the molecular interaction at the sensor surface.
Binding of analyte shifts the SPR angle from
position 1 to position 2. in the figures a) SPR is
observed as a dip in reflection intensity at an angle
which depends on the refractive index of the medium on
non-illuminated side of the surface; b) the SPR angle
shifts when biomolecules bind to the surface and change
the refractive index of the surface layer.c) the
sensorgram shows in real-time a quantitative measurement
of changes in surface concentration of the analyte.

58
Assuming that the relationship between response and mass
is the same for ligand and analyte (i.e. 1000RU-lng/mm2 for
proteins), the following equation may be applied.
#binding sites = (ligand response/ligandMW) (valence of the ligand)
The analytic response = analyte MW(#analyte molecules) at a
fully saturated ligand surface the #binding sites = #analyte
molecules. The response here would represent the maximum
possible resonance signal obtain for this system. Therefore,
we may substitute the expression for the #binding sites for
the #analyte molecules in the equation for analyte response
and obtain an estimate of Rmax
R, = (valence of the ligand) (analyte MW) (ligand response/ligand MW)
R. = nM.RL/ML
Where R* is the maximum binding capacity of the surface
ligands for this particular analyte in RU's.
Materials: BIACORE 2000 (Pharmacia Biosensor., Uppsala,
Sweden). Whole histones (Sigma Chemical Co.); Monoclonal
anti-histone antibodies were derived from WEHI hybridomas
obtained courtesy of Dr. Joel Schiffenbauer of the University
of Florida. Tresyl Star-PEG (Shearwater Polymers.,
Huntsville, Alabama USA)
Methods: Biospecific interaction analysis using either of
the BIACORE biosensors has many methodological similarities
with solid-phase immunoassays. The sensor chip is, however,
the signal transducer instead of the enzyme-antibody
conjugate that yields a signal in ELISA's after exposure to
substrate. The chip is a glass slide with a thin layer of
gold deposited on one side. The gold film is in turn covered
with a covalently bound matrix (e.g. carboxymethyl dextran)
onto which biomolecules can be immobilized.

59
The matrix serves four purposes.
It allows covalent immobilization of biomolecules using
well-characterized coupling chemistry.
It increases sensitivity by increasing the binding
capacity of the surface.
It provides a hydrophilic environment suitable for most
interactions of biological interest.
It provides a very low degree of non-specific binding to
the surface.
Multi-step methods rely on a series of binding steps to
analyze the interaction of interest. For multi-step methods,
real-time BIA has a marked advantage over many other surface
interaction techniques, in that surface binding is registered
on the sensorgram at each stage in the process. This gives
unique facilities for controlling the intermediate binding
events. Multi-step methods are typically used to study
multimolecular complex formation in sequential binding
experiments. This can give valuable information concerning
steric and allosteric effects, since the order of the
interaction and binding can be varied with ease. For an
example see the multi-binding experiment sensorgram shown in
fig. 2-6a and b.
Method: Start the immobilization sequence using an amine
coupling to bind proteins to the surface. The numbered steps
refer to the following stages in the immobilization
experiment.
1. Baseline for the unmodified sensor chip surface with
continuous flow buffer.
2. Injection of NHS/EDC to activate the surface gives an
increase in the SPR signal due to the change in bulk

60
refractive index. Resulting in a NHS ester on the surface
which readily reacts with primary amines.
3. Next we aminated the surface with 1M ethylenediaamine at
pH 8.3. Deactivation of unreacted NHS-esters using 1M
ethanolamine HCL, adjusted to pH 8.5 with NaOH. The
deactivation process also removes any remaining
electrostatically bound ligand.
4. 55 (IM tresyl-Star-PEG is immobilized on the surface (with
64-1 = 63 free arms for subsequent histone binding). Binding
curve was observed.
5. 1:10 whole histones in Hepes buffer was added. Binding
curve was observed.
6. Deactivation of unreacted tresyl groups using 1M
ethanolamine HCL, adjusted to pH 8.5 with NaOH. The increased
SPR signal is due to a change in the bulk refractive index.
The deactivation process also removes any remaining
electrostatically bound ligand.
7. Record the response (in RU's) of the immobilized ligand
after deactivation.
From the resulting sensorgram (see section 3.22) the
response and slope at each time point can be observed. These
parameters are then used to calculate the rate and
equilibrium constants of the histone-antihistone interaction.
2.2.5 Microscopic Analysis: Hollow Fiber Surface
Imaging using Atomic Force and Scanning Electron
Microscopies.
Materials and Method: We used the transmission electron
microscope to estimate the size range of the Star-PEG
polymers. A Star-PEG solution (2mg/mL Star-PEG in 50% MeOH) a
5(iL drop was evaporated on the grid then imaged by TEM. The
TEM images were performed courtesy of Mr. Scott Whitaker of
the University of Florida ICBR Microscopy Core Lab on a

61
Phillips EM-400T transmission electron microscope
(Hassborough, Oregon) at 6kV accelerating voltage.
Figure 2-6. Activation of a BIACORE sensorchip.
a) The sensor chip is activated with a pulse of EDC/NHS.
The resulting response curve displays (streptavidin)
immobilization to the surface. Residual activation sites
are quenched with ethanolamine b) Compares the response
curves of equal concentrations (4 0(ig/mL) of two ligands,
avidin and streptavidin.

62
We used both the SEM and "non-contact mode" AFM to
compare the chemically-modified PMMA with control hollow-
fibers and found organized structures in the size range of a
histone-Star i.e. between 150 and 250 nm. The SEM images were
performed courtesy of Mr. Scott Whitaker of the University of
Florida ICBR Microscopy Core Lab on a Jeol JSM 35CF scanning
electron microscope (Soquelec, Montreal) at 15kV accelerating
voltage. The AFM images were performed courtesy of Burleigh
Instruments, Inc. on a METRIS-NC-2000 scanning probe
microscope (Fishers, New York).
2.3 Specific Aims #3
2.3.1 Adsorption Capacity Experimental Methods: AHA
ELISA
An adsorption capacity experiment was performed to
assess the binding of antihistone antibodies (AHA) from
hybridoma supernatant solution (see fig. 2-7).
Materials Anti-Histone Immunoassay (ALPCO, Ltd. New
Hampshire USA); PMMA hollow fiber dialyzers were obtained
from TORAY Industries Inc. (Tokyo). PSF hollow fiber
dialyzers were obtained from Fresenius AG, Germany.
Hydrazine-monohydrate (Fluka Chemical Corp. New York USA).
CrCL2 was obtained from Fluka Chemical Corp. New York USA.
Tresyl Star-PEG was obtained from Shearwater Polymers
(Huntsville, Alabama USA)
Method: AHA ELISA
For the commercial kits the protocol followed was that
included in the package with one exception. Since the ALPCO
Anti-Histone Immunoassay was designed for the detection of
human AHA and we have designed experiments for the
immunoadsorption of murine (hybridoma) AHA, a anti-mouse IgG-

63
HRP conjugate was used instead in the detection step. In
house ELISA's were made according to method outlined in
section 2.14.
Figure 2-7. Adsorption/Incubation Experiment. In the
adsorption capacity experiments a thin film of the membrane
material (PMMA or PSF) was sheet cast onto the base of test
tubes at a well-defined surface area. The films were then
incubated for 15 mins.in hybridoma cell culture supernatant
containing a known concentration of AHA. The solutions were
sampled before and after each incubation then assay by
ELISA to determine AHA adsorption. The delta sub (i) values
represent the changes in AHA concentration during each
incubation and the sum of these changes was used to
estimate the effective adsorption capacity of these
chemically-modified membranes.

64
Method: Adsorption capacity of surface-modified PMMA films.
PMMA (0.8g) was dissolved in warm CHC13 (lOOmL). After
the slurry cooled to room temperature a portion of it was
covered with 50% Me0H/H20. Plastic cuvettes were then dipped
into the slurry layer and carefully raised up through the
aqueous layer repeatedly until a thin film of PMMA coated the
lower 13 cm2 of the cuvette. The cuvettes were then air dryed
for 4hrs. before being incubated in the hydrazine activator
solution as follows.
Ll = 69.0mg hr/cm2 polymer
L2 = 208.Omg hr/cm2 polymer
L3 = 690.Omg hr/cm2 polymer
L4 = 2081.Omg hr/cm2polymer
After each of the films were activated they were placed
with 250]iL of 12.5^iM tresylated Star-PEG in 2mL 50% MeOH (pH
8.0) for 12hrs. These "hydrophilized" PMMA films were then
incubated in 2mL of 2.5 mg/mL whole histones (PBS buffer)
overnight. The histone modified PMMA films are then placed in
PBS 0.1% BSA to block non-spacific binding sites then stored
at 4 C until used in the adsorption capacity experiments.
The AHA adsorption of a group of modified PMMA films
(L1-L4) was determined by incubating each of the films (13
cm2) in filtered hybridoma cell culture supernatant containing
approximately 1.3mg/3mL of antihistone antibodies as follows:
1. 3mL of the supernatant (1.3mg) was added to each of
twenty., 4mL reacti-vials (previously washed in buffer).
2. Then each of the four films (test tubes) were incubated
sequentially in each of five of these vials for 15 mins. The
initial concentration in each vial being approximately 1.32mg
the change in AHA concentration ACaha in each vial was
measured and plotted against incubation number (see fig.2-7).

65
Extrapolation of the curve yields an approximate adsorption
capacity of the films.
3. Summation of the ACaha values equals the AHA adsorption
capacity of the modified surface per unit area. Half (0.5)
milliliter samples of the supernatant were taken before and
after incubation and assayed for AHA concentration by ELISA
(see methods section 2.14).
4. The total adsorption for the four levels of PMMA surface
modification were calculated and the adsorption capacity per
unit area of each of the treated membranes was determined as
follows, (also see Results abd Discussion specific aim #3)
Q1 = 0.84mg/m2 polymer Q3 = 1.80mg/m2 polymer
Q2 = 0.335mg/m2 polymer Q4 = 0.066mg/m2 polymer
Method: Adsorption capacity of chlorosulfonated surface-
modified PSF film. The mechanism of polysulfone
chlorosulfonation theoretically takes place according to the
following reactions; see fig. 2-8 below [Gilbert 1965;
Pozniak et al. 1995].
1. ArH + C1S03H > ArS03H + HC1
2. ArS03H + CISChH <===> ArS02Cl + H2S04
(basic medium) (acidic medium)
3. ArS02Cl + H20 > AXSO3H + HC1
Figure 2-8. Chlorosulfonation of PSF.
Note: It is clear from the reaction scheme that polysulfone
amination should take place in acid medium free of intact H20.
The viscosity of the polysulfone solution also decreases with
increasing sulfonation rate, and the solvent affinity of
chlorosulfonated PSF is less than that of PSF resulting in

66
decreased solubility as the reaction proceeds. However,
chlorosulfonated PSF film may still be sheet cast onto the
base of glass cuvettes according to the following protocol.
Four levels of films were activated (one control)
Ll = control
L2 = 281 mg C1S03H /gm polymer
L3 = 562 mg C1S03H /gm polymer
L4 = 1124 mg C1S03H /gm polymer
After each of the activated films were sheet casted on
to the base of 11cm2 glass test tubes then incubated in a
solution of 1M diaminobutane for 2hrs. The aminated films
were then allowed to react with 250|iL of 12.5(iM tresylated
Star-PEG in 2mL 50% MeOH for 12hrs. These "hydrophilized" PSF
films were then incubated in 2mL of 2.5 mg/mL whole histones
overnight. The histone-modified PSF films were then placed in
PBS 0.1% BSA to block non-spacific binding sites then stored
at 4 C until used in the adsorption capacity experiments.
The aha adsorption of a group of modified PSF film (Ll-
L4) was determined by incubating each of the films (11 cm2) in
filtered hybridoma cell culture supernatant containing
approximately 88.32|ig/3mL of antihistone antibodies as
follows:
1. 3mL of the supernatant (88.32|ig) was added to each of
twenty, 4mL reacti-vials (previously washed in buffer).
2. Then each of the 4 films (test tubes) were incubated
sequentially in each of five of these vials for 15 mins. The
initial concentration in each vial being 88.32fig the change
in AHA concentration ACaha in each vial was measured and
plotted against incubation number (see fig. 2-7).
Extrapolation of the curve yields an approximate adsorption
capacity of the films.

67
3. Summation of the ACaha values equals the AHA adsorption
capacity of the modified surface per unit area. Half (0.5)
milliliter samples of the supernatant were taken before and
after incubation and assayed for AHA concentration by ELISA
(see methods section 2.14).
4. The total adsorption for the four levels of PMMA surface
modification were calculated and the adsorption capacity per
unit area of each of the treated membranes was determined,
(also see Results and Discussion specific aim #3)
Q1 = 0.0mg/m2 polymer Q3 = 1.06mg/m2 polymer
Q2 = 0.67mg/m2 polymer Q4 = 3.93mg/m2 polyme
Method: Adsorption capacity of nitration surface-modified
PSF film. Since chlorosulfonation of PSF requires the polymer
to be activated in organic solvents which dissolve the PSF
hollow fiber, methods of aminating the intact PSF hollow
fibers in aqueous were attempted. Nitration in aqueous
solution followed by reduction of the nitro-groups to the
corresponding amine was used to aminate PSF films sheet cast
at the base of plastic cuvettes. The aminated films were then
surface-modified by attaching whole histones to the surface
via a tresylated Star-PEG.
1. The nitration reagent contained
lOmL cone. HN03
20mL cone. H2SCh
30mL cone. Acetic acid
30mL distilled H,0
90mL nitration reagent.
The acetate is used to stabilize the nitronium ion [NO2]- in
the form of the mixed anhydride.

68
2. 13cm2 PSF films were sheet cast onto the base of plastic
cuvettes from the slurry and allowed to air dry for lhr, then
stored in distilled H20 before activation. The films were then
incubated in 2mL of the nitration reagent for 1, 3, 12, and
24 hrs to acheive four levels of nitration i.e. 1, 3, 12, and
24 hrs.
3. The nitrated PSF films were then each incubated for lhr in
4mL of reducting agent (5g CrCl2 in 250ml of 1M HC1). The
reduction was monitored by the change in color of the reagent
from green Cr2+ to Cr3+ which is blue.
4. The aminated films were then surface-modified by reacting
with 2mL of 14.2|iM tresyl Star-PEG and then 2mL of
8.6mg/100mL whole histone. The histone-modified films were
blocked against non-specific binding by storage in 0.1% BSA
PBS.
5. AHA adsorption capacity of the modified PSF films was then
estimated by incubating each of the treated films in
antihistone antibody hybridoma supernatant (1400U/2mL). The
concentration of the AHA in each reacti-vial was assayed
before and after the 15 min. incubation with the treated film
and the change in AHA concentration ACaha determined by ELISA
(see fig.2-7). The results of the experiment showed that
films activated in the nitration reagent for 1, 3, 12, and 24
hrs. respectively adsorbed the following amounts.
Q1 = 1.844 mg/m2 polymer Q3 = 2.29 mg/m2 polymer
Q2 = 1.190 mg/m2 polymer Q4 = 2.18 mg/m2 polymer
2.3.2 Extracorporeal Immunoadsorption Experimental
Methods.
We have used the in vitro model (described earlier) to
produce and evaluate chemically-modified hollow fiber
dialyzers as potential blood purification devices. Binding
agents were covalently immobilized onto the inside or outside

69
surface of polymethylmethacrylate (PMMA) and polysulfone
(PSF) hollow fibers by coupling reactions designed to remove
antihistone antibodies from donor compartment solutions over
the course of 4hrs.
Materials: Monoclonal anti-histone antibodies were derived
from WEHI hybridomas obtained courtesy of Dr. Joel
Schiffenbauer of the University of Florida. Monoclonal anti
histone antibody standards were obtained from Boehringer
Mannheim GmbH, (Indianapolis, USA). Anti-Histone Immunoassay
ALPCO, Ltd. (New Hampshire USA); PMMA hollow fiber dialyzers
were obtained from TORAY Industries, Inc. (Tokyo). PSF hollow
fiber dialyzers were obtained from Fresenius AG, Germany.
Hydrazine-monohydrate Fluka Chemical Corp. (New York USA).
CrCL2 was obtained from Fluka Chemical Corp. (New York USA).
Tresyl Star-PEG was obtained from Shearwater Polymers
(Huntsville, Alabama USA)
Methods:
1. The model circuit was rinsed throughly and then primed
with 50% MeOH/HjO. PMMA dialyzers were activated with 5.2mM
hydrazine-monohydrate (in 50% MeOH) for 4 hr. Polysulfone is
activated instead with "strong nitration reagent" then
reduced with "strong reducing agent" (5g CrCL2 in 250mL in
1M HCL).
2. The aminated hollow-fibers are then rinsed thoroughly with
50% Me0H/H20 then allowed to react with 14.2(iM tresyl-Star-PEG
for 8 hrs. Since each immobilized Star-PEG may bind to the
activated polymer surface by a single tresylate end group,
this would leave 64 1 = 63 PEG arm's free to bind proteins.
The immobilized star polymers dramatically increase the
active surface area of the hollow fibers.

70
3. After the Star-PEG coupling the circuit was washed
throughly with PBS. Then 0.4g (3.67|imol octamers) of histones
in PBS was injected into the circuit and allowed to bind onto
the star polymers. Hypothetically each arm of the immobilized
Star-PEG is available for each protein molecule.
4. After these modified dialyzers are rinsed with PBS 0.1%
albumin they were refrigerated at 8 C until ready for use in
an extracorporeal immunoadsorption experiment.
The clearance of antihistone antibody was tested in PMMA
hollow fiber dialyzers with whole histones immobilized on the
outside and inside lumen of the membrane. The results were
then compared against controls. The dialyzers were modified
as described earlier.
5. A known amount of antihistone antibody (0.289mg/mL) was
then injected into the donor circuit and allowed to mix (in
bypass) for 5 mins., after which time the circuit was taken
out of bypass and opened to one of the test dialyzers. Donor
compartment inlet and outlet samples (0.5mL) were then taken
at 5, 10, 30, 60, 120 and 240 minute intervals from the time
the donor circuit was opened. The samples were analyzed by
Anti-Histone Immunoassay for AHA content.
In the plasma studies the donor reservoir (475mL) was
filled with plasma and spiked with approximately 16.75mg of
murine AHA, then mixed in bypass at room temperature for
0.5hr. Modified or control dialyzers were then placed in the
extracorporeal circuit and perfused with the spiked plasma
for 4hrs. Plasma samples were taken from the dialyzer inlet
and outlet sampling port and assayed for AHA concentration.

CHAPTER 3
RESULTS AND DISCUSSION
3.1Specific Aim #1
3.1.1 An In Vitro Model for the Study of
Extracorporeal Therapy: Applications of the Model.
The results of the experiments that follow will
demonstrated the utility of the model in evaluating
hemodialyzer performance and in studying the effect of
dialysate additives on the clearance rate of "middle
molecules". Other potential applications of the model are
also shown, see fig.3-1.
3.1.2 The Effects of Dialysate Additives on the
Transport Properties of Donor Solutes across Hollow
Fiber Dialyzers.
It was found that the presence of antibodies in the
dialysate compartment potentiates the clearance of specific
antigens beyond steady-state concentrations only when the
permeability of the dialyzer membrane is sufficiently great
for antigen diffusion. In fig.3-2 the transport
characteristics of vit.B12 in a high flux PMMA dialyzer are
shown. The donor compartment concentration decreases from
2.75mg/100mL to a steady-state concentration of 1.85mg/100mL
in the first 60 minutes. After anti-vit.B12 antibodies were
added to the receiver compartment the mass transport was
increased again and donor compartment vit.Bi2 continued to
decrease from 1.85mg/100mL to 1.35mg/100mL within the next
60mins. The experiment shows that sink conditions for a
specific donor solute can be re-established by the presence
of complexing additives in the receiver compartment.
71

72
APPLICATIONS
DRUG METABOLISM: Immobilized microsomal enzymes may be used to study the
metabolism of new lead compounds in a physiologically controlled environment.
EXTRACORPOREAL IMMUNOADSORPTION: Immobilized self-antigens (i.e. histones) may
be use to study the adsorption of autoantibodies (i.e. anti-histone antibodies) or other
disease associated antigens.
MEMBRANE TRANSPORT: The transport properties of blood-borne solutes across hollow-
fibers may studied and/or manipulated by dialysate additives (i.e. antibodies,
cyclodextrins)
CELL CULTURE SUPERNATANT FRACTIONATION: By-products of cellular metabolism and
gene expression can be monitored and separated on-line in the in vitro model.
RADIOIMMUNOTARGET1NG : Targeting of radioimmunoconiugates to selected tissue-types
can be studied using a radiolabeled biotinylated-monoclonal antibody to tissue cell
antigens. After the radioimmunoconjugate has accumulated in the target tissue, the
circuit may be taken out of bypass so that the unbound radioimmunoconjugate can
adsorb to the surface modified (e g. avidin) hollow-fiber dialyser.
BLOOD COMPARTMENT DIALYSATE
COMPARTMENT
BLOOD PUMP DIALYSATE PUMP
Figure 3-1
In Vitro Model: applications

73
The mass clearance under equilibrium conditions was
(2.75-1.85)mg/100mL hr
=> (0.9mg/100mL hr)*(230mL)
= 2.07mg/hr
and that under "sink" conditions was
(1.85-1.35)mg/100mL hr
=>(0.5mg/100mL hr)*(230mL)
- 1.15mg/hr
This should be considered as a rate enhancement effect
against an equilibrium concentration gradient. Thus a
"clearance enhancement factor" can then be approximated for
anti-vit.Bi2 in this system which is equal to
1.15 + 2.07 = 3.22mg/hr
(3.22/2.07)-1 = 55.6%
A similar experiment was conducted using a high flux
cellulose acetate dialyzer see fig.3-3. The equilibrium and
sink condition clearances were equal to 1.84mg/hr and
1.27mg/hr.
The a "clearance enhancement factor" for anti-vit. B12 in
this system is equal to
(3.105/1.84)-l = 68.8%
In fig. 3-4, the transport of vit.Bi2 in a low flux
polysulfone dialyzer was studied. Equilibrium and sink
condition clearances were again determined and found equal to
0.644mg/hr and 0.575mg/hr, respectively. A clearance
enhancement factor of 89.3% was calculated here.

(IUJOOI/Buj) ooq
74
Transport ot vitamin B12 in a BK2.1P dialyzer
Figure 3-2. Mass transport curves of vit.Bi2 in a high
flux PMMA dialyzer. The solid () and hollow (o) circles
represent donor vit.B12 cone, before anti-vit.B12 antibody
administration to the receiver compartment at 60 mins.
The hollow triangles (A) represent receiver vit.B12
during the same time period. The hollow squares() and
solid triangles() represent donor vit.B22 inlet and
outlet concentrations respectively^, after anti-vit.B12
antibodies were injected into the receiver compartment.

Cone, (mg/100ml)
75
Transport ol vitamin B12 In a Altrex dlalyzer
Figure 3-3. Mass transport curves of vit.B12 in a high
flux cellulose acetate dialyzer. The hollow (o) circles
represent donor vit.Bu cone, before anti-vit.B22 antibody
administration to the receiver compartment at 120 mins.
The hollow triangles (A) represent receiver vit.B12
during the same time period. The hollow squares() inlet
and concentrations respectively, after anti-vit.B12
antibodies were injected into the receiver compartment.

76
Transport of vitamin B12 In a F8 dialyzer
Figure 3-4. Mass transport curves of vit.Bi2 in a low
flux PSF dialyzer. The solid () and hollow (o) circles
represent donor vit.B12 cone, before anti-vit.Bl2 antibody
administration to the receiver compartment at 60 mins.
The hollow triangles (A) represent receiver vit.B12
during the same time period. The hollow squares ()
represent donor vit.Bi2 inlet concentrations after anti-
vit.Bij antibodies were injected into the receiver
compartment.

P2-M Cone.(pg/100ml)
77
Similar transport studies were performed with the
polypeptides (32-M and CD25. Fig. 3-5 shows that for 02-M in
the high flux PMMA dialyzer, hydrophobic interactions were
responsible for the rapid decrease in donor concentration
from 10 to 0.7|ig/mL in the first 2hrs of the experiment. No
significant amounts of P2-M were measured in the receiver
compartment. Therefore, no clearance enhancement due to
dialysate additives (antibodies) was observed.
Transport of P2-M in a BK2.1 P dialyzer
Figure 3-5. Mass transport curves of p2-M in a high
flux PMMA dialyzer. The hollow (o) circles represent
donor vit.Bi2 cone, before anti-vit.B12 antibody
administration to the receiver compartment at 120 mins.
The hollow triangles (A) represent receiver vit.B12
during the same time period. The hollow squares()
represent donor vit.Bi2 inlet concentrations
respectively, after anti-vit.Bu antibodies were injected
into the receiver compartment.

P2-M Cone. (pg/100ml)
78
In fig. 3-6, the clearances of p2-M were studied in a
high flux cellulose acetate dialyzer. The equilibrium and
sink condition clearances were determined as 3.62(ig/mL and
3.22(ig/mL with a clearance enhancement factor of 8 8.9%.
Transport of P2-M in a Altrex dialyzer
Figure 3-6. Mass transport curves of P2-M in a high
flux cellulose acetate dialyzer. The hollow () circles
represent donor P2-M cone, before anti-P2-M antibody
administration to the receiver compartment at 120 mins.
The hollow triangles (A) represent receiver vit.Bl2
during the same time period. The hollow squares() inlet
and concentrations respectively, after anti-P2-M
antibodies were injected into the receiver compartment.

79
In fig. 3-7 (32-M transport in the low flux polysulfone
dialyzer was impeded due to the smaller pore size of the
hollow fibers and the larger molecular diameter of the
polypeptide. No clearance into the receiver compartment was
observed.
Transport of (32-M in a F8 dialyzer
E
o
o
O)
=L
d
c
o
O
C\J
CD_
Figure 3-7. Mass transport curves of (32-M in a low flux
PSF dialyzer^The solid () represent donor (32-M cone,
before anti-(32-M antibody administration to the receiver
compartment at 60 mins. The hollow triangles (A)
represent receiver (32-M during the same time period. The
hollow squares(Q) represent donor (32-M inlet and
outlet concentrations respectively, after anti-(32-M
antibodies were injected into the receiver compartment.

80
In fig. 3-8 CD25 in the high flux PMMA dialyzer, the
equilibrium and sink condition clearances were determined as
0.19ng/mL and 0.11ng/mL with a clearance enhancement factor
of 57.6%.
Transport of CD25 in a BK2.1P dialyzer
time mins.
Figure 3-8. Mass transport curves of CD25 in a high
flux PMMA dialyzer. The hollow () circles represent
donor CD25 cone, before anti-CD25 antibody
administration to the receiver compartment at 120 mins.
The hollow triangles (A) represent receiver CD25 during
the same time period. The hollow squares() represent
donor CD25 inlet concentrations respectively, after
anti-CD25 antibodies were injected into the receiver
compartment.

81
In fig.3-9 the clearances of CD25 were also studied in a
high flux cellulose acetate dialyzer.jlhe equilibrium and sink
condition clearances were determined as 0.172ng/mL and
0.422ng/mL with a clearance enhancement factor of 245%.
Transport of CD25 in a Altrex dialyzer
time mins.
Figure 3-9. Mass transport curves of CD25 in a high
flux cellulose acetate dialyzer. The solid () circles
represent donor CD25 cone, before anti-CD25 antibody
administration to the receiver compartment at 120 mins.
The hollow triangles (A) represent receiver CD25 during
the same time period. The hollow squares() represent
donor CD25 inlet concentrations respectively, after
anti-CD25 antibodies were injected into the receiver
compartment.

82
In fig.3-10 CD25 transport in the low flux polysulfone
dialyzer was also impeded due to the smaller pore size of the
hollow fiber membrane and the larger molecular diameter of
the polypeptide. No clearance into the receiver compartment
was observed.
Transport of CD25 in a F8 dialyzer
time mins.
Figure 3-10. Mass transport curves of CD25 in a low
flux PSF dialyzer. The hollow () circles represent
donor CD25 cone, before anti-CD25 antibody
administration to the receiver compartment at 120 mins.
The hollow triangles (A) represent receiver CD25
during the same time period.The hollow squares ()
represent donor CD25 inlet concentrations respectively
after anti-CD25 antibodies were injected into the
receiver compartment.

83
3.1.3 Summary.
We have shown that an in vitro model can be constructed
which mimics the extracorporeal circulatory system used in
conventional as well as novel dialysis therapies. The model
can be used to evaluate the clearance of a select donor
solute or a group of solutes under both equilibrium or sink
conditions. Passive adsorption of (32-M to PMMA membranes was
conclusively demonstrated in the model. New dialysate
formulations may be studied in the model as well. Here, we
found that dialysate additives which formed stable complexes
with a specific donor solute (i.e. antibodies) may increase
the clearance of that solute above and beyond steady-state
concentrations. For a summary of the experimental mass
transfer results see table 3-1.
Table 3-1. Summary of the experimental mass transfer
results. Percentages are clearance enhancement factors
described in the text.
Vitamin B,,
beta 2-M
CD25
BK2.1P
78.0%
22.06pg
57.6%
adsorbed
Altrex
68.8%
88.9%
245%
F8
89.3%
no clearance
no clearance
>mwco
>mwco
3.2 SPECIFIC AIM #2
3.2.1 Estimates of Tresylate Hydrolysis and
Reactivity.
Polyethylene glycol (PEG) can be coupled to other
molecules through its terminal hydroxide. One commonly used
coupling agent, tresyl chloride, activates these primary-OH's

84
into good leaving groups. The activated function can then be
displaced by a protein nucleophile, such as the amino-side
chain of lysine. See fig.3-11.
RO
+ ciso2-ch2-cf3
RO
n 0S02CH2CF
tresyl chloride
tresylated MPEG
Figure 3-11. Tresylated Star-PEG.
The (electrophilically) activated PEG-tresylate is a
very efficient derivative for coupling primary amines. The
product secondary amine is very stable toward hydrolysis and
the conjugate formed has similar charge (pi) to the native
protein, so that its bioreactivity is preserved; see fig.3-
12.
pH 7.5 8.5
PEG-0S02CH2CF3 + Protein-NH2 PEG-NH-Protein
Figure 3-12. PEG-protein coupling reaction.

85
Loss of activity by hydrolysis however, does occur but
at a slower rate than aminolysis. See fig. 3-13.
O
II
HOsch2cf3
o
Figure 3-13. Tresylate Hydrolysis.
Hydrolysis is most successful where the ester is that of a
strong acid i.e. -S030H > -S03CH2CF3 > -S03-ArR
Using a solid phase protein binding assay, the percent
hydrolysis of tresylated Star-PEG was determined 1) to be pH
dependent and 2) to be minimal at pH values 4.0 and 8.0; at
which tresylate hydrolysis was less than 35% in 0.5 hrs. See
fig. 3-14.

86
Hydrolysis ol Reactive End Groups
In hesylaled Slar-PEG's
Figure 3-14. pH Dependent Hydrolysis of Tresylated
Star-PEG Polymers. 300|iL of 1M diaminobutane was
incubated in a maleic anhydride activated polystyrene
microtiter plate for lhr. 300pL of the tresyl-Star-PEG
reagent was then coupled to the aminated wells. There
after wells 1-16 were incubated under nitrogen, while
the remaining wells were incubated with 300pL of 50%
MeOH pH buffers. For the control wells the average 405nm
absorbance was 0.465 as determined by a HRP conjugate
assay. Taking this value to represent the 100% protein
binding capacity of the immobilizedStar's; the percent
hydrolysis was determined from the equations below
% protein binding + % hydrolysis = %100; therefore the
% hydrolysis = 1 A405/0.465
3.2.2 Immunoreactivity of Immobilized Histones
Characterized using the BIACORE 2000.
We confirmed the immobilization of immunoreactive
histones on tresylated Star-PEG's and their subsequent
binding to circulating antihistone antibodies using the
BIACORE biosensor. A BIA sensorgram is a plot of the SPR
angle against time as it displays the progress of the
molecular interaction at the sensor surface. The magnitude of
the shift in SPR angle is proportional to the concentration
of proteins bound to the surface.

87
Estimates of the rate and equilibrium constants for the
binding of histones and antihistone antibodies were made
using a tresylated Star-PEG polymer as the reactive ligand on
the sensorchip. The response is shown in the sensorgram in
fig. 3-15. We first calculated the R,.x for the covalent
binding of histones on to tresylated Star-PEG using the
following data. See table 3-2.
Table 3-2. Data used in the Kinetic and Equilbrium
Analysis of the Histone-AntiHistone Interaction.
Ligand valency 64-1 arm per Star = 63
mol.wt.Histone octamer 109kDa
mol.wt. Star-PEG 700kDa
cone.histone in 0.22mM
5|iL flowcell
cone.Star-PEG in 0.22mM
5(iL flowcell
cone .antihistone in 0.093|iM
5(iL cell
mol.wt.antihistone
monoclonal
150kDa

sUod(sy UMW'iW'-'W
88
RU
TVno-
Window
AbsReso
SD
Slooe-
Baseline
ReJResp
Id
2ECS
50
135357"
010
oao
Yes.
a
Inject EDC/NHS.
7635
SO
13645.5
oor
-003
No
103.V
1M ethylenediamTne, pH 80
1233.5
50
133215
570
-018
No
3852
inject 1M ettianotarmne-. pH BO
1824.5
50
138757"
045
4X22
No
330.0
treeytated Stars.
2-04.5
50
141041
073
0.41
No
5674-
histones 1:10
23575
50
252753
1.S7"
-1.00
No
11733.1
histones bound
3134.5
50
25202.5
156
-073
No
116653
1M etoanoUmine; pH 80
35105
5Q
24242.4-
OB4-
045
Yw
10705r
aha monoclonal .
Figure 3-15.
Sensorgram.
AntiHistone-Histone Interaction

[eq. 3.01]
89
= Analyte m.w. (RJ(valence) = 136120.6
Ligand m.w.
The equilibrium affinity constant may now be calculated
based on the estimate
[eq. 3.02] KA = R^/C = 1036.5
R max R<\;
Now the rate constants k. and kd can be determined using
the calculated RBmx and the slope for histone binding.
[eq.3.03] dRA/dt = slope = 0.41 = ^CR^ (kaC+kd)RA
===>k = 3.10E-02 and k. = 2.99E-05
a a
Similarly for the binding of monoclonal antihistone antibodies to the
"histone stars". The R^was estimated from the antibody mol. wt., the mol.
wt. of histone octamers coupled to the Star-PEG and the response at
equilibrium for the AHA -"Histone- Star" interaction.
Rmax= Analyte m.w. (R^Cvalence) = 150kDa (25275.8X63) = 31565.5
Ligand m.w 63(109)+700
Applying equations 3.01 and 3.02 we obtained the kinetic and
equilibrium constants for AHA binding to the PEG immobilized histones, see
table 3-3 below.
Table 3-3. The kinetic and equilibrium constants for AHA binding to
the PEG immobilized histones.
Histone-tresvlated Star-PEG
5*
O
*
>
K
kd
^ max
9.65E-04 1036.5
3.10E-02
2.99E-05
136120.6
Anti-Histone-Histone Star-PEG
7^
O
7s
>
K
kd
2.81E-08 3.56E+07
1.5E+03
4.21E-05
31565.5

90
3.2.3 Surface Analysis of Surface-Modified Hollow
Fibers.
We used the transmission electron microscope to
estimate the size range of the Star-PEG polymers. Here a
Star-PEG solution (5|iL drop of 2mg/mL Star-PEG in 50% MeOH)
was evaporated on the grid then imaged by TEM see fig.3-16.
The diameter of these structures were found to be between 250
and 325nm. SEM and "non-contact mode" AFM were also used to
compare the chemically-modified PMMA with control hollow-
fibers. We found organized structures in the size range of an
histone-Star between 150 and 250 nm. It is speculated that a
slight reduction in the SEM image diameter has occurred due
to the heavy metal coating on the Star-PEG see fig.3-17. No
such artifacts occurred in the AFM images which clearly so
the presences of immobilized histone-Stars on the chemically-
modified hollow fiber see fig.3-18, 3-19, 3-20.
3.2.4 Summary.
A solid-phase binding assay was used to indirectly
determine an estimate for tresylate hydrolysis and reactivity
over a range of different pH values. It has been shown that
the stability of the leaving group is bimodal and has
greatest stability at pH 4.2 and 8.0. This is rationalized on
the basis that when HCL is used as the H+ ion donor the
released sulfonic acid group may become protonated at pH 4.2.
The loss of water and substitution by chloride, may be a
rapid Snl regeneration of the tresyl chloride which then
reactivates the terminal OH' s on the star. At pH 8.0 the
leaving group is less reactive simply because of the lower H+
ion availability, that is, in the absence of amines. The PEG-
tethered histones were shown to retain their
immunoreactivity. Kinetic and equilibrium analysis of a

91
BIACORE sensorgram generated during a multi-binding
experiment shows that the immobilized histones bind to
solution phase AHA with a KD of 2.81E-08. Microscopic analysis
of the surface-modified hollow fibers confirmed the presence
of organized structures in the size range of a polyethylene
glycol star. For the TEM and the AFM images it appears that
while all arms of the Star-PEG may not be attached to
histones, many of the globular clusters found in both imaging
techniques had both size and shape similarities to each
other. No such structures were found on the control hollow
fibers. From the SEM images of treated hollow fibers we also
found globular structures in the 150-250nm size range, but
suspect that reductions in diameter may be due to the heavy
metal coated used to prepare the specimen.
3.3 SPECIFIC AIM #3
3.3.1 Antihistone Adsorption Capacity of Surface-
Modified Polymethylmethacrylate and Polysulfone
Films.
Adsorption capacity experiments were performed to assess
the binding of antihistone antibodies (AHA) from hybridoma
supernatant solution. See fig. 2-7.
PMMA films where prepared according to the methods
described in section 2.31. PMMA films (13.0cm2) were treated
at four different levels of hydrazine-activation.
Ll = 69.0pg hr/cm2 polymer
L2 = 208.0|ig hr/cm2 polymer
L3 = 690.0|ig hr/cm2 polymer
L4 = 2081.Opg hr/cm2 polymer

92
Figure 3-16. Size Estimation of Star-PEG. a) The size
estimation of the Star-PEG was done using this TEM
image of an evaporated Star-PEG solution. The average
diameter was determined at the 19200 magnification to
be 260nm.

93
Figure 3-16. Size Estimation of Star-PEG. b
Star was observed at 80000 magnification. The
was determine to be 325nm.
A single
diameter

94
I 052. til
3/18/97 4:34:28 PM
1. 192nm histone Star
I 062.tif
3/18/97 4:34:12 PM
2. Hollow fiber lumen
I 072.tif
3/18/97 4:33:12 PM
3. 150nm histone Star
O 012.W
S118/97 4:39XPM
4. 56nm histone Star
Figure 3-17. SEM micrograph of the lumen of a treated
PMMA hollow fiber, a)Shows the lumen of a treated PMMA
hollow fiber, b)Shows an apparently collapsed Star at
70.OK magnification. c)This image shows a origanized
structure with a distinct globular shape in the size
range of a Histone-Star-PEG complex, d) More of these
strucures were found in the hollow fiber lumen and are
likely to be the histone star polymers under heavy metal
coatings. Note the reduced diameter of the "sputter
coated" samples (150nm).

95
Surface Histogram
Box Histogram
Image Extents: 2.00pm by 2.00pm
Area Analysis
Z-avg
Ra:
Rq:
Rp-p:
Skewness:
Kurtosis:
Image
0.31pm
0.05pm
0.07pm
0.39pm
-1.66
6.01
In box
0.36pm
0.01 pm
0.01 pm
0.05pm
-0.18
3.24
Box Dimensions
X:
Y:
dX:
dY:
0.91pm
0.54pm
0.37pm
0.65pm
0.32pm
Clear
Export.. |
OK
Erint
ASCII...
Figure 3-18. Atomic force micrograph scanned at 2.0pm
of a treated and untreated PMMA hollow fiber, a) Shows
atomic force micrograph scanned at 2.0pm of a treated
PMMA hollow fiber and its surface histogram.

96
1532
1226
919
613
306
0
O.OOnm 99.83nm 199.66nm 299.50nm 399.33nm 499.16nm
Bax Histogram
580 I
464-
348-
232-
116
0 1 1 | 1
O.OOnm 99.83nm 199,66nm 299.50nm 399.33nm 499.16nm
Figure 3-18. Atomic force micrograph scanned at 2.0jim
of a treated and untreated PMMA hollow fiber.b) Shown is
atomic force micrograph scanned at 2.0|rm of a untreated
PMMA hollow fiber and its surface histogram.
Image Extents: 2.00pm by 2.00pm
Area Analysis
Box Dimensions
Image
In box
Z-avg
321.13nm
463.83nm
X:
0.84pm
Ra:
11S.72nm
6.59nm
Y:
1,42pm
Rq:
131 64nm
8.1 Snm
dX:
0.38pm
Rp-p:
499.16nm
37.87nm
dY:
0.45pm
Skewness:
-0.43
-1.19
Kurtosls:
1.85
3.58
Z:
439.04nm
£lear
Erint
Surface Histogram

97
55.11nm
48.95nm-
42.80nm
36.6Snm-
30.50nm
24.34nm-
]
189.1 Onm
Trace Start Z
1 154.13nm 45.86nm
2 313.44nm 46 24nrn
:¡ 356.19nm 50.95nm
Select color to PLACE section cursor.
Length
82.89nm
71 24nrn
138.59nm
dZ
1 5lnm
0,38nrn
1 2.64nrn
Theta
1,04deg
0.30deg
5.21 deg
C Place Section
Statistics
C Statistics 2
( Section Analysis
Relative Angles
Clear AH
Clear Last
ASCII...
Export [
Print
OK
O.OOnm
I
378.21 nm
I
567.31 nm
7S6.41nm
945.51 nm
Figure 3-19. Atomic force micrograph scanned at 1.0|im
of a treated PMMA hollow fiber. Shown is the 1.0|i.m scan
and its cross sectional analysis features.

98
Figure 3-20. Atomic force micrograph scanned at 1.0|im
of a untreated PMMA hollow fiber. Shown is the 1.0|rm
scan and its cross sectional analysis features.
The AHA adsorption of the group of modified PMMA films
was determined by incubating each of them in hybridoma cell
culture supernatant containing approximately 29.44(ig/mL of
antihistone antibodies according to methods described in
section 2.3.1.

99
The total adsorption for the four levels of PMMA surface
modification were calculated and the adsorption capacity per
unit area of each of the treated membranes (see fig.3-21) was
determined as follows
Q1 = 0.840mg/m2 polymer Q3 = 1.806mg/m2 polymer
Q2 = 0.335mg/m2 polymer Q4 = 0.066mg/m2 polymer
PSF films where prepared according to the methods
described in section 2.31. PSF films (11.0cm2) were treated at
four different levels of chlorosulfonic acid-activation.
LI = control
L2 = 281 mg/gm polymer
L3 = 562 mg/gm polymer
L4 = 1124 mg/gm polymer
The AHA adsorption of the group of modified PSF films
was determined by incubating each of them in hybridoma cell
culture supernatant containing approximately 29.44(ig/mL of
antihistone antibodies according to methods described in
section 2.3.1.
The total adsorption at the four levels of PSF surface
modification were calculated and the adsorption capacity per
unit area was determined as follows:
Q1 = 0.0mg/m2 polymer Q3 = 1.04mg/m2 polymer
Q2 = 0.67mg/m2 polymer Q4 = 3.93mg/m2 polymer
See fig.3-22 on following page.

100
LI ADSORPTION
L3 ADSORPTION
0 1 2 3 4 S 6
Incubation #
L2 fiber adsorbed 0.335mg/m2 polymer
L4 ADSORPTION
Figure 3-21. Adsorption capacity of surf ace-modified
PMMA membranes.

101
L1 ADSCRPTICN L3 /sdscrpticn
L2 ADSCRPTICN
L4 ADSCRPT1CN
Figure 3-22. Adsorption capacity of chlorosulfonated
surface-modified PSF membranes.

102
Additional PSF films where prepared according to the
methods described in section 2.3.1. PSF films (13.0cm2) were
treated at four different levels of membrane amination by
nitration/reduction.
The AHA adsorption of the group of modified PSF films
was determined by incubating each them in hybridoma cell
culture supernatant containing approximately 46.2|ig/ml of
antihistone antibodies according to methods described in
section 2.31.
The results of the experiment showed that films
activated in the nitration reagent for 1, 3, 12, and 24 hrs
respectively adsorbed the following amounts.
Q1 = 1.844mg/m2 polymer Q3 = 2.29mg/m2 polymer
Q2 = 1.190mg/m2 polymer Q4 = 2.18mg/m2 polymer
See fig.3-23.
3.3.2 Extracorporeal Immunoadsorption of Circulating
Antihistone Antibodies from Saline Solution.
We used an in vitro model to produce and evaluate
therapeutically modified hollow-fiber dialyzers for
antihistone antibody extraction from saline. Whole histones
were covalently immobilized onto the inside or outside
surface of PMMA (Toray) B2-1.0H dialyzers using procedures
outlined in section 2.32.

103
Ll ADSORPTION
L4 ADSORPTION
L2 ADSORPTION L4 ADSORPTION
Figure 3-23. Adsorption capacity of nitration/
reduction surface-modified PSF membranes.

104
Solutions (230mL) containing mouse anti-histone
antibodies (1.0(ig/mL) were circulated through modified and
control dialyzers for 4hrs. The AHA concentration was
monitored in the circuit by ELISA. The resulting clearance of
AHA over time is shown in fig. 3-24.
Differences between peak and endpoint AHA concentrations
in the circuit were calculated and shown in table 3-4. The
inside treated dialyzer apparently removed all of the
available AHA in a single pass as shown by the flat line, no
detectable AHA eluted from the device during the course of
the experiment.
Table 3-
4. Total
mouse AHA removed
from PBS solution.
null(U/mL)
outside(U/mL)
inside(U/mL)
E ratio
0.0056
0.9168
0.6044
peak cone.
13.7
11.8
2.03
endpoint cone.
13.4
3.0
2.03
0.3
8.8

null(uq/mL)
outside(uq/mL)
inside(uq/mL)
E ratio
0.0056
0.9168
0.6044
peak cone.
0.9042
0.7788
1.0
endpoint cone.
0.8712
0.1980
0.1340
Total mass
7.59|ig
133.6|j.g
199.2|ig
removed

105
Effluent AHA
Concentration
U/ml
Extraction Ratio
Mouse Antihistone Immunoadsorption from Saline
Figure 3-24. Extracorporeal immunoadsorption of murine
antihistone antibodies from saline.

106
A study was done using the same devices after 1 week
storage to extract human antihistone antibodies from PBS. See
fig. 3-25. Total clearance results are shown in table 3-5
below.
Table 3-5. Total human AHA removed from PBS solution.
null(U/mL)
outside(U/mL)
inside(U/mL)
E ratio
0.1875
0.4259
0.4884
peak cone
9.823
4.926
7.267
endpoint
cone. 7.677
0.995
0.944
2.146
3.931
6.323
null(uq/mL)
outside(uq/mL)
inside(uq/mL)
E ratio
0.1875
0.4259
0.4884
peak cone
0.648
0.325
0.479
endpoint
cone. 0.507
0.066
0.062
Total mass 32.43)ig
removed
59.57(ig
95.91|ig
Three important observations can be noted: 1) Both of
the surface-modified dialyzers display a reduced extraction
ratio and AHA removal in the second immunoadsorption, this is
indicative of specific binding to the hollow fiber surface
since the binding sites are now apparently being depleted. 2)
The "inside" treated dialyzer again removals more AHA and at
a faster rate because of its greater surface area over the
"outside" dialyzer. 3) The null (control) dialyzer displays a
slight increase in extraction ratio and AHA removal during
the second experiment. This is indicative of non-specific
binding since the deposition (fouling) of solutes on to the
membrane surface after time can appear to create new binding
sites in an untreated device.

107
Human Antihistone Immunoadsorption from Saline
Effluent AHA
Concentration
U/ml
Extraction Ratio
Figure 3-25. Extracorporeal immunoadsorption of human
antihistone antibodies from saline.

108
3.3.3 Extracorporeal Immunoadsorption of Circulating
Antihistone Antibodies from Human Plasma.
Surface-modified commercial dialyzers (PMMA Toray* B2-
1.0H and PSF Fresenius@ F8) were treated as described in
section 2.32 and used to extract AHA from human plasma in the
in vitro model. The results of the immunoassays showed that
during the 4hr. experiment, the modified PMMA and PSF
dialyzers removed 3.17mg and 4.86mg of AHA from the plasma.
This was expected due to the larger surface area of the
F8(1.8m2) dialyzer compared to the B2-1. OH ( 1. Om2) .
Surprisingly, the unmodified (PSF) dialyzer removed 3.76mg
(Streicher E, and Schneider H 1983) during the same time
period; see fig.3-26.
This finding led us to suspect the selectivity of the
immunoassay and to check for interferences in the plasma
samples which may have produce "false positivies" in the
controls. Since human plasma contains numerous proteins which
may interfere with histone-antihistone binding two additional
experiments were done to try and resolve this problem.
Firstly, an elution experiment on the dialyzers was conducted
in wash buffer (PBS 0.05% Tween) and glycine buffer (2.5 pH).
The results of these elutions are shown in fig.3-27 and table
3-6.
Table 3-6. Total mouse AHA eluted from F8 dialyzers.
null(uq) modified(uq)
glycine 73.87 325.12
wash buffer. 959.33 80.99
Total mass
eluted
1033.2|ig
406. llfig

10 |-
16 -
14 -
2 -
0 -
8 -
6 -
-50
16
14 -
12 -
10 -
8 -
6 -
4 -
-5i
16
14
12
10
8
6
4
2
-5
109
Plasma Immunoadsorption of AHA
by Modified PMMA
Plasma Immunoadsorption of AHA
by Unmodified PSF
Plasma Immunoadsorption of AHA
by Modified PSF
0.6
0.4
02
0.0
02
0.4
0.6
0.8
-50 0 50 100 150 200 250 '00
Time mms.
V
Extraction Ratio
250 300
Time mins.
Figure 3-26. Extracorporeal immunoadsorption of murine
antihistone antibodies from plasma: preliminary study.

Elution of the
Umodfied F8 dialyzer
Unmodified F8 wash buffer elution Unmodified F8 glycine buffer elution
b)
Modified F8 wash buffer elution
Elution of the
Modfied F8 dalyzer
Modified F8 wash buffer elution
fraction #
Figure 3-27. Elution of AHA from a modified and a unmodified PSF dialyzer. a) Shows the
elution of AHA from a unmodified PSF dialyzer by wash and glycine buffers. Total AHA removed
equals 959.33 and 73.8|U.g respectively, b) Shows the elution of AHA from a modified PSF
dialyzer by wash and glycine buffers. Total AHA removed equals 80.99 and 325.12|ig
respectively.
110

111
Under these elution conditions more total AHA appears to
elute from the null.vs.the modified F8. This indicates that
AHA binding to the modified dialyzer is stronger and more
specific than to the non-modified F8. A majority of the
antibodies (93%) eluted from the nullF8 with only wash
buffer, while a much smaller portion (19.9%) elutes from the
modF8 with wash buffer. Therefore we might regard the glycine
eluate fraction as the more accurate reprensentation of AHA
binding. The percent specificity (%Sp) of each dialyzer can
be estimated as follows
%Sp = mass glycine fraction
mass of total eluate
For the nullF8 %Sp = 7.15% and for the modF8 = 80.0%.
This is consist with our claims that the modification produce
on these PSF dialyzers (i.e. immobilized histones) increase
the binding of the hollow fibers specificity toward
antihistone antibodies.
A second experiment was done in which the same plasma
samples were fractionated using a T-gel column. Each of the
fractions (5,10,30,60, etc...) were combined then assayed
again by AHA ELISA. The results below show that the sum of
the AHA concentrations for each of the 20 (5ml) fractions
collected, see fig.3-28a and 3-28b.

Plasma I mmunoadsorptl on null F8 Inlet
Plasma I mmunoadsorptl on modified F8 Inlet
b)
23 r 1 r
c
10 1 I 1 1 1- 1
-50 0 50 100 150 200 250 300
tl me ml ns
E
?
i
u
<
X
<
Plasma I mmunoadsorptl on null F8 outlet
t 1 1 1 1 r
3 l i I l i j
-50 0 50 100 150 200 250 300
time mins
Figure 3-28. T-Gel fractionation of plasma samples from preliminary study, a) Shows the
adsorption and extraction ratio of a unmodified PSF dialyzer after plasma T-Gel
fractionation. Total AHA removed equals 1.04mg b) Shows the adsorption and extraction ratio
of a surface-modified PSF dialyzer after plasma T-Gel fractionation. Total AHA removed
equals 4.39mg.
112

113
The dialyzer inlet and outlet concentrations were used
to calculate the extraction ratio throughout the course of
the experiment. The extraction ratio is given by the equation
Er= Ci-Co/Ci
when the extraction ratio of the null and the modifiied
dialyzers were compared the results indicate a positive Er for
the modified dialyzer at the beginning of the experiment
decreasing to zero toward the endpoint 240mins. This decrease
in Er indicates specific binding of AHA to the hollow fiber
surface and saturation after approximately 4hrs. In contrast,
the Er for the control dialyzer begins near zero and
approaches slightly more positive values toward the endpoint.
This increase in Er indicates non-specific binding to the
hollow fibers as if binding sites were apperently being
created rather than becoming depleted.
When the difference in peak values and the endpoint
donor reservoir AHA concentrations were calculated the net
removal of AHA from plasma was always found to be in favor of
the treated hollow fiber. See table 3-7.
Total mass removal from a 475ml plasma reservoir was
calculated and the average %AHA removed by the control
dialyzer was 18% while the average %AHA removed by modified
dialyzer modF8 = 34%
Table 3-7. Total AHA removed from human plasma
preliminary study.
outlet inlet E^
nullF8 4.73mg 1.04mg 0.0439
6.44mg 4.39mg
modF8
0.3811

114
3.3.4 Extracorporeal Immunoadsorption of Circulating
Antihistone Antibodies from Human Plasma by
Unmodified PMMA and F8 Dialyzer.
A series of AHA adsorption experiments were performed to
test the extraction of AHA by unmodified hollow fiber
dialyzers. Two types of membrane materials were used, a PSF
(F8 dialyzer) hollow fiber and a PMMA (B2-1.0H dialyzer).
Each of the dialyzers were tested in triplicate according to
methods discussed in section 2.33. Plots of the donor
reservoir AHA concentration versus time shows that little
adsorption of AHA occurred during the experiment for either
membrane under these experimental conditions. See fig.3-29a
and 3-29b (next page). The results are summarized in table 3-
8 below.
3.3.5 Extracorporeal Immunoadsorption of Circulating
Antihistone Antibodies from Human Plasma by Modified
PMMA and F8 Dialyzer.
A series of AHA adsorption experiments were also
performed to test the extraction of AHA by modified hollow
fiber dialyzers. Two types of membrane materials were used a
PSF (F8 dialyzer) hollow fiber and a PMMA (B2-1.0H dialyzer).
Each of the dialyzers were tested in triplicate according to
methods discussed in section 2.33. Plots of the donor
reservoir AHA concentration versus time shows that 40% of all
the AHA was adsorbed during the first 50mins. of the
experiment for either membrane (see fig.3-30a and 3-30b). The
results are summarized in table 3-8.

Anti-Histone Adsorption from Human Plasma
by an Unmodifed Bk 1.0H dialy zer
Figure 3-29a. The masa transfer curve shows the mean (n>*3)
AHA (outlet) concentration in U/ml (0.066 ug/U)
during a 4hr adsorption experiment.
Anti-Histone Adsorption from Human Plasma
by an Unmodifed BK1.0H dialyzer
Figure 3-29b. The mass transfer curve shows the mean (n*3)
AHA (inlet) concentration in U/ml (0.066 ug/U)
during a 4hr adsorption experiment.
Anti-Histone Adsorption from Human Plasma
by an Unmodifed F8 dialyzer
Figure 3-29c. The mass transfer curve shows the mean (n3)
AHA (outlet) concentration in U/ml (0.066 ug/U)
during a 4hr adsorption experiment.
Anti-Histone Adsorption from Human Plasma
by an Unmodifed F8 dialyzer
Figure 3-29d. The mass transfer curve shows the mean (n=3)
AHA (inlet) concentration in U/ml (0.066 ug/U)
during a 4hr adsorption experiment.
Figure 3-29. Extracorporeal immunoadsorption of murine antihistone antibodies from plasma
by unmodified dialyzers.

Anti-Histone Adsorption from Human Plasma
by a Modi fed BK1 .OH dialy zer
Figure 3-30a. The mass tranafer curve shows the mean (nm3)
fraction of initial AHA concentration in U/ml (0.066ug/U)
during a 4hr adsorption experiment.
Anti-Histone Adsorption from Human Plasma
by a Modifed BK1.0H dialy zer
Figure 3-30b. The masa tranafer curve ahowa the mean (n=3)
AHA (inlet) concentration in U/ml (0.066 ug/U)
during a 4hr adaorption experiment.
Anti-Histone Adsorption from Human Plasma
by an Modifed F8 dialyzer
Figure 3-30c. The maaa transfer curve shows the mean (n*3)
fraction initial AHA concentration in U/ml (0.066 ug/U)
during a 4hr adsorption experiment.
Anti-Histone Adsorption from Human Plasma
by an Modifed F8 dialyzer
Time mins.
Figure 3-30d. The maaa transfer curve ahowa the mean (n=3)
AHA (inlet) concentration in U/ml (0.066 ug/U)
during a 4hr adaorption experiment.
Figure 3-30. Extracorporeal immunoadsorption of murine antihistone antibodies from plasma
by surface-modified dialyzers.

117
Table 3-8
a modified
. Total AHA removed from human plasma by
dialyzer.
Max.Adsorbed
Avq.Adsorbed
Peak-endpoint
null Toray
n/d
<0.Omg
4.14mg
mod Toray
7.Olmg
5.01mg+l.08
n/d
null F8
n/d
n/d
0.397mg
mod F8
17.4 9mg
6.68mg+3.17
n/d
3.4 Summary
We have
determined that
the surface
immobilization
antigens onto
commercial
hollow fiber
dialyzers is
efficient method of preparing highly selective bioadsorbents
for the ex vivo extraction of autoantibodies from bodily
fluids. The effects of these surface-modifications can
furthermore, be demonstrated in an in vitro model. The use of
polyethylene glycol spacers in the immobilization procedure
significantly increases the surface area of the hollow
fibers, it apparently stabilizes the epitope recognized by
the autoantibodies and thereby increases binding. In addition
it prevents non-specific binding or fouling of the surface.
As shown in table 3-8 milligram quantities of immunoglobulin
can potentially be removed from complex mixtures and thus
therapeutic application of the devices are worthy of further
consideration. As shown in figure 3-26 a)the adsorption and
extraction ratio of a surface-modified PMMA dialyzer. Total
AHA removed equals 3.17mg. In b) the adsorption and
extraction ratio of a unmodified PSF dialyzer. Total AHA
removed equals 3.76mg and in c)the adsorption and extraction
ratio of a surface-modified PSF dialyzer. Total AHA removed
equals 4.86mg.

CHAPTER 4
CONCLUSIONS
Therapeutically modified hollow fiber dialyzers may be
used to treat a variety of immune or metabolic diseases. The
advantage of using this new approach to extracorporeal
therapy is that 1) the starting materials (dialyzers) are
safe and commercially available, 2) the method allows for the
extraction of the blood-borne molecule without prior plasma
separation. 3) the components are sterile, inexpensive and
disposable (one-time usage). 4) dialyzer modifications may be
individualized for the unique conditions of the patient on
short notice.
Adsorption capacity of the devices is not independent of
ligate concentration, however, but the efficiency of the
separation does increase with serum content. Thus, there may
be a therapeutic limit to which extraction can be optimized.
The orientation of the ligand molecules on the membrane is
responsible for differences between theoretical binding
activity (e.g. IgG's 2:1 mole ratio) and the actual binding
capacity. This occurs because the ligands are randomized
during the coupling procedure. As a result the epitopes of
some of them become inaccessible to the antibody. The random
orientation of immobilized antigens (histones) may be kept to
a minimum by the appropriate use of bifunctional spacers such
as gluteraldehyde or polyethylene glycol (PEG). Variations in
antivation site density on the hollow fibers may affect the
efficiency of adsorption also. The optimum amount of antibody
118

119
to be immobilized on an adsorbent is primarly determined by
the molecular size of the antigen. For the adsorption of
immunoglobulins the antigen tethered to a flexible polymeric
chain such as PEG has an enhancement effect on adsorption.
PEG also hydrophilizes the membrane so that non-specific
binding is reduced.
We have shown that an in vitro model can be constructed
which mimics the extracorporeal circulatory system used in
conventional as well as novel dialysis therapies. The model
was used to evaluate the clearance of a select donor solute
or a group of solutes under both equilibrium or sink
conditions. Passive adsorption of p2-M to PMMA membranes was
conclusively demonstrated in the model. New dialysate
formulations may be studied in the model as well. Here, we
found that dialysate additives which formed stable complexes
with a specific donor solute (i.e. antibodies) may increase
the clearance of that solute above and beyond steady-state
concentrations. For a summary of the experimental mass
transfer results, see table 3-1.
A solid-phase binding assay was used to indirectly
determine an estimate of tresylate hydrolysis and reactivity
over a range of different pH values. It has been shown that
the stability of the leaving group is bimodal and has
greatest stability at pH 4.2 and 8.0. This is rationalized on
the basis that when HCL is used as the H+ ion donor the
released sulfonic acid group may become protonated at pH 4.2.
The loss of water and substitution by chloride, may be a
rapid Snl regeneration of the tresyl chloride which then
reactivates the terminal OH' s on the star. At pH 8.0 the
leaving group is less reactive simply because of the lower H*

120
ion availability, that is, in the absence of amines.
The PEG-tethered histones were shown to still be
immunoreactive. Kinetic and equilibrium analysis of a BIACORE
sensorgram shows that the immobilized histones bind to
solution phase AHA with a KD of 2.81E-08. Further analysis of
the surface-modified hollow fibers confirmed the presence of
organized structures in the size range of a polyethylene
glycol star. For the TEM and the AFM images it appears that
while all arms of the Star-PEG were attached to histones,
many of the globular clusters found in both imaging
techniques had both size and shape similarities to each
other. No such structure were found on the control hollow
fibers. From the SEM images of treated hollow fibers we also
found globular structures in the 100-200nm size range, but
suspect that reductions in diameter may be due to the heavy
metal coated used to prepare the specimen.
During the extracorporeal immunoadsorption experiments
three important observation can be noted: 1) Both of the
surface-modified dialyzers display a reduced extraction in
the second immunoadsorption, this is indicative of specific
binding to the hollow fiber surface since the binding sites
are now apparently being depleted. 2) The "inside" treated
dialyzer again removals more AHA and at a faster rate because
of its greater surface area over the "outside" dialyzer. 3)
The null (control) dialyzer displays a slight increase in
extraction and AHA removal during the second experiment. This
is indicative of non-specific binding since the fouling of
the membrane surface can appear to create binding sites in an
untreated device. We have determined that the surface
immobilization of commercial hollow fiber dialyzers is a
potentially efficient method of preparing of highly selective

121
bioadsorbents for the ex vivo extraction of autoantibodies
from bodily fluids. The use of polyethylene glycol spacers in
the immobilization procedure significantly increases the
surface area of the hollow fibers, it apparently stabilizes
the epitope recognized by the antihistone antibodies and
thereby increases binding in additional its prevention of
non-specific binding to the surface.
As shown milligram guantities of immunoglobulin can
potentially be removed from complex mixtures (human plasma)
and thus therapeutic application of the devices are worthy of
further consideration.

APPENDIX I
MASS TRANSFER IN HOLLOW FIBER DIALYZERS
The removal of uremic toxins from the bloodstream by
conventional dialysis follows three mechanisms.
1. Small molecules <5kD diffuse from the blood into dialysate
through a membrane, along their concentration gradient,
according to Fick's laws. The diffusive permeability of many
membranes may be defined as the ratio between the diffusion
coefficient and membrane thickness, e.g. A 16mm cuprophan
membrane with a 1.6m2 surface area is found to clear urea
equally with a 5mm membrane with only 0.5m2 of surface area.
Thus the use of dialyzers with ultra thin membranes of this
type make it possible to efficiently clear blood solutes with
smaller surface areas. This has led to the manufacture of
high-efficiency dialysis from even low-flux membranes and
with Kuf>10ml/torr*hr
2. The second mechanism of transport is convection. This mode
of transport depends only on the pressure difference between
blood and dialysate compartments (TMP) and has its greatest
effect on the transport of "middle molecules of molecular
weight >5000kD
3. A third mechanism is adsorption. Blood proteins, uremic
toxins and drugs administered during dialysis can be
significantly adsorbed to the surface of dialysis membranes
of synthetic origin. Adsorption may take place by
electrostatic or by van der Waal forces which may have
significant impact on dialyzer performance (Reihanian
122

123
Robertson and Michaels 1983). While the adsorption of certain
proteins may be beneficial (i.e. C3a, C5a and others) the
adsorption of others is clinically undesirable as is the case
with EPO e.g. Several dialysis membranes remove (32-M not only
by means of filtration but by adsorption to an hydrophobic
surface. As a result (32-M can be cleared from the blood by
low-flux (PMMA). The sieving coefficient (SC) for a given
protein in hollow-fiber dialyzers is expressed as:
SC=Cf/(C.+Cv)
where C£ = cone.filtrate Note: Low concentrations in the
C. = cone, venous filtrate due to adsorption give rise
C, = cone, arteries lower SC value's.
As the adsorption sites for the protein are saturated it
will begin to pass through the membrane and can be detected
in the filtrate if the pore size is large enough. A detailed
analysis of the adsorption capacity of different dialyzer
membranes for (32-M been carried out by Klinke et al. 1989.
The overall efficiency of a dialyzer is determined by
two independent factors: 1) the ratio of the flow rates of
the two solutions and 2) the rate constant for solute
transport across the membrane. The rate constant may be
determined by membrane material and its dimensions; fluid
channel geometry; as well as local fluid velocities.
Schematic presentations simplify the flow patterns and solute
concentrations in a dialyzer. For a typical dialyzer
configuration see fig. 1-1. In the diagram, Q represents the
volumetric flow rates (mL/min) and C(mol/L) the solute
concentrations. Since mass is conserved, the rate of mass
transfer dm/dt is obtained
dm/dt= QdiCdi- QdoCdo = QboCbo-QbiCbi

124
The efficiency of a dialyzer is commonly expressed in terms
of its dialysance D, where D = dm/dt (1/Cbi-Cdi)
The extraction ratio E = D/Qb
and the clearance CLB = dm/Cbidt. Note that when Cdi = 0
clearance and dialysance are equal.
A Double Pipe Matt Exchanger
with Semipermeable Wall
Olaly >a|t Qdj Caj
i n
out
Q(jo ^do
Flowing
filocnj
\\\\\\\\\^|\\\\\\\\Wq Semipermeable
Membrane
Watt Product
Water
lowing Dlalyiale
Eiciu Eleclrolyle
Nllrogenout Metabolite
Trace Solute f
Figure 1-1. Hemodialysis Configuration.
The relationship between dialysance and the mass transfer
parameters; It may be assumed that the local overall mass
transfer coefficient, Kc is constant along the length of the
dialyzer and it therefore may be calculated from the sum of,
the mass transfer resistance of the membrane 1/P, and the
resistance of the two boundary layers 1/Kb and 1/Kd
1/K = 1/P + 1/Kd + 1/Kb
Note: Some important assumptions are made here first that the
flow rates are constant in each of the two compartment,

125
secondly that K0 represents a length average transfer constant
that holds for any changes in Kb and Kd along the length of
the dialyzer.
Let dm/dt = M*
Then the change in the rate of transport of a given solute
across a membrane per change in area dA; is described by
Fick's Law as
dM* / dA = K0 (Cb-Cd)
which may also be written dM* = K0(Cb-Cd)dA .
Since M* = dm/dt = QdiCdi QdoCdo = QboCboQbiCbi
The mass balance equation may be written in differential form
as
QBdCb = dM* = QDdCd Where
Qb = average of Qbo and Qbi
dCb = CboCbd
Qd = average of Qdo and Qdi
dCd = CdoCdi
If we let the concentration difference equal
dC = Cb-Cd or dCi-dCo and upon differentiating we get
d(dC) = dCb-dCd. Substituting we get d(dC)= dCb-dCd
= dM*/QB-dM*/Qd and that d(dC) = dM*/(QB-QD). Integrating
this equation
£ d(dC) = EdM*/(QB-QD) so that
dC = M*/ (Qb-Qd)
inserting this relation into the mass flow equation
dM* = K0(Cb-Cd)dA we get d(dC) = -dM*/(QB-QD) and
d(dC) = -K0(Cb-Cd)dA/ (Qb-Qd)
by dividing this equation through by d(dC) and integrating
across the dialyzer surface area we get
Ld(dC)/dc = £K0dA/(QB-QD)

126
equals In dC = -K0A/(QB-QD)=ln(CB/CD) or ln(Ci/C0)
Recall that d (dC ) = -dM*/(QB-QD) to integrate this
equation yields Z d(dC) = £ -dM*/ (QB-QD)
so dC = -M*/(Qb-Qd) = dCi-dCo, dividing by ln(Ci/C0) we get
-M*/ln(Ci/C0) (Qb-Qd) = dCi-dC0/In(Ci/CG)
we have from our previous relationship
M*/K0A(Qb-Qd)/(Qb-Qd) = dCi-dC/ln(Ci/C0) so that
M* = KoA dCi-dCo/ ln(Ci/C0) where the term (dCi-dC0)/ln(Ci/C0)
is called the log-mean concentration difference (Cm). The
equation M* = K0ACm holds for different flow geometries by
adjusting for the appropriated Cm. See table 1-1.
Table 1-1. Mass transfer parameters for different flow
geometries.
Flow Geometry
Cocurrent
0. *
0..
Couiitcicurrent
0
< Q.
Well mixed dialysa'c
_1_
1
0.
Cross-flow
~ir
E Sl NwM
0. T 0.
Estimation or Dialtsance
E l-ep[-NT UZ
E 5iC NliirZ)|zl
eiplNjd-ZiJ-Z
If Z l. E -
Nt + 1
l-exp|-NT)
WZrr-ep(-N,)l
£ rr Z[s.iN,)s.(Ntz)l
*T
w|iere S,(X) l-exp(-X)£
z_2
0.
Calculation of Overall
Mass Transfer Coefficient
from Experimental Data
In
I Qk.
All t Z) "' | 1 E(l
r 0. I -EZ
A "| 1 -E(ltZ)
I + Zll
CLm may be approximated by (C0-CD)/ln(CB/CD)
= (dCB-dCD)/ln(dCB/dCD) similarly Cto may be approximated by
(Ci-CJ/lnfCi/C,,) = (dCi-dCo) / In (dCi/dC,,)

The
127
Volumetric flow rate due to ultrafiltration:
ultrafiltration flow rate is given by
Qf = Qbi-Qbo = QdiQdo
when the ultrafiltration flow rate Qf is negligible as is the
case when little or no water is removed from the patient,
then Qf = 0 and Qbi = Qbo and Qdi = Qdo. Furthermore, the
definitions of clearance and dialysance take on simplier
forms.
CLb = QB(Cbi-Cbo)/Cbi = QD(Cdo-Cdi)/Cbl
Db = Qb(Cbi-Cbo) / (CbiCdi) = QD(Cdo-Cdi)/(Cbi-Cdi)
The effect of ultrafiltration on blood clearance CL^;
CLb = Qb (CbiCbo )/Cbi + QfCbo/Cbi
Since Cbo < Cbi the effect of Qf on CLB is always less than Qf
itself. Qf rarely exceeds 10mL/min so its impact is usually
negligble on the clearance of low molecular weight solutes.
However, the relative effect on the transport of high m.w.
solutes can be substantial.
The four most common contact geometries in dialysis
circuits;
1. If blood and dialysate flow in the same direction the
geometry is parallel or co-current.
2. If blood and dialysate flow in opposite directions the
geometry is counter-current.
3. When the blood and dialysate flow at right angles to each
other the geometry is cross-flow.
4. when the dialysate bath is being "stirred" in all
directions and the bulk dialysate solute concentration is
kept uniform the term "well mixed" geometry applies.

128
The log-mean concentration differences for the four most
common contact geometries in dialysis circuits are as
follows. Where Cim = (Ci-C0)/ln(Ci/C0) = (dCi-dC0)/ln(dCi/dC0)
For co-current and cross-flow geometries
dCi CbiCa and dC0 Cbo Cao
For counter-current flow geometries
dCi CbiCao and dC0 Cbo Cdi
For well-mixed flow geometries
dCi = Cbi Cdo and dCc = Cbo-Cdo
The dimensionless parameters E, N^, and Z:
The extraction ratio, E is often more meaningful than
dialysance alone. It represents the fraction of the maximum
solute concentration removed from the blood when its
concentration in contact with dialysate is equal to the inlet
concentration and constant across a membrane of infinite
area. The number of transfer units Nt, is a measure of the
mass transfer size of the dialyzer. Plots of extraction ratio
vs. the number of transfer units for each geometry can be
used to estimate dialyzer performance. From the curves it can
be seen that with all other parameters fixed the extraction
ratio decreases in the following order.
counter-current flow > cross-flow > co-current flow > well-mixed
dialysate.
The number of transfer units Nt provides a measure of the
"mass transfer size" of the dialyzer. Z, is the blood to
dialysate flow rate ratio it is important to note that as Qd
greatly exceeds Qb each of the expressions for the extraction
ratio reduce to the expression E = l-exp(-Nt). When this
equation applies the dialysate concentration is virtually
constant throughout the dialyzer and is equal to its inlet
value. Even under conditions in which it does not rigorously

129
hold this relationship is useful for estimating the maximum
possible dialysance for specific values of K0, A, and Qb.
For further information on the mathematics and
derivation of the above equations see Colton and Lowrie
(1981).

APPENDIX II
TUBING SELECTION
There are many factors which may determine the optimum
selection of tubing for any pumping application. When
selecting the best tubing for the work it is important to
consider fluid compatibility, motor rpm, fluid temperature,
duty cycle, system pressure, and maintenance pressure. Also
further consideration of the following characteristics is
also recommemneded; tubing, gas permeability, clarity,
flexiblility, purity, and cleaning/sterilization techniques.
All of the tubings used or considered in the above
experiments were manufactured according to GMP (Good
Manufacturing Practices) and are composites of formulations
approved in the Masterflex* Tubing Selection Guide.
The chemical compatibility of each tubing material has
been indicated by five general ratings. All ratings in the
chemical compatibility table A2-1 were determined at 70F.
Higher temperatures will decrease the chemical resistance of
all of the listed tubing materials.
Ratings for Tubing Compatibility
A = Excellent; No effect
B = Good; minor effects, slight corrosion or discoloration
C = Fair; not recommended for continuous use, some loss of strength
D = Not recommended for any use; Severe effects
- = No data
130

131
LEGEND. Tubing or Pump Head Material.
PN = PharMed*, Norprene* CF = C-FLEX*
S = Silicone T = Tygon*
TS = Tygon special V = Viton*
PSF = Polysulfone
PPS = Polyphenylene sulfide
PC = Polycarbonate
SS = Stainless steel
TABLE II-1. Tubing Selection and Compatibility
FLUID
TUBING
PUMP
HEAD
MATERIAL
PN
CF
S
T
TS
V
PSF
PC
EPS
Acetic acid, 5%
A
C
B
B
D

A
C
A
Acetic anhydride
A
-
D
D
D
D
D
D
A
Acetone
D
-
D
D
D
D
D
D
A
HCL(dil)
A
B
D
A
B
A
A
A
A
HCL(med)
-
B
D
A
B
B
A
D
A
HCL(cone)
-
B
D
A
A
B
A
B
A
MeOH
A
-
A
C
C
D
A
B
A
Nitric acid(dil)
A
A
C
A
A
A
A
B
A
Nitric acid(med)
C
-
D
A
B
A
C
C
-
Sulfuric acid (dil)
A
A
D
A
B
A
A
A
A
water
A
A
B
A
A
A
A
A
A
Adapted from MasterFlex Tubing selection table. Barnant Co.
Barrington, IL 60010 USA

APPENDIX III
A LEAST SQUARES REGRESSION METHOD
Least Squares Regression using matrix derived parameters
for the linear equation y = mx + b can be applied to any
pairwise data (x,y).
Step 1. Tabulate the data values as shown
X
y
x2
y2
xy
























Ex
Ey
Ex2
£y2
Exy
define the operator
5 =
Ex2 Ex
Ex n
then the slope
m = 1/5
Exy
Ex
Ey
n
and the y-intercept
b = 1/5
Ex2
Exy
Ex
En
Step 2. Plot the line through the pairwise data to obtain a
"curve fitting' estimate or use the equation to calculate
unknown quantities x or y.
132

133
APPENDIX IV
ANTIHISTONE ANTIBODY HYBRIDOMAS
Characterization of Antihistone Antibody Producing WEHI
Hvbridomas: The following procedures were used to obtained
antihistone antibodies from hybridomas cell lines.
1. WEHI hybridomas were obtained in a modified RPMI-1640
media which had been cultured for at least one week.
2. Two GIBCO media were purchased for cell culture transfer
and screening assays i.e. H-SFM (serum free) and PFHM-II
(protein free). These media transfers will simplify our assay
of the monoclonal antibodies by ELISA and UV spectroscopy
upon isolation from the suspension.
3. Place in each of two 175mm cell culture flasks 225ml of H-
SFM and 225ml of PFHM-II media (total 4 flasks). Put the
flasks in 5% C02 incubator. Allow them to environmentally
equilibrate for 4hr.
4. Remove 0.5ml of each hybridoma cell line and count cells
on a hemacytometer. Record #cells/ml.
5. After 4hr remove 25ml from each hybridoma and add to the 2
flasks of each media (asepsis technique) to achieve a 1:10
split as follows.
Media
PFHM-II
H-SFM
Hybridoma#l
25ml/225mL
25ml/225mL
Hybridoma#2
25ml/225mL
25ml/225mL

134
Antibody Screening of the Hybridoma Suspension
Take the remaining hybridoma suspension and spin-down the
cells at lOOOOxg for 15 mins.
Remove the supernatant and store at -20C. Resuspend the
pellet in cell culture freezing medium (CCFM) at a cell
density of 8-40xl06 cells/mL. Store at -20C
5. Take enough cell culture from each of the splits (flasks)
to make a 5mL suspension at i) 50 cells/mL and ii) 5
cells/mL. These will be used in antibody screening assays.
You should have (8) 5mL screening suspensions as follows
Media
Hybridoma#l
Hybridoma#l
PFHM-II
cell densities
cell densities
5cells/mL, 50cell/mL
5cells/mL, 50cell/mL
H-SFM
cell densities
cell densities
5cells/mL, 50cell/mL
5cells/mL, 50cell/mL
Place 200(iL of each screening suspension in 12 wells of a 96
well microtiter plate. (12x8 = 96) Incubate for 7-10 day in
5% C02 incubator. Then remove and observe cell clusters in
each of the viable wells. Remove 100|il from each well and
assay for antihistone antibodies by ELISA. Identify the
positive wells and expand into 24 then; 6 well plates and
finally in 175mm flask.
Preparation of MAb Suspension
Thaw frozen suspension at room temperature for 2hr then
centrifuge (bench top) for 5min. Filter supernatant through
0.5mm syringe filter, sample filtrate. Filter again through
0.2(im filter, sample filtrate. Assay the sample total protein

of the sample at 280nm. Assay the antihistone activity by
ELISA. These antibody protocal were adopted from [Hardy
1986].

APPENDIX V
STAR-PEG POLYMERS
Star-shaped polymers consists of a central core which
radiates a given number of chains or branches. Recently,
there has been increasing interest in star-PEG molecules due
to their compact structure and high segment flexibility.
Star-PEG molecules immobilized onto surfaces are more
effective at preventing protein adsorption than linear PEG'S.
In addition, Star-PEG molecules have more end groups than
their linear counterparts and therefore a greater number of
attachment sites for other molecules or to surfaces upon
which PEG grafting is desired. For many of these uses it is
necessary to employ a mono-dispersed sample of well
characterized polyethylene glycol star's.
Until recently the only Star-PEG molecules available
were those synthesized by a "living" anionic polymerization
using a polydivinyl benzene (PDVB) core. The cores were
synthesized first and the resulting star-PEG's were highly
polydispersed. It is now believed that this high
polydispersity is caused by variable growth rates and
functionality amongst the cores. In order to obtain mono-
dispersed samples of star-PEG molecules, researchers are
currently studying a variety of synthetic methods to prepare
these and other similarly useful materials.
136

137
The current "state of the art" of synthesizing PEG star
molecules includes the following approaches. See table V-l.
Table V-l. Synthetic Methods for Star-PEG Polymers.
I CORE-FIRST METHOD
poly(DVB) core
trifunctional initiator cores
dendritic PEG
II ARM-FIRST METHOD
plurifunctional electrophile deactivators
copolymerization with a dialkene monomer
III SEED-STAR METHODS
"three-step" polystyrene seed
"PAMAM dendrimer seed"
I CORE-FIRST METHOD
Poly(DVB) core; Stars that are made of polyvinylic arms
and whose synthesis is based on the core-first method are
often difficult to achieve because of the formation of
linear, mono and difunctional by-products. Some of the more
successful syntheses have been associated with the use of
multi-carbanionic initiators. These polyfunctional species
can be formed by reaction of DVB with Bu+Li' or (Naph)~K+ (see
fig.V-l).
Figure V-l. Core-First Method using poly(DVB).

138
The stars generated by this method are unfortunately
ill-defined with large fluctuations in the number of branches
which characterizes such stars.
In the poly(DVB) core method, a tight cylindrical
reactor is fitted with an argon inlet, magnetic stirrer and
temperature control and is connected with ampules containing
dry THF, DVB solution, ethylene oxide and the chosen
initiator. The synthesis of the cores is performed at -30C in
dilute THF by slowly dripping the required amount of DVB in
THF into the initiator solution. The mole ratio [DVB]/[M+]
should be kept between (1-2.5) After, the formation of the
cores, a calculated amount of oxirane is introduced into the
reaction medium at -30C then the temperature is raised to
room temperature. If the initiation reaction is fast the size
distribution of the PDVB branches does not noticeably effect
the polydispersity of the star polymer. It has been shown,
however, that the cores resulting from anionic polymerization
of small amounts of DVB under well defined conditions are
nontheless very broad and this is the major drawback of core-
first star polymer syntheses.
Trifunctional initiator cores; Trimethylolpropane (or
other trifunctional initiators) can yield three-arm star-
PEG's after deprotonation of its three alcohol functions and
polymerization of ethylene oxide. For an example of star
macromolecules obtained by trifunctional initiators, see fig.
V-2 below.
H,
c /.PSCH.-C-CI
ru.
CM,
CM,
Figure V-2. A trifunctional initiator: triisobutyl
benzene chloride.

139
Dendritic PEG: The synthesis of PEG dendrimers has been
reported recently [Six and Gnanou 1994]. The formulation of
PEG dendrimers relies on "living polymerization" techniques
as well, each generation requires two steps.
1. Growth of PEG branches from a substrate molecule which has
a precise number of alcoholic functions.
2. Modification of the PEG branch endings to introduce
reactive groups which will continue the growth of further
generations.
Although minimal, a certain fluctuation in size of the
arms generated still cannot be avoided even in this case, see
fig. V-3 below.
Et-CfCHjOH)
HCPh2,K4
Et-CjCHjO-.K4)
2nd generation
Figure V-3. A Dendritic PEG star by anionic
polymerization.

II ARM-FIRST METHOD
140
Plurifunctional electrophile deactivators; Molecules
with star-shape and precise numbers of branches can be made
by the deactivation of "living" mono- or di- carbanionic
precursors using a multifunctional electrophile. Chlorosilane
based compounds have been found to selectively deactivate
"living" polystyrenes and polydienes. Star molecules with 6,
12, and 18 branches have been prepared by this method using
reagents containing 2-, 4-, and 6-trichlorosilane functions,
respectively. Star polymers with arms consisting of two
different blocks have also been formed by the same strategy.
See fig. V-4.
e
J + scu
: ¡¡
Figure V-4. Plurifunctional electrophilic deactivator.
Copolymerization with a dialkene monomer; Copolymerization of
"living" carbanionic precusors with a di-unsaturated molecule
such as DVB or ethylene dimethyl acrylic ester, also results
in a star-like architecture. See fig.V-5.
Figure V-5. Stars by copolymerization with a
dialkenyl monomer.

141
A procedure which is easier to perform than the one
described above, exists but yields products that exhibit a
greater heterogeneity. The size of the cores generated during
copolymerization and therefore the number of chains linked to
it cannot be strictly controlled, mainly because the
formation of such tightly crosslinked bodies always involves
significant fluctuation. A number of factors are found to
affect the average dimensions of nodules; l)the size of the
"living" precursor, 2)the amount of the dinvinyl compound
with respect to the precursor, and 3)the overall
concentration of the medium.
Ill SEED-STAR METHODS
"Three-step" polystyrene seed method: The three-step
(in-out) method uses a polystyrene seed-star as an initiator
for the subsequent formation of star-PEG. Here the seed-stars
themselves arise from a classical "arm-first" process. This
method allows a better control over the functionality of the
cores.
The first step is the preparation of a linear living
polystyrene (PS) using a mole ratio of monomer to initiator
chosen so as to get the appropriate molecular weight
oligomer, usually <10% of the star-PEG. This living precursor
is then used to initiate the polymerization of a known amount
of DVB (arm-first). The synthesis of the living star-
polystyrene/polyDVB core should be carried out around -30C,
under efficient stirring; it is usually complete within a few
minutes. An aliquot of the reaction medium may be sampled for
the purpose of characterization. Although the average
functionality of the cores is not predictable, it has been
established that the fluctuations in core size and

142
functionality within a sample are far lower than poly(DVB)
core-first prepared structures. Each of the resulting seed-
stars contains as many carbanionic sites as there are
branches surrounding it. These sites are used to initiate the
polymerization of oxirane, in which an equal number of
branches (arms) of PEG are grown. The third step involves the
growth of PEG chains from the "living" cores of the PS star
molecules, from this point on the experimental procedure is
very similar to those used in the core-first methods. If the
polystyrene branches are very short, while the PEG branches
are chosen to become relatively large each branch of the
resulting star-shaped macromolecule is composed of an inner
PS oligomer attached to an outer PEG polymer chain. The
molecular weight of the resulting star is given by the
equation.
MW = MWps
See fig.V-6.
Centra] core
Branching points
Extent] functions
Branches
Figure V-6. Seed Star Methods.

143
"PAMAM dendrimer seed"; In the PAMAM methods dendrimers serve
as the cores and modified PEG are attached as arms to form
novel star molecules. The dendrimer is a dense, hyperbranched
polymer built up generation by generation. Polyamidoamine
(PAMAM) starburst dendrimers are also manufactured by
Dendritech* Inc. of Midland. Mich. Their synthesis results in
an approximately spherical molecule [Tommalia, DA et al 1990]
Starting with a center having four primary amines the
molecules are built up generation by generation; with each
successive generation having twice the numbers of primary
amines as the preceding generating. These primary amines
serve as points of attachment of the PEG chains that are to
be "arms" of the star (see fig.V-7).
H,N
Figure V-7. Preparation of dendritic polyamidoamine
(PAMAM) polymers.

144
Several different types of bifunctional PEG molecules
can be made at the outer ends of the PEG (see fig.V-7)
(i) HO-PEG-NHS () t-BOC-PEG-NHS (iii) PEG-(Tres).
Dendrimers of generation 2-6, having surface amino
groups ranging from 16 to 256 in number, can be reacted with
different PEG molecules to create star molecules varying in
size and functionality (see table V-2).
Table V-2. Efficiency of Dendritic Seed polymer
in Star-PEG Synthesis.
PEO Type
Mpco
Note
Dendrtner
fincbonafty (fa)
Note b
StvMn
found
Note c
P4
Note c
fa
Note d
MeO-PEO-ChtaCHiCONHS
5000
16
88700
1.09
17
MeO-PEO-CHsCHiCONHS
5000
32
161000
1.08
31
MeO-PEO-CHaCHsCONHS
5000
64
268000
1 01
51
MeO-PEO-CHsCPbCONHS
5000
128
496000
1.06
94
MeO-PEO-CJ-bCHiCONHS
5000
256
778000
103
144
HO-PEO-CHjCHiCONHS
4000
16
76800
1.15
18
HO-PEO-CHjCHjCONHS
4000
32
121000
1.2
29
HO-PEO-CH2CH3CONHS
4000
64
235000
1.24
56
HO-PEO-CH2CH3CONHS
4000
128
416000
1.05
97
HO-PEO-CH2CH3CONHS
4000
256
496000
1.04
112
t-boo-PEO-CHjChbCONHS
3400
64
177000
1.05
48
Notes: a)MPI0 is mol. wt. b)fd is the #amino functions
c)Mn is star polymer mol.wt. d)fs is star functionality
f.=Mn-Md where M^ is the dendrimer mol.wt. Adapted from
Yen and Merrill (1997).
As the number of surface amino groups increases to
values greater than 32, the molecular wt. of the star formed
is systematically lower than would be expected. Apparently
steric hindrances preclude the PEG'S from reacting with all
of the amines on the dendrimer, so that the production of
Star-PEG with up to 14 0 arms was found to be the maximum
limit acheived see Yen and Merrill Polymer Preprints v38
(1997) p351.

APPENDIX VI
NON-MODE AFM: BURLEIGH METRIS NC-2000
Much of the current research in improving the
performance of blood-contacting devices utilizes some type of
surface modified synthetic material. Different strategies
have to be followed for creating surfaces with enhanced
hemocompatibility e.g. immobilization of bioactive ligands,
grafting of hydrophilic or hydrophobic polymers onto the
surface, as well as designs which include polymeric
microdomains and/or molecular imprints.
In order to achieve a more complete understanding of the
molecular-level interactions between foreign materials and
biological systems, surface properties and structure must be
accurately elucidated. The surface structure of a material is
often reactive; even mobile in some cases, such that the
components of the surface reconstruct themselves in response
to the local environment. For this reason, highly sensitive
analytical techniques are required to provide a means to
understanding these data. See fig.1-13.
The simplest picture of a sample surface is provided by
the optical microscope. The sensitivity of this instrument
may be extended by the use of specialized methods which
include polarized microscopy, fluorescense and confocal
microscopy. A more detailed picture of the surface can also
be provided by scanning electron microscopes (SEM) which
produce a 3D-image of greater resolution and depth of field
than their optical counterparts. A major disadvantage of SEM,
145

146
however, is that non-conductive materials must be coated with
a thin, electrically grounded layer of metal to be observed
on the CRT screen. It is therefore the surface of the metal
coating not the sample that is imaged. Consequentially,
several aspects of the specimen surface are undetected.
AFM; concepts and modes of operation; The most advanced
methods of imaging surfaces currently used are a group of
instruments know collectively as "scanning probe
microscopes"(SPM). These microscopes not only produce highly
magnified images but also can measure important physical,
chemical and biological properties of a living specimen.
SPM's overcome many of the restrictions of optical and
electron microscopes. 1) they can produce extremely high
resolution images in air, liquid or under vacuum 2) they
can image conductive as well as nonconductive materials so
that gold plating is not always required 3) because no
electron beam is used, beam damage and charging of the sample
do not occur.
For these and other important reasons, SPM's can image
living cells in physiological fluids or anywhere chemical
reactions are taking place at a surface.
The first of these new types of microscopes to become
commercially available was the scanning tunneling microscope
(STM). It uses the tunneling current between a electrically
conductive sample surface and a scanning microprobe to
generate the sample image. A second type of SPM is known as
the atomic force microscope (AFM). In contrast to STM, AFM
does not require the specimen surface to be electrically
conductive, since it measures the interaction between a fine
tip probe (mounted on a flexible cantilever foil) and atoms
at the surface of the specimen. See fig VI-1.

147
Photodiode
Cantilever
Feedback and
xty,zScah
Control
x,y,z Piezo
Tube Scanner
Atomic Force Microscopy.
In NC-AFM the system vibrates a stiff cantilever near
its resonant frequency (100-400 KHZ) with an amplitude of (10-
100 A) The system then detects changes in resonant
frequency or vibrational amplitude when the tip encounters
changes in surface topography. The resonant frequency (v)

148
varies as the square root of cantilever spring constant v =
Q(Kf)1/2. The Kf varies as the tip approaches the sample
surface. Finally, the force gradient A (the derivative of
the force vs. distance curve) changes with tip/sample
separation.
As the piezoelectric device moves the sample under the
tip, any variation in interatomic forces cause the flexible
cantilever to move. A reflected laser beam from the surface
of the cantilever foil is detected by a photodiode and
movements of the foil result in current variation in the
detector which are subsequently used to produce an image on
the CRT screen.
Scanning modes include
Contact AFM: This is the normal mode of AFM, the tip atomic
orbitals overlap with those of the sample surface.
Lateral Force mode: Lateral or frictional force microscopy,
is a variant of AFM operation that detects the friction
between the probe and the sample surface. When the
cantilever/probe assembly is scanned laterally, friction will
twist the cantilever.
Constant Force mode: A feedback mechanism may be used to
move the tip as to keep the photodiode current constant. In
this case the variation in applied voltage to the
piezoelectric device is used to produce the image.
Variable Deflection mode: The feedback loop is open, such
that the cantilever undergoes a deflection proportional to
the change in the tip-sample interaction. Since the probe
height (piezo-z) is constant now, the surface image is
constructed from the deflection information. More suitable
for extremely smooth surfaces.

149
Non-contact AFM. As an alternative to contact mode
"attractive mode" or non-contact imaging may be used, here
the tip is brought within a few nanometers of the sample so
that weak intermolecular (van der Waals) forces between the
tip and the sample can be detected. Although the method does
avoid surface deformation by the tip, it is slow and
difficult to use outside of the research environment. To
obtain topographic information, the interaction is recorded
directly or used as a control parameter for a feedback
circuit that maintains the force derivative at a constant
value. For more details see [Porter, Sykes, and Caple 1994].
Tapping mode: A "semi-contact imaging method also used to
minimize the effects of friction and other lateral forces on
soft surfaces. In this force modulation mode, the tip is
lifted and lowered to the contact surface at a constant
frequency during the scanning procedure. A feedback loop
keeping the average force constant. At sufficiently high
amplitude it is found to eliminate frictional force entirely.
However, this mode is not very compatible with the fluid
environment wherein the oscillations become dampened
significantly.
Applications, advantages and disadvantages: After winning the
1986 Nobel Prize for inventing the STM, G. Binning developed
in collaboration with others the AFM. The AFM can image
three-dimensional surface structures and processes of
biological specimens in real-time at molecular and often
atomic resolution. Examples to date include
SURFACE IMAGING: large molecules such as DNA, plasmids,
membrane proteins, and a variety of organelles and living
cells.

150
MOLECULAR DYNAMIC AND SURFACE CHEMISTRY: biochemical
processes such as the polymerization of fibrinogen and
crystal growth. Physicochemical properties such as
elasticity, viscosity, and intermolecular forces on
biological specimens have been studied. In addition to
passive imaging and molecular dynamics,.
NANOARCHITECTURE: In addition to passive imaging and
molecular dynamics,. AFM has also been used to actively
manipulate sample surfaces in what is being called "molecular
machining".
Where AFM excels; it obtains topographical information
as to the surfaces of biological molecules i.e. it images the
surfaces where most of the regulatory biochemical and
physiological signals are directed. The AFM differs from the
SEM in that it can image living cells and molecules in an
aqueous environment at comparable or greater resolution. For
biological specimens in a normal imaging situation, the
resolution is -lOnm still higher than optical and comparable
to SEM. In biological specimens molecular resolution images
of membrane proteins have been obtained. The nature of the
sample as well as the methods of their preparation play a key
role in determining the resolution limits of the AFM. It has
recently been sugessted that for the high resolution of
biological samples an approach that increases rigidity would
be helpful e.g. operation at cryogenic temperatures. Perhaps
no other imaging technique has been adopted as rapidly by
biologists as AFM. The acceptance of AFM is attributed to a
number of factors a) AFM can be used for imaging at
atmospheric pressure and in physiological solutions b) it
covers an enormouos range of resolution from microns to
fractions of the nanometer c) it can be applied to

151
noncrystalline specimens. Lastly, the cost of the
instrumentation is not beyond the means of most individual
investigators.

APPENDIX VII
PURIFICATION OF ANTIBODIES THIOPHILIC CHROMATOGRAPHY:
T-GEL FRACTIONATION
A new method for antibody purification, termed
thiophilic adsorption chromatography, has recently been
developed. It provides a low cost, yet efficient alternative
to ammonium sulfate precipitation as a first step of a multi-
step immunoglobulin characterization scheme. Thiophilic
adsorption is a highly selective type of lyotropic salt-
promoted protein-ligand interaction, which has been
extensively studied by Porath and co-workers as well as other
researchers (Lihme and Heegaard 1990). This interaction is
termed thiophilic because it may be distinguished by proteins
that recognize a sulfone group in close proximity to a
thioether. The structure of a immobilized thiophilic ligand
is shown in fig.VII-1.
Figure VII-1. Thiophilic Adsorption Chromatography.
Thiophilic adsorption has some elements of both
hydrophobic and hydrophilic interaction. Salts that interact
with water molecules, such as potassium and ammoniuum
sulfate, promote binding of proteins to thiophilic supports.
152

153
The thiophilic adsorbent T-Gel has a high binding capacity
(~2 0mg of immunoglobulin per mL of gel) and it has broad
specificity toward immunoglobulins derived from various
species irrespective of the type of immunoglobulin or
immunoglobulin subclass.
Protocol
Step.l. Allow all components i.e. T-Gel adsorbent, samples,
and buffers to warm to room temperature.
Step.2. Open the top and bottom cap of the T-Gel adsorbent
column and allow the storage buffer to drain out.
Step.3. Equilibrate the column with 12mL T-Gel binding
buffer.
Step.4. Apply the sample(l-3mL) to the column and allow it to
completely enter the gel. Optional: Collect the 3mL effluent
as NB fraction.
Step.5. Wash the column with 30mL of binding buffer. You may
monitor protein concentration spectrophotometrically to
ensure that all material is washed from the column. Optional:
Collect and assay NB (non-bound) fractions.
Step 6. Elute the column with 36mL of T-Gel elution buffer.
Collect the column effluent as 3mL "bound" (B) fractions.
Measure the protein content of each fraction. The second,
third and fourth bound fractions should contain the highest
concentration of purified immunoglobulin. These purified
immunoglobulins may be used as is or desalted into other
buffers.
step 7. Regenerate the T-Gel column by adding 15mL of 8M
guanadine HCL solution to the column and allowing the column
to drain. Then rinse the column with 30mL of degassed, DI
water, followed by 6mL of storage buffer and allow the column
to drain. Install the lower cap on the column, add 3mL of

154
storage buffer then apply top cap and store the column,
refrigerated and in the upright position.
Sample Preparation for T-Gel Fractionation(for sera, plasma,
ascites or tissue culture supernatantO.
Add 87mg KSO, per mL of sample then mix well to achieve a
final concentration of 0.5M potassium sulfate in the sample.
Mix without foaming to avoid denaturation of the
immunoglobulins. When the potassium salt is fully dissolved
centrifuge the sample at 10,000xg for 20mins. Carefully
remove the clear supernatant and filter it through a 0.5|im
filter. One milliter of most animal sera applied to the 3mL
T-Gel column should result in the binding of all the
immunoglobulins present.
For larger or more concentrated samples, it is expected
that one or more of the NB fractions will contain some
immunoglobulins. These NB fractions may be pooled and tested
just as other samples in a subsequent run if it is desirable
to recover all of the immunoglobulin from the original
sample.

LIST OF REFERENCES
Amdur M, Doull J and Klaassen C (eds.) Casarett and
Doull's Toxicology. 4th ed. New York: Pergamon Press,
pp.176-179. (1991)
Aotsuka S, Funahashi T, Tani N, Okawa-Takatsuji M, Kinoshita
M and Yokohari R. Adsorption of anti-dsDNA antibodies
by immobilized polyanionic compounds. Clin.Exp.Immuno.
79:215-220 (1990)
Avanzi G, Cepparrone F, Lombardo R and Di Pietro, G.
Therapeutic plasma adsorption. A review. Intern.J.Artif.
Org. 16:S-5,160-164 (1993)
Balint J. US patent No. 4,801,449
Bergstrom J and Wehle B. Clinical implication of middle and
larger molecules. In: Nissenson AR, Fine RN and Gentile
DE (eds.). Clinical Dialysis. Norwalk, CT: Appleton &
Lange; 204-234 (1994)
Bonomini M, Fiederling B, Bucciarelli T, Manfrini V, Dillio C
and Albertazzi A. A new polymethylmeth-acrylate membrane
for hemodialysis. Art.Kid.and Dial. 19:4,232-239 (1996)
Bratt T and Ohlson S. The development of a Clq anti-Clq
immunoadsorbent for the removal of immune complexes
from plasma. J.Clin.Lab.Inmun. 27:191-195 (1988)
Bueno SMA, Haupt K and Vijayalakshmi MA. In vitro removal of
human IgG by pseudobiospecific affinity membrane filtra
tion on a large scale. A preliminary report. Int.J.
Art.Org. 18:7,392-396 (1995)
Burlingame, RW, Lowe WE, Wang BC, Hamlin R, Xuong NH and
Moudrianakis EN. Crystallographic structure of the
octameric histone core of the nucleosome at a
resolution of 3.3A. Science. 228:546 (1985)
Burlingame RW, Boey LM, Starkebaum G and Rubin L. The central
role of chromatin in autoimmune respones to histones and
DNA in systemic lupus erythematosus. J.Clin.Invest.
94:184-192 (1994)
Bygren P, Freiburghaus C, Lindholm T, Simonsen O, Thysell B
and Weislander J. Goodpasture's syndrome treated with
Staphylococcal protein A immunoadsorption.
Lancet 2:1295-1296 (1985)
155

156
Carlson JA, Hodder SR, Ucci AA and Madaio MP. Glomerular
localization of circulating single-stranded DNA in
mice. Dependence on the molecular weight of DNA.
J. Autoimmune. 1:231 ( 1988)
Casali P and Lambert PH. Purification of soluble immune
complexes from serum using PMMA beads coated with
conglutinin or Clq. Clin.Exp.Immuno. 37:295-309
(1979)
Chanard J, Vincent C, Caudwell V, Lavaud S, Toupance 0, Wong
T and Revillard M. Beta 2-microglobulin in uremic
patients who are undergoing dialysis. Kidney Int.
43:S83-S87 (1993)
Charlton B, Schindhelm K and Farrell, PC. Effect of extra
corporeal IgG removal on IgG Kinetics. Trans.Am.
Soc.Artif.Intern.Organs. 29:724-728 (1983)
Chen HS, Wu MS, Chen WY, Lee SH and Yang, CS. Serum Soluble
Interleukin-2 Receptor in Patients with End-Stage Renal
Disease on Hemodialysis. Dialysis and Transplantation.
24:8,459-464 (1995)
Choi SH, Zhang X and Stinson MW. Dynamics of Streptococcus
Histone Retention by Mouse Kidneys. Clin.Immuno.
Immunopatho. 76:1,68-74 (1995)
Cochrane CG and Koffler D. Immune complex disease in experi
mental animals and man. Adv.Immunol. 16:185-264
(1973)
Colton CK and Lowrie EG. Hemodialysis:physical principles and
technical considerations. In: Brenner, BM and Rector Jr.
FC (eds.). The Kidney. 2ed. Philadelphia: WB Saunders,
pp.2425-2489 (1981)
Descamps-Latscha B. Infection and Immunity in ESRD: In
Fundamentals of Dialysis. Heinrich, JF (ed.). New
York, NY: Plenum Pubis., pp.209-224 (1994)
Dixon FS, Feldman JD and Vazquez JJ. Experimental glomerulo
nephritis: The pathogenesis of a laboratory model
resembling the spectrum of human glomerulonephritis.
J.Exp.Med. 113:899-920 (1961)
El-Habib R, Laville M and Traeger J. Specific adsorption of
circulating antibodies by extracorporeal plasma perf
usions over antigen coated collagen flat-membranes:
applications to systemic lupus erythematosus. J.Clin.
Lab.Immunol. 15:111-117 (1984)
Feldman HI, Klag MJ, Chiapella AP and Whelton PK. End-Stage
Renal Disease in US Minority Groups. Amer.J.Nephro.
19:5,397-410 (1992)

157
Fernando-Salznero P, Bunch AW and Gutierrez-Merino C.
Immobilization of hepatic microsomal cytochrome P450
from phenobarbital treated rats in hollow fiber bio
reactors: In Progress in Membrane Biotechnology.
Gomez-Fernandez NC, Chapman D and Packer L (eds.).
Basel Switzerland: Birkhauser Verlag, pp.291-305 (1991)
Fessel WJ. Systemic lupus erythematosus in the community:
Incidence, prevalence, outcome, and first symptoms:
The high prevalence in black women. Arch.Intern.Med.
134:1027 (1974)
Fried JR. Polymer Science and Technology.Englewood
Cliffs, NJ: Prentice Hall (1995)
Funck-Brentano J, Sausse A and Man N. A new hemodialysis
treatment associating a membrane highly permeable to
middle molecules with a closed dialysate system. Proc.
Eur.Dial.Trans.Assoc. 9:55 (1972)
Gao CL, Li BA, Chen CZ, Yu YT, Yuan P and Song JC. Clinical
trials of immunoadsorbent in systemic lupus erythe
matosus therapy. Artif.Organs. 19:5,468-9 (1995)
Gilbert E. Sulfonation and Related Reactions.
New York: Interscience Publishers. (1965)
Gnanou Y. Design and Synthesis of New Model Polymers.
J.M.S.-Rev.Macromo1.Chem.Phys, C36(l) 77-117 (1996)
Gnanou Y, Lutz P and Rempp P. Synthesis of star-shaped poly
ethylene) oxide. Makromol.Chem. 189:2885-2892 (1988)
Gosling JP. A decade of development in immunoassay metho
dology. Clin.Chem. 36:1408-1427 (1990)
Hardy RR. Purification and characterization of monoclonal
antibodies, In: Handbook of Experimental Immunology,
Vol.lf Immunochemistry, Weir DM (ed.). Oxford: Blackwell
Scientific, Chpt. 13.1 (1986)
Hebert LA, Allhiser CL and Koethe SM. Some hemodynamic deter
minants of immune complex trapping by the kidney.
Kidney Int.14:452-465 (1978)
Heipe F. Development of a Clq-adsorbent for the selective
removal of circulating immune complexes. Biomat.Art.
Cells.Art.Org. 18:529-534 (1990)
Higgins RM, Streather CP, Buhler R and Hendry, BM. Relapse
of Systemic Lupus Erythematosus after Extracorporeal
Immunoadsorption. Nephron 69:183 (1995)
Hnilica LS. Histones and other Basic Nuclear Proteins.
Boca Raton, Florida: CRC Press, Inc. (1989)

158
Hochberg MC. The Epidemiology of Systemic Lupus Erythema
tosus: In Dubois' Lupus Erythematosus (4th ed.), DJ
Wallace and BH Hahn (eds.). Philadephia: Lea and
Febiger. pp.49-57 (1993)
Holmberg K, Bergstrom K and Stark MB. Immobilization of
Proteins Via PEG Chains In: Poly(Ethylene Glycol)
Chemistry. J. Milton Harris (ed.). New York: Plenum
Press. (1992)
Ikonomov V, Samtleben W, Schmidt B, Blumenstein M and Gurland
HJ. Adsorption profile of commercially available
adsorbents: An in vitro evaluation. Int.J.Art.Org.
15:312 (1992)
Iseka K, Miyasato F, Oura T, Uehara H, Nishime K and Fukiyama
K. An Epidemiologic Analysis of End stage Lupus
Nephritis. Amer.J.Kidney Dis. 23:4,547-554 (1994)
Jones, B. US patent No. 5,122,112.
Jones JV, Cumming RH, Bucknall RC, Asplin CM, Fraser ID,
Bothamley J, Davis P and Hamblin TJ. Plasmapheresis
in the management of acute systemic lupus erythematosus?
Lancet. 2: pp.709-710 (1976)
Klein E, Eichholz E and Yeager D. Affinity membranes prepared
from hydrophilic coatings on microporous polysulfone
fibers. J.Membrane Sci. 90:60-80 (1990)
Klinke B, Rockel A, Abdelhamid S, Fiegel D and Walb D.
Transmembranous transport and polysulfone, polyacrylo
nitrile, polymethylmethacrylate and cuprammonium rayon
membranes. Art.Kid.Dial. 12,2,697-702 (1989)
Klinkman H and Vienken J. Membranes for dialysis. Nephro.
Dial.Transplant. 10:suppl.3,39 (1995)
Kniker WT and Cochrane CG. Pathogenic factors in vascular
lesions of experimental serum sickness. J.Exp.Med.
122:83 (1965)
Kniker WT and Cochrane CG. Localization of circulating immune
complexes in experimental serum sickness. The role of
vasoactive amines and hydrodynamic forces. J.Exp.Med.
127:119-136 (1968)
Lihme A and Heegaard PM.Thiophilic adsorption chromatography:
the separation of serum proteins. Anal.Biochem. 192:64-
69 (1990)
Lowe C and Dean PDG. Affinity Chromatography. Chicester, GB:
Wiley, (1974)

159
Mahan JD, Herbert LA, McAllister CM, Birmingham DJ, Shen XP
and Coso FG. Platelet involvement in experimental
immune complex-mediated glomerulonephritis of primates.
Kidney Int. 44:716-725 (1993)
Malmqvist M. Biospecific interactions using biosensor
technology. Nature. 361:186 (1993)
Malmsten M, Lassen B, Holmberg K, Thomas V and Quash G.
Effects of hydrophilization and immobilization on
the interfacial behavior of immunoglobulins. J.
Colloid Interfac.Sci. 177:70-78 (1996)
Miedel MC, Hulmes JD, Weber DV, Pascal B and Pan YCE.
Structural analysis of recombinant soluble human
interleukin-2 receptor. Biochem.Biophys.Res.Comm.
154:5,372-379 (1988)
Minami YT. The IL-2 receptor complex:its structure, function
and target genes. Ann.Rev.of Immunology 11:165-190
(1993)
Morioka T, Woitas R, Fujigaki Y, Bats ford SR and Vogt A.
Histone mediates glomerular deposition of small size
DNA anti-DNA complex. Kidney Int. 45:991-997 (1994)
Morioka T, Fujigaki Y, Batsford SR, Woitas R, Oite T,
Shimizu F and Vogt A. Anti-DNA antibody derived from
a systemic lupus erythematosus (SLE) patient forms
histone-DNA-anti-DNA complexes that bind to rat
glomeruli in vivo complex. Clin.Exp.Immuno.
104:92-96 (1996)
Mosbach K and Nilsson K. Tresyl Chlorides. Methods in
Enzymology. 104:56 (1984)
Muller S. Characterization of a monoclonal antibody reacting
with antigenic structure of histone with histone H3.
FEBS Lett. 182:459-464 (1985)
Naitoh A. Removal of beta-2-microglobulin by diffusion alone
is feasible using highly permeable dialysis membranes.
ASAIO Trans. 34:630-634 (1988)
Palmer A, Cairns T, Dische F, Gluck G, Gjorstrup P, Parsons V
Welsh K and Taube D. Treatment of rapidly progressive
glomerulonephritis by extracorporeal immunoadsorption,
prednisolone and cyclophosphamide. Nephrol.Dial.
Transplant. 6:536-542 (1991)
Pinching AJ and Peters DK. Remission of Myasthenia Gravis
Following Plasma-Exchange. Lancet. 2:1373-1376(1976)
Porath J, Maisano F and Belew M. Thiophilic adsorption-a
new method for protein fractionation. FEBS Lett. 185,
306-310 (1985)

160
Porter TL, Sykes AG and Caple G. Scanning Probe Microscope
Imaging of Polyaniline Thin Films in Non-Contact Mode.
Surf.Interface Anal. 21:814 (1994)
Pozniak G, Krajewska B and Trochimczuk W. Urease immobilized
on modified polysulphone membrane: preparation and
properties. Biomaterials. 16:129-134 (1995)
Purdy R. A Mechanism-Mediated Model for Carcinogenicity
Model Content and Prediction of the Outcome of Rodent
Carcinogenicity Bioassays Currently Being Conducted on
25 Organic Chemicals.Environ.Health Persp 104:suppl.
5:1085-1094 (1996)
Ratner BD (ed.) Surface Characterization of Biomaterials
Amsterdam: Elsevier. (1988)
Reihanian H, Robertson CR and Michaels AS. Mechanisms of
polarization and fouling of ultrafiltration membranes
by proteins. J.Membrane.Sci. 16:237 ( 1983)
Reveille JD, Bartolucci A and Alarcon GS. Prognosis in
Systemic Lupus Erythematosus: negative impact of
increasing age at onset, black race, and thrombo
cytopenia, as well as causes of death. Arthri.Rheum.
33:37-48 (1990)
Roubinian JR, Talal N, Greespan JS, Goodman JR, and
Siiteri P. Effect of Castration and sex hormone
treatment on survival, anti-nucleic acid antibodies,
and glomerulonephritis in NZB/NZWF1 mice. J.Exp.Med.
147:1568 (1978)
Rubin L, Jay G and Nelson D. SIL2R binds to IL-2 efficiently.
J.Immuno. 137:3841 (1990)
Rubin LA and Nelson DL. The Soluble Interleukin-2 Receptor:
Biology, Function, and Clinical Application. Ann. Int.
Med. 113:619-627 (1990)
Rubin RL. Histone (H2A-H2B)-DNA Autoantibodies: In
Autoantibodies, Peter JB and Schoenfeld Y
(eds.). New York: Elsevier pp.364-373 (1996)
Ruffatti A. Anti-doublestranded DNA antibodies in the Eldery:
Prevalence and Characteristics. J.Clin.Immun. 10:6,300-
303 (1990)
Schneider KM. Plasmapheresis and immunoadsorption: Different
techniques and their current role in medical therapy.
Kidney Int. 53:suppl.64,S61-S65 (1998)
Sessa A, Meroni M, Battini G (eds.).Glomerulonephritis in the
Elderly, in Contrib. Nephrol. Basel: Karger, vol. 105
(1995)

161
Six J and Gnanou Y. From Star-Shaped to Dendritic Poly
ethylene oxide)s: Toward increasingly branched
architecures by anionic polymerization. Polym,Prep.
ACS. 35:488 (1994)
Sperling R and Wachtel E. The histones. Adv.Protein Chem.
34:1-60 (1981)
Stemmer C, Briand JP and Muller S.Mapping of linear epitopes
of human histone Hi recognized by rabbit anti-Hl/H5
antisera and antibodies from autoimmune patients. Mol.
Immun. 31:4,1037-1046 (1994)
Stemmer C and Muller S. Histone Autoantibodies other than
(H2A-H2B)-DNA Autoantibodies: In Autoantibodies,
Peter JB and Schoenfeld Y (eds.). New York: Elsevier,
pp.373-385 (1996)
Streicher E, and Schneider H. Polysulfone membrane mimicking
human glomerular basement membrane. Lancet. 2:1136
(1983)
Suenaga R and Abdou N. Anti(DNA-histone) antibodies in active
lupus nephritis. J.Rheumatol. 23:2,279-284 (1996)
Susuki K, Hara M, Harigai M, Ishizuka T, Hirose T, Matsuki Y,
Kawaguchi Y, Kitani A, Kawagoe M and Nakamura H. Cont
inuous removal of anti-DNA antibody using a new extra
corporeal immunoadsorption system in patients with
systemic lupus erythematosus. Art.Rheum. 34:12
1546-1552 (1991)
Susuki N, Otuka I, Harada T, Mizushima Y and SakaneT.
Preferential adsorption of cationic anti-DNA anti
bodies with immobilized polyanionic compounds,
dextran sulfate. Autoimmunity 19:105-112 (1994)
Susuki T, Burlingame RW, Casiano CA, Boey ML and Rubin RL.
Antihistone antibodies in systemic lupus eryth
ematosus: Assay dependency and the effects of
ubiquitination and serum DNA. J.Rheum. 21:1081-91
(1994)
Tan EM, Cohen AS, Fries JF, Masi AT, McShane DJ,Rothfield NF,
Schaller JG, Talal N and Winchester RJ. The 1982 revised
criteria for the classification of systemic lupus
erythematosus. Arthritis Rheum. 25:1271 (1982)
Termaat RM, Brinkman K, van Gompel F, van der Heuvel LPWJ,
Veerkamp JH, Smeenk RJT and Berden, JHM. Anti-DNA
antibodies can bind to the glomerulus via two
distinct mechanisms. Kidney Int. 42:1363-71 (1992)
Theofilopoulos AN and Dixon FJ. Autoimmune diseases:
Immunopathology and etiopathogenesis. Amer.J.Pathol.
108:321-365 (1982)

162
Tijssen P. Practice and Theory of Enzyme Immunoassays.
Amsterdam: Elsevier Science Publishers, (1985)
Tommalia DA, Naylor AM and Goddard WA. Starburst dendrimers:
molecular level control of size, shape, surface
chemistry, topology and flexibility from atoms to
macroscopic matter. Angew Chem.Int.Ed.Eng. 28:138-75
(1990)
Tron F and Bach JF. Relationships between antibodies to
native DNA and glomerulonephritis in systemic lupus
erythematosus. Clin.Exp.Immuno. 28,426 (1977)
Vallar L, Costa PP, Teixeira A, Pfister M, Costa A, Barrios
R, and Rivat C. Immunoadsorption procedure as a
potential method for the specific [32-microglobulin
removal from plasma of patients with chronic renal
failure. J.Chrom.B. 664:97-106 (1995)
Ward MH, Pyun E and Studenski S. Causes of Death in Systemic
Lupus Erythematosus. Arthritis Rheum. 38:10,1492-1499
(1995)
Wener MH and Mannik M. Mechanisms of immune deposit formation
in renal glomeruli. Springer Semin. Immunopathol. 9:219-
235 (1986)
Yamazaki Z. Extracorporeal immunoadsorption with IM-PH or IM-
TR column. Biomat.Art.Cells.Art.Org. 17:117-24 (1989)
Yen DR and Merrill EW. Synthesis of PEO star molecules based
on PAMAM dendrimer cores. Polymer Preprints. ACS.
38:1,531-532 (1997)
Yorioka N, Hirabayashi A; Kanahara K, Takemasa A, Oda H,
Joarder ZH and Yamakido M. Serum Soluble IL-2 Receptor
in patients with Glomerulonephritis. Amer.J.Nephro.
10:181-185 (1990)

BIOGRAPHICAL SKETCH
Michel Lee Branham was born in Chicago, Ill. in 1955.
After serving as a biomedical equipment repairman in the
United States Navy from 1983-1989, Mike worked for several
years as a laboratory technician in the immunoassay
development laboratories at Abbott Diagostics Division;
Abbott Park, Ill. He received a B.S. in chemistry from
Governors State University; University Park, Ill., in 1993
and began his studies at the University of Florida in June of
1995. His research interests include protein purification, as
well as the design and characterization of bioaffinity
adsorbent materials. He is a student member of the American
Association of Pharmaceutical Scientists (AAPS) and the
American Chemical Society (ACS).
163

I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality, as
a dissertation for the degree of Doctor of Philosophy.
Ian R. Tebbett, Chair
Professor of Medicinal
Chemistry
I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality, as
a dissertation for the degree of Doctor of Philosophy.
Chris Batich,
Professor of Materials
Science and Engineering
I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality, as
a dissertation for the degree of Doctor of Philosophy.
Edward Ross,
Associate Professor of
Medicine
I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality, as
a dissertation for the degree of Dpptor of Philosophy.
Kenneth Sloan,
Professor of Medicinal
Chemistry
I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality, as
a dissertation for the degree of Doctor of Philosophy.
Roger Tran-Son-Tay,
Associate Professor of
Aerospace, Engineering
Mechanics and
Engineering Science

I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality, as
a dissertation for the degree of Doctor of Philosophy.
)>vu f/y)i0 CQi
Donna Wielbo,
Assistant Professor of
Medicinal Chemistry
This dissertation was submitted to the Graduate faculty
of the College of Pharmacy and to the Graduate School and was
accepted as partial fulfillment of the requirements for the
degree of Doctor of Philosophy
Dean, College of Pharmacy
Dean, Graduate School