Citation
Targeting Angiogenic Growth Factors in Proliferative Diabetic Retinopathy

Material Information

Title:
Targeting Angiogenic Growth Factors in Proliferative Diabetic Retinopathy
Copyright Date:
2008

Subjects

Subjects / Keywords:
Angiogenesis ( jstor )
Cells ( jstor )
Endothelial cells ( jstor )
Gene therapy ( jstor )
Integrins ( jstor )
Messenger RNA ( jstor )
Receptors ( jstor )
Retina ( jstor )
RNA ( jstor )
Small interfering RNA ( jstor )

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Embargo Date:
7/24/2006

Downloads

This item has the following downloads:


Full Text












TARGETING ANGIOGENIC GROWTH FACTORS IN PROLIFERATIVE DIABETIC
RETINOPATHY















By

HAO PAN


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


2006

































Copyright 2006

by

Hao Pan

































This document is dedicated to the graduate students of the University of Florida.















ACKNOWLEDGMENTS

Since coming to the United States in August 2001, it has been five years. This

was the most challenging five years and there was happy and hard time. To study abroad,

especially in the United States, was one of my dreams when I was in high school. Now,

with the completed dissertation in hand, I can tell myself: Hao, you made it!

Language has been the biggest obstacle in my study. I was confident about my

English, but I came here and found that there is still so much to learn and it still takes

time. The study and life for me has been harder than most American students. But I am

happily seeing my improvement everyday. I composed my dissertation in English, gave

seminars in English and passed the final defense in English; all of these are making me

proud.

I thank my mentor, Dr. Maria Grant, for her patient and inspiring guidance in the

past four years. Every member in my committee, Dr. Alfred Lewin, Dr. Sean Sullivan

and Dr. Stratford May, has given me great suggestions for my dissertation work. I thank

everybody in the lab. Dr. Lynn Shaw instructed me in great details during my

experiments and dissertation writing. Dr. Aqeela Afzal was also a great help for my

bench work. And every other member in the lab has given me great support for my

defense.

I thank my parents. They are far away in China but I am sure they are proud and

as happy as I am now. They have done everything they could to provide me the best

education opportunities and they have always been there encouraging all the way along. I









am the only child in the family and I am thankful that they supported when I decided to

study abroad.

I thank Yao for her great help and support. Her love strengthened me during the

hardest time. Without her, I could not have overcome all the difficulties and successfully

graduated.

There is still a long way ahead, with more challenges and opportunities. I would

cherish everything I have had in the University of Florida. The orange and blue will

always be a source of courage and confidence. Go Gators!
















TABLE OF CONTENTS



A C K N O W L E D G M E N T S ................................................................................................. iv

LIST OF TABLES ......... ...... .................................... .. ...... .............

LIST OF FIGURES ......... ........................................... ............ xi

A B S T R A C T .......................................... ..................................................x v

CHAPTERS

1 B A C K G R O U N D ................................................................ ....... ................ .

Introduction and Project Aim ............................ ........................ .............1
T h e E y e ............... ........ ............................... ................................. . 2
T he A natom y of the E ye......... ...................... ................................ ...............
The Retina ................................................2
The B lood Supply to the R etina ........................................ ........................ 4
R etin op ath ies ......................... ... .............................. 5
Age-Related Macular Degeneration (ARMD) ...............................................6
Retinopathy of Prematurity (ROP)............................................... .................. 9
D iabetic R etinopathy (D R )............... ........... .................. ... ............ .11
Current Treatments for Retinopathies ...................................... ............... 13
Pathogenesis of Diabetic Retinopathy .......................... ............. .............. 16
Increased Polyol Pathway Flux ............. ...... .. ........ ............... 17
Production of AGE........................... .......... .... .............. 17
Generation of Reactive Oxygen Species.................................................19
Activation of Diacylglycerol and Protein Kinase C Isoforms...................19
How Does the Change in Retinal Blood Flow Occur?..............................20
What Causes Retinal Capillary Cell Death? ..........................................21
W hat Causes Retinal Ischem ia? ...................................... ............... 21
A ngiogenesis and G row th Factors.................................... ............................. ....... 22
V asculogenesis and A ngiogenesis.................................... ....................... 22
H ypoxia-Induced Factor (H IF)...................................... ................................ 23
Vascular Endothelial Growth Factor (VEGF)................. ............................25
V EGF Fam ily and Isoform s .............................................. ............... 25
VEGF Receptors ........................... ...... .... ..................27
VEGF Receptor Signaling................... ...................30
The Function of VEGF in Ocular Neovascularization..............................33









Basic Fibroblast Growth Factor (bFGF or FGF2).............................................35
A n g io p o ietin s .................................................................... .. 3 6
Platelet-Derived Growth Factor (PDGF).................................. ............... 36
In teg rin s ...................................... ............................................... 3 7
Integrin Signaling ............... .. ...... ....... .... ......... .. .......... ..........38
Relationships between Integrin and Other Growth Factor Receptors in
A n g io g en esis ................... .......................................................... 4 2
Pigment Epithelium-Derived Factor (PEDF) ............................................... 47
Insulin-Like Growth Factor (IGF)- ....................................... ............... 47
IGF-1 and IGF-1R................................ ............... ............... 48
IG F B P s an d A L S ..................... .... ....... ......... ...... ............ .............. 5 1
The Involvement of Insulin Receptor (IR) and IGF-2 in Angiogenesis ......56
R N A Silencing T technologies .......................................................... .....................57
A ntisense O ligonucleotides ......................................... ........................................59
R ib o zy m e s ...................................................... ................ 6 2
Self Splicing Introns ......................................... ............... ........ ...... .... 63
R N a se P ................................................................... ............... 6 5
H am m erhead Ribozym es ........................................ ......... ............... 66
H airpin R ibozym es ............................................... .. ...................... ... .. ..67
Hepatitis Delta Virus (HDV) Ribozymes and Neurospora Varkud
Satullite (V S) R ibozym es ........................................ ...... ............... 69
RN A Interference ...................................... ................. .... ....... 69
G ene Therapy O verview .................................................. ............................... 75
N on-Viral Gene D delivery ............................................................................. 76
V iral G ene D delivery ................................ .............. .............. ............. 78
Adeno-Associated Viral (AAV) Vectors ............................................. 79
A denovirus (A d) V ectors ........................................ ......... ............... 86
R etrovirus V ectors.......... ...... .............. .................... ......... ... ......... 88
Herpes Simplex Virus Type 1 (HSV-1) Vectors..............................89

2 M ETHODS AND M ATERIALS ........................................ .......................... 90

H am m erhead Ribozym e Target Sites ........................................ ...... ............... 90
A accessibility of Target Site ............................................... ............................. 91
Kinase of Target Oligonucleotides .............................. ...... ..... ...............92
Time Course of Cleavage Reactions for Hammerhead Ribozymes .........................93
M multiple T urnov er K inetics ........................................................................................94
Cloning of the Ribozymes into an rAAV Expression Vector ..............................95
Screening and Sequencing of the Clones......................................... ............... 97
HREC Tissue Culture .................. ...... ............. ... ............ .. ............. 98
Transfection of HRECs with Lipofectamine .... ........... .............................. 99
T total R N A E extraction ............................ ....................... .................. ............... 99
R elative Quantitative R T-PCR ............................ ...........................................100
Reverse Transcription-Real Time PCR................. ........ ................... 102
Total Protein Extraction.............. ............................ .................... ............... 103
W western Blotting ............. .. ........ ............ ............ .... ........ .............. ... 103
Flow Cytom etry ................................. ... .. .......... ................... .. 104









M ig ration A ssay ............ .................................................................... .. .... .. ... .. 10 5
Cell Proliferation A ssay (B rdU ) .................................... ............................. ........ 106
Tube formation Assay (Matrigel) ............................... ................................ 107
Proliferating Endothelial-Cell Specific Promoter Constructs................................107
Plasmid Formulation for Adult Mouse Eye Gene Transfer.................................... 107
A nim als .................................. ............... ..... ... ............. ............. 108
Intravitreal Injection into the Mouse Model of Oxygen-induced Retinopathy
(O IR ) ................ .... ..................... ......... .. ....... ...... .. ......................108
Intravitreal Injection into the Adult Mouse Model of Laser-Induced Retinopathy.. 110
Im m unohistological Studies ............................................................................... 111
S statistical A n aly sis ................................................................... ............... 1 1 1

3 RE SU LTS .................................. .................................... ......... 112

R ibozym e D design .............................................................. .. ........ .. 112
Target Site Selection .................. .............................. .. .. .. .. ........ .. 112
A accessibility of Target Site ................. .................. ............................. ........ 114
Sequences of the Ribozymes and the Targets .................................................. 116
In Vitro Testing of R ibozym es ........................................................... .. .............. 117
Tim e Course of Cleavage ............................................. ........................ 117
K inetic A naly sis .............. ............................................................. 119
Functional Analysis of Ribozymes in HRECs........................................ 120
Inhibition of mRNA Expression............................................... ...............120
Protein L evels.................................................. 121
M igration A ssay s............. ........................................................ .. .... ...... .. 123
Cell Proliferation A says ........................................................ ............. 124
Tube Form ation A says ......................................................... ............... 125
In Vivo A analysis of R ibozym es ........................................ .......................... 126
P rom other D evelopm ent.................................................................................... ... 128
Integrin Ribozyme Expression in vivo with the CMV/P-actin Enhancer
Prom oter............................................... ....... .. ......... ................129
The Proliferating Endothelial Cell-Specific Promoter ........... ............... 131
The New Promoter Tested in vivo................... ......... ..... ................133
The New Promoter Tested with Integrin Ribozyme............... ...................138

4 D ISCU SSION ............. ........................................... .. .. .... .. .. ....... ..... 141

Ribozyme Testing Results and Antisense Effect...............................................141
VEGFR-1 and VEGFR-2 Interactions ................................................................. 143
The Proliferating Endothelial Cell Specific Promoters ........................................145
Other Voices on Neovascularization in Diabetic Retinopathy ........... .................147
Final W ords on RN A Silencing................................................................... .. 148

LIST OF ABBREVIA TION S ........................ .. .................... ................. ...............156









L IST O F R E FE R E N C E S ......................................................................... ................... 160

BIOGRAPHICAL SKETCH ............................................................. ..................197
















LIST OF TABLES


Table p

2.1 Sequences of primer pairs and annealing temperatures used in relative
quantitative PCR. ................................ ... ... ........ ............... 102

2.2 Summary of primary and secondary antibodies used in western blottings. ...........104

3.1 Summary of ribozyme and target sequences............. ....... ............... 116

3.2 Summary of ribozyme kinetic data. ........... ...... ...... ...................... 120

3.3 Reduction in target mRNA levels in HREC by the ribozymes...........................121

3.4 Reduction in protein levels by the ribozymes. ............................... ......... ...... 123

3.5 All ribozym es tested in vivo .................... ........ ........... .. ............................128
















LIST OF FIGURES


Figure page

1.1 Basic structure of human eye (courtesy of National Eye Institute,
w w w .nei.nih.gov). ................................................. ....................... 3

1.2 Cross section of the retina (http://thalamus.wustl.edu/course/eyeret.html). ............3

1.3 Normal view vs. ARMD (courtesy of National Eye Institute, www.nei.nih.gov).....6

1.4 Fundus photograph and fluorescence angiogram of ARMD [11]...........................8

1.5 Five stages in ROP (courtesy of National Eye Institute, www.nei.nih.gov) ...........10

1.6 Normal view vs. DR (courtesy of National Eye Institute, www.nei.nih.gov). ........11

1.7 Fundus photograph and fluorescence angiogram of non-proliferative DR [11]. .....13

1.8 Fundus photograph and fluorescence angiogram of proliferative DR [11]. ............13

1.9 Photocoagulation (courtesy of National Eye Institute, www.nei.nih.gov) .............14

1.10 Cryotheropy (http://www.checdocs.org/dr treatment.htm) ............. ................. 15

1.11 Polyol Pathw ay [31] ............................... ..................... .. ............ 18

1.12 A G E form action [31] ............................................................ .................... .... 18

1.13 VEGF-A isoform s [92].......................................................................... 27

1.14 VEGF family ligands and their receptors [116] ............. ..... .................30

1.15 VEGF signaling via VEGFR-2 [92] ............................ ......... ............... 33

1.16 The activation of integrins can lead to the signal transduction in a number of
pathw ay s. [180]. .................................................... ................. 39

1.17 IGF-1 signaling transduction [216]. ........................................................................49

1.18 Proposed pathway of IGF-dependent IGFBP action [223]. ................................... 51

1.19 Overview of possible IGFBP-3 antiproliferation pathways [223]. .........................53









1.20 The crosstalk between IGF-1, IGF-2 and Insulin signalings [254].......................57

1.21 Overview of RNA silencing technologies [258]. ......... ..................................... 58

1.22 M modifications in antisense technology [258]. .................................. .................60

1.23 Self-cleaving and self-splicing reactions in ribozymes [263]. ................................62

1.24 Secondary structure and self splicing steps in group I intron [263].........................64

1.25 Secondary structures of natural and synthetic substrates for RNAse P[275]...........65

1.26 Structure of the hammerhead ribozyme. ............................................................67

1.27 Structure of the hairpin ribozym e................................ ........................ ......... 68

1.28 R N A interference [293]........................................................... ...........................70

1.29 Designing artificial shRNA for RNAi [303]. ................................ .................73

1.30 AAV internalization and intracellular trafficking [330]........................................81

1.31 AAV2 genome and the vector genome [330]................................ ............... 83

1.32 Helper virus -free systems in rAAV production [334] ........................................84

1.33 The 6 pDF helper plasmids in the two-plasmid system [330]..................................85

1.34 Ad genome and the vector genome [324] ...................... ...................................... 87

1.35 M LV genom e structure [329] ......... ................. .............................. ..... .......... 89

2.1 Typical structures of hammerhead ribozyme predicted by Mfold [257]..................92

2.2 The pTRUF21 expression and cloning vector and the orientation and position of
the hammerhead and hairpin ribozyme cassette............... .... .................96

2.3 Time course of OIR mouse model. ............. ...... .......................................... 109

2.4 Time course of the adult mouse model of laser-induced neovascularization......... 110

3.1 The human IR cDNA sequence with ribozyme target site highlighted................113

3.2 Mfold structures predicted for the human IR target region..................................114

3.3 Mfold predicted secondary structure of human IR ribozyme. ............................ 115

3.4 The 34-base ribozymes (black) annealed to the 13-base targets (red) for both
hum an and m house .......................................... ....... ........ .. ........ .. .. 116









3.5 Cleave time course of human IR ribozyme.................. ............................... 118

3.6 Summary of time courses cleavage of the ribozymes generated in this study....... 118

3.7 Multiple-turnover kinetic analysis of a human IR ribozyme ...............................19

3.8 Insulin receptor mRNA levels in HRECs. .................................. ............... 121

3.9 Western analysis of IR levels in cells expressing the human IR ribozyme...........122

3.10 HREC migration assays in response to IGF-1. ........................................ ...........124

3.11 Effect of the VEGFR-1 and VEGFR-2 ribozymes on HREC migration. ..............124

3.12 Effect of ribozyme expression on cell proliferation .................... .....................125

3.13 Effect of ribozymes on HREC tube formation..................................................126

3.14 Cross section of a mouse eye showing pre-retinal vessels................................... 127

3.15 Ribozyme reduction of pre-retinal neovascularization in the OIR model............ 127

3.16 Reduction of pre-retinal neovascularization in the OIR mouse model with
expression of the al or a3 integrin ribozymes. ....................... .......................... 129

3.17 Expression of al ribozyme in OIR model results in severe deformations of the
ey e ....................................................................................... 130

3.18 pLUC1297/1298 vectors and pLUC1297HHHP/1298HHHP clones ....................131

3.19 Verification of the cell specificity of the proliferating endothelial cell-specific
enhancer/prom other .................. ............................ .. ........ .. .......... 133

3.20 The proliferating endothelial cell-specific promoter limits expression of
luciferase to the actively proliferating blood vessels in the OIR model ..............135

3.21 Quantitative assessment of the IGF-1R ribozyme's ability to inhibit pre-retinal
neovascularization when expressed from the promoter. ......................................136

3.22 New promoter tested in adult mouse model of laser-induced neovascularization. 136

3.23 The expression of the IGF-1R ribozyme from the new promoter reduced
aberrant blood vessel formation in the adult laser model............... ............... 137

3.24 Expression of integrin ribozyme driven by proliferating endothelial cell-specific
promoter resulted in less eye deformation. ................................. .................139

3.25 Proliferating endothelial cell specific promoter with integrin ribozyme tested in
O IR m odel ............................................... ............... .................... 140















LIST OF OBJECTS

Object page

3.1 A blood vessel from the adult mouse model shows the luciferase expression is
specific for proliferating endothelial cells................................... ............... 137

3.2 The 3-D view of the blood vessel from the adult mouse model. .........................137















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

TARGETING ANGIOGENIC GROWTH FACTORS IN PROLIFERATIVE DIABETIC
RETINOPATHY

By

Hao Pan

May 2006

Chair: Maria B. Grant
Major Department: Pharmacology and Therapeutics

Proliferative diabetic retinopathy is the leading cause of blindness in the working

age adults. Pre-retinal angiogenesis is the hallmark of this disease and can lead to vessel

leaking, exudate accumulation, hemorrhage, or even retinal detachment. Many growth

factors have been identified to promote the vessel growth, physiologically and

pathologically. Inhibition of these growth factors can result in less abnormal angiogenesis

and potentially prevent the onset of vision impairment. One gene silencing technology,

hammerhead ribozyme, was used to inhibit the signaling of these growth factors.

Ribozymes are small RNA molecules that can recognize and cleave specific sequence in

the target mRNA. Ribozymes against the genes of a number of growth factor receptors,

including IGF-1R, insulin receptor, VEGF-R1, VEGF-R2, and multiple integrins, were

designed and tested in vitro and in vivo. All ribozymes were tested by cleavage time

courses, kinetic analysis and proved to be capable of cleaving synthetic RNA targets.

Then they were transfected in human retinal endothelial cells, and the mRNA levels and









protein levels of the growth factor receptors were reduced. Also the migration,

proliferation and tube formation of these cells were inhibited. We used the oxygen-

induced retinopathy mouse model to test the ribozymes in vivo. The expression of the

ribozymes induced significant reductions in the pre-retinal neovascularization levels. To

better target the proliferating endothelium in vivo, and to minimize the adverse effect of

ubiquitous ablation of targeted growth factor receptors, a proliferating endothelial cell-

specific promoter was designed. This new promoter was tested with IGF-1R ribozyme

and showed specific expression in the proliferating endothelium and significant reduction

in the pre-retinal neovascularization levels. Our results suggest that these ribozymes are a

useful tool to inhibit the angiogenesis in retinopathy, and the proliferating endothelial

cell-specific promoter adds the specificity without losing expression strength.














CHAPTER 1
BACKGROUND

Introduction and Project Aim

Vascular retinopathies, including retinopathy of prematurity, proliferative diabetic

retinopathy and age-related macular degeneration, are the leading cause of vision

impairment worldwide. Pre-retinal vessel growth is the hallmark for retinopathy of

prematurity and proliferative diabetic retinopathy. These new blood vessels are

abnormally positioned and are fragile, easy to leak, and can result in hemorrhage and

retinal detachment. Currently there is no cure for these diseases. The initiation and

maintenance of these pre-retinal blood vessels depend on the involvement of many

growth factors. In this project, with the help of a gene silencing technology, hammerhead

ribozyme, efforts have been made to target and inhibit the expression of a number of

growth factor receptors to reduce the growth factor signaling. Ribozymes are small RNA

molecules that can specifically bind to a sequence in the target mRNA and perform

cleavage. The genes of IGF-1R, VEGFR-1, VEGFR-2, integrins and insulin receptor

have been targeted and the inhibition effects were examined in vitro and in vivo. To better

target the proliferating endothelium in vivo, and to minimize the adverse effect of

ubiquitous ablation of targeted growth factor receptors, a proliferating endothelial cell-

specific promoter was designed and tested. In the basic science point of view, the

investigations on the involvement of the growth factors in the pre-retinal angiogenesis

can provide useful information about their signaling details; in the clinical application









point of view, this work could also imply new targets and methods for the disease

treatment in the future.

The Eye

The Anatomy of the Eye

Optically working like a film camera, the eyes of all the vertebrates are

structurally similar. The light enters the eye through the pupil and forms an inverted

image on the retina, the light-capturing component that functions like the film in a

camera. The cornea and the lens help to focus so that the clearest image is presented on

the retina. The white outer surface of the eye ball is termed sclera, which consists of

tough but flexible fibrous tissue and provides the mechanical support of the entire eye.

The choroid is a layer contained within the sclera, and it is a dense meshwork of blood

vessels and other tissues. One of the most important functions of the choroid is to provide

nutritional and metabolic support for the retina, which is a neuronal sheet that lies within

the choroid. The retina is the most inner surface at the back of the eye. Most of the space

in the eye is filled with a gelatinous body, called vitreous. It is surrounded by the lens and

the retina and the ciliary body. In the ciliary body, the cells secrete the aqueous fluid into

the eye, which contributes to the maintenance of the pressure within the eye.

The Retina

The retina, a layer about 0.4mm in thickness, is primarily composed of neural

tissue including five classes of neurons. It spreads out on the interior surface of the back

of the eye. The visual pathway is initiated when the light stimulates the photoreceptors

that are embedded in the outer retinal layers. The signal is transmitted to bipolar cells and

then to ganglion cells. The signal then travels along the axon of the ganglion cells lining









Vitreous gel

Optic nerve '






Reline



Figure 1.1. Basic structure of human eye (courtesy of National Eye Institute,
www.nei.nih.gov).



J choroid i

r m epithelium
outer segments
inner segments hotore
outer nuclear
layer (ON L)
outer plexiform
layer (OPL) horizontal cell
inner nuclear
layer (INL)
m 1~ inner plexiform
i t layer (IPL)

ganglion cell
layer (CCL)
optic fiber layer
l a (OFL)


Figure 1.2. Cross section of the retina (http://thalamus.wustl.edu/course/eyeret. html).

the inner surface of the retina to the optic nerve, which penetrates the retina and connects

to the brain. There are two more classes of neurons, horizontal cells and amacrine cells.

They are both interneuron and assist in signal processing. Horizontal cells primarily

contact with photoreceptor axons and bipolar cells in the outer plexiform layer and the









inner nuclear layer, respectively, while amacrine cells contact with bipolar axons

primarily in the inner plexiform layer.

Light passes through almost the whole thickness of the retina to be captured by

photoreceptors, or the outer segments of the photoreceptor in detail, where the visual

pigment molecules for light capturing are located. There are two types of photoreceptors,

rods and cones. Rods are specialized to convey variations in light intensity in dim

conditions, but they are not able to function in bright light. Cones are specialized for

bright light conditions, but they are not as sensitive as rods.

The retina cross section can be divided into multiple layers. The nuclear layers are

basically where cell nuclei are located, and the synaptic layers are the place where cells

communicate and transmit electric or chemical signals. The retinal pigment epithelium

(RPE) functions as the outer blood-retinal barrier (BRB) that shut off the diffusion of

large molecules from choroicapillaries. And the retinal vasculature doesn't grow beyond

the inner limiting membranes under normal physiological conditions.

The Blood Supply to the Retina

The metabolism in the retina performs in the highest rate in the body. For the

same mass of tissue, the metabolic needs of the retina are about seven times that of the

brain. In order to meet these high metabolic needs, two separate circulations are involved.

They are retinal and choroidal circulations. The larger arteries and veins of the retinal

circulation can be seen under an ophthalmoscope, and most of the retinal surface is

occupied with a meshwork of retinal capillaries. These capillaries form the inner BRB.

The endothelial cells at the capillaries are connected by tight junctions that prevent

leakage from the vessels. A lot of proteins or molecules work in the binding of the

adjacent cells. Because of the tight junctions, proteins and solutes have to pass through









the apical and the basal membranes of the endothelial cells in order to go into or out of

the circulation from or to surrounding tissues. Water, small molecules and dissolved

gases can do so, such as glucose, oxygen, carbon dioxide, and so on. But most large

molecules, including proteins, cannot pass through freely. The only possible way for

them to pass through is through a process of active transport with the help of the proper

membrane tunnel proteins. So basically the BRB provides a mechanism of keeping the

substance entering the retinal neural tissue in a controlled manner.

The central artery and vein of retinal circulation originate along the optical nerve

and extend into the retina from the center of the optical disc. While the choroidal arteries

and veins of pass through the sclera at multiple places around the optical nerve, and then

they branch into a meshwork of very large capillaries, called choroicapillaries. Large

capillaries increase the rate of blood passing through, which keeps the concentration of

oxygen high and the concentration of carbon dioxide low, and also quick removes the

heat from focused light on the eye bottom. The BRB is not maintained by choroidal

circulations, because the cells on the side facing the RPE are fenestrated, and there is no

tight junction between these cells. However, the RPE connecting with the choroid have

tight-junctions and provide the outer portion of the blood-retinal barrier.

Retinopathies

Retinopathies are diseases that affect the function of retinas. Usually they involve

the abnormalities in the vasculatures that nourish the retina. These abnormalities included

ectopic angiogenesis, rupture and leakage on the vessels, accumulation of exudates, retina

detachment caused by vessel and fibrous tissue contractions, and so on. There are three

types of retinopathy clinically identified: age-related macular degeneration (ARMD),

which occurs in the elderly people; diabetic retinopathy (DR), which occurs in the









working age people; and retinopathy of prematurity (ROP), which occurs in infants.

ARMD more involves the abnormalities in choroicapillaries, while DR and ROP are

basically related to the abnormalities of retinal vasculature.

Age-Related Macular Degeneration (ARMD)

ARMD is the leading cause of blindness among those aged over 65 in the western

world [1-3]. It affects the outer retina, RPE, Bruch's membrane and the choroids.

Thickening of Bruch's membrane is seen in this disease. Our understanding about the

pathogenesis has grown in the past decade, but still a lot remains unknown and the

current therapy is limited.














Figure 1.3. Normal view vs. ARMD (courtesy of National Eye Institute,
www.nei.nih.gov).

The clinical hallmark of ARMD is the appearance of drusen, localized deposits

lying between the basement membrane of the RPE and Bruch's membrane. Drusen can

be shown as semi-translucent punctuate or yellow-white deposits depending on the stage

of the disease. Morphologically drusen are classified as "hard" and "soft". Hard drusen

are pinpoint lesions; soft drusen are larger with vague edges and they are easy to become

confluent. Drusen can become calcified and they may also regress. Typically clustered

drusen are located in the central macula, so they can lead to deficits in macular function









such as color contrast sensitivity, central visual field sensitivity and spatiaotemporal

sensitivity [4]. Increased quantity and size of drusen are an independent risk factor for

visual loss in ARMD.

Geographic atrophy is also seen in ARMD, which refers to confluent areas of

RPE cell death accompanied by overlying photoreceptor atrophy [5]. Geographic atrophy

leads to vision impairment, especially the visual function in dark situations [6]. This loss

of function is probably because the RPE loss results in reduced nutrients for those

photoreceptors that are located in the RPE atrophy areas. Apoptosis in the corresponding

area are found [7].

Choroidal (or subretinal) neovascularization (CNV) is a major cause of vision loss

in ARMD. As the term itself indicates, CNV refers to the new blood vessel growth from

the choroids. It breaks through the Bruch's membrane into the space underneath RPE, or

it may further penetrate the RPE layer into the subretinal space. Usually CNV is

associated with leakage of fluid and blood. The repeated leakage of blood, serum, and

lipid can stimulate fibroglial organization leading to a cicatricial scar [4].

Drusen and CNV can cause irregular elevation of RPE, which can lead to RPE

detachment or even RPE tear. RPE detachment can cause visual loss in patients with

ARMD [8].

Depending on whether CNV is present, ARMD is classified into the dry form or

the wet form. The dry ARMD is nonexudative [4]. This is the early phase of ARMD, and

the earliest pathological changes are the appearance of basal laminar deposits between the

plasma membrane and basal lamina of the RPE, and the appearance of basal linear

deposits located in the inner collagenous zone of Bruch's membrane. The former deposits









are seen in an increase amount in ARMD [9], and the later deposits are only seen in

ARMD [10]. Approximately 10 percent of persons with AMD develop the exudative

form of the disease, or wet ARMD. Exudative AMD accounts for 80 to 90 percent of

cases of severe vision loss related to AMD. CNV occurrence is the hallmark of wet

ARMD. CNV is associated with abnormal vessels that leak fluid and blood in the macula,

resulting in blurred or distorted central vision. Figure 1.4 is the funds photograph (A)

and fluorescence angiogram (B) of an eye of a patient with exudative ARMD. Note

subretinal neovascularisation (A, asterisk) with surrounding hard exudates (arrowheads).

On the angiogram (B) the neovascularization is clearly stained by fluorescein (black

arrow) [11].













Figure 1.4. Fundus photograph and fluorescence angiogram of ARMD [11].

As for the pathogenesis of ARMD, shortly speaking, Campochiaro and coworkers

suggested that the age-related thickening of Bruch's membrane reduces the diffusion of

oxygen from the choroid to the RPE and retina [12], and recent evidence suggests that

VEGF plays an important role in the development of CNV. VEGF expression was found

to be increased in RPE cells of maculae of patients with age-related maculopathy, a

condition with a high risk of CNV occurrence [13] and in experimental animal models

[14]. VEGF levels in the vitreous of wet ARMD were found to be significantly higher









than healthy controls [15]. Chronic inflammation from drusen may be involved in the

development of ARMD [16], but the inflammatory contribution is still controversial.

Retinopathy of Prematurity (ROP)

ROP is an adverse effect of treating those premature neonates in respiratory

distress with high oxygen. The high oxygen helps these infants to survive, but it can

cause ROP, which will impair their vision. ROP mainly affects premature infants

weighing about 1250 grams or less that are born before 31 weeks of gestation. It is one of

the most common visual loss diseases in childhood. According to the National Eye

Institute, there are about 28,000 infants born weighing 1250 grams or less in the U.S., and

among them, 14,000-16,000 of the infants are affected by ROP to some degree. 10% of

them need medical treatment and 400-600 infants annually become legally blind of ROP.

The ROP complete progression can be divided into 5 stages. Stage 1 is

characterized by a demarcation line between the normal retina (near the optic nerve) and

vascularized retina. In stage 2, a ridge of scar tissue rises up from the retina due to growth

of abnormal vessels. This ridge forms in place of the demarcation line. In stage 3, the

vascular ridge grows due to spread of abnormal vessels and extends into the vitreous.

Stages 4 and 5 refer to retinal detachment; stage 4 refers to a partial retinal detachment

caused by contraction of the ridge, thus pulling the retina away from the wall of the eye;

and stage 5 refers to complete retinal detachment.

ROP is now considered as a two-phase process during the disease development.

In the first stage, the high oxygen condition will make the developing retinal blood

vessels and especially the developing capillary buds be more "pruned" to drop out. This

pathological vessel dropout is an exaggeration of the normal physiologic process, in

which there is a constant balance between developing and degenerating capillary buds, as
























Stage I KUP is characterized by a demarcation Stage 2 RUP; the white line is replaced by a
line. The orange vascular retina is on the left ridge of scar tissue R. The arrow shows a tuft
and the gray peripheral retina is on the right of new vessels.
scoartated by the white line.


Stage 3 ROP. The size of the ridge has Plus disease shows dilation and tortuosity of
increased (between arrows) and the growth of blood vessels near the optic nerve.
the ridge extends into the vitreous.

Figure 1.5. Five stages in ROP (courtesy of National Eye Institute, www.nei.nih.gov).

tissue demand changes [17]. In short, the hyperoxic vaso-obliteration occurs in the first

stage. When the high oxygen care is complete and the infants survive, they are taken out

the high oxygen environment and the second stage occurs. Because of the vessel loss, the

tissue becomes hypoxic and the ischemia-induced vaso-proliferation begins. The hypoxia

stimulates growth factors increases, especially VEGF. These growth factors play very

important roles in the vaso-proliferation. The vaso-proliferation is abnormal in that these

new vessels are fragile and leak, scarring the retina and pulling it out of position, which









will lead to retinal folds and retinal detachment. The term babies are less affected by

fluctuations in oxygen levels as once the vessels become developed and surrounded by

supportive matrix, thus they are no longer susceptible to pruning by hypoxia [18].

Diabetic Retinopathy (DR)

Diabetic retinopathy is one the three major complications of diabetes mellitus (the

other two are neuropathy and nephropathy) and occurs in both type I and type II diabetes.

DR primarily affects the working age people and is the leading cause of new-onset visual

loss in working people in the U.S. and other industrialized countries [19]. DR affects

approximately three-fourths of diabetic patients within 15 years after onset of the disease

[20]. Retinal neovascularization and macular edema are central features of DR and also

the two factors that cause vision loss. The newly-formed vessels are fragile and abnormal

and they can leak blood into the center of the eye, blurring vision. Macular edema usually

occurs as the disease progresses. The fluid leaks in the center of macular and makes the

macula swell, blurring vision. Other characteristics found in DR include basement

membrane thickening, pericyte loss, microaneurysms, and so on.














Figure 1.6. Normal view vs. DR (courtesy of National Eye Institute, www.nei.nih.gov).

In the beginning stage of DR, there are no clinically evident symptoms, but the

biochemical and cellular alterations are going on in the retinal vasculature. These









alterations include increased adhesion of leukocytes to the vessel wall, alterations in

blood flow, basement membrane thickening. These factors are involved in the blockage

of the retinal capillaries, which is thought to induce hypoxia and further trigger the

overexpression of the angiogenic factors. Other vascular alterations include death of

retinal pericytes, subtle increases in vascular permeability, or even the loss of vascular

endothelial cells. Following this, the blood and fluid leakage may come. The loss of

endothelial cells also leads to acellular capillaries worsening ischemia. With time, more

abnormal phenomena occur and they are clinically observable. These abnormalities

include microaneurysms, dot/blot hemorrhages, cotton-wool spots, venous beading and

vascular loops [20]. The blood and fluid leak out the vessels and accumulate in the retinal

tissue, giving rise to exudates. When this occurs in macula, patients will have macular

edema and impaired vision. This stage is also called nonproliferative retinopathy. With

the progression the disease, next stage is the proliferative retinopathy, featuring the

growth of new vessels on the surface of the retina. The new vessels are abnormal, fragile

and easy to break. The leaking blood can cloud the vitreous and further impair vision. In

more advanced stages, the exaggerated pre-retinal neovascularization can grow from the

retinal surface into the vitreous cavity. This can cause retinal detachment can lead to

blindness. Proliferative retinopathy typically develops in patients with type I diabetes,

whereas nonproliferative retinopathy with macular edema is more common in patients

with type II diabetes [20].

Figure 1.7 shows the funds photograph (A) and fluorescence angiogram (B) of

an eye of a patient with non-proliferative diabetic retinopathy. The arrowheads in Panel A









point to intra-retinal hard exudates surrounding areas of leaking microaneurysms (B,

white arrows) [11].













Figure 1.7. Fundus photograph and fluorescence angiogram of non-proliferative DR [11].

Fundus photograph (A) and fluorescence angiogram (B) of an eye of a patient

with proliferative diabetic retinopathy is shown in Figure 1.8. Note pre-retinal

neovascularization (black arrow) on the optic disc (A), which is extensively leaking

fluorescein (B. white arrows) [11].













Figure 1.8. Fundus photograph and fluorescence angiogram of proliferative DR [11].

Current Treatments for Retinopathies

Currently the clinical proved treatments for retinopathies are limited, and few

drug medications are available. The conventional treatments include laser

photocoagulation, cryotherapy, photodynamic therapy, scleral buckle, and vitrectomy.

All of them cannot cure the disease, but can only delay the disease progression.









In laser photocoagulation, the doctor places thousands, up to 3,500, small laser

burns on the retina. These burns will destroy the normal tissue and decrease the oxygen

needs of the retina. The treatment is usually effective, but at the cost of loss of normal

tissue, and it reduces peripheral vision, impair night vision and change color perception.

The laser photocoagulation is not a cure, as the disease can still progress in spite of

treatment. More treatments may be needed to further prevent vision loss. Laser treatment

is currently applied in all retinopathies, that is, ROP, DR, and ARMD. Laser is also used

to target at the leaking spots, like in severe macular edema, the laser burning is applied in

a focal way. When preventing abnormal vessel growth, as in proliferative DR, the laser

burning is applied in a scattered way.












The retina immediately after focal laser
treatment





I Ci : rr,-





Scatter laser threaten
Figure 1.9. Photocoagulation (courtesy of National Eye Institute, www.nei.nih.gov).









Cryotherapy is a procedure in which physicians use an instrument that generates

freezing temperature to briefly touch spots on the surface of the eye that overlie the

periphery of the retina. It also destroys the tissue and impairs the side vision. Cryotherapy

is more used for ROP. In Figure 1.10, cartoon is showing cryotherapy application to the

anterior avascular retina. A cold probe is placed on the sclera till an ice ball forms on the

retinal surface. Multiple applications are done to cover the entire vascular area. This

treatment thins the tissue under the retina and allows easier oxygen diffusion through the

retina.

















Figure 1.10. Cryotheropy (http://www.checdocs.org/dr treatment.htm).

In photodynamic therapy, a drug called verteporfin is injected i.v. and perfused to

the vasculature in the eye. The drug tends to "stick" to the surface of new blood vessels,

and then, a light is shined into the eye for about 90 seconds, and the light activates the

drug to destroy the new blood vessels. The advantage of this method is that the drug

doesn't destroy the normal tissue surrounding. But the patient needs to avoid bright light

for five days because the drug can be activated in their exposed body parts.

Photodynamic therapy is more used to treat wet ARMD.









In later stages of ROP, scleral buckle is another treatment option [21]. This

involves placing a silicone band around the eye and tightening it. This keeps the vitreous

from pulling on the scar tissue and allows the retina to flatten back down onto the wall of

the eye. The band will be removed later. In most severe conditions in retinopathies,

vitrectomy can be applied, in which the vitreous is removed, scar tissue on the retina

peeled back or cut away, and saline solution is replaced for vitreous. The retina

reattachment can be seen after this surgical treatment [22].

Pathogenesis of Diabetic Retinopathy

Diabetes mellitus is a serious disease leading to morbidity and mortality as it has

long-term complications include macrovascular and microvascular disease. Both type I

(characterized by no insulin production) and type II (characterized by insulin resistance)

diabetes can have these complications. Retinopathy is one of the microvascular

complications. It is believed that the chronic hyperglycemia has a strong relationship with

microvascular complications, and clinical research data demonstrates that improved

glycemic control contributes to significant microvascular risk reduction [23, 24].

Experiments on animal models also suggest that long-term hyperglycemia is necessary to

induce changes in the retinal vasculature [25].

In the retina, GLUT1, which is one of a family of glucose transporters, is

responsible for glucose transfer across BRB into the endothelial cell and retinal cells.

While in most other cells in the body, insulin assistance is required for internalize

glucose; this is not the case with the retina. Excessive transport of glucose through

GLUT1 [26], the involvement of GLUT1 in RPE cells [27], and increased density of

relocalized GLUT1 in inner BRB [28] have been proposed to be related with intracellular

hyperglycemia. Intracellular hyperglycemia in the early stages of diabetes causes









abnormalities in blood flow and increases in vascular permeability. The blood flow

changes come from decreased activity of vasodilators, such as nitric oxide, and increased

activity of vasoconstrictors such as angiotensin II and endothelin-1 [29]. The increase in

vascular permeability comes from VEGF functioning on endothelial cells and changes in

extracellular matrix. With time, hyperglycemia can further induce cell loss and

progressive capillary occlusion. All these changes will eventually lead to edema,

ischemia and hypoxia-induced neovascularization.

To date, there are several hypothesized theories on how hyperglycemia

contributes to microvascular damage, or retinopathy. The most common ones are polyol

pathway theory, advanced glycation end-products (AGE) theory, oxidative stress theory

and PKC activation theory.

Increased Polyol Pathway Flux

As shown in Figure 1.11, glucose is reduced to sorbitol by aldose reductase, and

at the same time, nicotinamide-adenine dinucleotide phosphate (NADPH) is oxidized to

NADP+. Then sorbitol is oxidized by sorbitol dehydrogenase to fructose, coupled with

the reduction of oxidized nicotinamide-adenine dinucleotide (NAD ) to NADH [29]. So

the intracellular high glucose level will result in excess sorbitol, fructose, NADH

accumulation and decrease in NADPH. Some damages caused by increase flux through

polyol pathway have been proposed to include: activation of protein kinase C [30],

contribution of AGE formation [30], decreased activity of Na/K-ATPase [29], and

increase in the formation of reactive oxygen species leading to oxidative stress [29].

Production of AGE

AGE are irreversibly cross-linked substances. Intracellular hyperglycemia is

possibly the primary initiating event in the formation of intracellular and extracellular










AGE [32, 33]. During formation of AGE, glucose reacts nonenzymatically with the

amino group of proteins and other macromolecular to form Schiff bases, which are

transformed into Amadori products that eventually lead to AGE formation [34].

When AGE bind their receptors, RAGEs, some abnormal cellular events can

occur, including: the stimulation of the production of the vasoconstrictor endothelin-1,

VEGF production that is associated with increased permeability, and production of

reactive oxygen species. The long-term effects induced by AGE and RAGEs are mostly

mediated by transcription factor KB to express cytokines and growth factors [29].


NADPH + H* NADP* NAD* NADH + H4

glucose sorbitol fructose
akdose reductase sdorbitol
dehydrogenase


I \^ AGEs


Lglutathione ImyoAGsitd
NO myolnito
increased osmotic stress nerve conduction
increased call permeability velocity
Soxidative stress LNa'/K-dependent
oxidative stress ATPaa activity
The Km of glucose for aldose reductase is high (70 rM); thus high concentrations
of this substrate are needed to bind to the enzyme.
Figure 1.11. Polyol Pathway [31].


Thlicked
Basnumnt

S-I duas embran
Glume Amdori Glycation iau S

*- Deactivarel Microvawcidr -
Hypelrision


Figure 1.12. AGE formation [31]

There are several adverse alterations in the micro vasculature associated with

AGE. AGE formation can contribute to thickening of the basement membrane and to

microvascular hypertension by inactivating nitric oxide [31]. The thickening and









hypertension can lead to microvascular leakage and occlusion. AGE can adversely affect

vascular permeability, alter the functions of matrix molecules, and alter the functions of

vessels, by decreasing the vessel elasticity, increasing fluid filtration across vessels [29],

decreasing endothelial cell adhesion [35], and so on.

Generation of Reactive Oxygen Species

The term oxidative stress refers to the imbalance between the production of

reactive oxygen species and the normal antioxidant protective mechanisms present to

guard tissues from oxidative damage [36]. As discussed above, both polyol pathway and

AGE formation can lead to the generation of reactive oxygen species. Glucose also has

pro-oxidant properties in the presence of heavy metals and the auto-oxidation of glucose

can form free radicals too. These reactive oxygen species can inactive or reduce nitric

oxide levels [37].

The reactive oxygen species can result in damaged protein and mitochondrial

DNA that have adverse effects on the microvasculature [38], especially leading to

increased microvascular permeability [39]. Oxidative stress has been shown to increase

intracellular calcium levels, which have been associated with endothelial

hyperpermeability of macromolecules [40].

Activation of Diacylglycerol and Protein Kinase C Isoforms

It has been shown that diacylglycerol (DAG) formation can be induced by glucose

in cell cultures, animal tissues, and diabetic patients [31]. DAG is very important in the

activation of various protein kinase c (PKC) isoforms, with the isoform 0 being thought

to be the most sensitive to changes in DAG levels. PKC-0 has been shown to be

increased in various vascular tissues following hyperglycemic exposure [41]. PKC-a,

PKC-P 1 and PKC-02 are seen to be elevated in the retina during acute and chronic









hyperglycemic states [42]. The consequences induced by PKC activation include

increased retinal permeability [43], increased basement matrix protein formation [44],

VEGF formation [44], and so on. So PKC may have adverse long-term effects in the

vasculature.

Based on the involvement of these pathways, a lot of pathological changes can

happen in diabetic retinopathy. Some of the most important changes are covered below.

They will lead to edema, ischemia and hypoxia in the retina, which all lead to abnormal

neovascularization.

How Does the Change in Retinal Blood Flow Occur?

Hyperglycemia induces changes in retinal blood flow via its effects on

vasodilators and vasoconstrictors. Nitric oxide (NO) is one of the most important

vasodilators. It is synthesized from L-arginine or L-citrulline in cells via activation of a

calcium-dependent nitric oxide synthase (NOS). The NOS isoform produced in

endothelium is called eNOS. NO functions by entering smooth muscle cells and

activating soluble guanylate cyclase, which will result in increased level of cyclic

guanosine 3', 5'-monophosphate (cGMP). cGMP can relax the smooth muscle cells

through a decrease in Ca2+ and dephosphorylation of myosin light chains [45]. In the

hyperglycemic environment, a couple of pathways mentioned above can lead to

decreased level of NO. In the polyol pathway as mentioned earlier, sorbitol is produced

coupled with the oxidation of NADPH and this reduces NADPH availability, and

NADPH is one of the cofactors for NO synthesis. AGE production can lead to subsequent

superoxide generation resulting in NO inactivation. PKC activation reduces the capacity

of a number of agonists to increase intracellular Ca2+ and to stimulate NO production; on

the other hand, the superoxide expression may also result from PKC activation.









Endothelin (ET)-1 is a powerful vasoconstrictor. At low concentrations, it induces

vasodilation. While at high concentrations, it causes the constrictive response by

interacting with its receptors on smooth muscle cells and pericytes in the retinal

vasculature. Hyperglycemia-induced PKC activation can enhance ET-1 transcription

level [46].

What Causes Retinal Capillary Cell Death?

Pericytes loss and endothelial cells loss are both seen in diabetic retina. The cell

death will inevitably lead to microaneurysms and vascular obstruction. Polyol pathway,

AGE pathway and oxidative stress are all thought to be associated with cell death.

Sorbitol accumulated in polyol pathway may cause hyperosmolality of the cells [47];

accumulated AGE production in the glycation pathway will form cross-links and to

generate oxygen-derived free radicals [48]; and the oxidative stress will inactivate NO

and cause abnormal chemical changes in DNA structure [49].

What Causes Retinal Ischemia?

Hyperglycemia causes ischemia via several possible mechanisms, including

thickened basement membrane, platelet aggregation, leukocyte activation and adherence.

Hyperglycemia is sufficient to increase the synthesis of basement membrane components,

like fibronectin [50], various types of collagens [51] and vitronectin [52]. Increased

number and size of platelet-fibrin thrombi in retinal capillaries have been found in the

retina of patients with diabetic retinopathy [53]. Hyperglycemia-induced PKC activation

will stimulate platelet-derived factor (PAF) production, which will activate platelets.

Activated platelets can produce platelet-derived microparticles, which are involved in the

thrombus formation [54]. PAF can also stimulate their receptors on leukocytes rolling on

the luminal endothelial membrane and activate them. p2 integrins on activated leukocytes









enable them to adhere tightly to the endothelial cells via binding intercellular adhesion

molecule-1 (ICAM-1), while as the same ICAM-1 is also unregulated by PKC activation.

And NO downregulation can allow leukocytes to escape from NO control, also leading to

leukocyte activation and adherence [54].

Angiogenesis and Growth Factors

Vasculogenesis and Angiogenesis

Small blood vessels consist only of endothelial cells (ECs), whereas larger vessels

are surrounded by mural cells (pericytes in medium-sized vessels and smooth muscle

cells (SMCs) in large vessels) [55]. Vessels can grow in several ways. Vasculogenesis

refers to the formation of blood vessels by endothelial progenitors [55]. It is a process by

which the initial vascular tree forms in the yolk sac and aortic arches, and begins

immediately following gastrulation when mesodermal cells aggregate into blood islands.

Blood islands contain the precursors of hematopoietic and vascular endothelial lineages

[56]. Angiogenesis refers to the formation of new vessels formation by sprouting from

pre-existing vessels and subsequent stabilization of these sprouts by mural cells.

Additional modes of vascular growth include intususception, bridge formation, and

vascular splitting, in which invaginations or extensions of the vessel wall form tubes that

connect or bifurcate parent vessels [56].

The traditional view is that vessels in the embryo developed from endothelial

progenitors, whereas sprouting of vessels in the adult resulted only from division of

differentiated ECs. However, recent evidence has shown that endothelial progenitors

contribute to vessel growth both in the embryo and in ischemic, malignant or inflamed

tissue in the adult. They can even be used therapeutically to stimulate vessel growth in

ischemic tissues, a progress called "Therapeutic Vasculogenesis" [57-59]. Although









retinal neovascularization has been thought to be due to proliferation of endothelial cells

by angiogenesis, Grant et al. showed that hematopoietic stem cells can enter the

circulation and reach the areas of angiogenesis, and clonally differentiate into endothelial

cells [60]. In another study, adult Lin(-) hematopoietic stem cells injected intravitreally

into neonatal mouse eyes have been shown to interact with retinal astrocytes that serve as

a template for retinal angiogenesis [61]. Blood vessels are being modified by endothelial

progenitor cells, hematopoietic stem cells or other stem cells, and these cells functionally

contribute to physiological and pathological angiogenesis.

Angiogenesis is usually inactivated or kept at low levels in normal tissue of an

adult, but may be activated to an excessive state in a number of diseases, such as cancer,

psoriasis, arthritis, retinopathy, obesity, asthma, atherosclerosis, and infectious diseases.

Cancer is another best known disease that involves pathological angiogenesis that can be

potentially targeted for therapy. In 1972 Folkman proposed that solid tumors are

dependent on angiogenesis for growth greater than a few millimeters in size, and that

increases in tumor diameter require a corresponding increase in vascularization [62]. A

critical step during angiogenesis is the local stimulation of endothelial cells by various

cytokines and growth factors. Stimulation causes the endothelial cells to lose their contact

inhibition, migrate and breach the basement membrane, proliferate, and differentiate to

organize into new vessels [63].

Hypoxia-Induced Factor (HIF)

Beyond a size limitation, simple diffusion of oxygen to metabolizing tissues

becomes inadequate, and specialized systems of increasing complexity have evolved to

meet the demands of oxygen delivery in higher animals [64]. One important role in the

systems is angiogenesis, to make new vessels sprouting into the location that blood









delivery is needed. So ischemia or hypoxia is one of the key factors that lead to the

initiation of angiogenesis. Exactly how hypoxia induces angiogenesis was however

poorly understood. The landmark of hypoxia study in the early 1990s showed that

hypoxia could induce expression of platelet-derived growth factor (PDGF) mRNA [65]

and vascular endothelial growth factor (VEGF) mRNA in tissue culture [66]. Both PDGF

and VEGF are thought to be important growth factors triggering angiogenesis. A large

number of genes are involved in different steps in angiogenesis and they are

independently responsive to hypoxia in tissue culture. Besides PDGF and VEGF, nitric

oxide synthase, fibroblast growth factor, angiopoietins, and matrix metalloproteinases are

involved [67-69]. Many of the individual phenotypic processes in angiogenesis such as

cell migration or endothelial tube formation can be induced by hypoxia tissue culture

[70]. Further study of hypoxia-induced angiogenesis leaded to the discovery of a key

transcriptional regulator, hypoxia-inducible factor (HIF)-1 [47, 68, 69, 71].

HIF-1 is a heterodimer DNA-binding factor. HIF-1 consists of an a and P

subunits, both of which have a number of isoforms. HIF-10 subunits are constitutive

nuclear proteins, while HIF-la subunits are hypoxia-inducible. There are three isoforms

for a subunit. HIF-1 a and HIF-2a appear closely related and are both able to interact with

hypoxia response elements (HREs) to induce transcriptional activity [72, 73]. In contrast,

HIF-3a appears to negatively regulate the response, through an alternatively spliced

transcript [74].

The molecular mechanism behind HIF-1 is a pathway that links oxygen

availability and the gene expression of various growth factors, especially VEGF. In

normoxia and hyperoxia oxygen-dependent prolyl hydroxylases hydroxylate HIF-1 a









proline residues, and this chemical modification leads to a HIF-1 capture by a ubiquitin

ligase complex that directs it to the proteasome for destruction. Under hypoxic

conditions, HIF-1 a is not hydroxylated, escapes ubiquitination, accumulates and directs

pro-angiogenic expression [75].

Vascular Endothelial Growth Factor (VEGF)

VEGF was originally discovered as the vascular permeability factor (VPF) that

increased the vascular permeability in the skin [76]. In 1989 Ferrara and Henzel

identified a growth factor for endothelial cells from bovine follicular pituitary cells and

named it VEGF [77], which was then proved to be identical to VPF [78, 79]. VEGF is the

most potent endothelial cell growth factor found to date. In the past two decades, this

growth factor has been studied extensively and its key roles in the proliferation,

migration, invasion, cell survival, differentiation of endothelial cells and other cell types

have been established. It is critical in the normal embryonic development of vasculature

and has essential functions in adults during normal physiological events such as would

healing, menstrual cycle, even though the mRNA levels of VEGF and its receptors

decrease significantly postnatally. Meanwhile, VEGF is also an important factor in

numerous pathological situations, many of which involve abnormal angiogenesis, for

example, inflammation, retinopathies, psoriasis, and cancer. Targeting VEGF signaling in

these diseases has been studied with enthusiasm and a number of novel drugs targeting

VEGF are being tested in clinical trials.

VEGF Family and Isoforms

The VEGF gene family consists of multiple variants, including VEGF-A

(hereafter referred to as VEGF), VEGF-B, VEGF-C, VEGF-D, VEGF-E, VEGF-F and

placental growth factor (P1GF-1 and P1GF-2 isoforms). They are secreted glycoproteins









that form homodimers, which belong to a structural superfamily of growth factors,

including the platelet derived growth factor (PDGF), characterized by the presence of

eight conserved cysteine residues [80, 81]. VEGF-A is believed to be the major

stimulator for vascular angiogenesis. VEGF-B is structurally similar to VEGF-A and

P1GF is highly abundant in heart, skeletal muscle and pancreas and may regulate

endothelial cell functions via a paracrine fashion [82]. VEGF-C and VEGF-D are

basically involved in lymphangiogenesis and induce the proliferation and cell survival of

lymphatic endothelial cells [83-85]. VEGF-E, encoded by the Orf virus, is structurally

similar to VEGF-A, specifically binds to VEGFR-2 and induces angiogenesis [86].

VEGF-F, as a collective name, summarized the variants isolated from snake venoms [87].

The term VEGF refers to a collection of related isoforms expressed from the same

gene [88]. The gene encoding VEGF, or VEGF-A, is located on the short arm of

chromosome 6 in humans [89] and on chromosome 17 in mice [90]. The vegfgene

consists of eight exons and seven introns, alternative splicing results in many isoforms.

The best studied isoforms in human are VEGF121, VEGF165 and VEGF189. In mice,

the homologous counterpart isoforms contain one less amino acid, so mVEGF164 is the

corresponding isoform for hVEGF165 [90], for example. In all isoforms, the transcrips of

exon 1-5 are all conserved and exon 6 and 7 are where the alternative splicing occurs.

Exon 3 and 4 encode the binding domains for VEGFR-1 and VEGFR-2 [91]. Exon 6 and

7 encode two heparin-binding domains, which influence receptor binding and solubility

[92]. VEGF 189, containing both the exon 6 and 7 transcripts, has high affinity for

heparin sulfate and is mostly associated with the cell surface and the extracellular matrix

[93]. On the contrary, VEGF165, lacking exon 6, is moderately diffusible; and









VEGF121, lacking both exon 6 and 7, is high diffusible [94]. Recently a new isoform

called VEGF165b, a variant of VEGF165, has been identified [95]. The C-terminus of

VEGF165b is encoded by exon 9, instead of exon 8 as in VEGF165 and other isoforms

[96]. VEGF165b binds to but does not trigger receptor phosphorylation, so it is actually

an endogenous inhibitory form of VEGF [96]. This is due to a missing exon 8-encoded

C-terminus, which has mitogenic signaling functions. Figure 1.13 [92] shows the

alternative splicing among VEGF isoforms.

Family Isoform Exons
















Figure 1.13. VEGF-A isoforms [92].
VEGF Receptors
189

165

165b

145

121

Figure 1.13. VEGF-A isoforms [92].

VEGF Receptors

VEGF binds to three cell surface receptor tyrosine kinases: VEGFR-1 (Flt-1),

VEGFR-2 (Flk-1/KDR) and VEGFR-3 (flt-4). VEGFR-1 and VEGFR-2 are primarily

located on vascular endothelium while VEGFR-3 is mostly found on lymphatic

endothelium. These receptors are structurally similar: all of them contain seven

extracellular immnoglobin (Ig)-like domains, a transmembrane domain, a regulatory

juxtamembrane domain, and a consensus tyrosine kinase domain interrupted by a kinase-

insert domain. The second and third Ig-like domains function as the high-affinity VEGF









binding domain, whereas the first and fourth Ig-like domains regulate ligand binding and

facilitate receptor dimerization, respectively [97-99].

VEGFR-1 has a molecular weight of 180 kDa and binds VEGF-A, VEGF-B and

P1GF. The affinity of VEGFR-1 for VEGF is ten-fold higher than VEGFR-2 but its

tyrosine kinase activity is ten-fold weaker than VEGFR-2 [92]. In the classical views, one

of the major functions for VEGFR-1 is to act as a decoy receptor restricting VEGF to

bind to VEGFR-2, which is more mitogenic [100]. VEGFR-1 is required for normal

blood vessel development during embryogenesis and a VEGFR-1 knock-out is lethal in

mice at embryonic day E8.5. The lethality was shown to be associated with an abnormal

increase in the number of endothelial progenitors, which is the phenotype as VEGF

hyperactivity, indicating a negative regulatory function of VEGFR-1 [101]. Supporting

this, a modified form of VEGFR-1 without the tyrosine kinase domain was constructed

and found to be compatible with normal vascular development and angiogenesis in

transgenic mice [102]. A naturally occurring soluble form of VEGFR-1, called sVEGFR-

1 or sFlt-1, is expressed from differential pre-mRNA splicing. sVEGFR-1 has the same

ligand affinity as VEGFR-1, but is missing the transmembrane and intracellular domains

[103, 104]. It binds to free VEGF and reduces its availability to VEGF receptors, which

further suggests its relative, VEGFR-1, as a negative regulator for VEGF signaling.

However, VEGFR-1 does mediate VEGF signaling in non-endothelial cells, especially

those cells that only express VEGFR-1 as the VEGF receptor, such as monocytes and

macrophages [105, 106]. A recent study showed that P1GF signaling mediated by

VEGFR-1 in monocytes is associated with the inflammatory reactions [107]. Besides









monocytes, VEGFR-1 signaling is also believed to be important for endothelial

progenitors and carcinoma cells.

VEGFR-2, a 230 kDa glycoprotein, is recognized as the primary mediator of

VEGF signaling. It regulates endothelial cell proliferation, migration, differentiation, cell

survival and vessel permeability and dilation. VEGFR-2 knock-out mice die between

E8.5 and E9.5 due to deficiency in blood vessel formation [108], indicating that VEGFR-

2 is also crucial for the functions of hematopoietic/endothelial progenitors. VEGFR-3,

170 kDa, binds to VEGF-C and VEGF-D. It is expressed in embryonic endothelial cells

but postnatally becomes restricted to the lymphatic endothelium [109].

Apart from these three VEGF receptors, neurophilins (NRPs) can also act as cell

surface receptor for VEGF, but in an isoform specific manner. NRP-1, originally

identified on neuron cells as a receptor for class 3 semaphorines/collapsins family of

neuronal guidance mediators [110], is also expressed on endothelial cells. It lacks the

intracellular tyrosine domain and needs to associate VEGFR-1 [111] and/or VEGFR-2

[112] to transduce a signal. It is suggested that NRP-1, as a co-receptor, can form a

receptor complex with VEGFR-2 to enhance the binding the signaling of VEGF 165 and

VEGFR-2 cannot sufficiently transducer the VEGF signaling without NRP-1 [113].

NRP-1 also binds VEGFR-1 forming a ligand-independent complex [111]. NRP-2,

lacking an intracellular domain like NRP-1, can also bind to VEGF. It can bind

VEGF121, VEGF145 and VEGF165, but NRP-1 cannot bind VEGF145. NRP-2 can also

bind to P1GF and can interact with VEGFR-1 [114]. In addition to NRPs, heparin sulfate

proteoglycans (HSPGs) can bind to the VEGF isoforms with the heparin binding

domains, such as VEGF165 and VEGF 189. HSPGs are abundant, highly conserved







components of the cell surface and the extracellular matrix of all cells and have been
reported to play a critical role in modulating the differential biological activities of VEGF
isoforms [115].
Figure 1.14 [116] demonstrates the binding of VEGF variants to the receptors. In
summary, VEGF-A binds to VEGFR-1, VEGFR-2 and the receptor heterodimer; VEGF-
C and VEGF-D bind to VEGFR-2 and VEGFR-3. Notably, P1GF and VEGF-B
exclusively bind to VEGFR-1 and VEGF-E exclusively binds to VEGFR-2, which is very
useful in receptor specificity studies.


~1


*VE


PIGF
VEGF-B




GFR-1


NRP-2


VEGFR-1


VEGF-A

h


NRP-1


VEGF-E VEGF-C
VEGF-F VEGF-D



S


II
VEGFR-2


NRP-1


VEGF exon 2-5
VEGF exon 7-8 -
HepaMn 0


IEGFR-3


Formation of blood vessels


Formation of lymphatic vessels


Figure 1.14. VEGF family ligands and their receptors [116].
VEGF Receptor Signaling
As mentioned above, VEGFR-2 is thought to be the major receptor for VEGF
signaling in endothelial cells. Upon binding of VEGF, VEGFR-2 is activated by


1


[Mimi"


E


H









autophosphorylation, and initiates a number of signaling cascades that induce cell

proliferation, migration, survival and/or increase in endothelium permeability.

The cell proliferation induced by VEGFR-2 signaling typically involves MAPK

pathways. Activation of VEGFR-2 recruits Grb-2 and activates it, which leads to the

activation of Sos, then the activation of Ras, eventually the stimulation of

Rafl/MEK/ERK signaling cascade [117]. Activated MAPK pathways will translocate to

the nucleus and regulate the gene expression and cell proliferation. VEGFR-2 can also

recruit PLCy-1, and the activation of PLCy-1 will induce phosphatidylinositol 4,5-

bisphosphate (PIP2) hydrolysis producing 1,2-diacylglycerol (DAG) and inositol 1,4,5-

trisphosphate (IP3). The activation of PKC can result from the production of DAG, which

further leads to the Ras-independent Raf activation and thus the stimulation of ERK

activity [118]. The data demonstrating the requirement of PI3 kinase in the VEGFR-2-

induced cell proliferation are conflicting, so the involvement of PI3 kinase is

controversial [119, 120]. Cells expressing VEGFR-1 are unable to activate MAPK [121].

VEGF can act as a chemoattractant for endothelial cells so that VEGF signaling is

believed to be involved in cell migration. Firstly, the signaling from activated VEGFR-2

can promote focal adhesion kinase (FAK) phosphorylation and recruit it to focal

adhesions, together with paxillin and actin-anchoring proteins like talin or vinculin [122,

123]. Therefore the cytoskeleton organization is modified and cell migration is promoted.

Secondly, the p38/MAPK pathway can be activated upon VEGF binding to VEGFR-2,

and thus may play a role in cell migration and p38 inhibitors can decrease cell migration

[124]. Thirdly, the PI3 kinase/Akt pathway can regulate the actin organization and cell

migration [125]. Besides VEGFR-2, VEGFR-1 and NRPs have all been implicated in









VEGF-mediated cell migration and invasion [92]. However VEGFR-2 is considered to be

the main mediator of cell migration. VEGFR-1 stimulates p38 phosphorylation and has

no effect on endothelial cell migration [126].

PI3 kinase/Akt pathway plays an important role in the VEGF-induced cell

survival. The phosphorylation of VEGFR-2 can lead to the activation of PI3 kinase and

Akt/protein kinase B (PKB). Akt is an anti-apoptotic factor and is sufficient to promote

cell survival. It has been reported that the inhibition of PI3 kinase abolished Akt

activation and the VEGF-mediated cell survival was also blocked [127]. VE-cadherin and

P-catenin can complex with VEGFR-2 and PI3 kinase and form a transient tetramer to

promote cell survival [128]. The expression of some anti-apoptotic factors can also be

induced by VEGF and contribute to cell survival, for instance, caspase inhibitors Bcl-1

and Al [129] and IAP (apoptosis inhibitors) family proteins [130]. VEGFR-1 cannot

associate with the VE-cadherin complex [128] and does not activate the PI3 kinase/Akt

pathway [127], so that it is thought to not be involved in VEGF-induced cell survival.

Originally discovered as a vascular permeability factor, VEGF can also increase

the vascular permeability. The administration of VEGF to endothelial cells is shortly

followed by the formation of some specialized regions in the cell membrane that are

highly permeable to macromolecules [131]. PI3 kinase and p38/MAPK have been

suggested to be involved in the increase of membrane permeability [132]. In the

established vessels, VEGF also regulates vascular permeability by affecting the

components of tight, adherence and gap junctions, such as VE-cadherin, p-catenin and

occludin [116]. Another aspect of this interaction is that endothelial NO synthase (eNOS)









can induce the activation of Akt, which further regulates the NO level and leads to vessel

dilation and permeabilization [133, 134].

Figure 1.15 [92] summarizes VEGF signaling via VEGFR-2.


VWEF Sigpaifg


Figure 1.15. VEGF signaling via VEGFR-2 [92].

The Function of VEGF in Ocular Neovascularization

VEGF is thought to play a central role in retinal angiogenesis as supported by data

from animal models and clinical investigation. VEGF is upregulated in the retina during

neovascularization in animal models with ischemia-induced retinopathy [135-138], and









the VEGF mRNA is increased by three-fold within 12 hours of the onset of relative

hypoxia and maintained for many days at higher levels until new vessels start to regress

[136]. Patients with active PDR were found to have increased levels of aqueous and

vitreous VEGF [139-145]. Higher levels of VEGF expression were also reported in

epiretinal neovascular membranes and retinas from PDR patients [146, 147]. However,

an interesting finding in the active PDR patients showed that there was a significant

decrease in VEGF levels after panretinal laser photocoagulation treatment [140]. Further

evidence supporting VEGF's major role in retinal neovascularization comes from VEGF

inhibition studies. VEGF receptor chimeric proteins, neutralizing antibodies, and

antisense oligonucleotides have successfully showed inhibition effects on

neovascularization [148-151].

Based on the evidence, it is widely accepted that VEGF is very important and

necessary for retinal neovascularization, but VEGF may not be sufficient for it. Repeated

intraocular injections of VEGF or sustained intravitreous release of VEGF in primates

results in severe changes to retinal vessels including dilation, leakage, and

microaneurysms, but no apparent retinal neovascularization [152, 153]. When VEGF

expression is driven by the retinal-specific rhodopsin promoter in the transgenic mice, the

development of neovascularization was produced in the deep capillary bed of the retina,

and high levels of VEGF expression can further cause retinal traction and detachment

[154]. The new vessels grew from the deep capillary bed into the subretinal space. The

close proximity of the deep capillary bed to the photoreceptor expressing VEGFs and

differential susceptibility of the vascular beds might be an explanation for this vascular

growth [155].









The role of VEGF in choroidal neovascularization (CNV) is less clear. Increased

VEGF expression was found in fibroblasts and RPE cells of choroidal neovascular

membranes surgically removed from patients [146, 156, 157]. And in the animal model

of laser-induced CNV, it has been shown that VEGF mRNAs were upregulated in the

neovascular lesions [158]. VEGF is thought to be necessary in CNV development

because several specific VEGF signaling inhibitors have shown reduced CNV [159-161].

But VEGF is not a sufficient stimulator of CNV because increased expression of VEGF

in photoreceptors or RPE cells does not lead to CNV [154, 162].

Basic Fibroblast Growth Factor (bFGF or FGF2)

FGF is a family of heparin-binding growth factors. bFGF has been localized in the

adult retina. In the mouse model of ischemia-induced retinopathy, bFGF level is elevated

during neovascularization [163]. In the animal model of laser-induced subretinal

neovascularization, RPE cells were found to be stained with aFGF and bFGF [164]. In

studies on clinical specimen, both elevated and non-significantly-changed levels of bFGF

have been reported in the vitreous sample of PDR patients [165, 166], which argues

against a major role in retinal neovascularization. Further evidence comes from animal

models. In the ischemia-induced retinopathy or laser-induced CNV mouse model,

transgenic mice deficient in bFGF developed the same amount of retinal or CNV as the

wild-type mice, respectively, indicating bFGF expression may not be necessary in

angiogenesis [167, 168]. It has been hypothesized that bFGF will manifest its angiogenic

potential when there is cell injury. It is found that bFGF can get access to the

extracellular compartment during photoreceptor damage and increased CNV can be

stimulated [169].









Angiopoietins

Angiopoietins and their receptors (Tie receptors) are another endothelial-specific

system that has been implicated in vascular growth and development. Current

understanding about the Tie receptors is that Tiel signaling is important for vascular

integrity and Tie2 signaling is important in remodeling of the developing vessels by

maximizing the interactions between endothelial and supporting cells [155]. The ligand

for Tiel has not been identified. Angiopoietin (Ang) 1 and 2 are ligands for Tie2

receptor. Angl binds with high affinity and initiates Tie2 phosphorylation and

downstream signaling. Ang2 also binds with high affinity, but does not stimulate

phosphorylation of Tie2. It looks like Ang2 is a naturally occurring antagonist for Angl

and Tie2. The interaction of Angl and Tie2 is essential for the remodeling function of

Tie2 on newly developing vessels. And it has been hypothesized that Ang2 might provide

a key destabilizing signal involved in initiating angiogenic remodeling. The Ang2

blockade of Tie2 signaling can disrupt "stabilizing" inputs to ECs, making ECs more

responsive to VEGF and thereby stimulating angiogenesis. But when there is no VEGF

present, those ECs are prone to apoptosis and the "destabilized" vessels regress [170].

Ang2 mRNA levels have been reported to increase in normal and pathological

retinal angiogenesis [171-174]. It has been shown that Ang2 can stimulate a significant

upregulation of proteinases in EC [174] that may be important for cell migration during

retinal neovascularization.

Platelet-Derived Growth Factor (PDGF)

PDGF, a dimer protein, a potent mitogen and a chemoattractant, has been

implicated in angiogenesis. Similar to VEGF, PDGF is another growth factor that is

elevated after hypoxia [65]. Recent findings about PDGF include: increased levels of









PDGF-AB was reported in vitreous samples of PDR patients [175]; overexpression of

PDGF-B in transgenic mice leads to proliferation of endothelial cells, pericytes and glial

cells resulting in traction retinal detachment [176-179]. It has been proposed that PDGF

may act in concert with VEGF in ischemic retinopathy [176-178].

Integrins

Integrins are a family of transmembrane proteins that are the major cell surface

receptors responsible for the attachment of cells to the extracellular matrix. Structurally,

integrins are heterodimeric receptors composed of two subunits, a and P. More than 20

different integrins are formed from the combination of 18 known a subunits and 8 known

p subunits. Each integrin binds to its own corresponding extracellular matrix (ECM)

and/or cell surface ligand. These include structural ECM proteins, such as collagens,

fibronectins, and laminins, as well as provisional ECM proteins that are deposited during

tissue remodeling and thrombotic events [180].The first integrin-binding site to be

identified was the sequence Arg-Gly-Asp, which is recognized by several integrins.

However, other integrins bind to other distinct peptide sequences. While integrins are one

of the most essential cell surface components in the body and are present in almost all

tissues, no cell expresses all integrins. Indeed the particular integrin types expressed are

dependent on the ECM ligands present within the local microenvironment. Even on a

given cell type, the specific integrins expressed are also altered to match the concurrent

changes within the local ECM. So the expression of integrin is spatially and temporally

regulated.

The integrins also function as an anchor for the cytoskeleton. The interaction

between the cytoskeleton and the extracellular matrix is responsible for the stability of

cell-matrix junctions. There are two categories of cell-matrix junctions: focal adhesion









and hemidesmosome. In focal adhesions the cytoplasmic domains of the 0 subunits of

integrins associate with bundles of actin filaments to anchor the actin cytoskeleton at the

cell-matrix junctions. While in hemidesmosome integrins interact with intermediate

filaments instead of actin. Hemidesmosome is mostly found in the anchorage of epithelial

cells to the basal lamina.

Integrin Signaling

Unlike many cell surface receptors that contain tyrosine kinases, integrins do not

contain intrinsic tyrosine kinase activity. Upon ligand binding, the integrins undergo a

conformational change into its activation state. The change in activation has been

assessed by showing evidence of polymerization, clustering, or the surface exposure of

different antibody binding epitopes [181]. Since the cytoplasmic domains of the integrins

can bind constitutively to cytoskeletal components such as talin, the conformational

change and activation of integrins can result in changes in cytoskeletal protein functions,

which will lead to major changes in cell shape and locomotion. On the other hand the

activation of integrins can initiate a series of signaling transductions, with the

involvement and assembly of a variety of signaling molecules.

A non-receptor protein tyrosine kinase called FAK (focal adhesion kinase) plays a

key role in integrin signaling. FAK is localized at the focal adhesion and is rapidly

tyrosine auto-phosphorylated following ligand binding by integrins. Besides FAK,

members of the Src family or non-receptor protein tyrosine kinases also associate with

focal adhesion and are involved in integrin signaling. Src and FAK probably interact with

each other, resulting from the binding of the Src SH2 domain to the auto-phosphorylated

sites of FAK. Src then phosphorylates additional sites on FAK. In addition to Src, the

binding sites for SH2 domain created during FAK phosphorylation are also taken









advantage of by other downstream molecules, for instance, PI-3 kinase and the Grb2-Sos

complex. These signaling molecules can form multicomponent signaling complexes that

recruit and include small GTPase proteins such as Ras, Rho, Rac. Their involvement and

activation will further lead to the activation of a number of signaling cascades. Figure

1.16 [180] demonstrates the integrin signaling via the Akt, ERK and JNK pathways.

These signals collaborate to regulate cellular proliferation, migration and survival. And

also, many small GTPases like Rho and Rac play critical roles in cytoskeletal remodeling

events [180].








Grb2 Shc
Sos
Ras
P13K (k
Nck
Raf. PDK/
I LK Rac
MEK PAK
AKT
ERK1/2 JNK


Proliferation
Survival
Migration


Figure 1.16. The activation of integrins can lead to the signal transduction in a number of
pathways. [180].

As mentioned above, integrins need to be activated to serve as a signaling

molecule. The activation involves a conformational change that results in an increase in

ligand-binding affinity. Proposed in the current model, the inactive form of integrins are









in a folded conformation in which the ligand-binding domain is adjacent to the

membrane. When activated, the affinity for the ligand is increased, and ligand occupancy

stabilizes the extended conformation of the integrin [182]. Simultaneously, the associated

topological change in the transmembrane and cytoplasmic domains makes them separate

and bind to intracellular signaling molecules to initiate downstream pathways [182].

According to this model, the conformational change in integrins that induces signaling is

the same as the one that is induced by activation. And this activation state can be

promoted by both extracellular ligands (so-called "outside-in" signaling) and intracellular

signaling molecules ("inside-out" signaling) [182]. The outside-in signaling is usually

triggered by ECM ligands and the inside-out signaling molecules are usually the effectors

of the activation of growth factor receptors. The ECM (local determiner) and growth

factors (systemic and local determiner) can work synergically to enhance the signaling

outcome induced by specific integrins in a given cell. Under certain circumstances it is

not sufficient to promote cell survival and proliferation until both proper ECM and

growth factors are both present.

The activation of integrins, especially those involving the interaction with growth

factor receptors, usually occur in lipid-raft microdomains, where cholesterol and

glycosphingolipids [183] and intracellular signaling molecules like Src family kinases

[184] are relatively concentrated in the cell membrane. These lipid-raft microdomains are

distinct from the surrounding membrane in that they restrict the diffusion of the contents.

It is suggested that the lipid-raft has other functions [182]. First, they could serve as a

physical concentration of pre-assembled molecules for signaling upstream or downstream

of the integrin, and the signaling inhibitory molecules could be excluded. Second,









different integrin pools could be separated so that their own distinct function could be

better performed. Third, the lipid-raft may also facilitate and/or maintain integrin

activation. In addition to help from concentrated pre-assembled molecules, the altered

membrane structure, due to the distinct chemical characteristics in the lipid-raft, may

favor conformational equilibrium between the inactive and the active forms. It is also

proposed that the active integrins might help to generate the lipid-rafts in other models

[185].

The integrins can regulate the signaling of growth factor receptors. First the

phosphorylation state of the growth factor receptors can be regulated. One example is the

interaction between av33 and the epidermal growth factor receptor (EGFR) on human

endothelial cells. The adhesion to the ECM mediated by integrin can lead to a low

phosphorylated state within the cell, resulting in the phosphorylation of four tyrosine

residues but not on the fifth tyrosine which is only phosphorylated by EGF binding. This

phosphorylated state is lower than in high concentration of EGF but ECM attachment

doesn't occur. This low phosphorylated state is sufficient to induce cell survival but not

proliferation. However, if only low concentrations of EGF are present, the ECM

attachment can promote the phosphorylation similar to high concentrations of EGF alone

[182]. Thus the phosphorylation of EGFR on endothelial cells is not only regulated by

ligand binding, but also regulated by integrins. The regulation on growth factor receptors

can also occur when integrins interfere with the receptor expression.

As for the inside-out signaling, the activation of growth factor receptors is usually

the source of signaling. Integrins can be regulated by growth factor receptors in many

aspects and cell behavior can be altered. The integrin expression level can be altered, for









instance, the expression level of a number of integrins on endothelial cells are increased

by angiogenic growth factors such as FGF-2 [186]. The phosphorylated state of integrins

can also be regulated by growth factor receptors. One example is the laminin receptor

a6p4, an essential component in the hemidesmosomes, influences epidermal cell

attachment to the underlying basal lamina. EGFR can induce the phosphorylation of the

cytoplasmic domain of 34 subunit. This results in the cytoplasmic recruitment of Shc, and

the activation MAPK and PI3K. More importantly, the change in the phosphorylation

state leads to release of the integrin from its ligand, thus the hemidesmosome

disassembles, which is a required step for cell proliferation and/or migration [187].

Besides the phosphorylation state, growth factor receptors can also alter the activation

state of integrins. For example, it has been shown that VEGF can activate avp3 on human

umbilical vein endothelial cells, thus the adhesion to ECM is promoted and cell migration

follows [188].

Relationships between Integrin and Other Growth Factor Receptors in
Angiogenesis

Among the over 20 integrins that have been discovered to date, two of them, avp3

and av35, are thought to be especially important for angiogenesis. These integrins are not

seen on normal epithelial cells in skin, but are highly expressed on endothelial cells

participating in angiogenesis [189]. Only avp3 was found in choroidal neovascular

membranes from ARMD patients, while both avp3 and av35 were found in epiretinal

membranes from DR patients [190]. Therefore, retinal and choroidal neovascularization

may differ in the integrin requirement. Inhibition studies on integrins further support this.

Agents that bind avp3 and/or av35 can suppress retinal neovascularization, even though









the effect is modest, but the inhibition of avp3 or av35 has no significant effect on

choroidal neovascularization [189].

Endothelial cells express at least eight different integrins including avp3 and av35

[191], each of them having their own specific ligand. For example, collagen is a ligand

for a2p31 while fibrin is a ligand for avp3, so that avp3 influences adhesion and signaling

events of the endothelial cells bound to fibrin [192] but not of those bound to collagen

[40, 193]. However, the endothelial cells will eventually become apoptotic when bound

with collagen alone via a2p1. The unligated avp3 receptors seem to cluster on the cell

membrane and colocalize with caspase activity, especially caspase 8 [194]. In addition to

avp3, many other unligated integrins are likely to induce cell death, this is why integrins

could be categorized as dependent receptors under a variety of circumstances.

avp3, expressed (although not exclusively) on endothelial cells, has been linked to

many angiogenic signaling pathways via the interaction with receptors for a number of

growth factors, such as VEGF, EGF, IGF-1, PDGF and insulin. Since VEGF and IGF-1

are the two most important growth factors involved in my dissertation work, I am

focusing on the interaction between avp3 and VEGFR and IGF-1R.

VEGFR-2 activation by phosphorylation is promoted by avp3 [195]. avp3 and

VEGFR-2 interact and the co-immunoprecipitation of these two receptors has been

demonstrated. However VEGFR-2 does not co-immunoprecipitation with the 31 or 35

subunits. VEGFR-2 phosphorylation and mitogenicity are enhanced in cells plated on

vitronectin, an avp3 ligand, compared with cells plated on fibronectin, an a531 ligand, or

collagen, an a23 1 ligand; further demonstrating a functional relationship between VEGR-

2 and avp3. Cell adhesion, migration, soluble ligand binding, and adenovirus gene









transfer mediated by avp3 are all enhanced by VEGFR-2 signaling. An anti-p3 integrin

antibody reduces VEGFR-2 phosphorylation and PI3 kinase activity suggesting that

VEGFR-2 signaling initiated by avp3 occurs through the PI3 kinase pathway.

Another molecule, p66 She (Src homology 2 domain containing), has been shown

to play a key role in the VEGF-avp3 interplay during tumor growth and vascularization

[196]. The activation state of avp3 integrin has a critical function in in vivo tumor growth

by influencing VEGF expression. By using a non-activable 03, a S752P mutant that

cannot cluster, it was found that the stimulation of VEGF expression also depends on

avp3 clustering. The recruitment of p66 She and phosphorylation of 33-associated p66

She are enhanced following avp3 clustering. The recruitment is not sufficient for avp3-

mediated effects on VEGF production and tumor vascularization but the phosphorylation

is necessary, in that a dominant-negative form of p66 Shc, which is phosphorylation-

defective, completely abolished integrin-induced VEGF expression.

IGF-1 is a classic endocrine hormone and systemically synthesized in liver and

transported to the peripheral tissues stimulating growth. In addition, IGF-1 is also

synthesized locally in peripheral tissue to promote growth in an autocrine/paracrine

manner. Similar to VEGF and other growth factors, the extracellular environment

contributes to influence the outcome of the hormone signaling. It has been shown that

many ECM proteins, such as collagen type I and type IV, fibronectin, thrombospondin,

and osteopontin, can modulate the response of various cell types to IGF-1 stimulation via

their integrin receptors [181]. The interactions between avp3 and IGF-1 on vascular

smooth muscle cells (SMC) have been illustrated in great detail and can be used as a

good example of how growth factors and integrin signaling influence each other.









When IGF-1 binds to the IGF l-R, IGF1-R will auto-transphosphorylate its two P

subunits, and further recruit signaling molecules such as insulin receptor substrate-1

(IRS-1) and Shc, which can transduce the singling into corresponding cascades, such as

the PI3K and MAPK pathways. Despite kinases, phosphatases also participate in the

signaling modulation. Phosphatases induce dephosphorylation reactions, which can result

in either activation or inactivation of signaling molecules. One phosphatase, Src

homology 2 containing tyrosine phosphatase (SHP-2), normally transfers to IGF-1R 20

minutes after IGF-1 stimulation, resulting in a decrease in the phosphorylation level of

the receptor and subsequent attenuation of MAKP and PI3K activation [181]. However, a

premature transfer at 5 minutes and premature attenuation has been found when the

ligand occupancy of avp3 is blocked [197]. So obviously the properly liganded and

activated avp3 is a necessary partner in IGF-1R signaling.

Normally when IGF1-R and av33 are activated after ligand binding, SHP-2 will

transfer to the phosphorylated 33 subunit first. An adaptor protein, DOK-1, facilitates the

transfer. DOK-1 is phosphorylated after IGF-1 stimulation, and the YXXL motifs within

its C-terminus domain become capable of binding to SHP-2 via SH-2 domains [198].

Also, DOK-1 contains a phosphotyrosine binding (PTB) domain, which allows it to bind

to p3 at a tyrosine that is phosphorylated after avp3 activation [199]. Thus DOK-1

mediates SHP-2/03 association. If the transfer of SHP-2 to 33 is impaired for any reason,

SHP-2 will be aberrantly transfer to IGF-1R instead and the premature dephosphorylation

of IGF-1R occurs [181].

One SHP substrate, SHPS-1, becomes phosphorylated after IGF-1R activation. It

is a single chain transmembrane protein and SHP-2 can bind to it via SH-2 domain. The









transfer of SHP-2 from p3 to phosphorylated SHPS-1 is a necessary step to maintain

optimal MAPK and PI3K activation [200]. SHPS-1 also recruits She to form a complex

that is critical for MAPK and PI3K activation. SHP-2 can activate a Src family kinase via

SH-2 domain binding, so that this Src family kinase is recruited to SHPS-1 and

phosphorylates She in the complex [181]. SHP-2 is further transferred to the appropriate

downstream signaling molecules to maintain MAPK and PI3K activation.

avp3 has several ECM ligands, such as osteopontin, thrombospondin and

vitronectin. For avp3 on SMC, the major ECM ligand is vitronectin. The heparin binding

domain and RGD (arginine-glycine-asparginine) sequence can both function as the avp3

binding site. It is believed that the heparin binding domain is the binding site triggering

p3 activation, in that the exposure of cells with the heparin binding domain peptide

results in avp3 phosphorylation and recruitment of SHP-2 to the plasma membrane [201].

Contrarily, binding of p3 to the RGD sequence has been found to induce the cleavage of

p3, thus also the premature recruitment of SHP-2 to IGF-1R and the premature IGF-1R

dephosphorylation [202].

Similar to the interaction between integrins and other receptors, it is believed that

avp3 and IGF1-R signaling occurs within a restricted compartment on the membrane.

Integrin-associated protein (IAP) facilitates the formation of this compartment. After

IGF-1 exposure, IAP is translocated to the regions where avp3 resides [181]. More

importantly, IAP can induce an increase in the affinity of avp3 for its ligands [203]. The

extracellular domain of IAP can associate with SHPS-1 and an antibody disrupting this

association prevents IGF-1 stimulation of SHPS-1 phosphorylation and SHP-2 transfer to









SHPS-2 [204]. Therefore, the clustering of avp3 and the assembly of a signaling

complex involving SHPS-1 may be a crucial in av33 and IGF-1R signaling.

Pigment Epithelium-Derived Factor (PEDF)

The vasculature is normally quiescent under physiological conditions, since there

is a balance between the pro-angiogenic and anti-angiogenic factors. Angiogenesis is

initiated when there is increase in pro-angiogenic factors and/or decrease in anti-

angiogenic factors. PEDF is one of the naturally occurring anti-angiogenic factors.

In the mouse model of retinopathy, it has been shown that hyperoxia results in a

decline of VEGF levels with a concomitant expression of PEDF, and the relative hypoxia

led to downregulation of PEDF during the angiogenesis process [205]. Systemic or

intravitreal administration of PEDF [206, 207] and gene transfer with adenoviral vectors

expressing PEDF [176-179] have been reported to decrease the ocular neovascularization

levels, In the clinical studies, The vitreous levels of PEDF from PDR patients were found

to be lower than normal [208], and the immunochemical staining of PEDF on retinas

from PDR patients are much less intense compared with non-PDR [208]. All of these

evidence supports that PEDF, an anti-angiogenic factor, may be involved in the

suppression of retinopathies.

Insulin-Like Growth Factor (IGF)-1

The discovery of a role of growth hormone (GH)/IGF-1 in DR can be traced back

to 1950s. The regression of retinal neovascularization was seen after pituitary infarction

[209], and pituitary ablation was even used as a therapeutic method for PDR. More

recently, in several studies in patients with PDR, elevated serum and vitreous levels of

IGF-1 have been associated with retinal neovascularization [210-212].









In a GH inhibition study, retinal neovascularization was suppressed in transgenic

mice expressing a GH antagonist gene and normal mice treated with an inhibitor of GH

secretion [213]. This inhibition of neovascularization could be reversed by exogenous

administration of IGF-1. IGF-1 also plays a necessary role in normal retinal vascular

development. In IGF-1 knockout mice, normal development of the retinal vasculature

was arrested despite the presence of VEGF [214]. This also supports the idea that VEGF

alone is not sufficient for the development of retinal vessels. Clinically it has been found

that the development of ROP in premature infants was strongly associated with a

prolonged period of low levels of IGF-1 [214]. This suggests that the critical role IGF-1

plays during normal retina vascular development. Lack of IGF-1 in the early neonatal

period leads to the development of avascular retina, and later the proliferative phase of

ROP [155]. The function of IGF-1 in CNV is still not clear.

The IGF system includes the IGF-1, IGF-2, the IGF-1 receptor (IGF-1R), and IGF

binding proteins (IGFBPs). IGF-1 can be expressed in the liver and utilized systemically

as an endocrine, or can be expressed at peripherals and function in autocrine/paracrine

mechanisms. The multiple physiologic and pathologic effects of IGF-1 are primarily

mediated by IGF-1R, and are also modulated by complex interactions with IGFBPs,

which themselves are also modulated at multiple levels.

IGF-1 and IGF-1R

IGFs are synthesized in almost all tissues and have important regulatory function

on cell growth, differentiation, and transformation. IGF-1 is the product of the IGF-1

gene, which has been mapped to chromosome 12 in humans and chromosome 10 in mice

[215]. IGF-1 functions in both prenatal and postnatal development and exerts all of its

known physiological effects through binding with IGF-1R. Circulating IGF-1 is









generated in the liver under the control of growth hormone [216], and bound with

IGFBPs as the endocrine form in the circulation. The IGF-1 produced in other organs and

tissues has a lower affinity for IGFBPs, representing autocrine and paracrine forms of

IGF-1. The IGF-1R gene is located on chromosome 15 in human [215], and IGF-1R is

expressed everywhere in the body. The mature receptor is a tetramer consisting of 2

extracellular a-chains and 2 intracellular 3-chains with the intracellular tyrosine kinase

domain. IGF-1R signaling involves autophosphorylation and subsequent tyrosine

phosphorylation of She and insulin receptor substrate (IRS) -1, -2, -3, and -4. IRS serves

as a docking protein and can activate multiple signaling pathways, including PI3K, Akt,

and MAPK. The activation of these signaling pathways will then induces numerous

biologic actions of IGF-1 (Figure 1.17 [216]).


Figure 1.17. IGF-1 signaling transduction [216].









The expression of IGF-1 in ECs is low, but it is expressed both in macrovessel

and microvessel ECs. IGF-1 stimulates vascular EC migration and tube formation. IGF-1

is important for promoting retinal angiogenesis, and an IGF-1R antagonist suppresses

retinal neovascularization in vivo by inhibiting vascular endothelial growth factor

(VEGF) signaling [217]. The effect of IGF-1 on ECs is mediated in different signaling

pathways. For example, IGF-1-induced nuclear factor-KB (NF-KB) translocation requires

both PI3K and extracellular-regulated kinase, while IGF-1-stimulated EC migration

requires only PI3K activation [218]. And the IGF-1 effects are also regulated by

endothelial nitric oxide synthase (eNOS) expression and VEGF signaling [217].

IGF-1 and IGF-1R are also expressed in vascular smooth muscle cells (VSMCs),

and their expressions are regulated by several growth factors in different pathways.

Thrombin and serum deprivation, tumor necrosis factor (TNF)-a, and estrogen

downregulate IGF-1 mRNA and protein levels; reactive oxygen species (ROS) increases

the levels; Ang2 and PDGF have been reported to both increase and decrease the levels.

IGF-1 functions as a potent mitogen and antiapoptotic factor and migration stimulator for

VSMCs [216]. As for the IGF-1R, its expression can be upregulated by Ang2 via the

activation ofNF-KB [219]; can be upregulated by fibroblast growth factor (FGF),

mediated by the transcriptional factor STAT1, STAT 3 [220]; and the Ras-Raf-MAPK

kinase pathway was shown to be required for both of the above growth factor. The cross-

talk between IGF-1R and other receptors can also regulate IGF-1 function. For instance,

blocking ligand occupancy of aV33 integrin receptor results in premature recruitment of

SHP-2 to the IGF-1R receptor and reduces IGF-1 signaling [200].








IGFBPs and ALS

At least 6 IGFBPs have been well characterized, and they function as transporter

proteins and as storage pools for IGF-1. The expression of IGFBPs is tissue- and

developmental stage-specific, and the concentrations of IGFBPs in different body

compartments are different. The functions of IGFBPs are regulated in multiple ways,

such as phosphorylation, proteolysis, polymerization [221], and cell or matrix association

[222] of the IGFBP. All IGFBPs have been shown to inhibit IGF-1 action, but IGFBP-1, -

3, and -5 are also shown to stimulate IGF-1 action [223]. Some of IGFBPs' effects might

be IGF-1 independent.


Proteolysis


IGF IGFO

0 k Extracellular
IGFRI

Intracellular



Inhibition
Proliferation

Figure 1.18. Proposed pathway of IGF-dependent IGFBP action [223].

The precursor forms of IGFBPs have secretary signal peptides and mature

proteins are all found extracellularly. They all have a conserved amino-terminal domain,

a conserved carboxyl-terminal domain and a non-conserved central domain. Both of the

amino-terminal and carboxyl-terminal contribute to IGF binding [223], which implies









IGF-binding pocket structure. The major IGF transport function can be attributed to

IGFBP-3, the most abundant circulating IGFBP. It carries 75% or more of serum IGF-1

and IGF-2 in heterotrimeric complexes that also contain the acid labile subunit (ALS)

[224]. Free or binary-complexes (without ALS) are believed to exit the circulation

rapidly, whereas ternary complexes appear to be essentially confined to the vascular

compartment. In addition to their effects derived form circulation, IGFBPs also have

local actions, both autocrine and paracrine. They have been documented to affect cell

mobility and adhesion [225, 226], apoptosis and survival, and cell cycle [227-229]. I will

concentrate on IGFBP-3 in this discussion.

IGFBP-3 have both potentiation and inhibition effect on IGF-1 actions. It is

thought that IGFBP-3 inhibits IGF-1-mediated effects via its high-affinity sequestration

of the IGF-1. But in contrast, preincubation of cells with IGFBP-3 before IGF-1

treatment can lead to the accumulation of cell-bound forms of IGFBP-3 with lowered

affinity for IGF [230], which may enhance the presentation of IGF-1 to IGF-1R. But It

was also found that cell-bound forms of IGFBP-3 could still attenuate IGF-1-mediated

IGF-1R signaling [231]. It has also been reported, based on competitive ligand-binding

studies, that IGFBP-3 can interact with IGF-1R, causing inhibition of IGF-1 binding to its

receptor [232]. Therefore, the interaction of IGFBP-3 with IGF-1 and IGF-1R signaling

system requires further study. Limited digestion from proteases on IGFBP-3 can release

IGF-1 from the complex and control the bioavailability of IGF-1. These specific

proteases include serine protease, cathepsins, and matrix metalloproteinases [223].

Proteolysis results in IGFBP-3 fragments with decrease affinity for IGF-1, but several

studies have shown the inhibition of IGF actions by IGFBP-3 fragments with low affinity









for IGFs [223]. It is not clear whether this inhibition comes form IGF-1 sequestration or

from its interaction with IGF-1R. IGFs themselves can also influence the production of

IGFBPs and IGFBP-specific proteases, or regulate the activity of these proteases [223].

Figure 1.18 [223] summarizes proposed IGFBP actions that depend on binding of IGFs

and modulation of IGF-1R.



DNA damage -- p534 baxd bd-2t

Antiestrogens 1 I
Retinoic acid I_____FBP-3 Apoptosis
Vitamin D
TNF-a 0 \t as
IGFBP-3K/ %%tPprESes

TGF-p r IGF-I
p21/wafl
Growth factors a
Mutation- ras_
Cell cycle Migration


Figure 1.19. Overview of possible IGFBP-3 antiproliferation pathways [223].

IGFBPs also have their own intrinsic bioactivity, without modulating IGF actions,

either in the absence of IGFs (IGF-independent effects) or in the presence of IGFs

without triggering IGF-1R signaling (IGF-1R-independent effects). Recently there has

been particular interest in IGFBP-3's function to induce apoptosis independently of

inhibiting the survival functions of IGF-1 [233-236]. Several studies using human breast

cancer cells have correlated the induction of IGFBP-3 mRNA and protein expression

with growth-inhibitory effects of various antiproliferative agents including TGF-3,

retinoic acid [237], antiestrogens [238], vitamin D analogs [239], and TNF-a [240].

IGFBP-3 expression is also upregulated by the transcription factor p53 in colon









carcinoma cells. And in the experiments using antisense IGFBP-3 or specific antibodies

to sequester the IGFBP-3, the antiproliferative effects of some of these factors and be

partially abrogated [223]. In addition, there is evidence showing that some proteolyzed

forms of IGFBP-3 also have IGF-independent effect, especially some IGFBP-3 amino-

terminal fragments [223], and they showed little or no affinity for IGFs. This supports the

existence of IGF-independent bioactivity. Figure 1.19 [223] summarized some of the

proposed pathways of IGFBP-3 independent functions.

IGFBP-3 has IGF-1 independent effects. Interactions of IGFBP-3 with known

signaling pathways have been demonstrated. The type V receptor for TGF-P (TORV) has

been shown to be bound with IGFBP-3 relative specifically and may be involved in

IGFBP-3 inhibitory signaling [241]. IGFBP-3 has been shown to stimulate the

phosphorylation of T3RI of the signaling intermediates Smad2 and Smad3 [242], while

TORV signaling does not involve Smad phosphorylation. All-tans-retinoic acid (RA) is a

potent inducer of IGFBP-3 in some cancer cells [223]. The growth-inhibitory effect of

RA requires the presence of RA receptor (RAR)-P and can be blocked by retinoid X

receptor (RXR)-specific retinoids. IGFBP-3 has been shown to inhibit RA signaling,

possibly through enhancing RXR signaling [223]. IGFBP-3 may also interact with PI3-

kinase pathway and MAPK pathway. LY294002, an inhibitor of PI3-kinase activity,

could block the effects of IGFBP-3 [243]; MAPK/ERK pathway inhibitor, PD98059, can

restore the inhibitory effect of IGFBP-3 on DNA synthesis, blocked in cells expressing

oncogenic ras, in breast epithelial cells [244]. Recently it has been shown that IGFBP-3

strongly up-regulate signal transducer and activator of transcription 1(STAT1) mRNA in

the process of chondrocyte differentiation, and phospho-STAT protein was shown to









increase and translocate to the nucleus, moreover, the antiproliferative effects of IGFBP-

3 in these cells can be ablated in the presence of STAT1 antisense oligonucleotide [245].

The acid-labile subunit (ALS), together with IGFBP-3 and IGF-1, forms the

ternary complex as the storage pool in the plasma. ALS is synthesized almost exclusively

by the liver, and predominantly stimulated by GH [246]. Presence of ALS after birth is

coincident with increased responsiveness to GH resulting from an increase in GH

secretion and hepatic GF receptors. After puberty, ALS concentrations basically remain

stable throughout adulthood [246]. ALS is a single copy gene, containing 2 exons and 1

intron. ALS has no affinity for free IGFs and very low affinity for uncomplexed IGFBP-

3, and even its affinity for binary complex (IGF-1 + IGFBP-3) is 300-1000 fold lower

that that of IGFBP-3 for IGFs [247]. The ability of ALS to form ternary complex is

irreversibly destroyed under acidic conditions. IGFBP-3 and IGFBP5 can both associate

with ALS, with the latter being much weaker [246]. The carboxyl-terminal domains of

IGFBP-3 and IGFBP5 are important for binding. The association is proposed to happen

within the negative-charged sialic acid on the glycan chains of ALS and an 18 amino acid

positive-charged domain in IGFBPs [246].

Besides liver, ALS local synthesis may occur in kidney, developing bone,

lactating mammary gland, thymus and lung [248, 249]. Their functions are to sequester

IGFs into ternary complex. A GH-responsive element of the ALS gene transcriptional

promoter was identified [250]. This sequence was called ALSGAS1 because of its

resemblance with the consensus sequence for y-interferon activated sequence (GAS). The

effects of GH on the ALS gene are mediated by the JAK-STAT pathway [251, 252]: the

tyrosine kinase JAK2 is recruited to the activated GH receptor complex and









phosphorylates signal transducers and activators of transcription (STAT)-5a and STAT-

5b. After dimerization, STATS isomers translocate to the nucleus, and activate ALS gene

transcription by binding to the ALSGAS1 element. The GH signaling pathway leading to

increased ALS gene transcription is critically dependent on the activation of STATS

isomers, and is independent of RAS activation.

One the of physiological significance of ALS is to extend the half-lives of IGFs

from 10 min when in free form, and 30-90 min when in binary complexes, to more than

12 hours when in ternary complexes [253]. The other important role of ALS is to prevent

the non-specific metabolic effects of the IGFs, given that serum IGF concentration is

-1000 fold that of insulin [246]. IGFs in ternary complexes cannot traverse capillary

endothelia and activate the insulin receptor, whereas free IGFs and IGFs bound as binary

complexes can do so. Incorporation of IGFs into ternary complexes therefore completely

restrains the intrinsic insulin-like effects of the IGFs. Null ALS mouse shows

significantly reduced circulating IGF-1 and IGFBP-3 concentrations [246], which proves

that ALS is absolutely necessary for serum accumulation of both IGF-1 and IGFBP-3.

The Involvement of Insulin Receptor (IR) and IGF-2 in Angiogenesis

IGF-1 primarily binds to IGF-1R, and insulin primarily binds to IR, while IGF-2

can bind to both of the two receptors and its own IGF-2R, as shown in Figure 1.20 [254].

Regarding retinopathy, insulin and IGF-1 have gained more attention. Kondo et at [255],

using the Cre-Lox knockout system, found that (1) the retinas of mice develop normally

in the absence of endothelial IR or IGF-1R. Presumably, sufficient growth factors (for

example, VEGF) are present to facilitate normal development. (2) Under conditions of

relative hypoxia and in the presence of endothelial IR/IGF-1R, VEGF, eNOS, and ET-1

are increased, leading to extra-retinal neovascularization. (3) Under conditions of relative








hypoxia and in the absence of endothelial IR or IGF-1R, VEGF, eNOS, and ET-1 are

reduced, possibly due to impaired HIF-1 activation or reduced PI3K activity related to

IG/IGF-1R [256]. Reduced neovascularization results from less IR/IGF-IR input. And in

their experiments, the reduction of VEGF, eNOS, and ET-1 are reduced to a greater

extend in IR knockout mouse than IGF-1R knockout mouse, which has brought more

emphasis on IR function in the retinopathy, while traditionally IGF-1R is thought to be

more important.

Ligands Insulin IGF-2 IGF-1 Other peptides



Receptors IR IGF-2/M-6-P IGF.1R IRR




Substrates IRS family, She, Crk, Gab


I
Metabolic and growth promoting responses

Figure 1.20. The crosstalk between IGF-1, IGF-2 and Insulin signalings [254].
RNA Silencing Technologies
The traditional method to inactive a gene is to create a gene knockout animal model. This

process has its advantages, in that it entirely abolishes a gene expression, however, the

disadvantages are that it is time consuming, expensive, labor-intensive, and subject to

possible failure due to embryonic lethality [257]. RNA silencing technologies, which









inhibit gene expression at the RNA level, are valuable tools to inhibit the


DNA



mRNA
AntisesI I N 1 small interfering RNA
Antisense Ribozyme / !ll NA
Oligonucleotide I DNA Enzyme


RISC
RNase H


X, Translation
blocked

Protein





Figure 1.21. Overview of RNA silencing technologies [258].

expression of a target gene in a sequence-specific manner, and may be used for functional

genomics, target validation and therapeutic purposes. Theoretically, RNA silencing could

be used to cure any disease that is caused by the expression of a deleterious gene [258].

There are three common types of anti-mRNA strategies. Firstly, the use of single

stranded antisense oligonucleotides; secondly, the triggering of RNA cleavage through

catalytically active oligoribonucleotides referred to as ribozymes; and thirdly, RNA

interference induced by small interfering RNA molecules. Figure 1.21 [258] basically

summarized the mechanisms of these three kinds of antisense technologies. This scheme

also demonstrates the difference between antisense approaches and conventional drugs,

most of which bind to proteins and thereby modulate their function. In contrast, RNA

silencing agents act at the mRNA level, preventing translation. Antisense-









oligonucleotides pair with their complementary mRNA, whereas ribozymes and DNA

enzymes are catalytically active oligonucleotides that not only bind, but can also cleave,

their target RNA. RNA interference is a highly efficient method of suppressing gene

expression in mammalian cells by the use of 21-23-mer small interfering RNA (siRNA)


molecules. These three RNA silencing methods are detailed below.

Antisense Oligonucleotides

The antisense oligonucleotides was first described by Zamecnik and Stephenson

who used a 13-mer DNA to inhibit Rous sarcoma virus expression in infected chicken

embryonic fibroblasts [259]. The antisense gene silencing naturally occurs in genomic

imprinting, in which only one copy of a gene in the mammalian genome is expressed

while the other is silenced. It could be the maternally inherited allele or the paternal

inherited allele.

Antisense oligonucleotides are complementary to the target mRNA and are

usually 15-20 nucleotides in length [258]. There are two major antisense mechanisms that

have been proposed [258]. First, RNase H cleaves RNA in the RNA-RNA heteroduplex

(or RNA:DNA heteroduplex for antisense DNA oligonucleotides), induced by binding of

the antisense oligonucleotides. This results in rapid degradation of the cleaved mRNA

products and a reduction in gene expression. Second, translation is arrested by steric

blocking the ribosome by the binding of antisense oligonucleotides. When the target

sequence is located within the 5' terminus of a gene, the binding and assembly of the

translation machinery can be prevented.

The first step in designing an antisense oligonucleotides is target selection and

verification of target site accessibility. Computer programs, like Mfold, perform mRNA









secondary structure analysis. This analysis can generate several mRNA secondary

structures centered on our target sequence. If the target is always contained within a

stable stem in every structure, this target should be eliminated. In addition to this type of

in silico analysis of RNA secondary structure, a number of in vitro methods have been

developed to examine secondary structure in solution. One way is to directly probe the

secondary structure of the target RNA with 1-cyclohexyl-(2-

morpholinoethylo)cabodiimide metho-p-toluene sulfonate (CMCT) [260]. CMCT will

mainly modify Us, and to a lesser extent Gs, in single-stranded regions of an RNA

molecule. CMCT modification is followed by reverse transcription. Modification of Us

and Gs will prevent read-through by reverse transcription, resulting in a pause or stop site

at the modified position. When these modification/reverse transcription reaction products

are separated on an appropriate electrophoresis gel next to DNA sequencing reactions of

the target mRNA region, accessible regions of the target RNA are easily identified. The

most sophisticated approach reported so far is to design DNA array to map an RNA for

hybridization sites of oligonucleotides [261].


B -- Base



O OH Ribose (2' OH group)
I
O =P--O Phosphate backbone

0,


Figure 1.22. Modifications in antisense technology [258].

When designing antisense oligonucleotides, there are some points to consider.

Four contiguous guanosine residues should be avoided due to the G-quartets formation









and CpG motifs should be avoided due to potential stimulation of the immune system. In

addition, a BLAST search for each oligonucleotide sequence is required to avoid

significant homology with other mRNAs that could cause unwanted gene silencing.

Unmodified oligonucleotides are rapidly degraded in biological fluids by

nucleases. So one of the major challenges for antisense RNA approaches is the

stabilization of RNA oligonucleotides. Chemical modifications of the bases and/or and

phospho sugar backbone have been developed to increase resistance against RNase

(Figure 1.22 [258]. The major representative of in the first generation modification is the

Phosphorothioate (PS) oligonucleotides, in which one of the nonbridge oxygen atoms in

the phosphodiester bond is replaced by sulfur [258]. The shortcomings include binding to

certain proteins, such as heparin-binding proteins, and their slightly reduced affinity to

the complementary RNA sequences [262]. In the second generation, most the emphasis

was placed on the 2' hydroxyl group. 2'-O-methyl and 2'-O-methoxyl-ethyl RNA are the

most common types of modifications [258]. However, RNase H cleavage can be

somewhat reduced or even blocked with these types of modifications, possibly due to the

steric blockade. One way to overcome this disadvantage is the gapmer technology [258],

in which the 2'-modified nucleotides are placed only at the ends of antisense

oligonucleotides. This protects the ends from degradation and a contiguous stretch of at

least four or five non-2'-modified residues in the center are sufficient for the activation of

RNase H. A variety of modified nucleotides have been developed in the third generation,

the antisense oligonucleotides properties such as target affinity, nuclease resistance and

pharmacokinetics have been improved [258]. The concept of conformational restriction

has been used widely to help enhance binding affinity and biostability.










Ribozymes

Ribozymes, or RNA enzymes, are catalytic molecules that can catalyze the

hydrolysis and phosphoryl exchange at the phosphodiester linkages within RNA resulting

in cleavage of the RNA strand. There are two types of chemical reactions that are

catalyzed during phosphate-group transfer by naturally occurring ribozymes: self-

cleaving and self-splicing reactions. The ribozymes that perform self-cleaving reactions

include hammerhead, hairpin, hepatitis delta virus (HDV) and Neurospora Varkud

satellite (VS) ribozymes. They are usually small RNAs of tens of nucleotides in length.

The ribozymes that perform self-splicing reactions include self-splicing introns and

RNase P. They are much larger in size and usually hundreds of nucleotides in length.

a Self-cleaving i- I


b Self-splicing
5'k
O N-1


(D OH
-O-P-O
I N+1
R-H'
0 OH
3'-


5,-
O-


O(H) 1 r
S 0O OH
'q


5'
O N-1




S0 0-

HO N+1


O OH
k3'

5'
O N-1


OH OH
0-
R-O-P=O
O N+1


O OH
%3'


Figure 1.23. Self-cleaving and self-splicing reactions in ribozymes [263].









As shown in Figure 1.23 [263], in the self-cleaving reactions, the RNAs catalyze a

reversible phosphodiester-cleavage reaction. The nucleophilic attack from the 2'-

hydroxyl group results in 5'-hydroxyl and 2'-3'-cyclic phosphate termini. The bridging

5'-oxygen is the leaving group. While in the self-splicing reactions, an exogenous

nucleophile attacks on the phosphorus generates a 5'-phosphate and a 3'-hydroxyl

termini. The bridging 3'-oxygen is the leaving group. In the first steps of group I intron

and group II intron self splicing and the RNase P-mediated cleavage of precursor of

tRNAs, the exogenous nucleophiles are, respectively, the 3'-hydroxyl group of

exogenous guanosine, the 2'-hydroxyl group of an adenosine in the intron, and the water.

They are indicated by the ROH in Figure 1.23 b. The transition states are shown in

brackets.

Self Splicing Introns

Self splicing introns can be divided into 2 classes based on the conserved

secondary structure and splicing mechanisms: Group I and Group II. Group I is found in

a variety of species, including prokaryotes and lower eukaryotes. Except for the

Tetrahymena large rRNA group I intron, all other known group I introns require a single

protein co-factor to provide a scaffold to hold the RNA in the catalytic reaction [264].

Group II introns are found within nuclear pre-mRNA and organelle pre-mRNA [265]. A

spliceosome consisting of proteins and small nuclear RNAs (SnRNA) is formed in the

catalytic reaction and high concentrations of magnesium and potassium are necessary

[265].

The splicing action of both group I and group II introns consists of two similar

consecutive transphosphoesterification reactions. In the first step, the 5'-end of the intron

is attacked by an exogenous nucleophile, which is the 3'-hydroxyl group of exogenous










guanosine in group I introns, or the 2'-hydroxyl group of an adenosine in group II

introns. This results in the cleavage at that site and the addition of the guanosine or

adenosine to the 5'-end of the intron. In the second step, the oxygen in the 3'-hydroxyl

group of the 3'-end of the up stream exon attacks the 3'-end of the intron. In group I

introns, it is a guanosine at the 3'-end of the intron that is attacked. This cleaves the 3'-

end of the intron, releasing the intron, and results in ligation of the upstream and

downstream exons. Figure 1.24 [263] shows the secondary structure and self splicing

steps of group I introns.

a P91 b CG -- 3'
A --- [P9.1-P9.2]
Hinge A G
A 320 u 5'- r
A*GACA U
G-C UA G
G-C 20G-C A
P5a C-G G U12 G
G-C U
-U-A P5 20- -U U
C-G A C
GC-G U.G J4/5 G.U* c C
AAU A .- U-A G ",---- 3'
AA*U A *-A
U A A P1 U- GC-w P9.0
U-A C-G
-A AC-G 5' 5'
P5o UuG-C C-G2 -Gste
P G5c GU 4 G-C-CA Gse
AAAGG C-G-U J6 G
C AI UA P7
SG A140 GAU-U I -AA
U G P6 "-A A 1
A-U 1 6a I A4A-4
loG-C J6/6a J3/4 G A
A-U I C
G-C 220 U A
U-A U -A J8/7 1003'
P5b UG C-G P 5 H

AA A [P2-P2.1] ___ ">---A
UG.UG P6a A P3

U -A C 280
C-G U
A-G, U A U 5.AIj

U-A C o2U
C-G
-A -U
P6b A-U U P8
C-G U
A-U
G -C 240 U1
GoA U U 5' 3'
AG AG -

P4-P6 P3-P9


Figure 1.24. Secondary structure and self splicing steps in group I intron [263].










RNase P

RNase P is a ribonucleoprotein complex that removes the 5' leader sequence form

precursor tRNAs (ptRNAs) via a hydrolysis reaction. It consists of a catalytic RNA

subunit (Ml RNA in E. coli) and a protein subunit (C5 protein in E. coli) [266, 267]. In

vitro, Ml RNA can cleave its ptRNA substrate without C5 protein, but the reaction

requires high concentrations of Mg2+. However, C5 protein can dramatically increase the

rate the cleavage, even at low concentration of Mg2+ [268]. In vivo, C5 protein is required

for RNase P activity and cell viability [266, 267]. Thus both the RNA subunit and the

protein subunit are essential for RNAse P function. It has been proposed that C5 protein

can facilitate the stabilization of the Ml RNA conformation and also enhance the enzyme

and substrate interaction [269, 270].

RNase P 5'
CCA3C CCA3
S -S'-ldrC CCA3'
Aoceptor stem
D stem-loop
Gude sequence (GS)
-T-loop

Anclidon sinm-loop Vaa'ble loop

0 3'tail
3'5'
ptRNA 4.5S RNA


Figure 1.25. Secondary structures of natural and synthetic substrates for RNAse P[275].

All the natural substrates of RNAse P (ptRNAs, precursor of 3.5S RNA and

several small RNAs [271-273] in E. coli) have a common feature in their secondary

structure which includes a 5' leader sequence, and acceptor-stem-like structure and a 3'-

CCA sequence. A synthetic external guide sequence (EGS) combined with a CCA

sequence has been designed to base pair with a targeted sequence to form a structure very









similar to the natural substrates of RNase P. The Ml RNA from E. coli can cleave at this

synthetic target site [274]. This EGS-based technology can be used to guide RNAse P to

cleave a targeted sequence. Figure 1.25 shows the secondary structures of ptRNA and the

3.8s RNA and the hybridization of the EGS with the targeted sequence [275].

Hammerhead Ribozymes

The hammerhead ribozyme was the first small self-cleaving RNA to be

discovered [276, 277], the first ribozyme to be crystallized [278, 279] and the smallest

naturally occurring catalytic RNA identified so far. It was found in several plant virus

satellite RNAs and is required for the rolling circle mechanism of virus replication [280].

The hammerhead ribozyme cleaves the multimeric transcripts of the circular RNA

genome into single genome length strands.

Hammerhead ribozymes are approximately 30-90 bases in length and cleave RNA

targets in trans. Annealing of the hammerhead ribozyme with the target sequence

produces a structure consisting of three stems, a tetra-loop and a conserved catalytic core

as shown in the Figure 1.26. Any mutation in the catalytic core will prevent catalytic

cleavage. The catalytic core has two functions: it destabilizes the substrate strand by

twisting it into a cleavable confirmation, and also binds the metal cofactor (Mg2+) needed

for catalysis [278]. The absolute requirement of the target sequence is a NUX cleavage

site, where N is any nucleotide and X is any nucleotide except G. The targeting arms of

the hammerhead ribozyme bind either side of the U of the NUX site forming stems I and

III. GUC has been shown to be the most efficient cleavage site [281], followed by CUC,

UUC and AUC. The advantages of hammerhead ribozymes include its small size, easy of

cloning and packaging into viral delivery systems, and versatility in target site selection.








In the traditional view, the Mg2+ and water are both required in the

transesterification reaction. The hydrated magnesium ion can help to provide an

environment to facilitate the nucleophilic attack, in which the Mg2+ acts as a Lewis acid

to coordinate directly with the 2'-hydroxyl and the 5'-leaving oxygen for activation of the

nucleophile and for stabilization of the environment. It has been also reported that some

monovalent cations (Li and NIH4) at higher concentration can substitute for Mg2+ [282].

There is another kind of antisense agent called DNA enzyme, which is similar to

the hammerhead ribozyme in structure and function but avoids the high susceptibility to

nucleases that is common to ribozymes. The best studied DNA enzyme, named "10-23"

[283], consists of a catalytic core of 15 nucleotide and two substrate recognition arms. It

is highly sequence-specific and can cleave any junction between a purine and a

pyrimidine, and its efficiency is similar to hammerhead ribozymes [283].

3' 5'
N-N

i N-N
N-N
N-N
GAAA- U

SC GGCI NNNNNN-3'
U GCCG UNNNNNN-5'
U A C
II GUAG III


Figure 1.26. Structure of the hammerhead ribozyme.

Hairpin Ribozymes

Similar to the hammerhead ribozyme, the hairpin ribozymes was first derived

from tobacco ring spot virus satellite RNA [284].When the hairpin ribozyme binds to the








substrate, a structure with 4 helices and 2 loops is formed. Helix 1 (6 base pairs) and

helix 4 (4 base pairs) are where the hairpin hybridizes to the target RNA. In loop A, a

BNGUC target sequence is required for cleavage, where B is G, C or U, and N is any

nucleotide [285]. Figure 1.27 shows the structure of the binding complex of the hairpin

ribozyme and its substrate.



UG 2
1 C A -
3'-NNNNNA NNNN-5'
1 1 1 11 1 1 I E E
5'-N N N N NU N N N NA-U-3'
A A C-G
GA
C-G
A A-U
GAG-CCA
G A
A U
U
A
B A
-A U
C A
A C-GGU
A-U
C-G
G-C 4
C-G
U-A
C-G
G A
UA

Figure 1.27. Structure of the hairpin ribozyme.

Hairpin and hammerhead ribozymes can also catalyze the ligation of the cleaved

products, which is the reverse of the cleavage reaction. The ligation efficiency is much

higher for the hairpin than the hammerhead. Another unique feature for the hairpin

ribozyme is that it does not require metal ions as cofactors [282, 286].









Hepatitis Delta Virus (HDV) Ribozymes and Neurospora Varkud Satullite (VS)
Ribozymes

HDV ribozymes and VS ribozymes also cleave the substrates via self-cleaving

reactions. HDV ribozymes are derived from the genomic and the anti-genomic RNAs of

HDV [287, 288]. Naturally, the HDV ribozyme cleaves its substrate during the rolling

circle replication mechanism of the circular RNA genome, like other self cleaving

ribozymes. The VS ribozyme was originated from the mitochondria of certain isolates of

Neurospora [289]. The self cleaving reactions require a divalent cation but it has also

been shown that monovalent cations are be sufficient for the ribozyme to catalyze

proficiently [282].

RNA Interference

RNA interference (RNAi) is a naturally occurring process and is a potent

sequence-specific mechanism for post-transcriptional gene silencing (PTGS). It was

described early in C. elegans [290] and then found to exist throughout nature as an

evolutionarily conserved mechanism in eukaryotic cells. RNAi has regulatory roles in

gene expression, such as genomic imprinting, translation regulation, alternative splicing,

X-chromosome inactivation and RNA editing [291]. In plants and lower organisms RNAi

also protects the genome from viruses and insertion of rogue genetic elements, like

transposons [292].

Figure 1.28 [293] shows the RNA interference pathways. Long double-stranded

(ds)RNA is cleaved by Dicer, an RNase III family member, into short interfering RNAs

(siRNAs) in an ATP-dependent reaction. These siRNAs contain an approximately 22-

nucleotide (nt) duplexed region and 2-nt unpaired and unphosphorylated 3'-ends. The 5'-

end is phosphorylated, which is a crucial requirement for further reactions. In fact, if the











-III

Drosha Dicer

S(Nucleus)


Pri-miRNA Pre-miRNA


dsRNA

1, Dicer
5' kinase
t11 11111111111 Iiiin1 < ; 1iiiiiiiiii
miRNA/siRNA Synthetic siRNA

4 Helicase



J Assembly of single-stranded
miRNA/ siRNA into RISC





miRNA-RISC siRNA-RISC


Target RNA Target RNA


Inhibition of translation Degradation of target
(No PROTEIN) (No RNA, No PROTEIN)
Figure 1.28. RNA interference [293].

siRNA is introduced into human cells as a synthetic molecule, its 5' hydroxyl gets

phosphorylated shortly after entry into the cells [294-296]. These siRNAs are then

incorporated into the RNA-inducing silencing complex (RISC). Although the uptake of

siRNAs by RISC is independent of ATP, the unwinding of the siRNA duplex requires

ATP. The unwinding favors the terminus with the lower melting temperature as the start

point. Thus the termini containing more A-U base pairs are preferred as the unwinding

start point. The strand whose 5'-end is at the start point will be used by RISC as the guide

sequence and the other strand is release and degraded. Once unwound, the guide strand

positions the RISC/siRNA complex with the mRNA that has a complementary sequence

to the siRNA, and the endonucleolytic cleavage of the target mRNA occurs. The target









mRNA is cleaved at the single site in the center of the duplex region between the guide

siRNA and the target mRNA. The microRNA (miRNA) pathway is another RNA

silencing pathway and is similar to siRNA. miRNA is also approximately 22-nt long, but

it is a product of a sequential processing on a single-stranded RNA by two enzymes of

the RNaseIII superfamily [297-299]. The long primary transcript (pri-miRNA) is cleaved

by a nuclear enzyme, named Drosha in human, into an approximately 70-nt long pre-

miRNA. The pri-miRNA is basically a short hairpin RNA (shRNA) and is further

processed in the cytoplasm by Dicer to produce the final miRNA. For both siRNAs and

miRNAs, the perfect or near-perfect match will lead to the degradation of the target upon

association with RICS, and mismatches will repress the translation. It now appears that at

least seven continuous complementary base pairs are required for cleavage [300].

Introduction of synthetic siRNAs as a mimic for the Dicer cleavage process

triggers the RNAi machinery. In addition, siRNAs or miRNAs produced form shRNA

expression cassettes can be cloned into RNA expression vectors to produce the self-

complementary hairpin sequences that induce the RNAi pathway. More importantly, the

shRNA expressed from a vector could establish long-term silencing of a targeted gene

expression. The transcription of shRNA from the vector is usually conducted using an

RNA polIII promoter such as the H1 or U6 promoter [301, 302]. U6 promoter strongly

favors a G residue at the first position of the transcribed sequence and H1 weakly prefers

an A residue [303]. The transcription mediated by polIII promoters terminates after the

second or third (less commonly) residue of a "TTTTT" stretch, which results in a 3'-UU

tail that forms the 3'-2-nt unpaired overhang end in the hairpin structure after self

complementarily annealing of the transcript. Both the preference of first residue and the









3'-2-nt unpaired UU end influence the target site selection. Similar to miRNAs, shRNAs

(in the nuclei) are bound by a complex consisting of the nuclear export factor Exportin 5

(Exp5) and the GTP-bound form of the cofactor Ran [304, 305]. For nuclear export, this

complex requires an RNA stem of 16 bp, a short 3'-overhang and a terminal loop of >

6 nt [304, 306]. The efficient cleavage by Dicer requires an RNA of >19 bp and a short

3'-overhang [307]. These prerequisites can be easily met when designing the shRNA

expression cassette.

Considering the strand preference of RISC during unwinding, the 3' end of the

guide strand in shRNA is designed tightly base-pair (higher CG contents) and the 5' end

of the guide strand is designed loose base-pair (higher AU contents). As an example

shown in Figure 1.29 [303], two GC base pairs at the 3'-end of guide strand (red) are

designed. More AU pairs at the 5- end of the guide strand would be appreciated for

correct unwinding and even a mismatch can be included. Actually bulges resulting from

mismatches are always present in natural pri-miRNAs and they may help to fine-tune the

cleavage sites used by Drosha and Dicer and/or may preclude activation of dsRNA-

responsive cellular signaling pathway like interferon responses [303]. During the design

of the shRNA, it is encouraged to include a bulge close to the 5' end of the guide strand,

which should be done by a introducing a mutation into the to-be-degraded (sense) strand,

not into the guide (antisense) strand. It has also been reported that an A residue at

position 3 and a U at position 10 of the sense strand can enhance siRNA function

significantly. And a G at position 13 of the sense strand may need to be avoided [308].

Figure 1.29 [303] shows the sequence of the designed shRNA, with the reference to









human pre-miR-1 sequence and structure. The blue strand is sense and the red strand is

antisense. Arrows mark the Dicer cleavage sites.


C A AUA
miR-1 5'-CCAUGCUUC UUGCAUUC AUA GUU U
3'-GAGGUAUGAAG AAUGUAAG UAU CGA
A G A ACU



N N AUA
shRNA 5'-GCANNNNNNU NNNNNNN NNN GUU U
3'-UUCGUNNNNNNA NNNNNNN NNN CGA
N N A ACU




mRNA target 5' -AAGCANNNNNNUNNNNNNNNNN-3'

Figure 1.29. Designing artificial shRNA for RNAi [303].

It has also been found that small dsRNA that are 25-30 nt in length requiring

RNAi processing appear to be more efficient in inducing RNAi than smaller 22 nt

siRNAs [309], which could be due to the fact that Dicer may direct endogenously

processed siRNAs and miRNAs to the RISC complex. This gives vector-expressed

shRNA an extra advantage over synthetic 22 nt siRNAs. Multiple shRNAs or siRNAs

can be introduced into the cell simultaneously, but it is worth keeping in mind that the

RNAi machinery can be limiting [310] so that the competence between exogenous

shRNAs and endogenous miRNAs, or between exogenous and endogenous siRNAs, for

limited amount of Dicer and RISC could occur, which would interfere with the cell's

endogenous RNAi pathways. Particularly when the cell is undergoing a cell division, the

RNAi machinery could be diluted and adversely affected by inhibiting the gene knock-

down mechanism.









RNAi is highly specific to its target and not toxic in almost all situations;

however, when designing an siRNA or shRNA, some off-target effects should be

considered and avoided. dsRNAs that are 30 nucleotides or longer tend to trigger at least

two cellular stress response pathways, both of which will lead to a general and non-

specific abrogation of protein synthesis, or even apoptosis [295, 311]. The IFN pathway

is usually a mechanism to eliminate virus-infected cells, in which the long dsRNA binds

to and activates the dsRNA-activated protein kinase (PKR). PKR can further

phosphorylate the translation initiation factor, eIF-2a, and induce global translation

inhibition and even apoptosis. In another pathway, dsRNA activates 2'-5' oligoadenylate

synthetase. The 2'-5' oligoadenylate will then be formed and bond to and activate RNase

I, resulting in non-specific degradation of RNAs. Although siRNA or shRNA, less than

30 nucleotides in length, usually do not activate these stress response pathways. In highly

sensitive cell lines and at high concentrations, a subset of interferon genes can be

activated [312-314]. In the designing of siRNAs or shRNA, the ones that have significant

homology to other irrelevant mRNAs should be avoided. As noted before, a seven

consecutive base pairing can be enough to activate the RISC-induced gene silencing.

Even the guide strand (antisense) has been designed to introduce RISC to the target site

after unwinding, it is still possible that unwinding could initiate from the 5' end of the

sense strand and thus sense strand would guide the RISC. The homology of the sense

strands should also be checked.

Vector-mediated expression of shRNA can lead to long-term RNAi and the

silencing effect has been observed even after two months [302]. The half-life of

unmodified siRNAs in vivo is only seconds to minutes [315]. The most important reason









for this short half-life is the rapid elimination by kidney filtration due to the small size

(-7 kDa). Endogenous serum RNases can degrade the siRNAs limiting the serum half-

life to 5-60 minutes. The half-life can be extended in a number of ways, for instance,

completing the siRNAs with other molecules or incorporating them into various types of

particles to bypass renal filtration [315-317], chemically modifying the ribose [316, 318-

320], or capping the ends of the siRNA [315, 320]. The modification on the ribose

usually takes place at the 2'-position; 2'-deoxyribose, 2'-O-metheyl and 2'-fluoro

substitutions/modifications have been reported [316, 318-320]. Usually the silencing

effects are affected more or less by these modifications but a modified siRNA, with two

2'-O-methyl at the 5' end and four methylated monomers at the 3' end, has been

demonstrated to be as active as its unmodified counterpart [321]. Even though siRNAs

have the potential to activate interferon pathways, no toxic effects after siRNA

application have been observed [258]. There is no strict specific sequence requirement in

RNA interference (although there are preferred bases at some positions), and, therefore,

the range of target for siRNA is greater that with ribozymes or antisense therapies.

Gene Therapy Overview

With the progress of Human Genome Project, people are reaching a new level of

understanding of many biological events, including the etiology of diseases with or

without proved treatment. Especially for those diseases currently without treatment,

finding the genes that are involved in the initiation and development of the diseases

provides new treatment targets.

The most common gene therapy targets are monogenic recessively inherited

diseases such as hemophilia [322]. In the treatment of these diseases, gene therapy is

designed to introduce a functional gene into a target cell to restore protein production that









is absent or deficient due to the genetic disorder. Conversely in monogenic dominantly

inherited diseases like hypercholesteroleamia [323], successful treatment requires the

aberrant gene to be silenced, and this is usually done by means of gene-silencing

technologies. Cancer, as an acquired genetic disease, is also a good candidate for gene

therapy. Apart from expressing functional tumor suppressor genes and silencing activated

oncogenes, gene therapy in cancer treatment has also been applied to introduce the

expression of immunopotentiation proteins, the expression of a toxic product in

transformed cells, and the expression of proteins in healthy cells helping the cell to be

resistant to higher doses of chemotherapy [324].

The methods to deliver a gene into cells can be roughly categorized into virus-

based system and non-viral system.

Non-Viral Gene Delivery

The gene transfer in non-viral system is in general inefficient and often transient

compared with viral vectors, but it has advantages such as low toxicity, simplicity of use

and ease of large-scale production. In addition, the transient expression of a therapeutic

gene would be desirable in the treatment of certain conditions, such as retinopathy of

prematurity. Basically there are three categories of methods for non-viral gene delivery:

naked DNA in the form of plasmid, liposomal packaging of the DNA and molecular

conjugates.

Naked DNA is the simplest way to delivery a gene. It is not very efficient and can

result in prolonged low levels of expression. The simplest way is to inject directly into

the tissue of interest or inject systemically from a vessel. The expression level and area

are usually limited in a systemic injection due to the rapid degradation by nuclease and

clearance by mononuclear phagocyte system. To facilitate the uptake of naked DNA,









several techniques, in addition to simple injection, have been developed. The Gene Gun

is a technology to shoot gold particles coated with DNA which allows direct penetration

through the cell membrane into the cytoplasm and even the nucleus, bypassing the

endosomal compartment [325]. Electroporation, the application of controlled electric

fields to facilitate cell permeabilization, is another way to facilitate DNA uptake. Skin

and muscle are ideal targets due to the ease of administration. Ultrasound can also

increase the permeability of cell membrane to macromolecules like plasmid DNA and has

been used to facilitate the gene transfer.

Liposomes are lipid bilayers entrapping a DNA fragment with a fraction of

aqueous fluid. It can naturally merge onto cell membrane and initiate the endocytosis

process. To improve transfection efficiency, target proteins recognized by cell surface

receptors have been included in liposome to facilitate uptake, for example, anti-MHC

antibody [326], transferring [327], and Sendai virus of its F protein [328], which help

DNA to escape from endosome into cytoplasm thus to increase DNA transportation to the

nucleus. The inclusion of a DNA binding protein on the liposome also enhances

transcription by bringing the plasmid DNA into the nucleus [328].

Molecular conjugates are usually a synthetic agent that can bind to DNA and a

ligand at the same time [324]. Thus the binding of the ligand to its receptor will initiate

the receptor-mediated endocytosis for the complex. This method is more specific for

different cell types and receptor types. The synthetic agent needs to be designed

accordingly, but this is useful in tissue-specific transfection. The transgene expression in

this method tends to be transient and limited by endosomal and lysosomal degeneration.









Viral Gene Delivery

Viral gene delivery systems are based on replicating viruses that can deliver

genetic information into the host cell. According to the existence status of the viruses, the

virus vectors can be divided into two categories: integrating and non-integrating [329].

Integrating virus include adeno-associated virus, retrovirus, and so on. These viruses can

integrate the viral genome into chromosomal DNA so that a life-long expression of

transgene could be possibly achieved. Adenovirus and herpes simplex virus fall into the

category of non-integrating viruses. They deliver viral genome into the nucleus of

targeted cell, however the viral genome remain episomal, so it is possible that the

transgene gets diluted during cell divisions.

Generally speaking, genomes of replicating viruses contain coding regions and

cis-acting regulatory elements. The coding sequences enclose the genetic information of

the viral structural and regulatory proteins and are required for propagation, whereas cis-

acting sequences are essential for packaging of viral genomes and integration into the

host cell. To generate a replication-defective viral vector, the coding regions of the virus

are replaced by a transgene, leaving the cis-acting sequences intact. When a helper

plasmid or virus providing the structural viral proteins in trans is introduced into the

producer cell, production of non-replicating virus particles containing the transgene is

established. An ideal viral vector should have these characteristics: 1) The virus genome

is relatively simple and easy to manipulate; 2) The viral transduction can yield high

vector concentration in the producer cells (>108 particles /ml); 3) The vector should have

no limitation in size capacity; 4) The viral vector can transduce dividing and non-dividing

cells; 5) The vector can deliver the transgene as integration in the host cell genome or as

segregation being an episome along with cell division so that sustained expression can be









established; 6) The vector has a naive or modified tissue specificity and the transgene

expression can be regulated; 7) The vector produces no or low immune response and

allows subsequent re-administration. [330]

The expression specificity can be regulated in many aspects. For tissue specificity,

we can pick the virus that has the right tropism specific to some tissue, and in addition

tissue-specific promoters can be added to further define the specificity. For spatial

specificity, radiation in conjugation with radiation-activated promoter (for example, ergl

promoter [331]) would be a good method. Of course the local delivery into the right place

is always preferred than systemic administration, if feasible. For temporal specificity,

drug-inducible promoters can provide a convenient way to switch the transgene

expression on and off. The drug can be used to work on transcription activation or

repressor elements to modulate the expression. There are many established drug-

regulated gene expression systems, such as rapamycin-regulated gene expression [332]

and RU486-regulated gene expression from GAL4 site [333]. And for promoters

containing binding site for hormone receptor, heavy metals or cytokines, these specific

hormone, heavy metals and cytokines can also be used to induce the expression.

Adeno-Associated Viral (AAV) Vectors

AAV is currently the virus closest to an ideal vector that is under study and

application. It belongs to the family of parvovirus; it is non-pathogenic and depends on

helper virus (usually adenovirus (Ad) or herpes virus) to proliferate. It is a non-enveloped

particle with a size of 20-25 nm and has a vector capacity of 4.7 kb [334]. AAV can

infect both dividing and non-dividing cells, with the transduction efficiency best in S-

phase of host cell cycle. The viral genome, coded in a single-stranded DNA, has two

open-reading frames (ORF). One is rep, which is responsible for viral structural proteins,









integration and replication proteins. The other is cap, coding for capsid proteins. There

are inverted terminal repeats (ITR) at both ends of the genome sized around 150 bp, T-

shaped and forming palindromic structure. TR is GC rich and contains a promoter. Due to

the integration into the host genome, AAV vector can potentially deliver a long term

expression of the transgene. Another advantage of AAV is that it induces overall low

immune response. Presence of circulating neutralizing antibodies is in the majority of

populations, but they don't prevent re-administration or shut down promoter activity

[329]. Small packaging capacity is the number one disadvantage of AAV vectors. Using

concatamers, formed by head-to-tail recombination in ITRs, up to 10 kb oftransgenes

can be packaged for delivery [335], by means of splitting promoter and transgenes

sequences over two AAV vectors. But this technology reduces transduction efficiencies.

The infection of a host cell starts when the viral particle binds to its receptor on

the cell membrane and initiates the endocytic pathway. The receptor type varies with

AAV serotypes. The AAV-2 serotype, the most studied and commonly used serotype, has

as its primary receptor heparin sulfate proteoglycans (HSPG) [336]. HSPG is widely

expressed in various tissues and this is why AAVs have a wide tropism. There are also

co-receptors for AAV-2 to facilitate endocytosis. Fibroblast growth factor receptor-1

(FGFR1), one of the co-receptors, can enhance the virus attachment to the cells [337].

Integrin avp3, another co-receptor, can facilitate endocytosis in the clathrin-mediated

process, and it may also activate Racl and further phosphorylate PIP3 Kinase [338],

which leads to microfilaments and microtubes rearrangement to support AAV2

trafficking to the nucleus. After entering the cell, the viral particle is released from the

endosome at low pH conditions. Low pH probably induces a conformational change of









viral proteins and thus helps with endosome release and nuclear entry [339]. The viral

particle is uncoated in the nucleus, and ssDNA is duplicated into dsDNA by either

annealing with a complementary DNA strand from a second AAV or by the host cell

machinery. The duplication from ssDNA to dsDNA is the rate-limiting step in AAV

transduction. To overcome this, self complementary vector (scAAV) has been designed

to expedite this process [340]. With the help of rep proteins, the viral genome or the

transgene is integrated to a specific site in chromosome 19 via a non-homologous

recombination and will be expressed by host cell transcriptional machinery. Some virus

may remain episomal and also get expressed. Figure 1.30 [330] summarizes major steps

in the AAV internalization and intracellular trafficking.


AAV Vector

Receptor Particle
Binding AZ L Transduced Cell


Figure 1.30. AAV internalization and intracellular trafficking [330].









AAV has a number of serotypes. AAV-1 and AAV-4 were isolated from simian

sources; AAV-2, -3, -5 were isolated from human clinical specimens; AAV-6 is thought

to be the recombination of AAV-1 and AAV-2 (AAV- 's 3' end recombined with AAV-

2's 5' end), and AAV-7 and AAV-8 were isolated from rhesus monkey [330]. They have

their own tropisms, which are determined by the capsid proteins. For example, AAV-2 is

preferred to use for infection of the human eye, spine, while AAV-1 has the highest

transduction efficiency in muscle and liver, and AAV-5 has high tropism for retina and is

able to transduce airway epithelial cells.

Among all the serotypes, AAV-2 is the most studied and commonly used. As with

all the AAV serotypes, the AAV-2 genome has two ORFs, rep and cap, which span over

90% of the genome. As shown in Figure 1.31 [330] Panel A, in the ORF of rep, there are

two promoters, p5 and p19, encoding four proteins. Rep 78 and its splicing variant,

Rep68 are transcribed from p5. They play important roles in replication, transcriptional

control and site-specific integration. Rep52 and its splicing variant, Rep40 are transcribed

form p19. They are important for the accumulation of single-stranded genome used for

packaging. The other ORF cap encodes for VP1, VP2 and VP3 which are transcribed

from p40. They are capsid proteins and have pivotal roles in tropism specificity. These

three proteins are expressed in the ratio of 1:1:20, making the capsid with icosahedral

symmetry. The ITRs at both ends of the viral genome have a couple of functions. The

detailed structure and sequence of ITR is shown in Figure 1.31 [330] Panel C. First, the

3' end of the ITR on the 5' end the genome serves as primer in the synthesis of a new

DNA strand. Second, ITRs contain Rep binding site (RBS) for Rep78 and Rep68 and






83


help them work as a helicase and an endonuclease. Third, the terminal resolution site

(TRS) is identical to a sequence in chromosome 19, serving as integration sequence [341].




A AAV Genome
ITR E ITR


ps psl p4
I .p7. I

epa68 f I
Fmp52 ..I____.._............ I
lep4 2 l ,
Mep42

VP1 (



I kb Pi


B Vector Genome
IT R Promoter Trunegene pA ITR




C
S| Inverted Terminal Repeat


cc
'1 "



Figure 1.1. AAV2 genome and the vector genome [30].





When making an AAV viral vector, the two ORFs and the viral promoter are all


replaced by a transgene and the only cis elements needed for AAV integration, packaging
and assembly are the ITRs. The vector genome is shown in Figure 1.31 [330] Panel B.
cc
SC



Figure 1.31. AAV2 genome and the vector genome [330].

When making an AAV viral vector, the two ORFs and the viral promoter are all

replaced by a transgene and the only cis elements needed for AAV integration, packaging

and assembly are the ITRs. The vector genome is shown in Figure 1.31 [330] Panel B.










rep and cap will be provided in trans in another plasmid, and helper virus gene products

(Ela, Elb, E2a, E4 and VA RNA from Ad) are also provided in trans. Originally the

vector production method is to co-transfect the HeLa cells with transgene plasmid, the

plasmid providing rep and cap, and wide type Ad, or to co-transfect human 293 cells

with the transgene plasmid, rep and cap plasmid, and El-deleted Ad, as the Elgene

products can be provided endogenously in 293 cells. Recently helper virus-free system

has been designed to minimize the safety issues. See Figure 1.32 [334].


recAAV



AAV helper,
eenome


Ad auxiliary
function


+



+
_I ""r vw '" lli a1 "


or or




wMd AdAl


HeLa celk 293 celv
Transitory packaging cells

Figure 1.32. Helper virus -free systems in rAAV production [334].

The helper virus-free system has the three-plasmid system and the two-plasmid

system [334]. In the three-plasmid system, besides AAV vector plasmid and AAV helper

plasmid providing rep and cap genes, an Ad helper plasmid is introduced to provide the

helper virus gene products (E2A, E4 and VA RNA from Ad) and human 293 cells are

used as the host cell to provide Ad El gene products. The best molar ratio for these three


EpHlpe




Full Text

PAGE 1

TARGETING ANGIOGENIC GROWTH FACT ORS IN PROLIFERATIVE DIABETIC RETINOPATHY By HAO PAN A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2006

PAGE 2

Copyright 2006 by Hao Pan

PAGE 3

This document is dedicated to the graduate students of the University of Florida.

PAGE 4

iv ACKNOWLEDGMENTS Since coming to the United States in A ugust 2001, it has been five years. This was the most challenging five years and th ere was happy and hard time. To study abroad, especially in the United Stat es, was one of my dreams when I was in high school. Now, with the completed dissertation in ha nd, I can tell myself: Hao, you made it! Language has been the biggest obstacle in my study. I was confident about my English, but I came here and found that there is still so much to learn and it still takes time. The study and life for me has been harder than most American students. But I am happily seeing my improvement everyday. I co mposed my dissertation in English, gave seminars in English and passed the final defe nse in English; all of these are making me proud. I thank my mentor, Dr. Maria Grant, for her patient and inspiring guidance in the past four years. Every member in my co mmittee, Dr. Alfred Lewin, Dr. Sean Sullivan and Dr. Stratford May, has given me great s uggestions for my disse rtation work. I thank everybody in the lab. Dr. Lynn Shaw instru cted me in great details during my experiments and dissertation wr iting. Dr. Aqeela Afzal was also a great help for my bench work. And every other member in th e lab has given me great support for my defense. I thank my parents. They are far away in China but I am sure they are proud and as happy as I am now. They have done everyt hing they could to provide me the best education opportunities and they have always been there encouragi ng all the way along. I

PAGE 5

v am the only child in the family and I am tha nkful that they suppor ted when I decided to study abroad. I thank Yao for her great help and support. Her love strengthened me during the hardest time. Without her, I could not have overcome all the difficu lties and successfully graduated. There is still a long way ahead, with mo re challenges and opportunities. I would cherish everything I have had in the Universi ty of Florida. The orange and blue will always be a source of courage and confidence. Go Gators!

PAGE 6

vi TABLE OF CONTENTS page ACKNOWLEDGMENTS.................................................................................................iv LIST OF TABLES...............................................................................................................x LIST OF FIGURES...........................................................................................................xi ABSTRACT....................................................................................................................... xv CHAPTERS 1 BACKGROUND..........................................................................................................1 Introduction and Project Aim.......................................................................................1 The Eye........................................................................................................................ .2 The Anatomy of the Eye........................................................................................2 The Retina.............................................................................................................2 The Blood Supply to the Retina............................................................................4 Retinopathies................................................................................................................5 Age-Related Macular Degeneration (ARMD)......................................................6 Retinopathy of Prematurity (ROP)........................................................................9 Diabetic Retinopathy (DR)..................................................................................11 Current Treatments for Retinopathies.................................................................13 Pathogenesis of Diabetic Retinopathy.................................................................16 Increased Polyol Pathway Flux....................................................................17 Production of AGE.......................................................................................17 Generation of Reactive Oxygen Species......................................................19 Activation of Diacylglycerol a nd Protein Kinase C Isoforms......................19 How Does the Change in Retinal Blood Flow Occur?.................................20 What Causes Retinal Ca pillary Cell Death?................................................21 What Causes Retinal Ischemia?...................................................................21 Angiogenesis and Growth Factors..............................................................................22 Vasculogenesis and Angiogenesis.......................................................................22 Hypoxia-Induced Factor (HIF)............................................................................23 Vascular Endothelial Gr owth Factor (VEGF).....................................................25 VEGF Family and Isoforms.........................................................................25 VEGF Receptors..........................................................................................27 VEGF Receptor Signaling............................................................................30 The Function of VEGF in Ocular Neovascularization.................................33

PAGE 7

vii Basic Fibroblast Growth Factor (bFGF or FGF2)...............................................35 Angiopoietins......................................................................................................36 Platelet-Derived Growth Factor (PDGF).............................................................36 Integrins...............................................................................................................37 Integrin Signaling.........................................................................................38 Relationships between Integrin and Other Growth Factor Receptors in Angiogenesis.............................................................................................42 Pigment Epithelium-Derived Factor (PEDF)......................................................47 Insulin-Like Growth Factor (IGF)-1...................................................................47 IGF-1 and IGF-1R........................................................................................48 IGFBPs and ALS..........................................................................................51 The Involvement of Insulin Recepto r (IR) and IGF-2 in Angiogenesis......56 RNA Silencing Technologies.....................................................................................57 Antisense Oligonucleotides.................................................................................59 Ribozymes...........................................................................................................62 Self Splicing Introns.....................................................................................63 RNase P........................................................................................................65 Hammerhead Ribozymes.............................................................................66 Hairpin Ribozymes.......................................................................................67 Hepatitis Delta Virus (HDV) Ribozymes and Neurospora Varkud Satullite (VS) Ribozymes.........................................................................69 RNA Interference................................................................................................69 Gene Therapy Overview.............................................................................................75 Non-Viral Gene Delivery....................................................................................76 Viral Gene Delivery............................................................................................78 Adeno-Associated Viral (AAV) Vectors.....................................................79 Adenovirus (Ad) Vectors.............................................................................86 Retrovirus Vectors........................................................................................88 Herpes Simplex Virus Type 1 (HSV-1) Vectors..........................................89 2 METHODS AND MATERIALS...............................................................................90 Hammerhead Ribozyme Target Sites.........................................................................90 Accessibility of Target Site........................................................................................91 Kinase of Target Oligonucleotides.............................................................................92 Time Course of Cleavage Reactions for Hammerhead Ribozymes...........................93 Multiple Turnover Kinetics........................................................................................94 Cloning of the Ribozymes into an rAAV Expression Vector.....................................95 Screening and Sequencing of the Clones....................................................................97 HREC Tissue Culture.................................................................................................98 Transfection of HRECs with Lipofectamine..............................................................99 Total RNA Extraction.................................................................................................99 Relative Quantitative RT-PCR.................................................................................100 Reverse TranscriptionReal Time PCR....................................................................102 Total Protein Extraction............................................................................................103 Western Blotting.......................................................................................................103 Flow Cytometry........................................................................................................104

PAGE 8

viii Migration Assay........................................................................................................105 Cell Proliferation Assay (BrdU)...............................................................................106 Tube formation Assay (Matrigel).............................................................................107 Proliferating Endothelial-Cell Sp ecific Promoter Constructs...................................107 Plasmid Formulation for Adult Mouse Eye Gene Transfer......................................107 Animals.....................................................................................................................108 Intravitreal Injection in to the Mouse Model of Oxygen-induced Retinopathy (OIR).....................................................................................................................108 Intravitreal Injection into the Adult Mouse Model of Laser-Induced Retinopathy..110 Immunohistological Studies.....................................................................................111 Statistical Analysis....................................................................................................111 3 RESULTS.................................................................................................................112 Ribozyme Design......................................................................................................112 Target Site Selection..........................................................................................112 Accessibility of Target Site...............................................................................114 Sequences of the Ribozymes and the Targets...................................................116 In Vitro Testing of Ribozymes.................................................................................117 Time Course of Cleavage..................................................................................117 Kinetic Analysis................................................................................................119 Functional Analysis of Ribozymes in HRECs..........................................................120 Inhibition of mRNA Expression........................................................................120 Protein Levels....................................................................................................121 Migration Assays...............................................................................................123 Cell Proliferation Assays...................................................................................124 Tube Formation Assays.....................................................................................125 In Vivo Analysis of Ribozymes................................................................................126 Promoter Development.............................................................................................128 Integrin Ribozyme Expression in vivo with the CMV/ -actin Enhancer Promoter.........................................................................................................129 The Proliferating Endothelia l Cell-Specific Promoter......................................131 The New Promoter Tested in vivo .....................................................................133 The New Promoter Tested with Integrin Ribozyme..........................................138 4 DISCUSSION...........................................................................................................141 Ribozyme Testing Results and Antisense Effect......................................................141 VEGFR-1 and VEGFR-2 Interactions......................................................................143 The Proliferating Endothelial Cell Specific Promoters............................................145 Other Voices on Neovasculariza tion in Diabetic Retinopathy.................................147 Final Words on RNA Silencing................................................................................148 LIST OF ABBREVIATIONS..........................................................................................156

PAGE 9

ix LIST OF REFERENCES.................................................................................................160 BIOGRAPHICAL SKETCH...........................................................................................197

PAGE 10

x LIST OF TABLES Table page 2.1 Sequences of primer pairs and anne aling temperatures used in relative quantitative PCR....................................................................................................102 2.2 Summary of primary and secondary an tibodies used in western blottings............104 3.1 Summary of ribozyme and target sequences..........................................................116 3.2 Summary of ribozyme kinetic data........................................................................120 3.3 Reduction in target mRNA levels in HREC by the ribozymes..............................121 3.4 Reduction in protein levels by the ribozymes........................................................123 3.5 All ribozymes tested in vivo ...................................................................................128

PAGE 11

xi LIST OF FIGURES Figure page 1.1 Basic structure of human eye (cou rtesy of National Eye Institute, www.nei.nih.gov).......................................................................................................3 1.2 Cross section of the retina (http://th alamus.wustl.edu/course/eyeret.html)...............3 1.3 Normal view vs. ARMD (courtesy of National Eye Institute, www.nei.nih.gov).....6 1.4 Fundus photograph and fluorescen ce angiogram of ARMD [11]..............................8 1.5 Five stages in ROP (courtesy of National Eye Institute, www.nei.nih.gov)............10 1.6 Normal view vs. DR (courtesy of National Eye Institute, www.nei.nih.gov).........11 1.7 Fundus photograph and fluorescence angiogr am of non-proliferative DR [11]......13 1.8 Fundus photograph and fluorescence angi ogram of proliferative DR [11].............13 1.9 Photocoagulation (courtesy of Na tional Eye Institute, www.nei.nih.gov)...............14 1.10 Cryotheropy (http://www.chec docs.org/dr_treatment.htm).....................................15 1.11 Polyol Pathway [31].................................................................................................18 1.12 AGE formation [31].................................................................................................18 1.13 VEGF-A isoforms [92].............................................................................................27 1.14 VEGF family ligands an d their receptors [116].......................................................30 1.15 VEGF signaling via VEGFR-2 [92].........................................................................33 1.16 The activation of integrins can lead to the signal transduction in a number of pathways. [180]........................................................................................................39 1.17 IGF-1 signaling transduction [216]..........................................................................49 1.18 Proposed pathway of IGF-de pendent IGFBP action [223]......................................51 1.19 Overview of possible IGFBP-3 an tiproliferation pa thways [223]...........................53

PAGE 12

xii 1.20 The crosstalk between IGF-1, IGF2 and Insulin signalings [254]..........................57 1.21 Overview of RNA silenc ing technologies [258]......................................................58 1.22 Modifications in antisense technology [258]...........................................................60 1.23 Self-cleaving and se lf-splicing reactions in ribozymes [263]..................................62 1.24 Secondary structure and self spli cing steps in group I intron [263].........................64 1.25 Secondary structures of natural and synthetic substrates for RNAse P[275]...........65 1.26 Structure of the hammerhead ribozyme...................................................................67 1.27 Structure of the hairpin ribozyme.............................................................................68 1.28 RNA interference [293]............................................................................................70 1.29 Designing artificial shRNA for RNAi [303]............................................................73 1.30 AAV internalization and intr acellular trafficking [330]...........................................81 1.31 AAV2 genome and the vector genome [330]...........................................................83 1.32 Helper virus free systems in rAAV production [334]............................................84 1.33 The 6 pDF helper plasmids in the two-plasmid system [330]..................................85 1.34 Ad genome and the vector genome [324]................................................................87 1.35 MLV genome structure [329]...................................................................................89 2.1 Typical structures of hammerhead ribozyme predicted by Mfold [257]..................92 2.2 The pTRUF21 expression and cloning vect or and the orientation and position of the hammerhead and hairpin ribozyme cassette.......................................................96 2.3 Time course of OIR mouse model.........................................................................109 2.4 Time course of the adult mouse mode l of laser-induced neovascularization.........110 3.1 The human IR cDNA sequence with ribozyme target site highlighted..................113 3.2 Mfold structures predicted fo r the human IR target region....................................114 3.3 Mfold predicted secondary stru cture of human IR ribozyme................................115 3.4 The 34-base ribozymes (black) annealed to the 13-base targets (red) for both human and mouse...................................................................................................116

PAGE 13

xiii 3.5 Cleave time course of human IR ribozyme............................................................118 3.6 Summary of time courses cleavage of th e ribozymes generated in this study.......118 3.7 Multiple-turnover kinetic analysis of a human IR ribozyme.................................119 3.8 Insulin receptor mRNA levels in HRECs..............................................................121 3.9 Western analysis of IR levels in ce lls expressing the human IR ribozyme............122 3.10 HREC migration assays in response to IGF-1.......................................................124 3.11 Effect of the VEGFR-1 and VEGF R-2 ribozymes on HREC migration...............124 3.12 Effect of ribozyme expre ssion on cell proliferation...............................................125 3.13 Effect of ribozymes on HREC tube formation.......................................................126 3.14 Cross section of a mouse eye showing pre-retinal vessels.....................................127 3.15 Ribozyme reduction of pre-retinal ne ovascularization in the OIR model..............127 3.16 Reduction of pre-retinal neovasculari zation in the OIR mouse model with expression of the 1 or 3 integrin ribozymes.......................................................129 3.17 Expression of 1 ribozyme in OIR model results in severe deformations of the eye..........................................................................................................................13 0 3.18 pLUC1297/1298 vectors and pLUC1297HHHP/1298HHHP clones....................131 3.19 Verification of the cell speci ficity of the proliferati ng endothelial cell-specific enhancer/promoter..................................................................................................133 3.20 The proliferating endoth elial cell-specific promoter limits expression of luciferase to the actively prolifera ting blood vessels in the OIR model................135 3.21 Quantitative assessment of the IGF-1R ri bozymes ability to inhibit pre-retinal neovascularization when expr essed from the promoter.........................................136 3.22 New promoter tested in adult mouse m odel of laser-induced neovascularization.136 3.23 The expression of the IGF-1R riboz yme from the new promoter reduced aberrant blood vessel formati on in the adult laser model.......................................137 3.24 Expression of integrin ribozyme driven by proliferating endot helial cell-specific promoter resulted in less eye deformation.............................................................139 3.25 Proliferating endothelial ce ll specific promoter with integrin ribozyme tested in OIR model..............................................................................................................140

PAGE 14

xiv LIST OF OBJECTS Object page 3.1 A blood vessel from the adult mouse model shows the luciferase expression is specific for proliferatin g endothelial cells..............................................................137 3.2 The 3-D view of the blood vessel from the adult mouse model............................137

PAGE 15

xv Abstract of Dissertation Pres ented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy TARGETING ANGIOGENIC GROWTH FACT ORS IN PROLIFERATIVE DIABETIC RETINOPATHY By Hao Pan May 2006 Chair: Maria B. Grant Major Department: Pharmacology and Therapeutics Proliferative diabetic retinopathy is the leading cause of blin dness in the working age adults. Pre-retinal angiogenesis is the hallm ark of this disease and can lead to vessel leaking, exudate accumulation, hemorrhage, or even retinal detachment. Many growth factors have been identified to promot e the vessel growth, physiologically and pathologically. Inhibition of these growth f actors can result in less abnormal angiogenesis and potentially prevent the onset of vision impairment. One gene silencing technology, hammerhead ribozyme, was used to inhibit the signaling of thes e growth factors. Ribozymes are small RNA molecules that can recognize and cleave sp ecific sequence in the target mRNA. Ribozymes against the genes of a number of growth factor receptors, including IGF-1R, insulin receptor, VEGF-R 1, VEGF-R2, and multiple integrins, were designed and tested in vitro and in vivo All ribozymes were tested by cleavage time courses, kinetic analysis a nd proved to be capable of cl eaving synthetic RNA targets. Then they were transfected in human retina l endothelial cells, and the mRNA levels and

PAGE 16

xvi protein levels of the growth factor re ceptors were reduced. Also the migration, proliferation and tube formation of thes e cells were inhibited. We used the oxygeninduced retinopathy mouse model to test the ribozymes in vivo The expression of the ribozymes induced significant reductions in the pre-retinal neovascul arization levels. To better target the proliferating endothelium in vivo and to minimize the adverse effect of ubiquitous ablation of targeted growth factor receptors, a proliferating endothelial cellspecific promoter was designed. This new pr omoter was tested with IGF-1R ribozyme and showed specific expression in the prol iferating endothelium a nd significant reduction in the pre-retinal neovascularization levels. Ou r results suggest that these ribozymes are a useful tool to inhibit the angiogenesis in retinopathy, and the proliferating endothelial cell-specific promoter adds the specif icity without losing expression strength.

PAGE 17

1 CHAPTER 1 BACKGROUND Introduction and Project Aim Vascular retinopathies, incl uding retinopathy of prematurit y, proliferative diabetic retinopathy and age-related macular degene ration, are the leading cause of vision impairment worldwide. Pre-retinal vessel growth is the hallmark for retinopathy of prematurity and proliferative diabetic retinopathy. These new blood vessels are abnormally positioned and are fragile, easy to leak, and can result in hemorrhage and retinal detachment. Currently there is no cu re for these diseases. The initiation and maintenance of these pre-retinal blood vessels depend on the involvement of many growth factors. In this project, with the he lp of a gene silencing technology, hammerhead ribozyme, efforts have been made to target and inhibit the expre ssion of a number of growth factor receptors to reduce the growth factor signaling. Ri bozymes are small RNA molecules that can specifically bind to a sequence in the target mRNA and perform cleavage. The genes of IGF-1R, VEGFR-1, VEGFR-2, integrins and insulin receptor have been targeted and the inhibition effects were examined in vitro and in vivo To better target the proliferating endothelium in vivo and to minimize the adverse effect of ubiquitous ablation of targeted growth factor receptors, a proliferating endothelial cellspecific promoter was designe d and tested. In the basic science point of view, the investigations on the involveme nt of the growth factors in the pre-retinal angiogenesis can provide useful information about their si gnaling details; in the clinical application

PAGE 18

2 point of view, this work could also imply new targets and methods for the disease treatment in the future. The Eye The Anatomy of the Eye Optically working like a film camera, the eyes of all the vertebrates are structurally similar. The light enters the eye through the pupil and forms an inverted image on the retina, the light -capturing component that f unctions like the film in a camera. The cornea and the lens help to focus so that the clearest image is presented on the retina. The white outer surface of the eye ball is termed sclera which consists of tough but flexible fibrous tissu e and provides the mechanical support of the entire eye. The choroid is a layer contained within the sclera, and it is a dense meshwork of blood vessels and other tissues. One of the most important functions of the choroid is to provide nutritional and metabolic support for the retina, which is a neuronal sheet that lies within the choroid. The retina is the most inner surface at the back of the eye. Most of the space in the eye is filled with a gelatinous body, cal led vitreous. It is surrounded by the lens and the retina and the ciliary body. In the ciliar y body, the cells secrete the aqueous fluid into the eye, which contributes to the maintena nce of the pressure within the eye. The Retina The retina, a layer about 0.4 mm in thickness, is primarily composed of neural tissue including five classes of neurons. It spreads out on the interi or surface of the back of the eye. The visual pathway is initiated when the light stimulates the photoreceptors that are embedded in the outer retinal layers. The signal is transmitted to bipolar cells and then to ganglion cells. The signal then travel s along the axon of the ganglion cells lining

PAGE 19

3 Figure 1.1. Basic structure of human eye (courtesy of National Eye Institute, www.nei.nih.gov ). Figure 1.2. Cross section of the retina ( http://thalamus.wustl.edu/course/eyeret.html ). the inner surface of the retina to the optic nerve, which pene trates the retina and connects to the brain. There are two more classes of neurons, horizontal cells and amacrine cells. They are both interneuron and assist in si gnal processing. Horizontal cells primarily contact with photoreceptor axons and bipolar cells in the outer pl exiform layer and the

PAGE 20

4 inner nuclear layer, respec tively, while amacrine cells c ontact with bipolar axons primarily in the inner plexiform layer. Light passes through almost the whole thic kness of the retina to be captured by photoreceptors, or the outer segments of the photoreceptor in deta il, where the visual pigment molecules for light capturing are lo cated. There are two types of photoreceptors, rods and cones. Rods are specialized to c onvey variations in li ght intensity in dim conditions, but they are not ab le to function in bright li ght. Cones are specialized for bright light conditions, but they are not as sensitive as rods. The retina cross section can be divided in to multiple layers. The nuclear layers are basically where cell nuclei are located, and the synaptic laye rs are the place where cells communicate and transmit electri c or chemical signals. The retinal pigment epithelium (RPE) functions as the outer blood-retinal ba rrier (BRB) that shut off the diffusion of large molecules from choroicapillaries. And the retinal vasculature doesnt grow beyond the inner limiting membranes under normal physiological conditions. The Blood Supply to the Retina The metabolism in the retina performs in the highest rate in the body. For the same mass of tissue, the metabolic needs of the retina are about seven times that of the brain. In order to meet these high metabolic n eeds, two separate circulations are involved. They are retinal and choroidal circulations. Th e larger arteries and veins of the retinal circulation can be seen under an ophthalmoscope, and most of the retinal surface is occupied with a meshwork of retinal capilla ries. These capillaries form the inner BRB. The endothelial cells at the capillaries are connected by tight junc tions that prevent leakage from the vessels. A lot of proteins or molecules work in the binding of the adjacent cells. Because of the tight junctions proteins and solutes have to pass through

PAGE 21

5 the apical and the basal membra nes of the endothelial cells in order to go into or out of the circulation from or to surrounding tissu es. Water, small molecules and dissolved gases can do so, such as glucose, oxyge n, carbon dioxide, and so on. But most large molecules, including protei ns, cannot pass through freel y. The only possible way for them to pass through is through a process of act ive transport with th e help of the proper membrane tunnel proteins. So basically th e BRB provides a mechanism of keeping the substance entering the retinal neural tissue in a controlled manner. The central artery and vein of retinal ci rculation originate al ong the optical nerve and extend into the retina from the center of the optical disc. While the choroidal arteries and veins of pass through the sclera at multip le places around the optical nerve, and then they branch into a meshwork of very larg e capillaries, called choroicapillaries. Large capillaries increase the rate of blood passi ng through, which keeps th e concentration of oxygen high and the concentration of carbon di oxide low, and also quick removes the heat from focused light on the eye bottom. The BRB is not maintained by choroidal circulations, because the cells on the side f acing the RPE are fenestrated, and there is no tight junction between these cells. However, the RPE connecting with the choroid have tight-junctions and provide the outer portion of the bl ood-retinal barrier. Retinopathies Retinopathies are diseases that affect the function of re tinas. Usually they involve the abnormalities in the vasculatures that nou rish the retina. These abnormalities included ectopic angiogenesis, rupture a nd leakage on the vessels, accum ulation of exudates, retina detachment caused by vessel and fibrous tissu e contractions, and so on. There are three types of retinopathy clinical ly identified: age-related macular degeneration (ARMD), which occurs in the elderly people; diabet ic retinopathy (DR), which occurs in the

PAGE 22

6 working age people; and retinopa thy of prematurity (ROP), which occurs in infants. ARMD more involves the abnor malities in choroicapillari es, while DR and ROP are basically related to the abnorma lities of retinal vasculature. Age-Related Macular Degeneration (ARMD) ARMD is the leading cause of blindness among those aged over 65 in the western world [1-3]. It affects the outer retina, RPE, Bruchs membrane and the choroids. Thickening of Bruchs membrane is seen in this disease. Our understanding about the pathogenesis has grown in the past decad e, but still a lot remains unknown and the current therapy is limited. Figure 1.3. Normal view vs. ARMD (c ourtesy of National Eye Institute, www.nei.nih.gov ). The clinical hallmark of ARMD is the a ppearance of drusen, localized deposits lying between the basement membrane of the RPE and Bruchs membrane. Drusen can be shown as semi-translucent punctuate or ye llow-white deposits depending on the stage of the disease. Morphologically drusen are cl assified as hard and soft. Hard drusen are pinpoint lesions; soft drusen are larger with vague edges and they are easy to become confluent. Drusen can become calcified and they may also regress. Typically clustered drusen are located in the central macula, so th ey can lead to deficits in macular function

PAGE 23

7 such as color contrast sensitivity, central visual field sensitivity and spatiaotemporal sensitivity [4]. Increased quantity and size of drusen are an independent risk factor for visual loss in ARMD. Geographic atrophy is also seen in ARMD which refers to confluent areas of RPE cell death accompanied by overlying phot oreceptor atrophy [5]. Geographic atrophy leads to vision impairment, especially the visual function in dark situations [6]. This loss of function is probably because the RPE lo ss results in reduced nutrients for those photoreceptors that are located in the RPE atrophy areas. Apoptosis in the corresponding area are found [7]. Choroidal (or subretinal) ne ovascularization (CNV) is a major cause of vision loss in ARMD. As the term itself indicates, C NV refers to the new blood vessel growth from the choroids. It breaks through the Bruchs membrane into the space underneath RPE, or it may further penetrate the RPE layer into the subretinal space. Usually CNV is associated with leakage of fluid and blood. The repeated leakage of blood, serum, and lipid can stimulate fibroglial organization leading to a cicatricial scar [4]. Drusen and CNV can cause irregular elev ation of RPE, which can lead to RPE detachment or even RPE tear. RPE detachment can cause visual loss in patients with ARMD [8]. Depending on whether CNV is present, ARMD is classified into the dry form or the wet form. The dry ARMD is nonexudative [4]. This is the early phase of ARMD, and the earliest pathological changes are the a ppearance of basal laminar deposits between the plasma membrane and basal lamina of the RPE, and the appearance of basal linear deposits located in the inner collagenous zone of Bruchs membrane. The former deposits

PAGE 24

8 are seen in an increase amount in ARMD [9 ], and the later deposits are only seen in ARMD [10]. Approximately 10 percent of persons with AMD develop the exudative form of the disease, or wet ARMD. E xudative AMD accounts for 80 to 90 percent of cases of severe vision loss related to AMD. CNV occurrence is the hallmark of wet ARMD. CNV is associated with abnormal vessels that leak fluid and blood in the macula, resulting in blurred or di storted central vision. Figure 1.4 is the fundus photograph (A) and fluorescence angiogram (B) of an eye of a patient with exudative ARMD. Note subretinal neovascularisation (A, asterisk) with surrounding hard exudates (arrowheads). On the angiogram (B) the neovascularization is clearly stained by fluorescein (black arrow) [11]. Figure 1.4. Fundus photograph and fluores cence angiogram of ARMD [11]. As for the pathogenesis of ARMD, shor tly speaking, Campochiaro and coworkers suggested that the age-relate d thickening of Bruchs membra ne reduces the diffusion of oxygen from the choroid to the RPE and retina [12], and recent evidence suggests that VEGF plays an important role in the deve lopment of CNV. VEGF expression was found to be increased in RPE cells of maculae of patients with age-related maculopathy, a condition with a high risk of CNV occurrence [13] and in experimental animal models [14]. VEGF levels in the v itreous of wet ARMD were f ound to be significantly higher

PAGE 25

9 than healthy controls [15]. Chronic inflamma tion from drusen may be involved in the development of ARMD [16], but the inflammatory contributi on is still controversial. Retinopathy of Prematurity (ROP) ROP is an adverse effect of treating t hose premature neonates in respiratory distress with high oxygen. The high oxygen helps these infant s to survive, but it can cause ROP, which will impair their vision. ROP mainly affects premature infants weighing about 1250 grams or less that are born be fore 31 weeks of gestation. It is one of the most common visual loss diseases in childhood. According to the National Eye Institute, there are about 28,000 infants born weighing 1250 grams or less in the U.S., and among them, 14,000-16,000 of the infants are aff ected by ROP to some degree. 10% of them need medical treatment and 400-600 infant s annually become legally blind of ROP. The ROP complete progression can be divided into 5 stages. Stage 1 is characterized by a demarcation line between the normal retina (near the optic nerve) and vascularized retina. In stage 2, a ridge of scar tissue rises up from the retina due to growth of abnormal vessels. This ridge forms in plac e of the demarcation line. In stage 3, the vascular ridge grows due to spread of abnor mal vessels and extends into the vitreous. Stages 4 and 5 refer to retinal detachment; st age 4 refers to a partial retinal detachment caused by contraction of the ridge, thus pulling the retina away from the wall of the eye; and stage 5 refers to comp lete retinal detachment. ROP is now considered as a two-phase process during the disease development. In the first stage, the high oxygen cond ition will make the developing retinal blood vessels and especially the developing capillary buds be more pruned to drop out. This pathological vessel dropout is an exaggeration of the normal physiologic process, in which there is a constant balance between de veloping and degenerating capillary buds, as

PAGE 26

10 Figure 1.5. Five stages in ROP (courtesy of National Eye Institute, www.nei.nih.gov ). tissue demand changes [17]. In short, the hype roxic vaso-obliteration occurs in the first stage. When the high oxygen care is complete and the infants survive, they are taken out the high oxygen environment and the second stag e occurs. Because of the vessel loss, the tissue becomes hypoxic and the ischemia-induc ed vaso-proliferation begins. The hypoxia stimulates growth factors increases, especi ally VEGF. These growth factors play very important roles in the vaso-proliferation. The vaso-proliferation is abnormal in that these new vessels are fragile and leak, scarring th e retina and pulling it out of position, which

PAGE 27

11 will lead to retinal folds and retinal detach ment. The term babies are less affected by fluctuations in oxygen levels as once the vessels become developed and surrounded by supportive matrix, thus they are no longe r susceptible to pruning by hypoxia [18]. Diabetic Retinopathy (DR) Diabetic retinopathy is one the three major complications of diabetes mellitus (the other two are neuropathy and nephr opathy) and occurs in both type I and type II diabetes. DR primarily affects the working age people an d is the leading cause of new-onset visual loss in working people in the U.S. and other industrialized countri es [19]. DR affects approximately three-fourths of diabetic patien ts within 15 years afte r onset of the disease [20]. Retinal neovascularizati on and macular edema are central features of DR and also the two factors that cause vision loss. The ne wly-formed vessels are fragile and abnormal and they can leak blood into the center of the eye, blurring vision. Macular edema usually occurs as the disease progresses. The fluid leaks in the center of macular and makes the macula swell, blurring vision. Other charac teristics found in DR include basement membrane thickening, pericyte lo ss, microaneurysms, and so on. Figure 1.6. Normal view vs. DR (cour tesy of National Eye Institute, www.nei.nih.gov ). In the beginning stage of DR, there are no clinically evident symptoms, but the biochemical and cellular alterations are goi ng on in the retinal vasculature. These

PAGE 28

12 alterations include increased adhesion of le ukocytes to the vessel wall, alterations in blood flow, basement membrane thickening. Thes e factors are involved in the blockage of the retinal capillaries, which is tho ught to induce hypoxia and further trigger the overexpression of the angiogenic factors. Othe r vascular alterations include death of retinal pericytes, subtle increases in vascul ar permeability, or even the loss of vascular endothelial cells. Following th is, the blood and fluid leak age may come. The loss of endothelial cells also leads to acellular capillarie s worsening ischemia. With time, more abnormal phenomena occur and they are clinically observable. These abnormalities include microaneurysms, dot/blot hemorrhages cotton-wool spots, venous beading and vascular loops [20]. The blood and fluid leak out the vessels and accumu late in the retinal tissue, giving rise to exudates. When this occurs in macula, patients will have macular edema and impaired vision. This stage is al so called nonproliferative retinopathy. With the progression the disease, next stage is the proliferative reti nopathy, featuring the growth of new vessels on the surface of the retina. The new vessels are abnormal, fragile and easy to break. The leaking blood can cloud the vitreous and further impair vision. In more advanced stages, the exaggerated pre-re tinal neovascularization can grow from the retinal surface into the vitreous cavity. This can cause retinal detachment can lead to blindness. Proliferative retinopat hy typically develops in pati ents with type I diabetes, whereas nonproliferative reti nopathy with macular edema is more common in patients with type II diabetes [20]. Figure 1.7 shows the fundus photograph (A) and fluorescence angiogram (B) of an eye of a patient with nonproliferative diabetic retinopat hy. The arrowheads in Panel A

PAGE 29

13 point to intra-retinal hard exudates surr ounding areas of leaking microaneurysms (B, white arrows) [11]. Figure 1.7. Fundus photograph and fluorescence a ngiogram of non-proliferative DR [11]. Fundus photograph (A) and fluorescence angi ogram (B) of an eye of a patient with proliferative diabetic retinopathy is shown in Figure 1.8. Note pre-retinal neovascularization (black arrow) on the optic disc (A), which is extensively leaking fluorescein (B. white arrows) [11]. Figure 1.8. Fundus photograph and fluorescence a ngiogram of proliferative DR [11]. Current Treatments for Retinopathies Currently the clinical proved treatments for retinopathies are limited, and few drug medications are available. The c onventional treatments include laser photocoagulation, cryotherapy, photodynamic th erapy, scleral buckle, and vitrectomy. All of them cannot cure th e disease, but can only de lay the disease progression.

PAGE 30

14 In laser photocoagulation, the doctor pl aces thousands, up to 3,500, small laser burns on the retina. These burns will dest roy the normal tissue a nd decrease the oxygen needs of the retina. The treatment is usually effective, but at the cost of loss of normal tissue, and it reduces peripheral vision, impair night vision and change color perception. The laser photocoagulation is not a cure, as the disease ca n still progress in spite of treatment. More treatments may be needed to further prevent vision loss. Laser treatment is currently applied in all retinopathies, that is, ROP, DR, and ARMD. Laser is also used to target at the leaking spots, like in severe macular edema, the laser burning is applied in a focal way. When preventing abnormal vessel growth, as in proliferative DR, the laser burning is applied in a scattered way. Figure 1.9. Photocoagulation (courte sy of National Eye Institute, www.nei.nih.gov ).

PAGE 31

15 Cryotherapy is a procedure in which physic ians use an instrument that generates freezing temperature to briefly touch spots on the surface of the eye that overlie the periphery of the retina. It also destroys th e tissue and impairs the side vision. Cryotherapy is more used for ROP. In Figure 1.10, cartoon is showing cryothera py application to the anterior avascular retina. A cold probe is pl aced on the sclera till an ice ball forms on the retinal surface. Multiple app lications are done to cover th e entire vascular area. This treatment thins the tissue under the retina a nd allows easier oxygen diffusion through the retina. Figure 1.10. Cryotheropy ( http://www.checdocs.org/dr_treatment.htm ). In photodynamic therapy, a drug called verte porfin is injected i.v. and perfused to the vasculature in the eye. The drug tends to stick to the surf ace of new blood vessels, and then, a light is shined into the eye for about 90 seconds, and th e light activates the drug to destroy the new blood vessels. The a dvantage of this method is that the drug doesnt destroy the normal tissue surrounding. But the patient needs to avoid bright light for five days because the drug can be activated in their exposed body parts. Photodynamic therapy is more used to treat wet ARMD.

PAGE 32

16 In later stages of ROP, scleral buckle is another treatment option [21]. This involves placing a silicone band around the eye and tightening it. This keeps the vitreous from pulling on the scar tissue and allows the retina to flatten back down onto the wall of the eye. The band will be removed later. In most severe conditions in retinopathies, vitrectomy can be applied, in which the vitr eous is removed, scar tissue on the retina peeled back or cut away, and saline so lution is replaced for vitreous. The retina reattachment can be seen after this surgical treatment [22]. Pathogenesis of Diabetic Retinopathy Diabetes mellitus is a serious disease lead ing to morbidity and mortality as it has long-term complications include macrovascular and microvascular di sease. Both type I (characterized by no insulin pr oduction) and type II (charact erized by insulin resistance) diabetes can have these co mplications. Retinopathy is one of the microvascular complications. It is believed that the chroni c hyperglycemia has a st rong relationship with microvascular complications, and clinical re search data demonstrates that improved glycemic control contributes to signifi cant microvascular risk reduction [23, 24]. Experiments on animal models also suggest th at long-term hyperglycemia is necessary to induce changes in the retin al vasculature [25]. In the retina, GLUT1, which is one of a family of glucose transporters, is responsible for glucose transf er across BRB into the endothelial cell and retinal cells. While in most other cells in the body, insulin assistance is required for internalize glucose; this is not the case with the re tina. Excessive transport of glucose through GLUT1 [26], the involvement of GLUT1 in RPE cells [27], and increased density of relocalized GLUT1 in inner BRB [28] have been proposed to be related with intracellular hyperglycemia. Intracellular hyperglycemia in the early stages of diabetes causes

PAGE 33

17 abnormalities in blood flow and increases in vascular permeability. The blood flow changes come from decreased activity of vasodi lators, such as nitric oxide, and increased activity of vasoconstrictors such as angiot ensin II and endothelin-1 [29]. The increase in vascular permeability comes from VEGF functioning on endothelial cells and changes in extracellular matrix. With time, hypergly cemia can further induce cell loss and progressive capillary occlusion. All these changes will eventually lead to edema, ischemia and hypoxia-induced neovascularization. To date, there are several hypothesized theories on how hyperglycemia contributes to microvascular damage, or retinopathy. The most common ones are polyol pathway theory, advanced glycation end-pr oducts (AGE) theory, oxi dative stress theory and PKC activation theory. Increased Polyol Pathway Flux As shown in Figure 1.11, glucose is redu ced to sorbitol by aldose reductase, and at the same time, nicotinamide-adenine di nucleotide phosphate (NADPH) is oxidized to NADP+. Then sorbitol is oxidized by sorbitol dehydrogenase to fructose, coupled with the reduction of oxidized nicoti namide-adenine dinucleotide (NAD+) to NADH [29]. So the intracellular high glucose level will result in excess sorbitol, fructose, NADH accumulation and decrease in NADPH. Some damages caused by increase flux through polyol pathway have been proposed to incl ude: activation of prot ein kinase C [30], contribution of AGE formation [30], decr eased activity of Na /K-ATPase [29], and increase in the formation of reactive oxygen species leading to oxidative stress [29]. Production of AGE AGE are irreversibly cross-linked subs tances. Intracellular hyperglycemia is possibly the primary initiating event in the formation of intracellular and extracellular

PAGE 34

18 AGE [32, 33]. During formation of AGE, gluc ose reacts nonenzymatically with the amino group of proteins and other macromol ecular to form Schiff bases, which are transformed into Amadori products that even tually lead to AGE formation [34]. When AGE bind their receptors, RAGEs, some abnormal cellular events can occur, including: the stimulation of the pr oduction of the vasocons trictor e ndothelin-1, VEGF production that is associated with increased permeability, and production of reactive oxygen species. The long-term eff ects induced by AGE and RAGEs are mostly mediated by transcription factor B to express cytokines a nd growth factors [29]. Figure 1.11. Polyol Pathway [31]. Figure 1.12. AGE formation [31] There are several adverse al terations in the micro vasc ulature associated with AGE. AGE formation can contribute to thicke ning of the basement membrane and to microvascular hypertension by inactivating nitric oxide [31]. The thickening and

PAGE 35

19 hypertension can lead to micr ovascular leakage and occlusio n. AGE can adversely affect vascular permeability, alter the functions of ma trix molecules, and alter the functions of vessels, by decreasing the vessel elasticity, increasing fluid filt ration across vessels [29], decreasing endothelial cell adhesion [35], and so on. Generation of Reac tive Oxygen Species The term oxidative stress refers to the imbalance between the production of reactive oxygen species and th e normal antioxidant protective mechanisms present to guard tissues from oxidative damage [36]. As discussed above, both polyol pathway and AGE formation can lead to the generation of reactive oxygen species. Glucose also has pro-oxidant properties in the pr esence of heavy metals and the auto-oxidation of glucose can form free radicals too. These reactive oxyg en species can inactiv e or reduce nitric oxide levels [37]. The reactive oxygen species can result in damaged protein and mitochondrial DNA that have adverse effects on the microva sculature [38], especially leading to increased microvascular permeability [39]. Oxidative stress has been shown to increase intracellular calcium levels, which have been associated with endothelial hyperpermeability of macromolecules [40]. Activation of Diacylglycerol a nd Protein Kinase C Isoforms It has been shown that di acylglycerol (DAG) formation can be induced by glucose in cell cultures, animal tissues, and diabetic patients [31]. DAG is very important in the activation of various protein kinase c (PKC) isoforms, with the isoform being thought to be the most sensitive to changes in DAG levels. PKChas been shown to be increased in various vascular tissues following hyperglycemic exposure [41]. PKC, PKC1 and PKC2 are seen to be elevated in the retina during acute and chronic

PAGE 36

20 hyperglycemic states [42]. The conseque nces induced by PKC activation include increased retinal permeability [43], increased basement matrix protein formation [44], VEGF formation [44], and so on. So PKC ma y have adverse long-term effects in the vasculature. Based on the involvement of these pathwa ys, a lot of pathological changes can happen in diabetic retinopathy. Some of the most important change s are covered below. They will lead to edema, ischemia and hypoxia in the retina, which all lead to abnormal neovascularization. How Does the Change in Retinal Blood Flow Occur? Hyperglycemia induces changes in re tinal blood flow via its effects on vasodilators and vasoconstric tors. Nitric oxide (NO) is one of the most important vasodilators. It is synthesized from L-arginine or L-citrullin e in cells via activation of a calcium-dependent nitric oxide synthase (NOS). The NOS isoform produced in endothelium is called eNOS. NO functi ons by entering smooth muscle cells and activating soluble guanylate cyclase, which will result in increased level of cyclic guanosine 3, 5-monophosphate (cGMP). cGMP can relax the smooth muscle cells through a decrease in Ca2+ and dephosphorylation of myosin light chains [45]. In the hyperglycemic environment, a couple of pathways mentioned above can lead to decreased level of NO. In th e polyol pathway as mentioned earlier, sorbitol is produced coupled with the oxidation of NADPH a nd this reduces NADPH availability, and NADPH is one of the cofactors for NO synthe sis. AGE production can lead to subsequent superoxide generation result ing in NO inactivation. PKC ac tivation reduces the capacity of a number of agonists to increase intracellular Ca2+ and to stimulate NO production; on the other hand, the superoxide expression may also result from PKC activation.

PAGE 37

21 Endothelin (ET)-1 is a powerful vasoconstr ictor. At low concen trations, it induces vasodilation. While at high concentrations it causes the constrictive response by interacting with its receptors on smooth mu scle cells and pericytes in the retinal vasculature. Hyperglycemia-induced PKC activation can enhance ET-1 transcription level [46]. What Causes Retinal Capillary Cell Death? Pericytes loss and endothelial cells loss are both seen in diabetic retina. The cell death will inevitably lead to microaneurysms and vascular obstruction. Polyol pathway, AGE pathway and oxidative stress are all t hought to be associated with cell death. Sorbitol accumulated in polyol pathway may cause hyperosmolality of the cells [47]; accumulated AGE production in the glycati on pathway will form cross-links and to generate oxygen-derived free radicals [48]; and the oxidative stress will inactivate NO and cause abnormal chemical changes in DNA structure [49]. What Causes Retinal Ischemia? Hyperglycemia causes ischemia via several possible mechanisms, including thickened basement membrane, platelet aggr egation, leukocyte activation and adherence. Hyperglycemia is sufficient to increase the synthesis of basement membrane components, like fibronectin [50], various types of collagens [51] and vitronectin [52]. Increased number and size of platelet-fibrin thrombi in retinal capillaries have been found in the retina of patients with diabetic retinopat hy [53]. Hyperglycemia-induced PKC activation will stimulate platelet-deriv ed factor (PAF) production, wh ich will activate platelets. Activated platelets can produce platelet-derived microparticles, which are involved in the thrombus formation [54]. PAF can also stim ulate their receptors on leukocytes rolling on the luminal endothelial membrane and activate them. 2 integrins on activated leukocytes

PAGE 38

22 enable them to adhere tightly to the endothelial cells via binding intercellular adhesion molecule-1 (ICAM-1), while as the same IC AM-1 is also unregulated by PKC activation. And NO downregulation can allow leukocytes to escape from NO control, also leading to leukocyte activation a nd adherence [54]. Angiogenesis and Growth Factors Vasculogenesis and Angiogenesis Small blood vessels consist only of endotheli al cells (ECs), whereas larger vessels are surrounded by mural cells (pericytes in medium-sized vessels and smooth muscle cells (SMCs) in large vessels ) [55]. Vessels can grow in several ways. Vasculogenesis refers to the formation of blood vessels by endot helial progenitors [55]. It is a process by which the initial vascular tree forms in th e yolk sac and aortic arches, and begins immediately following gastrulation when mes odermal cells aggregate into blood islands. Blood islands contain the precursors of hemat opoietic and vascular endothelial lineages [56]. Angiogenesis refers to the formation of new vessels formation by sprouting from pre-existing vessels and subs equent stabilization of th ese sprouts by mural cells. Additional modes of vascular growth incl ude intususception, bridge formation, and vascular splitting, in which inva ginations or extensions of th e vessel wall form tubes that connect or bifurcate parent vessels [56]. The traditional view is that vessels in the embryo developed from endothelial progenitors, whereas sprouting of vessels in the adult resu lted only from division of differentiated ECs. However, recent eviden ce has shown that endothelial progenitors contribute to vessel growth bot h in the embryo and in ischemic, malignant or inflamed tissue in the adult. They can even be used th erapeutically to stimulate vessel growth in ischemic tissues, a progress called Thera peutic Vasculogenesi s [57-59]. Although

PAGE 39

23 retinal neovascularization has been thought to be due to proliferation of endothelial cells by angiogenesis, Grant et al. showed that hematopoietic stem cells can enter the circulation and reach the areas of angiogenesis, and clonally differentia te into endothelial cells [60]. In another study, adult Lin(-) hemato poietic stem cells inje cted intravitreally into neonatal mouse eyes have been shown to in teract with retinal astrocytes that serve as a template for retinal angiogenesis [61]. Bl ood vessels are being modified by endothelial progenitor cells, hematopoietic stem cells or other stem cells, and th ese cells functionally contribute to physiological and pathological angiogenesis. Angiogenesis is usually inactivated or kept at low levels in normal tissue of an adult, but may be activated to an excessive st ate in a number of diseases, such as cancer, psoriasis, arthritis, retinopathy, obesity, asthma atherosclerosis, and infectious diseases. Cancer is another best known disease that involves pathologi cal angiogenesis that can be potentially targeted for therapy. In 1972 Folkman proposed that solid tumors are dependent on angiogenesis for growth greater than a few millimeters in size, and that increases in tumor diameter require a corres ponding increase in vascularization [62]. A critical step during angiogenesis is the loca l stimulation of endothelial cells by various cytokines and growth factors. Stimulation causes the endothelial cells to lose their contact inhibition, migrate and breach the basement me mbrane, proliferate, and differentiate to organize into new vessels [63]. Hypoxia-Induced Factor (HIF) Beyond a size limitation, simple diffusi on of oxygen to metabolizing tissues becomes inadequate, and specialized systems of increasing complexity have evolved to meet the demands of oxygen delivery in highe r animals [64]. One important role in the systems is angiogenesis, to make new vesse ls sprouting into the location that blood

PAGE 40

24 delivery is needed. So ischemia or hypoxia is one of the key factor s that lead to the initiation of angiogenesis. Exactly ho w hypoxia induces angiogenesis was however poorly understood. The landmark of hypoxia st udy in the early 1990s showed that hypoxia could induce expression of platelet-derived growth factor (PDGF) mRNA [65] and vascular endothelial grow th factor (VEGF) mRNA in ti ssue culture [66]. Both PDGF and VEGF are thought to be important growth factors triggering angiogenesis. A large number of genes are involved in differ ent steps in angiogenesis and they are independently responsive to hypoxia in tissue culture. Besides PDGF and VEGF, nitric oxide synthase, fibroblast growth factor, angi opoietins, and matrix metalloproteinases are involved [67-69]. Many of the individual phenot ypic processes in angiogenesis such as cell migration or endothelial tube formati on can be induced by hypoxia tissue culture [70]. Further study of hypoxia-induced angiogene sis leaded to the discovery of a key transcriptional regulato r, hypoxia-inducible factor (HIF)-1 [47, 68, 69, 71]. HIF-1 is a heterodimer DNA-binding factor. HIF-1 consists of an and subunits, both of which have a number of isoforms. HIF-1 subunits are constitutive nuclear proteins, while HIF-1 subunits are hypoxia-inducible There are three isoforms for subunit. HIF-1 and HIF-2 appear closely related and ar e both able to interact with hypoxia response elements (HREs) to induce tran scriptional activity [ 72, 73]. In contrast, HIF-3 appears to negatively regulate the re sponse, through an alternatively spliced transcript [74]. The molecular mechanism behind HIF-1 is a pathway that links oxygen availability and the gene expression of vari ous growth factors, especially VEGF. In normoxia and hyperoxia oxygen-dependent prolyl hydroxylases hydroxylate HIF-1

PAGE 41

25 proline residues, and this chemical modifi cation leads to a HIF-1 capture by a ubiquitin ligase complex that directs it to the proteasome for destruction. Under hypoxic conditions, HIF-1 is not hydroxylated, escapes ubiqui tination, accumulates and directs pro-angiogenic expression [75]. Vascular Endothelial Growth Factor (VEGF) VEGF was originally discovered as the vascular permeability factor (VPF) that increased the vascular permeability in the skin [76]. In 1989 Ferrara and Henzel identified a growth factor for endothelial ce lls from bovine follicular pituitary cells and named it VEGF [77], which was then proved to be identical to VPF [78, 79]. VEGF is the most potent endothelial cell growth factor f ound to date. In the past two decades, this growth factor has been studied extensivel y and its key roles in the proliferation, migration, invasion, cell survival, differentiati on of endothelial cells and other cell types have been established. It is critical in the normal embryonic development of vasculature and has essential functions in adults during normal physiologi cal events such as would healing, menstrual cycle, even though the mRNA levels of VEGF and its receptors decrease significantly postna tally. Meanwhile, VEGF is also an important factor in numerous pathological situ ations, many of which involve abnormal angiogenesis, for example, inflammation, retinopathies, psoriasi s, and cancer. Targeting VEGF signaling in these diseases has been studied with ent husiasm and a number of novel drugs targeting VEGF are being tested in clinical trials. VEGF Family and Isoforms The VEGF gene family consists of multiple variants, including VEGF-A (hereafter referred to as VEGF), VEGF-B, VEGF-C, VEGF-D, VEGF-E, VEGF-F and placental growth factor (PlGF-1 and PlGF-2 is oforms). They are secreted glycoproteins

PAGE 42

26 that form homodimers, which belong to a st ructural superfamily of growth factors, including the platelet derive d growth factor (PDGF), char acterized by the presence of eight conserved cysteine residues [80, 81]. VEGF-A is believed to be the major stimulator for vascular angiogenesis. VEGF-B is structurally similar to VEGF-A and PlGF is highly abundant in heart, skelet al muscle and pancreas and may regulate endothelial cell functions via a paracrin e fashion [82]. VEGF-C and VEGF-D are basically involved in lymphangiogenesis and i nduce the proliferation and cell survival of lymphatic endothelial cells [83-85]. VEGF-E, encoded by the Orf virus, is structurally similar to VEGF-A, specifically binds to VEGFR-2 and induces angiogenesis [86]. VEGF-F, as a collective name, summarized the variants isol ated from snake venoms [87]. The term VEGF refers to a collection of related isoforms expressed from the same gene [88]. The gene encoding VEGF, or VEGF-A, is located on the short arm of chromosome 6 in humans [89] and on chromosome 17 in mice [90]. The vegf gene consists of eight exons and seven introns, al ternative splicing resu lts in many isoforms. The best studied isoforms in human ar e VEGF121, VEGF165 and VEGF189. In mice, the homologous counterpart isoforms contai n one less amino acid, so mVEGF164 is the corresponding isoform for hVEGF1 65 [90], for example. In all isoforms, the transcrips of exon 1-5 are all conserved and exon 6 and 7 are where the alternative splicing occurs. Exon 3 and 4 encode the binding domains fo r VEGFR-1 and VEGFR-2 [91]. Exon 6 and 7 encode two heparin-binding domains, which influence receptor binding and solubility [92]. VEGF189, containing both the exon 6 and 7 transcripts, has high affinity for heparin sulfate and is mostly associated w ith the cell surface and th e extracellular matrix [93]. On the contrary, VEGF165, lacking e xon 6, is moderately diffusible; and

PAGE 43

27 VEGF121, lacking both exon 6 and 7, is high diffusible [94]. Recently a new isoform called VEGF165b, a variant of VEGF165, has been identified [95]. The C-terminus of VEGF165b is encoded by exon 9, instead of e xon 8 as in VEGF165 and other isoforms [96]. VEGF165b binds to but does not trigger receptor phosphorylation, so it is actually an endogenous inhibitory form of VEGF [9 6]. This is due to a missing exon 8-encoded C-terminus, which has mitogenic signa ling functions. Figure 1.13 [92] shows the alternative splicing among VEGF isoforms. Figure 1.13. VEGF-A isoforms [92]. VEGF Receptors VEGF binds to three cell surface receptor tyrosine kinases: VEGFR-1 (Flt-1), VEGFR-2 (Flk-1/KDR) and VEGFR-3 (flt-4) VEGFR-1 and VEGFR-2 are primarily located on vascular endothelium while VEGFR-3 is mostly found on lymphatic endothelium. These receptors are structura lly similar: all of them contain seven extracellular immnoglobin (Ig )-like domains, a transmembrane domain, a regulatory juxtamembrane domain, and a consensus tyrosine kinase domain interrupted by a kinaseinsert domain. The second and third Ig-like domains function as the high-affinity VEGF

PAGE 44

28 binding domain, whereas the first and fourth Ig-like domains regulate ligand binding and facilitate receptor dimeri zation, respectively [97-99]. VEGFR-1 has a molecular weight of 180 kDa and binds VEGF-A, VEGF-B and PlGF. The affinity of VEGFR-1 for VEGF is ten-fold higher than VEGFR-2 but its tyrosine kinase activity is ten-fold weaker th an VEGFR-2 [92]. In the classical views, one of the major functions for VEGFR-1 is to ac t as a decoy receptor restricting VEGF to bind to VEGFR-2, which is more mitogeni c [100]. VEGFR-1 is required for normal blood vessel development during embryogenesis and a VEGFR-1 knock-out is lethal in mice at embryonic day E8.5. The lethality was sh own to be associated with an abnormal increase in the number of endothelial proge nitors, which is the phenotype as VEGF hyperactivity, indicating a negative regulat ory function of VEGFR-1 [101]. Supporting this, a modified form of VEGFR-1 without the tyrosine kinase domain was constructed and found to be compatible with normal va scular development and angiogenesis in transgenic mice [102]. A naturally occurri ng soluble form of VEGFR-1, called sVEGFR1 or sFlt-1, is expressed from differentia l pre-mRNA splicing. sVEGFR-1 has the same ligand affinity as VEGFR-1, but is missing th e transmembrane and intracellular domains [103, 104]. It binds to free VEGF and reduces its availability to VEGF receptors, which further suggests its relative, VEGFR-1, as a negative regulator for VEGF signaling. However, VEGFR-1 does mediate VEGF signali ng in non-endothelial cells, especially those cells that only express VEGFR-1 as the VEGF receptor, such as monocytes and macrophages [105, 106]. A recent study showed that PlGF signaling mediated by VEGFR-1 in monocytes is associated with the inflammatory reac tions [107]. Besides

PAGE 45

29 monocytes, VEGFR-1 signaling is also beli eved to be important for endothelial progenitors and carcinoma cells. VEGFR-2, a 230 kDa glycoprotein, is rec ognized as the primary mediator of VEGF signaling. It regulates e ndothelial cell proliferation, migration, differentiation, cell survival and vessel permeability and dila tion. VEGFR-2 knock-out mice die between E8.5 and E9.5 due to deficiency in blood ve ssel formation [108], indicating that VEGFR2 is also crucial for the functions of he matopoietic/endothelial progenitors. VEGFR-3, 170 kDa, binds to VEGF-C and VEGF-D. It is expressed in embryoni c endothelial cells but postnatally becomes restricted to the lymphatic endothelium [109]. Apart from these three VEGF receptors, neurophilins (NRPs) can also act as cell surface receptor for VEGF, but in an is oform specific manner. NRP-1, originally identified on neuron cells as a receptor for class 3 semaphor ines/collapsins family of neuronal guidance mediators [110], is also expressed on endothelial cells. It lacks the intracellular tyrosine domain and needs to associate VEGFR-1 [111] and/or VEGFR-2 [112] to transduce a signal. It is suggeste d that NRP-1, as a co-receptor, can form a receptor complex with VEGFR-2 to enhan ce the binding the signaling of VEGF165 and VEGFR-2 cannot sufficiently transducer the VEGF signaling w ithout NRP-1 [113]. NRP-1 also binds VEGFR-1 forming a lig and-independent complex [111]. NRP-2, lacking an intracellular domain like NR P-1, can also bind to VEGF. It can bind VEGF121, VEGF145 and VEGF165, but NRP-1 ca nnot bind VEGF145. NRP-2 can also bind to PlGF and can interact with VEGFR-1 [114]. In addition to NRPs, heparin sulfate proteoglycans (HSPGs) can bind to the VE GF isoforms with the heparin binding domains, such as VEGF165 and VEGF189. H SPGs are abundant, highly conserved

PAGE 46

30 components of the cell surface and the extracellu lar matrix of all cells and have been reported to play a critical role in modulating the differential biological activities of VEGF isoforms [115]. Figure 1.14 [116] demonstrates the binding of VEGF varian ts to the receptors. In summary, VEGF-A binds to VEGFR-1, VEGFR2 and the receptor heterodimer; VEGFC and VEGF-D bind to VEGFR-2 and VEGFR-3. Notably, PlGF and VEGF-B exclusively bind to VEGFR-1 and VEGF-E excl usively binds to VEGFR-2, which is very useful in receptor specificity studies. Figure 1.14. VEGF family ligands and their receptors [116]. VEGF Receptor Signaling As mentioned above, VEGFR-2 is thought to be the major receptor for VEGF signaling in endothelial cel ls. Upon binding of VEGF, VEGFR-2 is activated by

PAGE 47

31 autophosphorylation, and initiates a number of signaling cascades that induce cell proliferation, migration, survival and/or increase in endot helium permeability. The cell proliferation induced by VEGF R-2 signaling typica lly involves MAPK pathways. Activation of VEGFR2 recruits Grb-2 and activat es it, which leads to the activation of Sos, then the activation of Ras, eventually the stimulation of Raf1/MEK/ERK signaling cascade [117]. Activat ed MAPK pathways will translocate to the nucleus and regulate the gene expressi on and cell proliferation. VEGFR-2 can also recruit PLC -1, and the activation of PLC -1 will induce phosphatidylinositol 4,5bisphosphate (PIP2) hydrolysis producing 1,2-diacylg lycerol (DAG) and inositol 1,4,5trisphosphate (IP3). The activation of PKC can result from the production of DAG, which further leads to the Ras-independent Raf ac tivation and thus the stimulation of ERK activity [118]. The data dem onstrating the requirement of PI3 kinase in the VEGFR-2induced cell proliferation ar e conflicting, so the involve ment of PI3 kinase is controversial [119, 120]. Cells expressing VE GFR-1 are unable to activate MAPK [121]. VEGF can act as a chemoattractant for e ndothelial cells so th at VEGF signaling is believed to be involved in cel l migration. Firstly, the sign aling from activated VEGFR-2 can promote focal adhesion kinase (FAK) phosphorylation and recruit it to focal adhesions, together with paxil lin and actin-anchoring proteins like talin or vinculin [122, 123]. Therefore the cytoskeleton organization is modified and cell migration is promoted. Secondly, the p38/MAPK pathway can be activated upon VEGF binding to VEGFR-2, and thus may play a role in cell migration and p38 inhibitors can decrease cell migration [124]. Thirdly, the PI3 kinase/Akt pathway can regulate the actin organization and cell migration [125]. Besides VEGF R-2, VEGFR-1 and NRPs have all been implicated in

PAGE 48

32 VEGF-mediated cell migration and invasion [92] However VEGFR-2 is considered to be the main mediator of cell migration. VE GFR-1 stimulates p38 phosphorylation and has no effect on endothelial cell migration [126]. PI3 kinase/Akt pathway plays an im portant role in the VEGF-induced cell survival. The phosphorylation of VEGFR-2 can l ead to the activation of PI3 kinase and Akt/protein kinase B (PKB). Akt is an anti-a poptotic factor and is sufficient to promote cell survival. It has been re ported that the inhibition of PI3 kinase abolished Akt activation and the VEGF-mediate d cell survival was also bl ocked [127]. VE-cadherin and -catenin can complex with VEGFR-2 and PI3 kinase and form a transient tetramer to promote cell survival [128]. The expression of some anti-apoptotic factors can also be induced by VEGF and contribute to cell surviv al, for instance, caspase inhibitors Bcl-1 and A1 [129] and IAP (apoptosis inhibitors ) family proteins [130]. VEGFR-1 cannot associate with the VE-cadherin complex [128] and does not activate the PI3 kinase/Akt pathway [127], so that it is thought to not be involved in VEGF-induced cell survival. Originally discovered as a vascular pe rmeability factor, VEGF can also increase the vascular permeability. The administration of VEGF to endothelial cells is shortly followed by the formation of some specialized regions in the cell membrane that are highly permeable to macromolecules [131]. PI3 kinase and p38/MAPK have been suggested to be involved in the increase of membrane permeab ility [132]. In the established vessels, VEGF also regulates vascular permeability by affecting the components of tight, adherence and ga p junctions, such as VE-cadherin, -catenin and occludin [116]. Another aspect of this intera ction is that endothelial NO synthase (eNOS)

PAGE 49

33 can induce the activation of Ak t, which further regulates th e NO level and leads to vessel dilation and permeabilization [133, 134]. Figure 1.15 [92] summarizes VE GF signaling via VEGFR-2. Figure 1.15. VEGF signaling via VEGFR-2 [92]. The Function of VEGF in Ocular Neovascularization VEGF is thought to play a central role in retinal angiogenesis as supported by data from animal models and clini cal investigation. VEGF is upre gulated in the retina during neovascularization in animal models with ischemia-induced retinopathy [135-138], and

PAGE 50

34 the VEGF mRNA is increased by three-fold within 12 hours of the onset of relative hypoxia and maintained for many da ys at higher levels until new vessels start to regress [136]. Patients with active PDR were found to have increased levels of aqueous and vitreous VEGF [139-145]. Higher levels of VEGF expression were also reported in epiretinal neovascular membranes and reti nas from PDR patients [146, 147]. However, an interesting finding in the active PDR pa tients showed that there was a significant decrease in VEGF levels after panretinal laser photocoagulation tr eatment [140]. Further evidence supporting VEGFs major role in re tinal neovascularizati on comes from VEGF inhibition studies. VEGF receptor chimeric proteins, neutralizing antibodies, and antisense oligonucleotides have succe ssfully showed inhibition effects on neovascularization [148-151]. Based on the evidence, it is widely accep ted that VEGF is very important and necessary for retinal neovascularization, but VE GF may not be sufficient for it. Repeated intraocular injections of VEGF or sustained intravitreous release of VEGF in primates results in severe changes to retinal vessels including di lation, leakage, and microaneurysms, but no apparent retinal neovascularization [152, 153]. When VEGF expression is driven by the reti nal-specific rhodopsin promoter in the transgenic mice, the development of neovascularization was produced in the deep capillary bed of the retina, and high levels of VEGF expression can furt her cause retinal traction and detachment [154]. The new vessels grew from the deep cap illary bed into the s ubretinal space. The close proximity of the deep capillary bed to the photorec eptor expressing VEGFs and differential susceptibility of the vascular beds might be an explanation for this vascular growth [155].

PAGE 51

35 The role of VEGF in choroidal neovascu larization (CNV) is less clear. Increased VEGF expression was found in fibroblasts and RPE cells of choroidal neovascular membranes surgically removed from patie nts [146, 156, 157]. And in the animal model of laser-induced CNV, it has been shown th at VEGF mRNAs were upregulated in the neovascular lesions [158]. VEGF is thought to be necessary in CNV development because several specific VEGF signaling inhi bitors have shown reduced CNV [159-161]. But VEGF is not a sufficient stimulator of CNV because increased expression of VEGF in photoreceptors or RPE cells do es not lead to CNV [154, 162]. Basic Fibroblast Growth Factor (bFGF or FGF2) FGF is a family of heparin-binding growth factors. bFGF has been localized in the adult retina. In the mouse model of ischemia -induced retinopathy, bFGF level is elevated during neovascularization [163] In the animal model of laser-induced subretinal neovascularization, RPE cells were found to be stained with aFGF and bFGF [164]. In studies on clinical specimen, both elevated a nd non-significantly-changed levels of bFGF have been reported in the vitreous samp le of PDR patients [165, 166], which argues against a major role in retinal neovasculariz ation. Further evidence comes from animal models. In the ischemia-induced retinopat hy or laser-induced CNV mouse model, transgenic mice deficient in bFGF developed the same amount of retinal or CNV as the wild-type mice, respectively, indicating bF GF expression may not be necessary in angiogenesis [167, 168]. It has been hypothesize d that bFGF will manifest its angiogenic potential when there is cell injury. It is found that bFGF can get access to the extracellular compartment during photorec eptor damage and increased CNV can be stimulated [169].

PAGE 52

36 Angiopoietins Angiopoietins and their receptors (Tie re ceptors) are another endothelial-specific system that has been implicated in va scular growth and development. Current understanding about the Tie receptors is that Tie1 signaling is important for vascular integrity and Tie2 signaling is important in remodeling of the developing vessels by maximizing the interactions between endothe lial and supporting cells [155]. The ligand for Tie1 has not been identified. Angiopoiet in (Ang) 1 and 2 are ligands for Tie2 receptor. Ang1 binds with high affinity and initiates Tie2 phosphorylation and downstream signaling. Ang2 also binds with high affinity, but does not stimulate phosphorylation of Tie2. It l ooks like Ang2 is a naturally occurring antagonist for Ang1 and Tie2. The interaction of Ang1 and Tie2 is essential for the remodeling function of Tie2 on newly developing vessels. And it has been hypothesized that Ang2 might provide a key destabilizing signal involved in in itiating angiogenic remodeling. The Ang2 blockade of Tie2 signaling can disrupt sta bilizing inputs to ECs, making ECs more responsive to VEGF and thereby stimulating angiogenesis. But when there is no VEGF present, those ECs are prone to apoptosis a nd the destabilized ve ssels regress [170]. Ang2 mRNA levels have been reported to increase in normal and pathological retinal angiogenesis [171-174]. It has been shown that Ang2 can stimulate a significant upregulation of proteinases in EC [174] that may be important for cell migration during retinal neovascularization. Platelet-Derived Growth Factor (PDGF) PDGF, a dimer protein, a potent mitogen and a chemoattractant, has been implicated in angiogenesis. Similar to VEGF PDGF is another grow th factor that is elevated after hypoxia [65]. Recent findings a bout PDGF include: increased levels of

PAGE 53

37 PDGF-AB was reported in vitreous sample s of PDR patients [ 175]; overexpression of PDGF-B in transgenic mice leads to prolifera tion of endothelial cells, pericytes and glial cells resulting in tract ion retinal detachment [176-179]. It has been proposed that PDGF may act in concert with VEGF in ischemic retinopathy [176-178]. Integrins Integrins are a family of transmembrane proteins that are the major cell surface receptors responsible for the at tachment of cells to the extr acellular matrix. Structurally, integrins are heterodimeric recep tors composed of two subunits, and More than 20 different integrins are formed from the combination of 18 known subunits and 8 known subunits. Each integrin binds to its own corresponding extracel lular matrix (ECM) and/or cell surface ligand. These include stru ctural ECM proteins, such as collagens, fibronectins, and laminins, as well as provisi onal ECM proteins that are deposited during tissue remodeling and thrombotic events [180] .The first integrin-binding site to be identified was the sequence Arg-Gly-Asp, wh ich is recognized by several integrins. However, other integrins bind to other distin ct peptide sequences. Wh ile integrins are one of the most essential cell su rface components in the body and are present in almost all tissues, no cell expresses all integrins. Indeed the particular integrin types expressed are dependent on the ECM ligands present with in the local microenvironment. Even on a given cell type, the specific integrins expresse d are also altered to match the concurrent changes within the local ECM. So the expressi on of integrin is spatially and temporally regulated. The integrins also function as an anc hor for the cytoskeleton. The interaction between the cytoskeleton and th e extracellular matrix is resp onsible for the stability of cell-matrix junctions. There are two categorie s of cell-matrix junctions: focal adhesion

PAGE 54

38 and hemidesmosome. In focal adhesi ons the cytoplasmic domains of the subunits of integrins associate with bundles of actin filame nts to anchor the actin cytoskeleton at the cell-matrix junctions. While in hemidesmosom e integrins interact with intermediate filaments instead of actin. Hemidesmosome is mostly found in the an chorage of epithelial cells to the basal lamina. Integrin Signaling Unlike many cell surface receptors that c ontain tyrosine kinases, integrins do not contain intrinsic tyrosine kinase activit y. Upon ligand binding, th e integrins undergo a conformational change into its activation st ate. The change in activation has been assessed by showing evidence of polymerizati on, clustering, or the surface exposure of different antibody binding epitopes [181]. Since the cytoplasmic domains of the integrins can bind constitutively to cytoskeletal com ponents such as talin, the conformational change and activation of integrins can result in changes in cytoskeletal protein functions, which will lead to major changes in cell shape and locomotion. On the other hand the activation of integrins can initiate a seri es of signaling tran sductions, with the involvement and assembly of a variety of signaling molecules. A non-receptor protein tyrosine kinase cal led FAK (focal adhesion kinase) plays a key role in integrin signaling. FAK is loca lized at the focal a dhesion and is rapidly tyrosine auto-phosphorylated following lig and binding by integrins. Besides FAK, members of the Src family or non-receptor prot ein tyrosine kinases also associate with focal adhesion and are involved in integrin si gnaling. Src and FAK prob ably interact with each other, resulting from the binding of th e Src SH2 domain to the auto-phosphorylated sites of FAK. Src then phosphorylates additi onal sites on FAK. In addition to Src, the binding sites for SH2 domain created dur ing FAK phosphorylation are also taken

PAGE 55

39 advantage of by other downstream molecules, fo r instance, PI-3 kinase and the Grb2-Sos complex. These signaling molecules can form multicomponent signaling complexes that recruit and include small GTPase proteins su ch as Ras, Rho, Rac. Their involvement and activation will further lead to the activati on of a number of signaling cascades. Figure 1.16 [180] demonstrates the in tegrin signaling via the Akt, ERK and JNK pathways. These signals collaborate to regulate cellular pr oliferation, migration and survival. And also, many small GTPases like Rho and Rac play critical roles in cytoskeletal remodeling events [180]. Figure 1.16. The activation of inte grins can lead to the signal transduction in a number of pathways. [180]. As mentioned above, integrins need to be activated to serve as a signaling molecule. The activation involves a conformational change that results in an increase in ligand-binding affinity. Proposed in the current model, the inactive fo rm of integrins are

PAGE 56

40 in a folded conformation in which the lig and-binding domain is adjacent to the membrane. When activated, the affinity for the ligand is increased, and ligand occupancy stabilizes the extended conformation of the integrin [182] Simultaneously, the associated topological change in the transmembrane and cytoplasmic domains makes them separate and bind to intracellular signaling molecule s to initiate downstream pathways [182]. According to this model, the conformational ch ange in integrins that induces signaling is the same as the one that is induced by ac tivation. And this ac tivation state can be promoted by both extracellular lig ands (so-called outside-in signaling) and intracellular signaling molecules (inside-out signaling) [182]. The outside-in signaling is usually triggered by ECM ligands and the inside-out si gnaling molecules are usually the effectors of the activation of growth factor receptors. The ECM (l ocal determiner) and growth factors (systemic and local determiner) can work synergically to enhance the signaling outcome induced by specific integrins in a gi ven cell. Under certain circumstances it is not sufficient to promote cell survival a nd proliferation until both proper ECM and growth factors are both present. The activation of integrins, especially t hose involving the interaction with growth factor receptors, usually o ccur in lipid-raft microdomains, where cholesterol and glycosphingolipids [183] and intracellular signaling molecules like Src family kinases [184] are relatively concentrated in the cel l membrane. These lipid-raft microdomains are distinct from the surrounding membrane in that they restrict the diffusion of the contents. It is suggested that the lipid -raft has other functions [182]. First, they could serve as a physical concentration of pre-assembled mo lecules for signaling upstream or downstream of the integrin, and the si gnaling inhibitory molecules could be excluded. Second,

PAGE 57

41 different integrin pools could be separated so that their ow n distinct function could be better performed. Third, the lipi d-raft may also facilitate and/or maintain integrin activation. In addition to help from concentr ated pre-assembled molecules, the altered membrane structure, due to the distinct ch emical characteristics in the lipid-raft, may favor conformational equilibrium between the inactive and the active forms. It is also proposed that the active integrin s might help to generate the lipid-rafts in other models [185]. The integrins can regulate the signaling of growth factor re ceptors. First the phosphorylation state of the grow th factor receptors can be regulated. One example is the interaction between v 3 and the epidermal growth fact or receptor (EGFR) on human endothelial cells. The adhesi on to the ECM mediated by integrin can lead to a low phosphorylated state within the cell, resulti ng in the phosphorylation of four tyrosine residues but not on the fifth tyrosine which is only phosphor ylated by EGF binding. This phosphorylated state is lower than in high concentration of EGF but ECM attachment doesnt occur. This low phosphorylated state is sufficient to induce ce ll survival but not proliferation. However, if only low con centrations of EGF ar e present, the ECM attachment can promote the phosphorylation sim ilar to high concentrations of EGF alone [182]. Thus the phosphorylation of EGFR on endothelial cells is not only regulated by ligand binding, but also regulated by integrins. The regulation on growth factor receptors can also occur when integrins inte rfere with the receptor expression. As for the inside-out signaling, the activa tion of growth factor receptors is usually the source of signaling. Integrins can be re gulated by growth factor receptors in many aspects and cell behavior can be altered. The integrin expression leve l can be altered, for

PAGE 58

42 instance, the expression level of a number of integrins on endothelial cells are increased by angiogenic growth factors such as FGF-2 [186]. The phosphorylated state of integrins can also be regulated by growth factor rece ptors. One example is the laminin receptor 6 4, an essential component in the hemi desmosomes, influences epidermal cell attachment to the underlying basal lamina. EGFR can induce the phosphorylation of the cytoplasmic domain of 4 subunit. This results in the cyto plasmic recruitment of Shc, and the activation MAPK and PI3K More importantly, the change in the phosphorylation state leads to release of the integrin from its liga nd, thus the hemidesmosome disassembles, which is a required step for cell proliferation and/ or migration [187]. Besides the phosphorylation state, growth fact or receptors can also alter the activation state of integrins. For example, it ha s been shown that VEGF can activate v 3 on human umbilical vein endothelial cells thus the adhesion to ECM is promoted and cell migration follows [188]. Relationships between Integrin and Other Growth Factor Receptors in Angiogenesis Among the over 20 integrins that have b een discovered to date, two of them, v 3 and v 5, are thought to be especia lly important for angiogenesi s. These integrins are not seen on normal epithelial cells in skin, but are highly expressed on endothelial cells participating in angiogenesis [189]. Only v 3 was found in choroidal neovascular membranes from ARMD patients, while both v 3 and v 5 were found in epiretinal membranes from DR patients [190]. Therefore, retinal and choroida l neovascularization may differ in the integrin requirement. Inhibi tion studies on integrin s further support this. Agents that bind v 3 and/or v 5 can suppress retinal neov ascularization, even though

PAGE 59

43 the effect is modest, but the inhibition of v 3 or v 5 has no significant effect on choroidal neovascularization [189]. Endothelial cells express at least eight different integrins including v 3 and v 5 [191], each of them having their own specifi c ligand. For example, collagen is a ligand for 2 1 while fibrin is a ligand for v 3, so that v 3 influences adhesion and signaling events of the endothelial cells bound to fibr in [192] but not of those bound to collagen [40, 193]. However, the endothelial cells w ill eventually become apoptotic when bound with collagen alone via 2 1. The unligated v 3 receptors seem to cluster on the cell membrane and colocalize with caspase activity, especially caspase 8 [194]. In addition to v 3, many other unligated integrins are likely to induce cell death, this is why integrins could be categorized as dependent recep tors under a variety of circumstances. v 3, expressed (although not exclusively) on endothelial cells, has been linked to many angiogenic signaling pathways via the in teraction with recepto rs for a number of growth factors, such as VEGF, EGF, IGF1, PDGF and insulin. Since VEGF and IGF-1 are the two most important growth factor s involved in my dissertation work, I am focusing on the interaction between v 3 and VEGFR and IGF-1R. VEGFR-2 activation by phosphor ylation is promoted by v 3 [195]. v 3 and VEGFR-2 interact and the co -immunoprecipitation of these two receptors has been demonstrated. However VEGFR-2 does not co-immunoprecipitation with the 1 or 5 subunits. VEGFR-2 phosphorylation and mitogeni city are enhanced in cells plated on vitronectin, an v 3 ligand, compared with cells plated on fibronectin, an 5 1 ligand, or collagen, an 2 1 ligand; further demonstrating a f unctional relationship between VEGR2 and v 3. Cell adhesion, migration, soluble ligand binding, and adenovirus gene

PAGE 60

44 transfer mediated by v 3 are all enhanced by VEGFR-2 signaling. An anti3 integrin antibody reduces VEGFR-2 phosphorylation and PI3 kinase activity suggesting that VEGFR-2 signaling initiated by v 3 occurs through the PI3 kinase pathway. Another molecule, p66 Shc (Src homology 2 domain containing), has been shown to play a key role in the VEGFv 3 interplay during tumor gr owth and vascularization [196]. The activation state of v 3 integrin has a critical function in in vivo tumor growth by influencing VEGF expressi on. By using a non-activable 3, a S752P mutant that cannot cluster, it was found that the stimul ation of VEGF expre ssion also depends on v 3 clustering. The recruitment of p66 Shc and phosphorylation of 3-associated p66 Shc are enhanced following v 3 clustering. The recruitment is not sufficient for v 3mediated effects on VEGF production and tumo r vascularization but the phosphorylation is necessary, in that a dominant-negative form of p66 Shc, which is phosphorylationdefective, completely abolished integrin-induced VEGF expression. IGF-1 is a classic endocrine hormone a nd systemically synthesized in liver and transported to the peripheral tissues stimulating growth. In addition, IGF-1 is also synthesized locally in peripheral tissue to promote growth in an autocrine/paracrine manner. Similar to VEGF and other growth factors, the extra cellular environment contributes to influence the outcome of th e hormone signaling. It has been shown that many ECM proteins, such as collagen type I and type IV, fibr onectin, thrombospondin, and osteopontin, can modulate the response of various cell types to IGF-1 stimulation via their integrin receptors [181] The interactions between v 3 and IGF-1 on vascular smooth muscle cells (SMC) have been illustrated in great detail and can be used as a good example of how growth factors and in tegrin signaling influence each other.

PAGE 61

45 When IGF-1 binds to the IGF1-R, IGF1-R will auto-transphosphorylate its two subunits, and further recruit signaling molecule s such as insulin re ceptor substrate-1 (IRS-1) and Shc, which can transduce the singling into correspondi ng cascades, such as the PI3K and MAPK pathways. Despite kina ses, phosphatases also participate in the signaling modulation. Phosphatases induce deph osphorylation reactions which can result in either activation or inactivation of signaling molecules. One phosphatase, Src homology 2 containing tyrosine phosphatase (SHP-2), normally transfers to IGF-1R 20 minutes after IGF-1 stimulat ion, resulting in a decrease in the phosphorylation level of the receptor and subsequent at tenuation of MAKP and PI3K activation [181]. However, a premature transfer at 5 minutes and prem ature attenuation has been found when the ligand occupancy of v 3 is blocked [197]. So obvious ly the properly liganded and activated v 3 is a necessary partne r in IGF-1R signaling. Normally when IGF1-R and v 3 are activated after ligand binding, SHP-2 will transfer to the phosphorylated 3 subunit first. An adaptor protein, DOK-1, facilitates the transfer. DOK-1 is phosphorylated after IG F-1 stimulation, and the YXXL motifs within its C-terminus domain become capable of bi nding to SHP-2 via SH-2 domains [198]. Also, DOK-1 contains a phosphotyr osine binding (PTB) domain, which allows it to bind to 3 at a tyrosine that is phosphorylated after v 3 activation [199]. Thus DOK-1 mediates SHP-2/ 3 association. If the transfer of SHP-2 to 3 is impaired for any reason, SHP-2 will be aberrantly tran sfer to IGF-1R instead and the premature dephosphorylation of IGF-1R occurs [181]. One SHP substrate, SHPS-1, becomes phos phorylated after IGF-1R activation. It is a single chain transmembrane protein a nd SHP-2 can bind to it via SH-2 domain. The

PAGE 62

46 transfer of SHP-2 from 3 to phosphorylated SHPS-1 is a necessary step to maintain optimal MAPK and PI3K activation [200]. SHPS -1 also recruits Shc to form a complex that is critical for MAPK a nd PI3K activation. SHP-2 can activate a Src family kinase via SH-2 domain binding, so that this Src fa mily kinase is recruited to SHPS-1 and phosphorylates Shc in the complex [181]. SHP-2 is further transferre d to the appropriate downstream signaling molecules to main tain MAPK and PI3K activation. v 3 has several ECM ligands, such as osteopontin, thrombospondin and vitronectin. For v 3 on SMC, the major ECM ligand is vitronectin. The heparin binding domain and RGD (arginine-glycine-asparg inine) sequence can both function as the v 3 binding site. It is believed that the heparin binding domain is the binding site triggering 3 activation, in that the exposure of cells with the heparin binding domain peptide results in v 3 phosphorylation and recruitment of SHP2 to the plasma membrane [201]. Contrarily, binding of 3 to the RGD sequence has been found to induce the cleavage of 3, thus also the premature recruitment of SHP-2 to IGF-1R and the premature IGF-1R dephosphorylation [202]. Similar to the interaction between integrin s and other receptors, it is believed that v 3 and IGF1-R signaling occurs within a restricted compartment on the membrane. Integrin-associated protein (IA P) facilitates the formation of this compartment. After IGF-1 exposure, IAP is transl ocated to the regions where v 3 resides [181]. More importantly, IAP can induce an increase in the affinity of v 3 for its ligands [203]. The extracellular domain of IAP can associate with SHPS-1 and an antibody disrupting this association prevents IGF-1 stimulation of SHPS-1 phosphorylation and SHP-2 transfer to

PAGE 63

47 SHPS-2 [204]. Therefore, the clustering of v 3 and the assembly of a signaling complex involving SHPS-1 may be a crucial in v 3 and IGF-1R signaling. Pigment Epithelium-Derived Factor (PEDF) The vasculature is normally quiescent under physiological conditions, since there is a balance between the proangiogenic and anti-angiogeni c factors. Angiogenesis is initiated when there is increase in pro-a ngiogenic factors and/or decrease in antiangiogenic factors. PEDF is one of the na turally occurring antiangiogenic factors. In the mouse model of retinopathy, it has been shown that hyperoxia results in a decline of VEGF levels with a concomitant expression of PEDF, and the relative hypoxia led to downregulation of PEDF during the angiogenesis process [205]. Systemic or intravitreal administration of PEDF [206, 207] and gene transfer w ith adenoviral vectors expressing PEDF [176-179] have been reported to decrease the ocul ar neovascularization levels, In the clinical studies, The vitreous levels of PEDF from PDR patients were found to be lower than normal [208], and the im munochemical staining of PEDF on retinas from PDR patients are much less intense co mpared with non-PDR [208]. All of these evidence supports that PEDF an anti-angiogenic factor may be involved in the suppression of retinopathies. Insulin-Like Growth Factor (IGF)-1 The discovery of a role of growth hormone (GH)/IGF-1 in DR can be traced back to 1950s. The regression of retinal neovascular ization was seen afte r pituitary infarction [209], and pituitary ablation was even used as a therapeutic method for PDR. More recently, in several studies in patients with PD R, elevated serum and vitreous levels of IGF-1 have been associated with retinal neovascularization [210-212].

PAGE 64

48 In a GH inhibition study, retinal neovascu larization was suppre ssed in transgenic mice expressing a GH antagonist gene and norma l mice treated with an inhibitor of GH secretion [213]. This inhibition of neovasc ularization could be reversed by exogenous administration of IGF-1. IGF-1 also plays a necessary role in nor mal retinal vascular development. In IGF-1 knockout mice, norma l development of the retinal vasculature was arrested despite the presence of VEGF [214]. This also supports the idea that VEGF alone is not sufficient for the development of retinal vessels. Clinic ally it has been found that the development of ROP in premature infants was strongly associated with a prolonged period of low levels of IGF-1 [214]. This suggests that the critical role IGF-1 plays during normal retina vascular developm ent. Lack of IGF-1 in the early neonatal period leads to the development of avascular retina, and late r the proliferative phase of ROP [155]. The function of IGF1 in CNV is still not clear. The IGF system includes the IGF-1, IGF2, the IGF-1 receptor (IGF-1R), and IGF binding proteins (IGFBPs). IGF-1 can be expressed in the liver and utilized systemically as an endocrine, or can be expressed at pe ripherals and function in autocrine/paracrine mechanisms. The multiple physiologic and pa thologic effects of IGF-1 are primarily mediated by IGF-1R, and are also modulated by complex interactions with IGFBPs, which themselves are also modulated at multiple levels. IGF-1 and IGF-1R IGFs are synthesized in almost all tiss ues and have important regulatory function on cell growth, differentiation, and transf ormation. IGF-1 is the product of the IGF-1 gene, which has been mapped to chromosome 12 in humans and chromosome 10 in mice [215]. IGF-1 functions in both prenatal and postnatal development and exerts all of its known physiological effects through binding with IGF-1R. Circulating IGF-1 is

PAGE 65

49 generated in the liver under the contro l of growth hormone [216], and bound with IGFBPs as the endocrine form in the circula tion. The IGF-1 produced in other organs and tissues has a lower affinity for IGFBPs, re presenting autocrine and paracrine forms of IGF-1. The IGF-1R gene is located on chromosome 15 in human [215], and IGF-1R is expressed everywhere in the body. The mature receptor is a tetramer consisting of 2 extracellular -chains and 2 intracellular -chains with the intracellular tyrosine kinase domain. IGF-1R signaling involves autophosphorylation a nd subsequent tyrosine phosphorylation of Shc and insulin receptor subs trate (IRS) -1, -2, 3, and -4. IRS serves as a docking protein and can activate multiple signaling pathways, including PI3K, Akt, and MAPK. The activation of these signali ng pathways will then induces numerous biologic actions of IGF1 (Figure 1.17 [216]). Figure 1.17. IGF-1 signaling transduction [216].

PAGE 66

50 The expression of IGF-1 in ECs is low, but it is expressed both in macrovessel and microvessel ECs. IGF-1 stimulates vascul ar EC migration and tube formation. IGF-1 is important for promoting retinal angiogene sis, and an IGF-1R antagonist suppresses retinal neovascularization in vivo by inhibiting vascular e ndothelial growth factor (VEGF) signaling [217]. The effect of IGF-1 on ECs is mediated in different signaling pathways. For example, IGF-1-induced nuclear factorB (NFB) translocation requires both PI3K and extracellular-regulated kina se, while IGF-1-stimu lated EC migration requires only PI3K activation [218]. And th e IGF-1 effects are also regulated by endothelial nitric oxide synthase (eNOS) expression and VEGF signaling [217]. IGF-1 and IGF-1R are also expressed in vascular smooth muscle cells (VSMCs), and their expressions are regul ated by several growth factors in different pathways. Thrombin and serum deprivation, tumor necrosis factor (TNF), and estrogen downregulate IGF-1 mRNA and protein levels ; reactive oxygen species (ROS) increases the levels; Ang2 and PDGF have been reporte d to both increase and decrease the levels. IGF-1 functions as a potent mitogen and antiap optotic factor and migration stimulator for VSMCs [216]. As for the IGF-1R, its expr ession can be upregul ated by Ang2 via the activation of NFB [219]; can be upregulated by fi broblast growth factor (FGF), mediated by the transcriptional factor ST AT1, STAT 3 [220]; and the Ras-Raf-MAPK kinase pathway was shown to be required fo r both of the above growth factor. The crosstalk between IGF-1R and other receptors can also regulate IGF-1 function. For instance, blocking ligand occupancy of V 3 integrin receptor results in premature recruitment of SHP-2 to the IGF-1R receptor a nd reduces IGF-1 signaling [200].

PAGE 67

51 IGFBPs and ALS At least 6 IGFBPs have been well charac terized, and they function as transporter proteins and as storage pools for IGF-1. The expression of IGFBPs is tissueand developmental stagespecific, and the con centrations of IGFBPs in different body compartments are different. The functions of IGFBPs are regulated in multiple ways, such as phosphorylation, proteoly sis, polymerization [221], and cell or matrix association [222] of the IGFBP. All IGFBPs have been shown to inhibit IGF-1 action, but IGFBP-1, 3, and -5 are also shown to stimulate IGF-1 action [223]. Some of IGFBPs effects might be IGF-1 independent. Figure 1.18. Proposed pathway of IG F-dependent IGFBP action [223]. The precursor forms of IGFBPs have secretary signal peptides and mature proteins are all found extracellularly. They a ll have a conserved amino-terminal domain, a conserved carboxyl-terminal domain and a non-conserved central domain. Both of the amino-terminal and carboxyl-terminal contri bute to IGF binding [223], which implies

PAGE 68

52 IGF-binding pocket structure. The major IGF transport function can be attributed to IGFBP-3, the most abundant circulating IGFBP. It carries 75% or more of serum IGF-1 and IGF-2 in heterotrimeric complexes that also contain the aci d labile subunit (ALS) [224]. Free or binary-complexes (without AL S) are believed to exit the circulation rapidly, whereas ternary complexes appear to be essentially confin ed to the vascular compartment. In addition to their effects de rived form circulation, IGFBPs also have local actions, both autocrine a nd paracrine. They have been documented to affect cell mobility and adhesion [225, 226], apoptosis a nd survival, and cell cycle [227-229]. I will concentrate on IGFBP-3 in this discussion. IGFBP-3 have both potentiation and inhibi tion effect on IGF-1 actions. It is thought that IGFBP-3 inhibits IGF-1-mediated effects via its high-affinity sequestration of the IGF-1. But in contrast, preincuba tion of cells with IGFBP-3 before IGF-1 treatment can lead to the accumulation of cell-bound forms of IGFBP-3 with lowered affinity for IGF [230], which may enhance th e presentation of IGF1 to IGF-1R. But It was also found that cell-bound forms of IGFB P-3 could still attenuate IGF-1-mediated IGF-1R signaling [231]. It has also been reported, based on competitive ligand-binding studies, that IGFBP-3 can inte ract with IGF-1R, causing inhi bition of IGF-1 binding to its receptor [232]. Therefore, the interaction of IGFBP-3 with IGF-1 and IGF-1R signaling system requires further study. Limited digest ion from proteases on IGFBP-3 can release IGF-1 from the complex and control the bi oavailability of IGF-1. These specific proteases include serine pr otease, cathepsins, and matr ix metalloproteinases [223]. Proteolysis results in IGFBP-3 fragments with decrease affinity for IGF-1, but several studies have shown the inhibition of IGF acti ons by IGFBP-3 fragments with low affinity

PAGE 69

53 for IGFs [223]. It is not clear whether this inhibition comes form IGF-1 sequestration or from its interaction with IG F-1R. IGFs themselves can al so influence the production of IGFBPs and IGFBP-specific proteases, or regula te the activity of these proteases [223]. Figure 1.18 [223] summarizes proposed IGFBP actions that depend on binding of IGFs and modulation of IGF-1R. Figure 1.19. Overview of possible IGFBP3 antiproliferation pathways [223]. IGFBPs also have their own intrinsic bi oactivity, without modulating IGF actions, either in the absence of IGFs (IGF-independent effects) or in the presence of IGFs without triggering IGF-1R signaling (IGF-1R-i ndependent effects). Recently there has been particular interest in IGFBP-3s f unction to induce apopto sis independently of inhibiting the survival functi ons of IGF-1 [233-236]. Several studies using human breast cancer cells have correlated the induction of IGFBP-3 mRNA and protein expression with growth-inhib itory effects of vari ous antiproliferative agents including TGF, retinoic acid [237], antiestrogens [238 ], vitamin D analogs [239], and TNF[240]. IGFBP-3 expression is also upregulated by the transcription factor p53 in colon

PAGE 70

54 carcinoma cells. And in the experiments using antisense IGFBP-3 or specific antibodies to sequester the IGFBP-3, the antiproliferativ e effects of some of these factors and be partially abrogated [223]. In addition, there is evidence s howing that some proteolyzed forms of IGFBP-3 also have IGF-independent effect, especially some IGFBP-3 aminoterminal fragments [223], and they showed little or no affinity for IGFs. This supports the existence of IGF-independent bioactivity. Figure 1.19 [223] summarized some of the proposed pathways of IGFBP-3 independent functions. IGFBP-3 has IGF-1 independent effects. Interactions of IG FBP-3 with known signaling pathways have been demons trated. The type V receptor for TGF(T RV) has been shown to be bound with IGFBP-3 relati ve specifically and may be involved in IGFBP-3 inhibitory signa ling [241]. IGFBP-3 has been shown to stimulate the phosphorylation of T RI of the signaling intermediate s Smad2 and Smad3 [242], while T RV signaling does not involve Smad phosphoryl ation. All-tans-retinoic acid (RA) is a potent inducer of IGFBP-3 in some cancer ce lls [223]. The growth-inhibitory effect of RA requires the presence of RA receptor (RAR)and can be blocked by retinoid X receptor (RXR)-specific retinoids. IGFBP-3 ha s been shown to inhibit RA signaling, possibly through enhancing RXR signaling [223] IGFBP-3 may also interact with PI3kinase pathway and MAPK pathway. LY294002, an inhibitor of PI3-kinase activity, could block the effects of IGFBP-3 [243] ; MAPK/ERK pathway inhibitor, PD98059, can restore the inhibitory effect of IGFBP-3 on DNA synthesis, blocked in cells expressing oncogenic ras in breast epithelial ce lls [244]. Recently it has been shown that IGFBP-3 strongly up-regulate signal tran sducer and activator of tran scription 1(STAT1) mRNA in the process of chondrocyte differentiation, and phospho-STAT1 protein was shown to

PAGE 71

55 increase and translocate to the nucleus, more over, the antiproliferative effects of IGFBP3 in these cells can be ablate d in the presence of STAT1 an tisense oligonucleotide [245]. The acid-labile subunit (ALS ), together with IGFBP-3 and IGF-1, forms the ternary complex as the storage pool in the pl asma. ALS is synthesized almost exclusively by the liver, and predominantly stimulated by GH [246]. Presence of ALS after birth is coincident with increased responsiveness to GH resulting from an increase in GH secretion and hepatic GF receptors. After pube rty, ALS concentrations basically remain stable throughout adulthood [246]. ALS is a single copy gene, containing 2 exons and 1 intron. ALS has no affinity for free IGFs and very low affinity for uncomplexed IGFBP3, and even its affinity for binary comp lex (IGF-1 + IGFBP-3) is 300-1000 fold lower that that of IGFBP-3 for IG Fs [247]. The ability of ALS to form ternary complex is irreversibly destroyed under acidic conditions. IGFBP-3 and IGFBP5 can both associate with ALS, with the latter being much w eaker [246]. The carboxyl-terminal domains of IGFBP-3 and IGFBP5 are important for bindi ng. The association is proposed to happen within the negative-charged sialic acid on the glycan chains of ALS and an 18 amino acid positive-charged domain in IGFBPs [246]. Besides liver, ALS local synthesis may occur in kidney, developing bone, lactating mammary gland, thymus and lung [ 248, 249]. Their functions are to sequester IGFs into ternary complex. A GH-responsive element of the ALS gene transcriptional promoter was identified [250]. This se quence was called ALSGAS1 because of its resemblance with the consensus sequence for -interferon activated sequence (GAS). The effects of GH on the ALS gene are mediated by the JAK-STAT pathway [251, 252]: the tyrosine kinase JAK2 is recruited to the activated GH receptor complex and

PAGE 72

56 phosphorylates signal transducers and activat ors of transcription (STAT)-5a and STAT5b. After dimerization, STAT5 isomers tran slocate to the nucl eus, and activate ALS gene transcription by binding to the ALSGAS1 elem ent. The GH signaling pathway leading to increased ALS gene transcription is critically dependent on the activation of STAT5 isomers, and is indepe ndent of RAS activation. One the of physiological significances of AL S is to extend the half-lives of IGFs from 10 min when in free form, and 30-90 min when in binary complexes, to more than 12 hours when in ternary complexes [253]. The other important role of ALS is to prevent the non-specific metabolic effects of the IGFs given that serum IGF concentration is ~1000 fold that of insulin [246]. IGFs in ternary complexes cannot traverse capillary endothelia and activate the insu lin receptor, whereas free IGFs and IGFs bound as binary complexes can do so. Incorporation of IGFs in to ternary complexes therefore completely restrains the intrinsic insulin-like eff ects of the IGFs. Null ALS mouse shows significantly reduced circulating IGF-1 and IG FBP-3 concentrations [246], which proves that ALS is absolutely necessary for se rum accumulation of both IGF-1 and IGFBP-3. The Involvement of Insulin Receptor (IR) and IGF-2 in Angiogenesis IGF-1 primarily binds to IGF-1R, and insu lin primarily binds to IR, while IGF-2 can bind to both of the two receptors and its own IGF-2R, as shown in Figure 1.20 [254]. Regarding retinopathy, insulin and IGF1 have gained more attention. Kondo et a t [255], using the Cre-Lox knockout system, found that (1) the retinas of mice develop normally in the absence of endothelial IR or IGF-1R. Presumably, suff icient growth factors (for example, VEGF) are present to facilitate normal developmen t. (2) Under conditions of relative hypoxia and in the pr esence of endothelial IR/IGF -1R, VEGF, eNOS, and ET-1 are increased, leading to extra-retinal neovascularization. (3) Under conditions of relative

PAGE 73

57 hypoxia and in the absence of endothelial IR or IGF-1R, VEGF, eNOS, and ET-1 are reduced, possibly due to impaired HIF-1 activ ation or reduced PI3K activity related to IG/IGF-1R [256]. Reduced neovascularization results from less IR/IGF-IR input. And in their experiments, the reduction of VEGF, eNOS, and ET-1 are reduced to a greater extend in IR knockout mouse than IGF-1R knockout mouse, which has brought more emphasis on IR function in the retinopathy, wh ile traditionally IGF-1R is thought to be more important. Figure 1.20. The crosstalk between IGF-1, IGF-2 and Insulin signalings [254]. RNA Silencing Technologies The traditional method to inactive a gene is to create a gene knockout animal model. This process has its advantages, in that it entir ely abolishes a gene expression, however, the disadvantages are that it is time consuming, expensive, labor-intensive, and subject to possible failure due to embryonic lethality [257]. RNA silencing technologies, which

PAGE 74

58 inhibit gene expression at the RNA leve l, are valuable tools to inhibit the Figure 1.21. Overview of RNA s ilencing technologies [258]. expression of a target gene in a sequence-sp ecific manner, and may be used for functional genomics, target validation and therapeutic purposes. Theoretically, RNA silencing could be used to cure any disease that is caused by the expression of a deleterious gene [258]. There are three common types of anti-mR NA strategies. Firstly, the use of single stranded antisense oligonucle otides; secondly, the triggeri ng of RNA cleavage through catalytically active oligori bonucleotides referred to as ribozymes; and thirdly, RNA interference induced by small interfering R NA molecules. Figure 1.21 [258] basically summarized the mechanisms of these th ree kinds of antisense technologies.This scheme also demonstrates the difference between an tisense approaches and conventional drugs, most of which bind to proteins and there by modulate their function. In contrast, RNA silencing agents act at the mRNA leve l, preventing translation. Antisense-

PAGE 75

59 oligonucleotides pair with their comple mentary mRNA, whereas ribozymes and DNA enzymes are catalytically active oligonucleotid es that not only bind, but can also cleave, their target RNA. RNA interf erence is a highly efficient method of suppressing gene expression in mammalian cells by the use of 21 23-mer small interfering RNA (siRNA) molecules. These three RNA silenc ing methods are detailed below. Antisense Oligonucleotides The antisense oligonucleotides was first described by Zamecnik and Stephenson who used a 13-mer DNA to inhibit Rous sarc oma virus expression in infected chicken embryonic fibroblasts [259]. The antisense gene silencing naturally occurs in genomic imprinting, in which only one copy of a gene in the mammalian genome is expressed while the other is silenced. It could be the maternally inherited allele or the paternal inherited allele. Antisense oligonucleotides are comple mentary to the target mRNA and are usually 15-20 nucleotides in le ngth [258]. There are two majo r antisense mechanisms that have been proposed [258]. First, RNase H cleaves RNA in the RNA-RNA heteroduplex (or RNA:DNA heteroduplex for antisense DNA o ligonucleotides), induced by binding of the antisense oligonucleotides. This results in rapid degrad ation of the cleaved mRNA products and a reduction in ge ne expression. Second, transla tion is arrested by steric blocking the ribosome by the binding of antis ense oligonucleotides. When the target sequence is located within the 5 terminus of a gene, the binding and assembly of the translation machinery can be prevented. The first step in designing an antisense oligonucleotides is ta rget selection and verification of target site accessibility. Computer program s, like Mfold, perform mRNA

PAGE 76

60 secondary structure analysis. This analys is can generate several mRNA secondary structures centered on our target sequence. If the target is always contained within a stable stem in every structure, this target shou ld be eliminated. In addition to this type of in silico analysis of RNA secondary structure, a number of in vitro methods have been developed to examine secondary structure in solution. One way is to directly probe the secondary structure of the ta rget RNA with 1-cyclohexyl-(2morpholinoethylo)cabodiimide methop -toluene sulfonate (CMCT) [260]. CMCT will mainly modify Us, and to a lesser extent Gs, in single-stranded regions of an RNA molecule. CMCT modification is followed by reverse transcription. Modification of Us and Gs will prevent read-through by reverse tran scription, resulting in a pause or stop site at the modified position. When these modifi cation/reverse transcri ption reaction products are separated on an appropriate electrophoresis gel next to DNA sequencing reactions of the target mRNA region, accessible regions of the target RNA are easily identified.The most sophisticated approach reported so far is to desi gn DNA array to map an RNA for hybridization sites of oligonucleotides [261]. Figure 1.22. Modifications in an tisense technology [258]. When designing antisense oligonucleotides there are some points to consider. Four contiguous guanosine residues should be avoided due to the G-quartets formation

PAGE 77

61 and CpG motifs should be avoided due to poten tial stimulation of the immune system. In addition, a BLAST search for each oligonucle otide sequence is required to avoid significant homology with other mRNAs that could cause unwanted gene silencing. Unmodified oligonucleotides are rapidl y degraded in biological fluids by nucleases. So one of the major challenges for antisense RNA approaches is the stabilization of RNA oligonucleotides. Chemi cal modifications of the bases and/or and phospho sugar backbone have been developed to increase resistance against RNase (Figure 1.22 [258]. The major repr esentative of in the first generation modification is the Phosphorothioate (PS) oligonuc leotides, in which one of the nonbridge oxygen atoms in the phosphodiester bond is replaced by sulfur [258]. The shortcomings include binding to certain proteins, such as hepa rin-binding proteins, and their slightly reduced affinity to the complementary RNA sequences [262]. In the second generation, most the emphasis was placed on the 2 hydroxyl group. 2-O-m ethyl and 2-O-methoxyl-ethyl RNA are the most common types of modifications [258] However, RNase H cleavage can be somewhat reduced or even blocked with thes e types of modifications, possibly due to the steric blockade. One way to overcome this disadvantage is the gapmer technology [258], in which the 2-modified nucleotides are placed only at the ends of antisense oligonucleotides. This protects the ends from degradation a nd a contiguous stretch of at least four or five non2-modified residues in the center ar e sufficient for th e activation of RNase H. A variety of modified nucleotides ha ve been developed in the third generation, the antisense oligonucleotides properties such as target affinity, nuc lease resistance and pharmacokinetics have been improved [258]. Th e concept of conformational restriction has been used widely to help enha nce binding affinity and biostability.

PAGE 78

62 Ribozymes Ribozymes, or RNA enzymes, are cataly tic molecules that can catalyze the hydrolysis and phosphoryl exchange at the pho sphodiester linkages within RNA resulting in cleavage of the RNA strand. There are tw o types of chemical reactions that are catalyzed during phosphate-group transfer by naturally occurring ribozymes: selfcleaving and self-splicing reactions. The riboz ymes that perform self-cleaving reactions include hammerhead, hairpin, he patitis delta virus (HDV) and Neurospora Varkud satellite (VS) ribozymes. They are usually sma ll RNAs of tens of nucleotides in length. The ribozymes that perform self-splicing re actions include self-splicing introns and RNase P. They are much larger in size a nd usually hundreds of nucleotides in length. Figure 1.23. Self-cleaving a nd self-splicing reactions in ribozymes [263].

PAGE 79

63 As shown in Figure 1.23 [263], in the self-c leaving reactions, the RNAs catalyze a reversible phosphodiester-cl eavage reaction. The nucleophilic attack from the 2hydroxyl group results in 5-hydroxyl and 2-3 -cyclic phosphate termini. The bridging 5-oxygen is the leaving group. While in th e self-splicing reactions, an exogenous nucleophile attacks on the phosphorus gene rates a 5-phosphate and a 3-hydroxyl termini. The bridging 3-oxygen is the leavi ng group. In the first st eps of group I intron and group II intron self splicing and the RNas e P-mediated cleavage of precursor of tRNAs, the exogenous nucleophiles are, respectively, the 3-hydroxyl group of exogenous guanosine, the 2-hydroxyl group of an adenosine in the intron, and the water. They are indicated by the ROH in Figure 1.23 b. The transition states are shown in brackets. Self Splicing Introns Self splicing introns can be divided into 2 classes based on the conserved secondary structure and splicing mechanisms : Group I and Group II. Group I is found in a variety of species, including prokaryotes and lower eukaryotes. Except for the Tetrahymena large rRNA group I intron, all other know n group I introns require a single protein co-factor to provide a scaffold to hold the RNA in the catalytic reaction [264]. Group II introns are found within nuclear pre-mRNA and organell e pre-mRNA [265]. A spliceosome consisting of proteins and sma ll nuclear RNAs (SnRNA) is formed in the catalytic reaction and high concentrations of magnesium and potassium are necessary [265]. The splicing action of both group I and gr oup II introns consists of two similar consecutive transphosphoesterific ation reactions. In the first step, the 5-end of the intron is attacked by an exogenous nucleophile which is the 3-hydroxyl group of exogenous

PAGE 80

64 guanosine in group I introns, or the 2-hydroxyl group of an adenosine in group II introns. This results in the cleavage at that site and the additio n of the guanosine or adenosine to the 5-end of the intron. In th e second step, the oxygen in the 3-hydroxyl group of the 3-end of the up stream exon attacks the 3-end of the intron. In group I introns, it is a guanosine at the 3 -end of the intron that is attacked. This cleaves the 3end of the intron, releasing the intron, and results in ligation of the upstream and downstream exons. Figure 1.24 [263] shows th e secondary structure and self splicing steps of group I introns. Figure 1.24. Secondary structure and self sp licing steps in group I intron [263].

PAGE 81

65 RNase P RNase P is a ribonucleoprotein complex that removes the 5 lead er sequence form precursor tRNAs (ptRNAs) via a hydrolysis reaction. It consists of a catalytic RNA subunit (M1 RNA in E. coli ) and a protein subunit (C5 protein in E. coli ) [266, 267]. In vitro M1 RNA can cleave its ptRNA substrat e without C5 protei n, but the reaction requires high concentrations of Mg2+. However, C5 protein can dramatically increase the rate the cleavage, even at low concentration of Mg2+ [268]. In vivo C5 protein is required for RNase P activity and cell viability [ 266, 267]. Thus both the RNA subunit and the protein subunit are essential for RNAse P func tion. It has been proposed that C5 protein can facilitate the stabilizati on of the M1 RNA conformation and also enhance the enzyme and substrate interaction [269, 270]. Figure 1.25. Secondary structures of natural a nd synthetic substrates for RNAse P[275]. All the natural substrates of RNAse P (ptRNAs, precursor of 3.5S RNA and several small RNAs [271-273] in E. coli ) have a common feature in their secondary structure which includes a 5 leader sequen ce, and acceptor-stem-like structure and a 3CCA sequence. A synthetic external guide sequence (E GS) combined with a CCA sequence has been designed to base pair with a targeted sequence to form a structure very

PAGE 82

66 similar to the natural substrates of RNase P. The M1 RNA from E. coli can cleave at this synthetic target site [274]. This EGS-based technology can be used to guide RNAse P to cleave a targeted sequ ence. Figure 1.25 shows the secondary structures of ptRNA and the 3.8s RNA and the hybridization of the EG S with the target ed sequence [275]. Hammerhead Ribozymes The hammerhead ribozyme was the first small self-cleaving RNA to be discovered [276, 277], the first ribozyme to be crystallized [278, 279] and the smallest naturally occurring catalytic R NA identified so far. It was found in several plant virus satellite RNAs and is required for the rolling circle mechanis m of virus replication [280]. The hammerhead ribozyme cleaves the multimeric transcripts of the circular RNA genome into single genome length strands. Hammerhead ribozymes are approximately 30-90 bases in length and cleave RNA targets in trans Annealing of the hammerhead ri bozyme with the target sequence produces a structure consisting of three stems, a tetra-loop and a conserved catalytic core as shown in the Figure 1.26. Any mutation in the catalytic core will prevent catalytic cleavage. The catalytic core has two func tions: it destabilizes the substrate strand by twisting it into a cleavable confirmation, and also binds the metal cofactor (Mg2+) needed for catalysis [278]. The absolute requiremen t of the target sequence is a NUX cleavage site, where N is any nucleotide and X is any nucleotide except G. The targeting arms of the hammerhead ribozyme bind either side of the U of the NUX site forming stems I and III. GUC has been shown to be the most effi cient cleavage site [281], followed by CUC, UUC and AUC. The advantages of hammerhead ribozymes include its small size, easy of cloning and packaging into viral delivery systems, and versatili ty in target site selection.

PAGE 83

67 In the traditional view, the Mg2+ and water are both required in the transesterification reaction. The hydrated magnesium ion can help to provide an environment to facilitate the nuc leophilic attack, in which the Mg2+ acts as a Lewis acid to coordinate directly with the 2-hydroxyl and the 5-leaving oxygen for activation of the nucleophile and for stabilization of the environment. It has be en also reported that some monovalent cations (Li+ and NH4 +) at higher concentrati on can substitute for Mg2+ [282]. There is another kind of antisense agent called DNA enzyme, which is similar to the hammerhead ribozyme in structure and f unction but avoids the high susceptibility to nucleases that is common to ribozymes. The best studied DNA enzyme, named -23 [283], consists of a catalytic core of 15 nuc leotide and two substrate recognition arms. It is highly sequence-specific and can cleav e any junction between a purine and a pyrimidine, and its efficiency is si milar to hammerhead ribozymes [283]. Figure 1.26. Structure of the hammerhead ribozyme. Hairpin Ribozymes Similar to the hammerhead ribozyme, th e hairpin ribozymes was first derived from tobacco ring spot virus satellite RNA [284].When the hairpin ribozyme binds to the

PAGE 84

68 substrate, a structure with 4 helices and 2 l oops is formed. Helix 1 (6 base pairs) and helix 4 (4 base pairs) are wh ere the hairpin hybridizes to th e target RNA. In loop A, a BNGUC target sequence is required for cleavage, where B is G, C or U, and N is any nucleotide [285]. Figure 1.27 shows the structur e of the binding complex of the hairpin ribozyme and its substrate. Figure 1.27. Structure of the hairpin ribozyme. Hairpin and hammerhead ribozymes can al so catalyze the ligation of the cleaved products, which is the reverse of the cleavag e reaction. The ligation efficiency is much higher for the hairpin than the hammerhead. Another unique feature for the hairpin ribozyme is that it does not require metal ions as cofactors [282, 286].

PAGE 85

69 Hepatitis Delta Virus (HDV) Ribozymes and Neurospora Varkud Satullite (VS) Ribozymes HDV ribozymes and VS ribozymes also cl eave the substrates via self-cleaving reactions. HDV ribozymes are derived from th e genomic and the anti-genomic RNAs of HDV [287, 288]. Naturally, the HDV ribozyme cl eaves its substrat e during the rolling circle replication mechanism of the circ ular RNA genome, like other self cleaving ribozymes. The VS ribozyme was originated from the mitochondria of certain isolates of Neurospora [289]. The self cleaving reactions require a divale nt cation but it has also been shown that monovalent cations are be sufficient for the ribozyme to catalyze proficiently [282]. RNA Interference RNA interference (RNAi) is a naturally occurring proce ss and is a potent sequence-specific mechanism for post-transcri ptional gene silencing (PTGS). It was described early in C. elegans [290] and th en found to exist throughout nature as an evolutionarily conserved mechanism in euka ryotic cells. RNAi has regulatory roles in gene expression, such as genomic imprinting, translation regulation, alternative splicing, X-chromosome inactivation and RNA editing [2 91]. In plants and lower organisms RNAi also protects the genome from viruses and insertion of rogue genetic elements, like transposons [292]. Figure 1.28 [293] shows the RNA interf erence pathways. Long double-stranded (ds)RNA is cleaved by Dicer, an RNase III fam ily member, into short interfering RNAs (siRNAs) in an ATP-dependent reaction. Th ese siRNAs contain an approximately 22nucleotide (nt) duplexed region and 2-nt unpaired and unphosphorylated 3-ends. The 5end is phosphorylated, which is a crucial requirement for furthe r reactions. In fact, if the

PAGE 86

70 Figure 1.28. RNA interference [293]. siRNA is introduced into human cells as a synthetic molecule, its 5 hydroxyl gets phosphorylated shortly after entry into th e cells [294-296]. These siRNAs are then incorporated into the RNA-inducing silenc ing complex (RISC). Although the uptake of siRNAs by RISC is independent of ATP, the unwinding of the siRNA duplex requires ATP. The unwinding favors the terminus with the lower melting temperature as the start point. Thus the termini containing more A-U base pairs are prefe rred as the unwinding start point. The strand whose 5-end is at the start point will be used by RISC as the guide sequence and the other strand is release and degraded. Once unwound, the guide strand positions the RISC/siRNA complex with the mRNA that has a complementary sequence to the siRNA, and the endonucleolytic cleavag e of the target mRNA occurs. The target

PAGE 87

71 mRNA is cleaved at the single site in the center of the duplex re gion between the guide siRNA and the target mRNA. The microR NA (miRNA) pathway is another RNA silencing pathway and is similar to siRNA. miRNA is also approxi mately 22-nt long, but it is a product of a sequen tial processing on a single-stranded RNA by two enzymes of the RNaseIII superfamily [297-299]. The long pr imary transcript (pri-miRNA) is cleaved by a nuclear enzyme, named Drosha in huma n, into an approximately 70-nt long premiRNA. The pri-miRNA is basically a s hort hairpin RNA (shRNA) and is further processed in the cytoplasm by Dicer to pr oduce the final miRNA. For both siRNAs and miRNAs, the perfect or near-perfect match will lead to the degradation of the target upon association with RICS, and mism atches will repress the translat ion. It now appears that at least seven continuous complementary base pairs are required for cleavage [300]. Introduction of synthetic siRNAs as a mimic for the Dicer cleavage process triggers the RNAi machinery. In addition, siRNAs or miRNAs produced form shRNA expression cassettes can be cloned into RNA expression v ectors to produce the selfcomplementary hairpin sequences that induc e the RNAi pathway. More importantly, the shRNA expressed from a vector could establis h long-term silencing of a targeted gene expression. The transcription of shRNA from the vector is usually conducted using an RNA polIII promoter such as the H1 or U6 promoter [301, 302]. U6 promoter strongly favors a G residue at the first position of th e transcribed sequence and H1 weakly prefers an A residue [303]. The transcription mediat ed by polIII promoters terminates after the second or third (less commonly) residue of a TTTTT stretch, which results in a 3-UU tail that forms the 3-2-nt unpaired overha ng end in the hairpin structure after self complementarily annealing of the transcript. Both the preference of first residue and the

PAGE 88

72 3-2-nt unpaired UU end influen ce the target site selection. Similar to miRNAs, shRNAs (in the nuclei) are bound by a complex consisti ng of the nuclear expor t factor Exportin 5 (Exp5) and the GTP-bound form of the cofactor Ran [304, 305]. For nuc lear export, this complex requires an RNA stem of 16 bp, a short 3-overhang and a terminal loop of 6 nt [304, 306]. The efficient cleav age by Dicer requires an RNA of 19 bp and a short 3-overhang [307]. These prerequisites can be easily met when designing the shRNA expression cassette. Considering the strand preference of RI SC during unwinding, the 3 end of the guide strand in shRNA is designed tightly ba se-pair (higher CG cont ents) and the 5 end of the guide strand is designe d loose base-pair (higher AU contents). As an example shown in Figure 1.29 [303], two GC base pair s at the 3-end of gui de strand (red) are designed. More AU pairs at the 5end of the guide strand would be appreciated for correct unwinding and even a mismatch can be included. Actually bu lges resulting from mismatches are always present in natural primiRNAs and they may help to fine-tune the cleavage sites used by Drosha and Dicer and/or may preclude activation of dsRNAresponsive cellular signaling pa thway like interferon responses [303]. During the design of the shRNA, it is encouraged to include a bulge close to the 5 end of the guide strand, which should be done by a introducing a mutati on into the to-be-degraded (sense) strand, not into the guide (antisense) strand. It has also been repo rted that an A residue at position 3 and a U at position 10 of the se nse strand can enhance siRNA function significantly. And a G at position 13 of the se nse strand may need to be avoided [308]. Figure 1.29 [303] shows the sequence of the designed shRNA, with the reference to

PAGE 89

73 human pre-miR-1 sequence and structure. The blue strand is sense and the red strand is antisense. Arrows mark the Dicer cleavage sites. Figure 1.29. Designing artificia l shRNA for RNAi [303]. It has also been found that small dsR NA that are 25-30 nt in length requiring RNAi processing appear to be more effici ent in inducing RNAi than smaller 22 nt siRNAs [309], which could be due to the fact that Dicer may direct endogenously processed siRNAs and miRNAs to the RI SC complex. This gives vector-expressed shRNA an extra advantage over synthetic 22 nt siRNAs. Multiple shRNAs or siRNAs can be introduced into the cell simultaneousl y, but it is worth keeping in mind that the RNAi machinery can be limiting [310] so that the competence between exogenous shRNAs and endogenous miRNAs, or betw een exogenous and endogenous siRNAs, for limited amount of Dicer and RISC could occu r, which would interfere with the cells endogenous RNAi pathways. Particularly when the cell is undergoi ng a cell division, the RNAi machinery could be diluted and adve rsely affected by inhi biting the gene knockdown mechanism.

PAGE 90

74 RNAi is highly specific to its target and not toxic in almost all situations; however, when designing an siRNA or shR NA, some off-target effects should be considered and avoided. dsRNAs that are 30 nu cleotides or longer tend to trigger at least two cellular stress response pathways, both of which will lead to a general and nonspecific abrogation of protein synthesis, or even apoptosis [295, 311]. The IFN pathway is usually a mechanism to eliminate virus-in fected cells, in which the long dsRNA binds to and activates the dsRNA-activated pr otein kinase (PKR). PKR can further phosphorylate the translation initiation factor, eIF-2 and induce global translation inhibition and even apoptosis. In another path way, dsRNA activates 2-5 oligoadenylate synthetase. The 2-5 oligoadenylate will then be formed and bond to and activate RNase I, resulting in non-specific degradation of RNAs. Although siRNA or shRNA, less than 30 nucleotides in length, usually do not activat e these stress response pathways. In highly sensitive cell lines and at high concentra tions, a subset of in terferon genes can be activated [312-314]. In the design ing of siRNAs or shRNA, the ones that have significant homology to other irrelevant mRNAs shoul d be avoided. As noted before, a seven consecutive base pairing can be enough to activate the RISC-induced gene silencing. Even the guide strand (antisense) has been desi gned to introduce RISC to the target site after unwinding, it is still possi ble that unwinding could initiat e from the 5 end of the sense strand and thus sense strand would gui de the RISC. The homology of the sense strands should also be checked. Vector-mediated expression of shRNA can lead to long-term RNAi and the silencing effect has been observed even af ter two months [302]. The half-life of unmodified siRNAs in vivo is only seconds to minutes [315]. The most important reason

PAGE 91

75 for this short half-life is the rapid elimina tion by kidney filtration due to the small size (~7 kDa). Endogenous serum RNases can degr ade the siRNAs limiting the serum halflife to 5-60 minutes. The half-l ife can be extended in a num ber of ways, for instance, complexing the siRNAs with other molecules or incorporating them into various types of particles to bypass renal filtr ation [315-317], chemically modifying the ribose [316, 318320], or capping the ends of the siRNA [315, 320]. The modification on the ribose usually takes place at the 2-position; 2 -deoxyribose, 2-O-metheyl and 2-fluoro substitutions/modifications have been re ported [316, 318-320]. Usually the silencing effects are affected more or less by these m odifications but a modifi ed siRNA, with two 2-O-methyl at the 5 end and four methyl ated monomers at the 3 end, has been demonstrated to be as active as its unm odified counterpart [321]. Even though siRNAs have the potential to activate interfer on pathways, no toxic effects after siRNA application have been observed [258]. There is no strict specific se quence requirement in RNA interference (although there are preferred bases at some positions), and, therefore, the range of target for siRNA is greater th at with ribozymes or antisense therapies. Gene Therapy Overview With the progress of Human Genome Proj ect, people are reaching a new level of understanding of many biological events, including the etio logy of diseases with or without proved treatment. Especially for t hose diseases currently without treatment, finding the genes that are invol ved in the initiation and de velopment of the diseases provides new treatment targets. The most common gene therapy targets are monogenic recessively inherited diseases such as hemophilia [322]. In the tr eatment of these diseases, gene therapy is designed to introduce a functional gene into a target cell to restore protein production that

PAGE 92

76 is absent or deficient due to the genetic disorder. Conversely in monogenic dominantly inherited diseases like hypercholesteroleam ia [323], successful treatment requires the aberrant gene to be silenced, and this is usually done by means of gene-silencing technologies. Cancer, as an acquired genetic disease, is also a good candidate for gene therapy. Apart from expressing functional tumo r suppressor genes and silencing activated oncogenes, gene therapy in cancer treatment has also been applied to introduce the expression of immunopotentiation proteins, the expression of a toxic product in transformed cells, and the expression of protei ns in healthy cells he lping the cell to be resistant to higher doses of chemotherapy [324]. The methods to deliver a gene into cells can be roughly categorized into virusbased system and non-viral system. Non-Viral Gene Delivery The gene transfer in non-vira l system is in general ine fficient and often transient compared with viral vectors, but it has advant ages such as low toxi city, simplicity of use and ease of large-scale production. In addition, the transient expression of a therapeutic gene would be desirable in the treatment of certain conditions, such as retinopathy of prematurity. Basically there are three categor ies of methods for non-viral gene delivery: naked DNA in the form of plasmid, liposom al packaging of the DNA and molecular conjugates. Naked DNA is the simplest way to delivery a gene. It is not very efficient and can result in prolonged low levels of expression. Th e simplest way is to inject directly into the tissue of interest or inject systemically from a vessel. The expression level and area are usually limited in a systemic injection due to the rapid degrad ation by nuclease and clearance by mononuclear phagocyte system. To facilitate the uptake of naked DNA,

PAGE 93

77 several techniques, in addition to simple in jection, have been developed. The Gene Gun is a technology to shoot gold particles coated with DNA whic h allows direct penetration through the cell membrane into the cytopl asm and even the nucleus, bypassing the endosomal compartment [325]. Electroporation, the application of c ontrolled electric fields to facilitate cell permeabilization, is another way to facilitate DNA uptake. Skin and muscle are ideal targets due to the ease of administration. Ultrasound can also increase the permeability of cell membrane to macromolecules like plasmid DNA and has been used to facilitate the gene transfer. Liposomes are lipid bilayers entrappi ng a DNA fragment with a fraction of aqueous fluid. It can naturally merge onto cel l membrane and initiate the endocytosis process. To improve transfection efficiency, target proteins rec ognized by cell surface receptors have been included in liposome to facilitate uptake, for example, anti-MHC antibody [326], transferrin [327], and Sendai virus of its F protein [328], which help DNA to escape from endosome into cytoplasm t hus to increase DNA transportation to the nucleus. The inclusion of a DNA binding protein on the liposome also enhances transcription by bringing the plasmid DNA into the nucleus [328]. Molecular conjugates are us ually a synthetic agent that can bind to DNA and a ligand at the same time [324]. Thus the binding of the ligand to its receptor will initiate the receptor-mediated endocytosis for the complex. This method is more specific for different cell types and receptor types. Th e synthetic agent needs to be designed accordingly, but this is useful in tissue-spec ific transfection. The transgene expression in this method tends to be transient and lim ited by endosomal and lysosomal degeneration.

PAGE 94

78 Viral Gene Delivery Viral gene delivery systems are based on replicating viruses that can deliver genetic information into the host cell. According to the existence status of the viruses, the virus vectors can be divided into two categ ories: integrating and non-integrating [329]. Integrating virus includ e adeno-associated virus, retrovi rus, and so on. These viruses can integrate the viral genome into chromosoma l DNA so that a life-long expression of transgene could be possibly achieved. Adenovi rus and herpes simplex virus fall into the category of non-integrating viruses. They deliver viral genome in to the nucleus of targeted cell, however the vi ral genome remain episomal, so it is possible that the transgene gets dilute d during cell divisions. Generally speaking, genomes of replica ting viruses contain coding regions and cis -acting regulatory elements. The coding sequ ences enclose the genetic information of the viral structural and regul atory proteins and are requi red for propagation, whereas cis acting sequences are essential for packaging of viral genomes and integration into the host cell. To generate a replication-defective viral vector, the coding regions of the virus are replaced by a transgene, leaving the cis -acting sequences intact. When a helper plasmid or virus providing th e structural viral proteins in trans is introduced into the producer cell, production of non-replicating vi rus particles containi ng the transgene is established. An ideal viral vector should have these characteristics: 1) The virus genome is relatively simple and easy to manipulat e; 2) The viral transduction can yield high vector concentration in the producer cells (>108 particles /ml); 3) The vector should have no limitation in size capacity; 4) The viral vector can tran sduce dividing and non-dividing cells; 5) The vector can deliver the transgene as integration in the host cell genome or as segregation being an episome along with cell division so that sustained expression can be

PAGE 95

79 established; 6) The vector has a nave or modified tissue specificity and the transgene expression can be regulated; 7) The vect or produces no or low immune response and allows subsequent re-administration. [330] The expression specificity can be regulate d in many aspects. For tissue specificity, we can pick the virus that has the right trop ism specific to some tissue, and in addition tissue-specific promoters can be added to further define the sp ecificity. For spatial specificity, radiation in conjuga tion with radiation-activated pr omoter (for example, ergl promoter [331]) would be a good method. Of course the local delivery into the right place is always preferred than systemic administration, if feasible. For temporal specificity, drug-inducible promoters can provide a convenient way to switch the transgene expression on and off. The drug can be used to work on transcription activation or repressor elements to modulate the expr ession. There are many established drugregulated gene expression systems, such as rapamycin-regulated gene expression [332] and RU486-regulated gene expression from GAL4 site [333]. And for promoters containing binding site for hormone receptor, he avy metals or cytokines, these specific hormone, heavy metals and cytokines can al so be used to induce the expression. Adeno-Associated Viral (AAV) Vectors AAV is currently the virus closest to an ideal vector that is under study and application. It belongs to th e family of parvovirus; it is non-pathogenic and depends on helper virus (usually adenovirus (Ad) or herpes virus) to prol iferate. It is a non-enveloped particle with a size of 20 -25 nm and has a vector capacity of 4.7 kb [334]. AAV can infect both dividing and non-divi ding cells, with the transduc tion efficiency best in Sphase of host cell cycle. The viral genome coded in a single-stranded DNA, has two open-reading frames (ORF). One is rep which is responsible for viral structural proteins,

PAGE 96

80 integration and replicatio n proteins. The other is cap coding for capsid proteins. There are inverted terminal repeats (ITR) at both ends of the genome sized around 150 bp, Tshaped and forming palindromic structure. TR is GC rich and contai ns a promoter. Due to the integration into the host genome, AAV vector can potentially deliver a long term expression of the transgene. Another advantag e of AAV is that it induces overall low immune response. Presence of circulating neut ralizing antibodies is in the majority of populations, but they dont prevent re-admin istration or shut down promoter activity [329]. Small packaging capacity is the number one disadvant age of AAV vectors. Using concatamers, formed by head-to-tail recombin ation in ITRs, up to 10 kb of transgenes can be packaged for delive ry [335], by means of splitti ng promoter and transgenes sequences over two AAV vectors. But this t echnology reduces transduction efficiencies. The infection of a host cell starts when th e viral particle binds to its receptor on the cell membrane and initiates the endocytic pathway. The receptor type varies with AAV serotypes. The AAV-2 serotype, the most studied and commonly used serotype, has as its primary receptor heparin sulfate pr oteoglycans (HSPG) [336] HSPG is widely expressed in various tissues a nd this is why AAVs have a wi de tropism. There are also co-receptors for AAV-2 to faci litate endocytosis. Fibroblas t growth factor receptor-1 (FGFR1), one of the co-receptors, can enhance the virus attachment to the cells [337]. Integrin v 3, another co-receptor, can facilitate endocytosis in the clathrin-mediated process, and it may also activate Rac1 a nd further phosphorylate PIP3 Kinase [338], which leads to microfilaments and mi crotubes rearrangement to support AAV2 trafficking to the nucleus. After entering the cell, the viral particle is released from the endosome at low pH conditions. Low pH probably induces a conformational change of

PAGE 97

81 viral proteins and thus helps with endosom e release and nuclear entry [339]. The viral particle is uncoated in the nucleus, a nd ssDNA is duplicated into dsDNA by either annealing with a complementary DNA strand from a second AAV or by the host cell machinery. The duplication from ssDNA to dsDNA is the rate-limiting step in AAV transduction. To overcome this, self comple mentary vector (scAAV) has been designed to expedite this process [340]. With the help of rep proteins, the viral genome or the transgene is integrated to a specific site in chromosome 19 via a non-homologous recombination and will be expressed by host cell transcriptional machinery. Some virus may remain episomal and also get expresse d. Figure 1.30 [330] summarizes major steps in the AAV internalization a nd intracellular trafficking. Figure 1.30. AAV internalization and intr acellular trafficking [330].

PAGE 98

82 AAV has a number of serotypes. AAV-1 a nd AAV-4 were isolated from simian sources; AAV-2, -3, -5 were isolated from hu man clinical specimens; AAV-6 is thought to be the recombination of AAV-1 and AAV-2 (AAV-1s 3 end recombined with AAV2s 5 end), and AAV-7 and AAV8 were isolated from rhes us monkey [330]. They have their own tropisms, which are determined by the capsid proteins. For example, AAV-2 is preferred to use for infection of the huma n eye, spine, while AAV-1 has the highest transduction efficiency in muscle and liver and AAV-5 has high tropism for retina and is able to transduce airway epithelial cells. Among all the serotypes, AAV-2 is the most studied and commonly used. As with all the AAV serotypes, the AAV-2 genome has two ORFs, rep and cap which span over 90% of the genome. As shown in Figur e 1.31 [330] Panel A, in the ORF of rep there are two promoters, p5 and p19, encoding four proteins. Rep 78 and its splicing variant, Rep68 are transcribed from p5. They play importa nt roles in replica tion, transcriptional control and site-specific inte gration. Rep52 and its splicing va riant, Rep40 are transcribed form p19. They are important for the accumula tion of single-stranded genome used for packaging. The other ORF cap encodes for VP1, VP2 and VP3 which are transcribed from p40. They are capsid proteins and have pivotal roles in tropi sm specificity. These three proteins are expressed in the ratio of 1:1:20, making the capsid with icosahedral symmetry. The ITRs at both ends of the vi ral genome have a couple of functions. The detailed structure and sequence of ITR is s hown in Figure 1.31 [330] Panel C. First, the 3 end of the ITR on the 5 end the genome se rves as primer in the synthesis of a new DNA strand. Second, ITRs contain Rep bindi ng site (RBS) for Rep78 and Rep68 and

PAGE 99

83 help them work as a helicase and an endonuc lease. Third, the terminal resolution site (TRS) is identical to a sequence in chromosome 19, serving as integration sequence [341]. Figure 1.31. AAV2 genome and the vector genome [330]. When making an AAV viral vector, the two ORFs and the viral promoter are all replaced by a transgene and the only cis elements needed for AAV integration, packaging and assembly are the ITRs. The vector genome is shown in Figure 1.31 [330] Panel B.

PAGE 100

84 rep and cap will be provided in trans in another plasmid, and helper virus gene products (E1a, E1b, E2a, E4 and VA RNA from Ad) are also provided in trans Originally the vector production method is to co-transfect the HeLa cells with transgene plasmid, the plasmid providing rep and cap and wide type Ad, or to co-transfect human 293 cells with the transgene plasmid, rep and cap plasmid, and E1-deleted Ad, as the E1gene products can be provided endogenously in 293 ce lls. Recently helper virus-free system has been designed to minimize the safety issues. See Figure 1.32 [334]. Figure 1.32. Helper virus free sy stems in rAAV production [334]. The helper virus-free system has the three-plasmid system and the two-plasmid system [334]. In the three-plasmid system besides AAV vector plasmid and AAV helper plasmid providing rep and cap genes, an Ad helper plasmi d is introduced to provide the helper virus gene products (E2A, E4 a nd VA RNA from Ad) and human 293 cells are used as the host cell to provide Ad E1 gene products. The best molar ratio for these three

PAGE 101

85 plasmids is 1:1:1, or 1:1:3 in mass [342] The two-plasmid system combines the AAV helper plasmid and the Ad helper plasmid into one. The most recent version (called pDF) has 6 different helper plasmids, common in rep gene (AAV-2 rep ) but varying in cap genes (AAV-1 to AAV-6 cap ) [330]. Whats more, different fluorescence protein gene is incorporated in the different plasmids. Because there are 5 kinds of most frequently used fluorescence protein genes, the plasmid having AAV-1 cap and the plasmid having AAV-6 cap use the same fluorescence protein gene cyan fluorescence protein (CFP). The other four plasmids encode green fl uorescence protein (GFP), yellow fluorescence protein (YFP), blue fluores cence protein (BFP), red fluores cence protein (RFP). This two-plasmid system is called Helper virus fr ee, Optically controllable, Two-plasmid-base, or HOT [343]. This system not only minimizes safety issues and simplifies operation, but also adds controllable tropism Figure 1.33 [330] summarizes the rep and cap genes, fluorescence and preferred tropism fo r these 6 pDF helper plasmids. Figure 1.33. The 6 pDF helper plasmids in the two-plasmid system [330]. After the production of the AAV vectors, the vector purification can be accomplished in several ways, such as CsCl gradient ultracentrifugation, iodixanol

PAGE 102

86 discontinuous gradient ultracentrifugation, heparin affinity column (for AAV-2), and HPLC [330]. The combined use of heparin affinity column and HPLC for AAV-2 can result in more than 50% recovery and more than 99% purity [344, 345]. Beyond a vectors own tropism, if we want to regulate the transgene expression more specifically or more precisely, some modification on AAV vector can be done. A linker molecule between the viru s and the target cell can be in corporated as an indirect modification. The linker molecule can be bispecific antibody or streptarvidin. One example is the antibody F(ab )2 which is used to help AAV-2 capsid to target 2 3 integrin [346]. A direct modi fication would be the modifi cation on capsid proteins, for instance, the serpin receptor ligand has been incorporated into AAV-2 capsid gene [347]. Besides the modification on external su rface of the vectors, tissue specific promoters/enhancers can also be used to regu late the transgene expression is specific tissue. Adenovirus (Ad) Vectors Adenoviruses are non-enveloped, dsDNA viruse s. Ad vectors can capacitate up to 30 kb of transgene [324], which is larger than AAV vectors. Ad is non-integrating virus and remains as an episomal element in the nuc leus. Ad is very efficient at transducing, in vivo and in vitro dividing and non-dividing cells and can produce very high titers (>1011/ml). Ad vector administration usually induces strong immune response. After intravenous injection, 90% of the vector is degraded in the liver and the remaining viruses have their promoter inactivated [348]. The persisting antibody prevents subsequent administration. Transit immunosuppre ssive therapies may be needed with Ad vector administration.

PAGE 103

87 The Ad genome has four early transcrip tional units (E1, E2, E3, E4), which have regulatory functions, and one la te transcript, which encodes the structural proteins. The gutless vector contains the inverted term inal repeats (ITR), the packaging sequences around the transgene, and additional stuffe r DNA to maintain the optimum package size. The necessary viral genes are provided in trans by helper virus. Figure 1.34 [324] shows a simple structure of the Ad genome, vector genome and helper virus. Figure 1.34. Ad genome and the vector genome [324]. During transduction, the fiber protein in Ad binds to cellular receptors, which are usually MHC class I molecules [349] and coxsackievirus-adenovirus receptors (CAR) [350]. Next the penton base protein on Ad binds to the co-receptor, v 3 and v 5 integrins, and internalize via clathrin-mediate d endocytosis. After tran sport to the nucleus, the transcription of early genes is initiated and interferes with the antiviral defense of host cells. DNA replication is initiate d by E2 products. In the late phase, structural proteins are highly expressed and vi rus assembly starts. Simila r to AAV, specific cellular promoters/enhancers are used to direct tissue specific transgene expression.

PAGE 104

88 Retrovirus Vectors Retroviruses are enveloped, ssRNA viru ses. ssRNA needs to be reversely transcribed into dsDNA and then the viral genom e can integrate into th e host gene but the virus can only target dividing cells. The vector capacity is around 7.5 kb [324]. The vector is easily inactivated by c1 complement protein and antigalaetosyl epitope antibody, both of them are pr esent in human sera [351, 352] The biggest disadvantage for retrovirus vectors is that insertiona l mutagenesis can possibly occur, because retrovirus (with its own proto-oncogene remove d) can transform cells by integrating near a cellular protoncogene and dr ive inappropriate expression from its 3 long terminal repeats (LTR); or disrupt a tumor suppressor gene. There are three categories of retrovirus: onc oretrovirus, lentivirus and spumavirus [329]. Oncoretrovirus is the simplest in st ructure and is the most commonly used. The oncorectrovirus genome contains three genes: gag encoding the core proteins, pol encoding the reverse transcriptase, env encoding the envelope pr oteins and determining tropism. There are LTR at both ends of the ge nome. The LTR is comprised of 3 regions, which are U3, R and U5. The LTR is essentia l for reverse transcri ption, integration and transcriptional activation as it contains a viral promoter/enhancer. located between the 5 LTR and the viral genes. This sequence re quired for packaging. In retrovirus vectors, LTRs and are retained and the viral genome is replaced by a transgene. Transgene expression can be driven by the viral prom oter/enhance in the 5 LTR or by other exogenous promoters. Figure 1.35 [329] shows the genomic struct ure of MLV DNA, which is the most frequently used vector in oncoretravirus.

PAGE 105

89 Figure 1.35. MLV genome structure [329]. Herpes Simplex Virus Ty pe 1 (HSV-1) Vectors HSV-1 are dsDNA, neurotropic viruses, a nd are good for neural gene transfers. The capacity of HSV-1 is about 40-50 kb [324] HSV-1 has two life cycles. In the lytic life cycle, the viral genes get expressed s hortly after transducti on and new viruses are packed and released. In latent life cycle, th e virus remains an intranuclear episome and the infected cell functions normally. Structurally, beside s envelope, capsid and viral genome, the HSV-1 contains tegument, which is a protein layer be tween capsid and the envelope. Tegument is essential for viral in ternalization, resistance from the host cell defense system and transcription activati on [329]. The HSV-1 genome has three classes of genes: immediate-early gene, early genes, and late genes. One kind of plasmid vector derived from HSV-1 is called amplicon. It contains Col E1 ori (an Ecoli origin of replication), Ori S (HSV-1 origin of replication) and HSV1 packing sequence [324]. The transgene is under the control of an immediate-early promoter. The expression is dependent on help er virus or helper plasmid containing the necessary genes from helper virus. The othe r kind of HSV-1 vector is the replicationdeficient HSV-1, which is made by deletion of one immediate-early gene [324]. The deleted immediate-early gene is provided in trans

PAGE 106

90 CHAPTER 2 METHODS AND MATERIALS Hammerhead Ribozyme Target Sites The hammerhead ribozymes designed for th is study were all 34 bases in length. They formed three stem structures when bound to target. The targeting arms of the ribozyme bound to either side of the X of th e target NUX sequence to form stems I and III. The cleavage target was, therefore, 13 base s in length. Stem II was four base pairs in length and formed a stabilizing tetraloop within the folded ribozyme structure. An internal loop, formed with in the ribozyme when it was bound to target, contained the catalytic core of the hammerhead ribozyme. Choosing a target site in the mRNA seque nce was the first step in ribozyme design. Shimayama et al. [281] refined the NUX rule of hammerhead ribozyme target site selection and demonstrated that a GUC site is the most efficient site for cleavage. Then Fritz and colleagues found GUCUU or GUCUA was more efficient cleaved [257]. The initial step in designing a ribozyme wa to s earch for potential target sites within the mRNA sequence of the target gene. The so ftware package Vector NTi (Invitrogen, Carlsbad, CA) can be used for this purpos e. The target mRNA sequence was downloaded into Vector NTi from GeneBank and the sequence was searched for the presence of GUCUU and GUCUA sites. Once potential target si tes were identified the next step was to determine target accessibility.

PAGE 107

91 Accessibility of Target Site All the selected sites were only potential ta rget sites because in the real world the secondary and tertiary stru cture of mRNA would affect the sequence accessibility significantly. To select the most accessibl e sites we used Zukers Mfold program ( http://www.bioinfo.rpi.edu/a pplications/mfold/old/rna/ ) to examine target accessibility. This program was used to predict the sec ondary structure of 200 bases of the target mRNA centered on the NUX target sequence. In most cases the program would generate several possible structures. Si nce the structure of RNA in solution is dynamic, it is possible that the target regi on would exist in a dynamic equilibrium made up of the structures generated by Mfold a nd other structures also. For these studies, target sites which were completely or partially accessible within loop structures or at the end of a stem structure were considered accessible. The accessibility and thermodynamic stability of the ribozyme is also very important. After the most accessible targ et sites were found by using Mfold, the corresponding ribozyme secondary structure was then examined using the same program. Generally ribozymes fold into one the four t ypes of secondary struct ures shown in Figure 2.1 [257]. Type A has its two targeting ar ms completely accessible and forming no internal secondary structure with the rest of the ribozym e. Ribozymes that form only structure A, by Mfold analysis, typically have high catalytic ac tivity. Type B and C structures do have internal secondary stru ctures formed within one or both of the targeting arms but have a higher dG than structure A and, therefore, should have relatively good accessibility to bind to the ta rget and possess relatively high catalytic activity. Ribozymes that form structures like D have lower dG values than structure A

PAGE 108

92 and are more stable. Ribozymes that form st ructure D may have very low or no catalytic activity. Figure 2.1. Typical structures of hammerhead ribozyme predicted by Mfold [257]. Once an accessible target site and ribozyme were identified the next step was to purchase RNA oligonucleotides corresponding to both the target and the ribozyme in order to perform in vitro cleavage analysis to determine the kinetic parameters of the ribozyme. Kinase of Target Oligonucleotides For the in vitro assays ribozymes and target RNA oligonucleotides were purchased from Dharmacon (Boulder, CO). They were deprotected according to manufacturers protocols and suspended to a concentration of 300 pmol/l in TE or water

PAGE 109

93 and stored at -70C. The ta rget RNA oligonucleotide was radioactively labeled with 32P at the 5 end. 2 l of RNA oligo (10 pmol/l, 20 pmole total) was mixed with 1 l 10X polynucleotide kinase buffer (Promega, Madiso n, WI), 1 l RNasin (Promega, Madison, WI), 1 l 0.1M DTT (Sigma, St. L ouis, MO), 3 l water, 1 l [ 32P] (10Ci) (ICN, Santa Clara, CA) and 1 l T4 polynucleotide kina se (5 units) (Promega, Madison, WI). Reactions were incubated at 37C for 30 minutes, 65 l of water was added, and the mixture was extracted with 100 l of pheno l/chloroform/isoamyl alcohol solution. The aqueous layer was added to a pre-packed G-50 fine spin column to separate the labeled target oligonucleotide from th e unincorporated label. Sample s were collected in a 1.5 ml Eppendorf tube (Fisher, Suwanee, GA) and st ored at 4C. Samples are usable for one week but best when used within 24 hours. Time Course of Cleavage Reac tions for Hammerhead Ribozymes 1 l of ribozyme (2 pmole total, dilute d from 300 pmol/l stock) was mixed with 13 l of 400 mM Tris-HCl (Fisher, Suwanee, GA), pH 7.4-7.5, and 88 l of water. The mixture was incubated at 90C for 2 minutes to denature the ribozyme and then held at room temperature for 10 minutes. Next 13 l of 1:10 RNasin:0.1M DTT and 13 l of 200 mM MgCl2 (20 mM final) (Sigma, St. Louis, MO) was added and the mixture was incubated at 37C for 10 minutes. Cleav age was initiated by addition of 2 l of target oligonucleotide (1 l 32P-kinased target plus 1 l of cold target (20 pmole, 150 nM final)). The reaction was inc ubated at 37C, and time point s were taken at 0, 1, 2, 5, 10, 15, 30 and 60 minutes. For each time point 10 l of the reaction was added to 10 l of formamide dye mix (90% formamide (Sigma, St. Louis, MO), 50 mM ethylenediamine tetra acetic acid (EDTA) pH 8 (Fisher, Suwanee, GA), 0.05% bromophenol blue (Sigma,

PAGE 110

94 St. Louis, MO) and 0.05% xylene cyanol (Sigma St. Louis, MO)). Samples were placed on ice then stored at -20C. The samples we re heat denatured at 90C for 5 minutes, cooled on ice, and applied to a 10% PAGE-8 M urea gel (6 l sample loaded each well) to separate the reaction products. The gel wa s held at 33 mA until the bromophenol blue moved 60% of the length of the gel. The ge l was fixed and dried. The gel was analyzed on a Molecular Dynamics PhosphoImager (Ammersham, Sunnyvale, CA). The time point when 15% of the target was cleaved was dete rmined and used as the endpoint for the multiple turnover ki netic reactions. Multiple Turnover Kinetics Reactions were done in a final volume of 20 l. Ribozyme (0.3 pmol/l, 15 nM final) in 40 mM Tris-HCl (pH 7.5) was incuba ted at 65C for 2 minutes and then at 25C for 10 minutes. The reactions were supplem ented with DTT (20 mM final) and MgCl2 (20 mM final) and 4 units of RNasin, incuba ted at 37C for 10 minutes. Adding gradient concentrations of the target oligonucleotide (0-3 00 pmol/l; 0-1500 nM final) initiated the cleavage reactions. The reaction tubes were incubated at 37C for a fixed interval determined in the time course analysis of cleavage. This experiment could also be done with incubation in 1 mM MgCl2 at 25C. The addition of 20 l of formamide stopped buffer terminated the reactions. Samples were in itially held on ice then stored at -20C. Later the samples were heat denatured at 90 C for 5 minutes, placed on ice and cleavage products were separated on 10% polyacrylamide-8 M urea gels The gels were analyzed on a Molecular Dynamics PhosphoImager. Treating ribozymes as classical catalytic enzymes, this experiment was done to determine the kinetic parameters (Vmax, Km, kcat). The addition of gr adient concentrations of the target oligonucleotide initiated the cleavage reaction and high concentrations of

PAGE 111

95 target saturate the ribozymes catalytic capability. This fixed interval for reaction incubation was usually how much time it needs to reach the 15% cleavage of the target, at which the cleavage reaction was linear. Thus the average cleavage velocity determined in this interval could be used as the initial reaction velocity. This velocity was determined using the amount of cleavage product divided by the fixed interval of time. The amount of the target cleavage product was dete rmined by autoradiograph. By plotting the cleavage velocity versus the corresponding ta rget concentration, a saturation curve was generated that is a typical Michaelis-Men ten equation curve for the enzyme kinetics study. A double-reciprocal plot (also called: Lineweaver-Bur ke plot) could be further generated to graphically determine the kine tic parameters. This double-reciprocal plot was linear; the slope equaled Km/Vmax and the intercept on the Y-axis equaled 1/Vmax. Thus Vmax and Km could be determined. kcat was the turnover number and assigned as the value of Vmax divided by enzyme total concentr ation. The enzyme (ribozyme) total concentration was a known value. The two equations are listed below. ] [ ] [max 0S K S V vm Michaelis-Menten Equation; max max 01 ] [ 1 1 V S V K vm Lineweaver-Burke Plot; where v0 is the initial reaction velocity, and [S] is the concentr ation of substrate (oligonucleotide target). Cloning of the Ribozymes into an rAAV Expression Vector The rAAV cloning vector was p21NewHp (Figure 2.2) and was described by Shaw and coworkers [353]. This vector was modified from the pTRUF21 plasmid (obtained from the UF Vector Core) by insert ion of a self-cleavi ng hairpin ribozyme

PAGE 112

96 between SpeI and Nsil sites. These sites we re immediately downstream of the position where the hammerhead ribozyme is located. The hairpin ribozyme cassette included a cleavage site at its 5-end that was recognized by the hairpin ribozyme. During transcription, the cytomegalovirus (CMV)/chicken -actin chimeric enhancer/promoter drived the transcription through the hairpi n ribozyme. Self cleavage by the hairpin ribozyme liberated the 3-end of the hammerhead ribozyme with an additional eight bases at the hammerheads 3-end. Cleavage at this position eliminated downstream sequences that could interfere with the hammerh ead ribozyme annealing to its target. Figure 2.2. The pTRUF21 expressi on and cloning vector and the orientation and position of the hammerhead and ha irpin ribozyme cassette.

PAGE 113

97 For the cloning of one of the hamm erhead ribozymes, two synthetic complimentary DNA oligonucleotides with phos phates at the 5-ends that code for a single hammerhead ribozyme were purchased from Invitrogen (Carlsbad, CA). They were annealed by incubating at 90C for 3 minut es then slow cooled to room temperature for about 40 minutes. The annealed produc t was a double-stranded DNA oligonucleotide with a cut HindIII site at 5 end and a cut SpeI site at 3 end. This double stranded DNA fragment was cloned into the HindI II and SpeI site of p21NewHp. The plasmid was digested by HindIII and SpeI restriction endonucleases (Promega, Madison, WI) according to manufact urers protocols. Then the annealed oligonucleotide product was ligated into the Hi ndIII and SpeI sites in the plasmid using DNA T4 Ligase (Promega, Madison, WI) acco rding to manufacturers protocols. The ligated products were transformed into SURE competent cells (Stratag ene, La Jolla, CA) using electroporation. SURE cells were used in order to maintain the integrity of the inverted terminal repeats (TRs). Screening and Sequencing of the Clones The ligation mixture-transformed SRUE cells were grown in terrific broth (TB, Sigma, St. Louis, MO) supplemented with ampicillin at 37oC for 16 hours or less. The plasmid DNA was purified using Genelute HP Plasmid Maxiprep Kit (Sigma, St. Louis, MO). The purified plasmids were digested with PstI restriction endonuclease (Promega, Madison, WI) according to manufacturers protoc ols to monitor the integrity of TRs. The insertion of the hammerhead ribozyme eliminated a PstI site and loss of this site was used as a diagnostic indicatior of ha mmerhead insertion into the vector. In addition, we also performed SmaI digests on the plasmids to de termine the integrity of the TRs. Plasmids that lacked the PstI site and still retained intact TRs were then sequenced to verify the

PAGE 114

98 presence of the hammerhead ribozyme in sert. The plasmids were sequenced using Ladderman Dideoxy Sequencing Kit (TaK aRa Shuzo Co, Japan) according to manufacturers protocol or sequenced at ICRB Core sequencing facility at UF. HREC Tissue Culture Human eyes were obtained from Nationa l Disease Research Interchange within 36 hours of death. HRECs were prepared and ma intained as previously described [354]. The eyes were placed on a sterile gauze pad (Johnson and Johnson Medical Supplies, Arlington, Texas) in a laminar flow hood and wa shed with 5 ml of betadine (Fisher, Suwanee, GA), and dissected with sterile scalpels (No. 1, Feather Industries Limited, Japan) and tweezers. Neural retina was isolat ed from the posterior portion of both eyes clear of RPE layer. The retina was placed on a 53 micron mesh nylon membrane (Tetko Inc, Lab Pack, Kansas City, MO) to separa te the endothelial cells. Phosphate buffered saline (PBS) mixed with 2% antibiotic/antim ycotic mix (ABAM) (Sigma, St. Louis, MO) was used to wash the retina. While washing, the retina was ground over the nylon membrane by a sterile wooden spatula. Then th e remaining retina was transferred to a 20 ml Erlenmeyer flask containing 10 ml of PB S with antibiotics, using a sterile 10 ml pipette. Approximately 1 mg of collagena se (342 units/mg, Worthington Biomedical Corporation, Lakewood, NJ) was added to the flask. The flask was incubated in a 37 C water bath for 15 minutes, and the contents we re stirred every 5 mi nutes to dissolve the collagenase. Then 20 ml of complete endotheli al cell media (250 ml Dulbelcos Modified Eagle Medium (DMEM) low glucose, 250 ml HAMs F12, 10% fetal bovine serum, 15% endothelial cell growth supplement, 15% insu lin/transferring/selen ium, 2% L-glutamic acid, 2% antibiotic/antimycotic mix) was added to the flask. The cells were washed twice with media and placed into a T25 flask (Fishe r, Springfield, NJ) coat ed with 1% gelatin

PAGE 115

99 (Sigma, St. Louis, MO). For the next 48-72 hour s the cells were kept undisturbed so that they could grow and attach to the flask. Th en the media was changed and fresh antibiotics were added. The cells were passaged upon reach ing confluence. They were washed twice with PBS, then washed with 5 ml of trypsin (Sigma, St. Louis, MO) and then incubated in CO2 for 45 seconds. The trypsin was neutraliz ed with 2X volume of the complete endothelial cell media. The cells were centrifuged at 1000 rpm in an Eppendorf CT 5810R, resuspended in 6 ml of complete endot helial cell media and transferred to a T75 flask (Fisher, Suwanee, GA). Next 15 ml of complete endothelial cell growth media and fresh antibiotics were added to the T75 flask to culture the cells. Confluent cells used for the studies were those of passages 3-4 and were ascertained positive for acetylated LDL. Transfection of HRECs with Lipofectamine HRECs were grown on 150 mm plates to confluence before transfection with the ribozyme expressing plasmids using Lipofect amine 2000 (Invitrogen, Carlsbad, CA). For the transfection 728 l of Opti-MEM I wa s mixed with 52 l Lipofectamine 2000 and kept at room temperature fo r 5 minutes. A second aliquot of 780 l of Opti-MEM I was mixed with 13 g of plasmid DNA and held at room temperature for 5 minutes. These two solutions were combined and held at room temperature for 20 minutes. The combined solutions were then added to cell cultures. After 24 hours the media was replaced. Cell cultures were harvest after 72 hours. Tr ansfection efficiency was determined using a rAAV plasmid expre ssing green fluorescent protein (GFP). Total RNA Extraction Trizol LS reagent (Invitrogen, Carlsbad, CA) was used to isolate total RNA from HRECs. Experiments were done following manufacturers protocol. For the 12-well plate, after media removed, 1.5 ml Trizol was added to each well. The cells were

PAGE 116

100 resuspended and then transferred to 1.5 ml eppendorf tubes and incubated at room temperature for 5 minutes. 80 l chloroform was added and the mixture was extracted for 30 seconds on a vortex mixer. After incubation at room temperature for 3 minutes, cells were centrifuged at 7400 rpm for 15 minutes. Th e aqueous layer was then transferred to a fresh tube and 190 l isopropanol was added. The mixture was incubated for 10 minutes at room temperature, and centrifuged at 7400 rpm for 15 minutes. The supernatant was discarded and the pellet was washed with 380 l of 75% ethanol. The cells were mixed on a vortex mixer for 15 seconds and centr ifuge at 7400 rpm at 15 minutes and the supernatant was discarded. The pellet was th en air dried for 10 minutes, dissolved in 25 l RNase-free water or TE. Th e product was stored at -70oC, aliquoted at 8 l in 3 tubes for future use. Relative Quantitative RT-PCR Relative quantitative RT-PCR was perfor med on RNA isolated from HRECs transfected with plasmids expressing ri bozymes and the p21NewHp vector expressing no ribozyme. Reverse transcription (RT) re actions were performed using reverse transcriptase (SuperScript from Invitr ogen, Carlsbad, CA) and a random hexamer according to manufacturers protocol. In brief, 8 l of RNA isolated from transfected HRECs and 2 l of random hexamer were mixed and incubated at 90oC for 3 minutes and then held on ice for 5 minutes. Then 4 l 5X RT buffer, 10 mM dNTP and 1 l RNasin, 0.1 M DTT and the reverse transcriptase were added into the reacti on. After a series of incubations at 25oC for 10 minutes, 42oC for 60 minutes and 95oC for 10 minutes, the RT product was complete and stored at -20oC. For the relative quantitative PCR, the linear range of the amplification of the RT product was determined by using a PCR master mix (1 l RT product/50 l, 200 M

PAGE 117

101 dNTPs, 1 mM MgCl2, 0.4 M PCR oligonucleotides, 1x Taq DNA polymerase buffer (Sigma-Aldrich, St. Louis, MO), 2 U Taq DNA polymerase (RED Taq ; Sigma-Aldrich, St. Louis, MO, 0.5 Ci/50 l [ 32P]-dATP (ICN, Irvine, CA)). 50 l of the master mix was separated into eight 0.2 ml tubes, and amplification wa s performed with an annealing temperature of 61C. Samples we re removed at even-numbered cycles starting at cycle 26. For each PCR sample, 5 l was removed and 2 l of formamide dye mix was added. The samples were heat denatured at 95C for 3 minutes, cooled on ice, and applied to a 6% polyacrylamide-8 M urea electrophoresis gel. Dried gels were analyzed on the phosphorescence imager to determine the linear range of amplification. Cycle 34 and cycle 36 was determined to be within the li near range of amplification for IGF-1R and integrin 1 mRNAs, respectively. In the relative quantitative RT-PCR assays the level of target mRNA was determined within each sample relative to an internal -actin standard. -actin mRNA levels were determined with a -actin primer/competimer oligonucleotide set (QuantumRNA) from Ambion (Austin, TX ). The competimer oligonucleotide pair from the -actin primer set annealed to the same targets as the primer oligonucleotide pair, but they were blocked at their 3' ends to prevent extension. This primer/competimer oligonucleotide set allowed us to determine the ratio of primer to competimer that yields a -actin PCR fragment that is approximately equimolar to the IGF-1R PCR product. To determine the ratio of the primer/competim er oligonucleotide set necessary to achieve this, PCR reactions were perf ormed as described earlier, and amplification proceeded for 34 cycles (IGF-1R) or 36 cycles ( 1-integrin). The ratio of primer to competimer

PAGE 118

102 oligonucleotide was determined to be 10:1 at a final concentration of 0.4 M for the combined primer/competimer mixture. Table 2.1. Sequences of primer pairs and a nnealing temperatures used in relative quantitative PCR. mRNA Primer Pairs Annealing C IGF-1R AGGACGGCTACCTTTA CCCGGCACAATTAC ATCAACAGGACAGC GACGGGCAGAG 61 Integrin 1 GAAAAACTCAATGACT TTCAGCGGC CCAGTTGTGTAATGC AAATGTCCACA 54 Integrin 3 CGTCGTCTCCGCCTTC AACCTGGAT GGCCACAGTCACTCC AAGCCACATG 60 Integrin 5 ACCCAGGGTCGGGGG CTTCAACTTA GCCCCGAACCACTG CAAGGACTTGT 61 Integrin v CGCTTCTTCTCTCGGG ACTCCTGCT CAGATGCTCCAAAC CACTGATGGGA 58 PCR reactions were then perf ormed for IGF-1R, integrin 1 and -actin simultaneously to determine the relative amount of IGF-1R to -actin, using the above conditions. Table 2.1 lists the sequences of pr imer pairs and the a nnealing temperatures. PCR products were later separated on 6% pol yacrylamide-8 M urea gels and analyzed on the phosphorescence imager. Reverse TranscriptionReal Time PCR For each reverse transcription (RT) react ion, 4 l of total RNA isolated from HRECs was used with iScr ipt cDNA Synthesis Kit (BioRad, Hercules, CA) following manufacturers protocol in a 20 l reaction. 4 l from the 20l RT-reaction product was used to perform real-time PCR using iQ SYBR Green Supermix (BioRad, Hercules, CA) according to manufacturers protocol. Primer s from manufacturers were resuspended to 7.5 pmol/l before addition to the PCR react ion mix. All reactions were performed in duplicate. -actin PCR primers (Ambion, Austin, TX) were used as the internal normalization control. Real-time PCR was performed on a DNA engine Opticon system

PAGE 119

103 with continuous fluorescence detector ( MJ Research, Waltham, MA). Opticon monitor analysis software (MJ Research, Waltham, MA) was used for analysis. For IR, the sequences of primer pairs used were GATGCACCGTCATCAACGGGAGTCTGATC and GGCGCCCCTTGGTTCCTGAAA CTTC, and annealing temperature was 58 C. For VEGFR-1 and VEGFR-2, the primers were pr e-synthesized from manufacturer (R&D, Minneapolis, MN) and the annealin g temperatures were both 55 C. Total Protein Extraction HRECs grown on 150 mm tissue culture plates were washed with PBS and scraped in ice cold phenol-free Hanks bala nced salt solution (HB SS) containing 1 mM EDTA. The cells were centrifuged at 1000 rp m for 5 minutes at 4C in Eppendorf 5810R centrifuge and 30 l of lysis buffer (150 mM Tris-H Cl, 150 mM NaCl, 1 mM EDTA, 1% Igepal CA-630 (Sigma St. Louis, MO), 1% pr otease inhibitor cocktail (Sigma, St. Louis, MO) and 1 mM DTT (Fisher, Suwanee, GA)) was added. The mixture was sonicated for 2 seconds and then centrifuged at 13,200 rpm for 15 minutes at 4C. Protein levels in the supernatant were determined using a bicinchon inic acid (BCA) protei n assay kit (Pierce, Rockford, IL) according to manufacturers protocol. Western Blotting 80 g of total protein was loaded on a 4%-15% gradient polyacrylamide gel (Criterion; BioRad, Richmond, CA). The gel was electrophoresed at 120 V for 20 minutes to allow for stacking of the samples and then 140 V for 65 minutes to separate proteins. The gel was transfe rred to a nitrocellulose memb rane (Millipore, Bedford, MA) using a blot cell apparatus (BioRad, Rich mond, CA) at 80 V for 5 hours on ice in 4C cold room. The membrane was blocked in TBS containing 0.05% Tween (Sigma St. Louis, MO) and 5% milk for 1 hour at room temperature. Then the membrane was

PAGE 120

104 incubated with primary antibody at 4C overnig ht. The membrane was washed again with TBS containing 0.05% Tween and 5% milk for 5 minutes and then incubated with secondary antibody for 1 hour at room temperat ure. The membrane was washed twice for 5 minutes and twice for 10 minutes with TBS containing 0.05% Tween. Usually the same membrane was also used to detect the internal protein control, -actin or cofilin. An enhanced chemiluminescence (ECL) Western bl ot Detection Kit (Amersham Biosciences Ltd., Amersham, UK) was used to visualize the western bands. Standard molecular weight markers (BioRad, Richmond, CA) were loaded on the same gel and used to determine the target proteins molecular we ight. The band intensity was analyzed using Scion Image (Scion, Frederick, MD). Table 2.2 lists the concentration of primary and secondary antibodies and molecular sizes of probed proteins. Table 2.2. Summary of primary and secondary antibodies used in western blottings. Protein Primary Antibody Secondary Antibody Molecular Size IGF-1R 1:2000 rabbit polyclonal anti-human IGF-1R subunit IgG antibody (Santa Cruz Biotechnology, Santa Cruz, CA) 1:2000 horseradish peroxidase (HRP)conjugated mouse antirabbit antibody (Santa Cruz) 95.2 kDa IR 1:100 rabbit polyclonal anti-human IR subunit IgG antibody (Santa Cruz) 1:1000 mouse anti-rabbit IgG-HRP (Santa Cruz) 95 kDa -actin 1:5000 mouse monoclonal anti-actin antibody (Sigma) 1:7500 HRP-conjugated rabbit-anti-mouse IgG antibody (Sigma) 42 kDa Cofilin 1:2000 rabbit anti-cofilin (Cytoskeleton) 1:1000 mouse anti-rabbit IgG-HRP (Santa Cruz) 18 kDa Flow Cytometry The protein levels of VEGFR-1 and VEGFR-2 were determined using flow cytometry rather than western blotting. Transf ected cells were harvested into single cell

PAGE 121

105 suspensions 48 hours post tran sfection. After centrifugation at 1500 rpm at 4 C for 10 minutes, the cell pellets were suspended in 1 ml of buffer (0.1% BSA in 10 mM NaCl and kept on ice). 10 g of either VEGF R-1 antibody (Santa Cruz Biotechnology Inc., Santa Cruz, CA) or VEGFR-2 antibody (Neo markers, Fremont, CA) was added to the cells. After an incubation on i ce for 30 minutes, the cells were washed twice in the buffer and incubated with the seconda ry antibody for 30 minutes in the dark (22.5 g of goat anti-rabbit-FITC antibody (Jack son Immuno Research, West Gr ove, PA) in 1 ml of 0.1% BSA). The cells were washed twice in buffer and 5000 cells were analyzed on a FACScan (BD Biosciences, San Jose, CA). Migration Assay The modified Boyden chamber assay was used to assay the cell ability to migrate to increasing concentrations of growth f actors. Trypsin (Trypsin-EDTA solution for endothelial cell culture, Sigma-Al drich) was used to detach the transfected HRECs into a single cell suspension. After the trypsin was in activated, cells were washed three times with PBS and suspended in DMEM to a final concentration of 1000 cells/l. 30,000 cells (30 l) were added per lower well in the blind-well chemotaxis chamber. A porous polyvinyland pyrrolidone-free polycarbonate membrane (12 m pores) coated with 10% bovine collagen was applied on the wells and the chamber was fully assembled. The chemotaxis chamber was inverted and held in 5% CO2 and room air at 37C to allow cell attachment to the membrane. After 4 hours, ch ambers were then placed upright. 50 l of a cocktail containing VEGF (25 ng/ml), bFGF (25 ng/ml), and various concentration s (1 ng/ml, 10 ng/ml, or 100 ng/ml) of the specifi c growth factor required to stimulate migration was added to the upper wells. For exam ple, IGF-1 was used to test the IGF-1R ribozyme, PlGF was used to test the VEGFR1 ribozymes and VEGF-E was used to test

PAGE 122

106 the VEGFR-2 ribozyme. The chamber was incubated in 5% CO2 and room air at 37C overnight. Next, cells on the attachment si de (lower wells) were scraped from the membrane, and only those cells that migrated through the pores of the membrane into the upper wells were left. The cells were fixed in methanol and then stained with a modified Wright-Giemsa stain (LeukoStat solution; Fisher Scientific, Springfield, NJ), and finally mounted on glass slides. DMEM served as a negative control of random cell migration, and DMEM with 10% FBS was used as a positive control in each experiment. A minimum of six replicate wells were assaye d for each condition. A light microscope was used to count the cells, and the average of the number of cells count ed in three separate, high-power (400X) fields was used as a quantita tive reflection of th e number of migrated cells per well. Cell Proliferation Assay (BrdU) The BrdU-incorporation assay was perfor med following manufacturers protocol (Roche Applied Science, Indianapolis, IN). For a 96-well plate, tr ansfected cells were added in a final volume of 100 l/well. The cells were incubated for 48 hours. BrdU was added in the wells to a final concentration of 10 M, and the cells were incubated for 2 hours. The cell media was removed and 200 l/well of FixDenat from the kit was added to the cells and the plate was incubated at room temperature for 30 minutes. The FixDenat was removed and 100 l/well of BrdU antibody-conjugate was added and the plate was incubated for 90 minutes at r oom temperature. The antibody solution was discarded and the cells were washed with 300 l/well of washing solution from the kit. The washing solution was removed and 100 l/w ell of substrate solution was added. The plate was held at room temperature until color development was sufficient for photometric detecti on (5-30 minutes).

PAGE 123

107 Tube formation Assay (Matrigel) Cells were transfected as detailed abov e. 24 hours after tranfection the cells were trypsinized and 5000 cells were seeded on matrigel (BD Biosciences, San Jose, CA) and incubated at 37oC in 5% CO2 environment. The cells were photographed every 24 hours. Proliferating Endothelial-Cell Specific Promoter Constructs The pLUC1297 and pLUC1298 plasmids were transformed into and isolated from DH5 E.Coli Bacteria (Qiagen Mega Kit). Thes e plasmids contained the proliferating endothelial-cell specific promot er and a luciferase reporter gene followed by a polyA site. The promoter was composed of a 4X (1297) or 7X (1298) 46-mer of the endothelin enhancer upstream of the human cdc6 promoter. The pLUC1297HHHP and pLUC1298 plasmids contained the IGF-1R hammerhead ribozyme/hairpin ribozyme cassette. A variant of pLUC1298HHHP designated pG E1298HHHP was missing the luciferase reporter gene. The cloning of IGF-1R ha mmerhead ribozyme with immediatedownstream processing hairpin ribozyme into the vectors was performed as detailed above, but ligation was into the Xba1 a nd Xho1 sites. The ligation product was transformed in StblII cells and the anti-kan amycin clones were selected. The isolated plasmid DNA (Giga Prep Kit-Qiagen) was sequenced to confirm ribozyme sequences. Plasmid Formulation for Adult Mouse Eye Gene Transfer Plasmid DNA was isolated and purified from StblII bacteria (Invitrogen). A cationic lipid (Lipid 89 Genzyme Corporation) in a molar ration of 3:1 lipid:plasmid was used to condense the DNA in 40% ethanol/5% dextrose. A helper lipid mixture composed of lysophospatidylcholine: monoacylglycerol: free fatty acid (1:4:2) (mole/mole/mole) was mixed with the cationic lipid. The lengths of the acyl chain helper lipids were 18:2

PAGE 124

108 and the ratio of cationic lipid to helper lip id mixture was 10:90. Ethanol was removed by dialysis against PBS and the final DNA concentration was 0.5 mg/ml. Animals All animals were treated in accordance w ith the ARVO statement for the use of animals in Ophthalmic and Vision Research a nd with the Guide for the Care and Use of Laboratory Animals. All protoc ols were approved by the IA CUC of the University of Florida. C57BL6/J timed pregnant mice a nd adult mice were obtained from Jackson Laboratories (Bar Harbor, ME). The mice we re housed in the University of Florida Health Science Center Animal Resources facilities. Ketamine (70 mg/kg body weight) and xylazine (14 mg/kg body weight) mixture was i.p. injected to anesthetize mice before laser treatments or euthanizat ion. Intravitreal injection in to 24-hour-old mouse pups was accomplished by placing the pups on a plastic sh ield on ice in order to anesthetize the pups before injection. Intravitreal Injection into the Mouse Mo del of Oxygen-induced Retinopathy (OIR) Shown in Figure 2.3 is the time course of the mouse model of oxygen-induced retinopathy (OIR). The newborn mouse pups we re injected intrav itreally with 0.5 l plasmid (2 mg/ml) in the right eye within 24 hours of birth. Left eyes were used as uninjected controls. Seven days after birth the pups were placed in a chamber that maintained a 75% oxygen environment. After 5 days the 12-day-old pups were returned to normal room air. Return to normal air si mulated a hypoxic response that resulted in the onset of retinopathy. This process mimicked human retinopathy of prematurity, and the aberrant neovascularization was very similar to what is seen in diabetic retinopathy patients. Neovascularization was initiated upon return to normal room air on day 12 and peaks on day 17 when the mice were sacrificed and the eyes were removed for further

PAGE 125

109 analysis. Selected animals were perfused with 1.5 ml of 4% paraformaldehyde for immunohistological studies. Figure 2.3. Time course of OIR mouse model. The average number of pre-retinal nuclei per retinal cross section was determined and used as a quantitative measure of the extent of abnormal neovascularization. To prepare the eyes for analysis the eyes were fixed in 4% pa raformaldehyde for one hour or in TRUMPS over night, washed in PBS, a nd embedded in paraffin. Each eye was sliced into 300 serial sections (6 m thick) sagita lly through the cornea para llel to the optic disc. Every thirtieth section was placed on slide and stained with hematoxylin-eosin (H&E). Three blinded individuals count ed the number of nuclei in the cross-sections of preretinal blood vessels that grew beyond the retinal inner limiting membrane into the vitreous space. The total number of nuclei was used as the indication of neovascularization levels.

PAGE 126

110 Intravitreal Injection into the Adult Mo use Model of Laser-Induced Retinopathy Figure 2.4 shows the time course of this mouse model. Six to eight week old C57BL/6 mice were intravitreally injected in right eyes wi th 2 l of AAV-VEGF, which expressed VEGF and induced neovascularizati on. Left eyes were uninjected controls. Figure 2.4. Time course of the adult mouse m odel of laser-induced neovascularization. Four weeks later the AAV-VEGF injected eyes were subjected to with laser occlusion of the large venous vessels in the retina. An ar gon green laser system (HGM Corporation, Salt Lake City, UT) with a 78diopter lens was used for retinal vessel photocoagulation. The blue-green argon lase r (wavelength 488-514 nm) was applied to selected venous sites next to the optic nerv e. Laser occlusion was performed at the setting of 1 second duration, 50 m spot size and an average of 600 mW intensity with an average of 44 burns per retina. Immediately af terwards we injected 2 l of the ribozyme expressing plasmid intravitreally in the same eye that had received laser treatment. Plasmids were formulated in a final concentr ation of 0.3 mg/ml with a cationic lipid that can facilitate transfec tion of the adult mouse endothelium The mice were sacrificed three weeks following laser treatment. 3 ml of 4% paraformaldehyde was perfused in each

PAGE 127

111 mouse. The eyes were enucleated and the retinas removed for immunohistological studies. Immunohistological Studies Retinas were placed in 96-well plates and permeabilized with PBS containing 0.2% Triton X-100, 0.1% BSA and 0.1% rabbit serum for 24 hours at 4C. Then they were washed in PBS for 24 hours at 4C. Anti-luciferase pAb (1:50 dilution of goat polyclonal IgG) (Promega, Madison, WI) wa s the primary antibody. After a wash in PBS the retina were treated with 0.1% rabbit serum for 24 hours at 4C. The secondary antibody was FITC-conjugated ra bbit-anti-goat-IgG (green) (1:4000 dilution in PBS) (Sigma, St. Louis, MO). The blood vessels in the flat-mounted retina were labeled with endothelial cell-specific aggl utinin conjugated to rhodamine (red) (1:1000 dilution in permeabilization solution, Vector Laboratories, Burlingame, CA). The retinas were imaged using a MRC-1024 Confocal Laser Scanning Microscope at the Optical Microscopy Facility at the Univ ersity of Florida (Gainesvi lle, FL). ImageJ (ImageJ 1.32j. Wayne Rasband, National Institutes of Health, USA, http://rsb.info.nih.gov/ij/ ) was used to analyze the images. Statistical Analysis Statistical analysis of the data was perf ormed using the Students t-test (Excel; Microsoft, Redmond, WA). Results were re ported in mean SE. P < 0.05 is deemed significant.

PAGE 128

112 CHAPTER 3 RESULTS In this project a number of cell surface protein receptors were selected as our ribozyme targets. These ribozymes were tested in vitro in HRECs and/or in vivo in mouse models of retinal neovasculariz ation. Inactive versions of se lected ribozymes were also produced and tested. A proliferating endotheli al cell-specific promoter was developed and tested with selected ribozymes. Ribozyme Design All hammerhead ribozymes designed in this study were 34 bases in length. They bound to targeted mRNA, formed a three stem st ructure, and cleaved at the 3 end of the X in the NUX triplet within the target sequenc e. The target sequences were all 13 bases long with six bases on either side of the X forming ribozyme stems I and III. The design of a ribozyme began with the GeneBank sear ch of the proteins full cDNA sequence and the selection of a 13-base-long target site. Target Site Selection Taking the insulin receptor (IR) as an example, the human and mouse IR gene sequences were first retrieved from GeneBank (human accession number NM_000208 and mouse accession number NM_010568). We sear ched for GUC sites as the candidates of cleavage sites. All these ca ndidate targets were examined in silico for accessibility. We eventually chose the 5-UUACGUCUGAUUC-3 sequence in the human gene and the 5-GCUUGUCUGAAAU-3sequence in the mouse ge ne as cleavage targ et sites. Figure

PAGE 129

113 3.1 shows the full cDNA sequence of human IR gene and the selected target site is highlighted. Figure 3.1. The human IR cDNA sequence with ribozyme target site highlighted.

PAGE 130

114 Accessibility of Target Site The tertiary structure of the mRNA su rrounding the target site will affect the accessibility. The mRNA sequence from 200bp upstream and downstream of the NUX target site was examined with Zukers Mfold program ( http://www.bioinfo.rpi.edu/a pplications/mfold/old/rna/ ). Figure 3.2 shows some of the possible secondary structures of the huma n IR target region predicted by the Mfold program. As shown, the target site (red arro ws) is partially within the loops, which indicates, at least, partial accessibility of the site. The ideal situation suggesting complete accessibility of the target would be location of the target completely within loops. Figure 3.2. Mfold structures predicte d for the human IR target region.

PAGE 131

115 After determining if the target site was accessible, it was then necessary to examine the potential folding of the riboz yme by Mfold. The sequence of the ribozyme was generated from the target sequence, whic h gave the sequence of the 5and 3-six base targeting arms that form stems I and III and the catalytic core and tetraloop sequence that was used in all of the hammerhead ri bozymes in this study. Figure 3.3 shows the only Mfold structure predicted for the human IR ribozyme. This structure is a typical type A structure, based on the nome nclature of Fritz, et al. [ 257]. In this structure both targeting arms are completely accessible and the internal tetraloop has been formed by stem II hybridization. Only this structure bei ng predicted by Mfold indicates that this ribozyme is completely accessible for target arm binding to the mRNA target sequence. Figure 3.3. Mfold predicted secondary structure of human IR ribozyme. The mouse target and ribozyme selecti on and design were performed as above. Figure 3.4 shows the technical structure of the bound co mplex of the human/mouse insulin receptor ribozymes and their target sequences. The stem I, II, and III in this complex are also marked out.

PAGE 132

116 Figure 3.4. The 34-base ribozymes (black) anneal ed to the 13-base targets (red) for both human and mouse. Sequences of the Ribozymes and the Targets Table 3.1 shows all ribozyme sequences a nd their 13-nucleotide target sequences. Table 3.1. Summary of ribozyme and target sequences. Ribozyme Ribozyme Sequence (5 3) Target Sequence (5 3) Mouse: CUUCGU C UUUGCG Rat: CUU UGU C UUUGC A IGF-1R Rz1 CGCAAA CUGAUGAGCCG UUCGCGGCGAAACGAAG Human: CUUCGU C UUUGC A Mouse: GUAUGU C UUCCAU Rat: GUAUGU C UUCCAU IGF-1R Rz2 AUGGAA CUGAUGAGCCG UUCGCGGCGAAACAUAC Human: GUAUGU C UUCCAU IR human GAAUCA CUGAUGAGCCG UUCGCGGCGAAACGUAA UUACGU C UGAUUC IR mouse AUUUCA CUGAUGAGCCG UUCGCGGCGAAACAAGC GCUUGU C UGAAAU Mouse: GGGUGU C UAUAGG VEGFR-1 CCUAUA CUGAUGAGCCG UUCGCGGCGAAACACCC Human: AGGUGU C UAU CAC VEGFR-2 ACAGAA CUGAUGAGCCG UUCGCGGCGAAACCAUG CAUGGU C UUCUGU Integrin 1 mouse CUUAUA CUGAUGAGCCG UUCGCGGCGAAACAUCU AGAUGU C UAUAAG Integrin 3 mouse CAUGAA CUGAUGAGCCG UUCGCGGCGAAACAUAG CUAUGU C UUCAUG Integrin 5 mouse GUGGCA CUGAUGAGCCG UUCGCGGCGAAACAGGA UCCUGU C UGCCAC Integrin v mouse AACUUG CUGAUGAGCCG UUCGCGGCGAAACCAUU AAUGGU C CAAGUU

PAGE 133

117 Some mRNA sequences shared significan t homology between species so the same ribozyme can target and cleave multiple specie s. Cleavage occurred on the 3 side of the boxed C and the ribozyme and target complime ntary sequences are underlined. The part of the target sequence that is not underlined did not base -pair with the ribozyme. The remainder of the ribozyme sequence formed th e catalytic core and stem II of ribozyme. In Vitro Testing of Ribozymes The first step in testing a ribozyme was to examine its in vitro cleavage activity. 13-base 32P-labeled RNA oligonucleotides were used as the cleavage ta rgets to examine the cleavage activity of the ribozymes. Time Course of Cleavage Time course of cleavage analysis gave the first indication of the level of catalytic activity of a ribozyme. To examine a ribozym es time course of cleavage, synthetic 13nucleotide-long target RNA oligonucle otides were 5-end-labeled with 32P and cleaved by ribozyme in vitro The 5 cleavage product was 7 nucleotides in length. Panel A in Figure 3.5 is the autoradiograph of a 10% polya crylamide-8M urea gel used to separate cleavage products of the hu man IR ribozyme on the human RNA target oligonucleotide. Panel B is the graphical represen tation of the data obtained by analysis of the gel in panel A using a PhosphorImage. The IR ribozyme had good catalytic activity since over 80% of the cleavable RNA oligonucleotide was cl eaved within 2 minutes. Notice, however, that only approximately 60% of the targ et RNA was cleaved in this reaction. The remaining target remained uncleaved presum ably due to a fraction of the ribozyme was misfolded and therefore inactiv e. Some of our other ribozy mes exhibited more complete cleavage reactions that could reach 90% to 100% cleavage of the ta rget oligonucleotide. The situation for this IR ribozyme was not unusual.

PAGE 134

118 Figure 3.5. Cleave time course of human IR ribozyme. Summary of Cleavage Time Course of Ribozymes0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 0510152025303540 Time (minutes)Fraction of target cleaved IGF-1R IR VEGFR1 by R1 Rz VEGFR2 by R2 Rz VEGFR1 by R2 Rz VEGFR2 by R1 Rz alpha 1 alpha 3 alpha 5 alpha v Figure 3.6. Summary of time course s cleavage of the ribozymes generated in this study. Other ribozymes were test ed similarly and their time courses of cleavage are shown in Figure 3.6. For IGF-1R ribozyme, th e inactive forms were also tested. As expected, they did not show any cleavage ac tivity (data not shown). We tested VEGFR-1 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 051015202530 Time (minutes)Fraction of target cleaved Target 5 cleavage product IR Rz human 0 1 2 5 15 30 60 120 180 A B

PAGE 135

119 ribozyme on its own target and also the targ et of VEGFR-2 ribozyme. Similarly VEGFR2 ribozyme was also tested with both targ ets. Both VEGFR-1 and VEGFR-2 ribozymes showed high catalytic activity on their respec tive target RNA oligonucleotides. Over 90% of the target RNA oligonucleotides were cleav ed within 5 minutes. On the other hand, we did not see any cleavag e of the VEGFR-1 ribozyme on the VEGFR-2 target or of the VEGFR-2 ribozyme on the VEGFR-1 target. This demonstrated the specificity of these two ribozymes. The cleavage rates of integrin ribozymes varied from lower than 20% cleavage to around 90% cleavage within 10 minutes. Kinetic Analysis After the time course cleavage analysis, multi-turnover kinetic analysis was performed to determine the kinetic parameters (Vmax, Km, kcat). Figure 3.7 shows the saturation curve and the double-reciprocal plot of the human IR ribozyme kinetic analysis. The slope and intercep tion of the double-reci procal plot were used to determine the kinetic parameters. Saturation Curve0 20 40 60 80 100 120 140 05000100001500020000 Substrate (nM)Velocity (nM/min) y = 33.849x + 0.0278 R2 = 0.1361 0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 0.16 0.18 0.20 -1.E-03-5.E-040.E+005.E-041.E-032.E-032.E-03 1/[S]1/v Figure 3.7. Multiple-turnover ki netic analysis of a human IR ribozyme. Table 3.2 summarizes the kinetic parameters (Vmax, Km, kcat) of the ribozymes.

PAGE 136

120 Table 3.2. Summary of ribozyme kinetic data. Ribozyme Vmax (nM/min) Km (M) kcat (min-1) IGF-1R Rz1 7.0.3 47.1.7 0.47.01 IGF-1R Rz2 2.8.60 1.8.1 0.2.04 IR 35.97 1217.59 2.39808 VEGFR-1 227.3.8 5.4.58 15.2.2 VEGFR-2 356.9.2 7.6.35 29.6.6 Integrin 1 57.05.7 25.6.0 3.8.3 Integrin 3 6.1.1 41.6.41 57.05.07 Integrin 5 322.6.6 81.1.2 21.5.8 Integrin v 33.4.7 5.2.2 2.2.2 Functional Analysis of Ribozymes in HRECs Following successful in vitro analysis of the ribozyme s, we cloned the ribozymes into the p21NewHp plasmid and transfected thes e plasmids into HRECs. This allowed for the determination of target mRNA levels, ta rget protein expressi on levels and for the testing of cellular activ ities related to normal HREC physio logical functions such as cell migration, cell proliferation a nd endothelial tube formation. Inhibition of mRNA Expression Messenger RNA was the target of a ribozyme thus mRNA levels were measured after transfection. As an example, IR mRNA levels are shown in Figure 3.8. The mRNA levels were determined using reverse tr anscription on isolated total cellular mRNA followed by real time PCR on the cDNA products w ith primer pairs specific for the target mRNAs. The levels of the target were normalized to -actin mRNA levels. The mRNA levels in mock-transfected cells were se t as 100%. Cells expressing the human IR ribozyme showed a signification reducti on of 42.4.1% in IR mRNA level ( P =0.014). The non-transfection or transformation with the p21NewHp vector showed no significant difference in IR mRNA levels compar ed to the mock-transfected cells ( P =0.35 for NT and P =0.16 for vector)

PAGE 137

121 Insulin Receptor Ribozyme0 20 40 60 80 100 120 NTMockVectorIR Rz% IR mRNA l evels relaive to beta actin Figure 3.8. Insulin receptor mRNA levels in HRECs. Table 3.3 summarizes the reduction in mRNA levels in HRECS after transfection with plasmids expressing the indicated ri bozymes. Inactive versions of the IGF-1R ribozyme 1 and 2 and the 1 integrin ribozyme were also tested. As expected, expression of these inactive ribozymes resulted in no reduction of target mRNA levels ( P >0.1, data not shown). Table 3.3. Reduction in target mRNA levels in HREC by the ribozymes. Ribozyme Reduction in mRNA levels P value RT-PCR method IGF-1R Rz1 39.5.1% 0.003 Relative Quantitative IGF-1R Rz2 12.7.7% 0.003 Relative Quantitative IR 42.4.4% 0.014 Real-time VEGFR-1 71.1.1% 0.0002 Real-time VEGFR-2 85.1.9% 0.0008 Real-time Integrin 1 32.4.0% <0.01 Relative Quantitative Protein Levels IR protein levels were also investigat ed after transfection of HRECs with the plasmids. The -subunit of the IR appeared in tw o bands (around 200 kDa and 90 kDa) as

PAGE 138

122 precursor and mature forms (Figure 3.9). Th e protein level in non-transfected HRECs was set to 100%. Vector-transfected cells s howed no significant reduction. Expression of the human IR ribozyme resulted in a reduction of 20.9.1% ( P =0.006) in IR protein levels. IR Ribozyme0 20 40 60 80 100 120 NTVectorIR Rz%IR protein levels relative to Cofili n Figure 3.9. Western analysis of IR levels in cells expressi ng the human IR ribozyme. While expression of inactive forms of the ribozymes resulted in no reduction in mRNA levels, there was a significant reducti on in protein levels (30.8.6% for inactive IGF-1R Rz1). This reduction resulted from th e antisense binding of the ribozyme to the target mRNA. VEGFR-1 and VEGFR-2 prot eins levels were measured by flow cytometry rather than western analysis. As expected, the VEGFR-1 ribozyme reduced both VEGFR-1 mRNA and prot ein levels and the VEGFR-2 ribozyme reduced both VEGFR-2 mRNA and protein levels. In a ddition the VEGFR-1 ribozyme reduced the levels of VEGFR-2 mRNA and protein and the VEGFR-2 riboz yme reduced the levels of VEGFR-1 mRNA and protein. Th ese results demonstrated that there is co-regulation between these two receptors.

PAGE 139

123 Table 3.4. Reduction in protein levels by the ribozymes. Ribozyme Reduction in protein levels P value IGF-1R Rz1 active 47.7.6% 5.4x10-5 IGF-1R Rz1 inactive 30.8.6% 4.6x10-5 IR 20.9.1% 0.006 VEGFR-1 R-1 protein: 66.7% R-2 protein:34.9% <0.01 VEGFR-2 R-1 protein: 15.4% R-2 protein:41.9% <0.01 VEGFR-1 + VEGFR-2 R-1 protein: 64.1% R-2 protein:27.9% <0.01 Migration Assays Figure 3.10 examines the ability of transf ected HRECs to migr ate in response to increasing concentrations of IGF-1. The cell migration was examined for HRECs transfected with the vector, or plasmids e xpressing the IGF-1R ri bozyme 1 or ribozyme 2, or the inactive ribozyme 1. Migration assa ys were performed in a modified Boyden chamber. The active IGF-1R ribozyme 1 and 2 demonstrated a reducti on in migration of approximately 91% and 58%, respectively. Inac tive ribozyme 1 also showed approximate 51% reduction in cell migrations. This re duction possibly resulted from antisense inhibition of the IGF-1R protein. The effect of the VEGFR-1 or VEGF R-2 ribozymes on migration was also examined (Figure 3.11). For these assays VEGF -E or placental growth factor (PlGF) was used to stimulate migration of HRECs. VE GF-E specifically binds to VEGFR-2 while PlGF specifically binds to VEGFR-1. The abil ity of HRECs, transfected with vector DNA, to migrate across a membrane to soluti ons containing either VEGF-E, PlGF or the heterodimer VEGF-E/PlGF was measured. HRECs expressing the VEGFR-1 ribozyme did not migrate toward PlGF, suggesting th at they lacked VEGFR-1. HRECs expressing the VEGFR-2 ribozyme did not migrate toward a VEGF-E suggesting that they lacked VEGFR-2.

PAGE 140

124 0 50 100 150 200 250 020406080100 IGF1 (ng/ml)Number of Migrating Cells/ High-Powered Field Control Inactive Rz1 Acitve Rz1 0 50 100 150 200 250 300 350 400 450 020406080100 IGF1 (ng/ml)Number of Migrating Cells/ High-Powered Field Control Active Rz2 Figure 3.10. HREC migration assa ys in response to IGF-1. 0 50 100 150 200 250 300 VectorVEGFR1 RzVEGFR2 Rz% Number of migrating cells per high power field VEGF-E PlGF VEGF-E/PlGF Figure 3.11. Effect of the VEGFR-1 and VE GFR-2 ribozymes on HREC migration. Cell Proliferation Assays Cell proliferation was measured by cellu lar incorporation of bromo-uridine (BrdU). Cells transfected with the vector or plasmids expressing the VEGFR-1 or

PAGE 141

125 VEGFR-2 ribozymes or the IGF-1R ribozym e were examined. Results are shown in Figure 3.12. The incorporation of BrdU in v ector-transfected cells was set to 100%. VEGFR-1 ribozyme expression redu ced incorporation by 42.7.5% ( P =5.1x10-4), VEGFR-2 ribozyme expression redu ced incorporation by 50.25.9% ( P =1.3x10-5), and IGF-1R ribozyme expression re duced incorporation by 83.7.7% ( P =6.8x10-7). Ribozymes Inhibit BrdU Incorporation 0 20 40 60 80 100 120 140 MockVectorVEGFR1 RzVEGFR2 RzIGF-1R Rz% BdrU invorporation (RFU ) Figure 3.12. Effect of ribozyme e xpression on cell proliferation. Tube Formation Assays The ability of HRECs to form tubes on Matrigel is anothe r basic function of endothelial cells. HRECs woul d form honeycomb-like structur es consisting endothelial tubes naturally when cultured on Matrigel. Wh en cells were transfected with the plasmid expressing the IGF-1R ribozyme, or the VE GFR-1 ribozyme or the VEGFR-2 ribozyme; tube formation was completely inhibited (Figure 3.13). The empty vector transfected cells were used as the control.

PAGE 142

126 Figure 3.13. Effect of ribozymes on HREC tube formation. In Vivo Analysis of Ribozymes The in vivo effects of the ribozymes were examined in the mouse model of oxygen-induced retinopathy (OIR). Figure 3.14 is a cross section of mouse pup eye, stained with H&E, from the OIR model. All major anatomical parts of the eye are shown in this figure. Pre-retinal blood vessels (green arrows) grew beyond the retinal inner limiting membrane into the vitreous space. They are the representation of abnormal neovascularization. The measure of aberrant neovascularization was determined by the average number of pre-retinal blood vessel nuclei per section.

PAGE 143

127 Figure 3.14. Cross section of a mouse eye showing pre-retinal vessels. IR Ribozyme Tested in OIR Mouse Model0 20 40 60 80 100 120 140 Control IR RzIGF-1R RzCombo Rzs% Average Nuclei per Section Figure 3.15. Ribozyme reduction of pre-retinal neovasculariza tion in the OIR model. Table 3.5 summarizes the results of the OI R mouse model assays on all ribozymes tested. Rows separated by solid lines ar e different groups of mice and dotted lines separate different test litter s in the same group. Inactive ri bozymes led to reductions in pre-retinal neovascularization to some extent. However, the reduction found with inactive

PAGE 144

128 IGF-1R ribozyme 1 and 2 were mi nimal or close to significant ( P <0.05 is considered significant), while the reductions resulted from inactive Integrin ribozymes were significant. Table 3.5. All ribozymes tested in vivo Ribozyme Reduction in average nuclei per section P value IGF-1R Rz1 active 64.7.6% 2.7x10-5 IGF-1R Rz1 inactive 17.3.1% 0.03 IGF-1R Rz2 active 51.7.2% 2.3x10-5 IGF-1R Rz2 inactive 10.1.8% 0.09 VEGFR-1 47.0.0% 5.3x10-4 VEGFR-2 75.5.0% 7.5x10-8 Integrin 1 active 88.8.4% 5.44x10-7 Integrin 1 inactive 46.2.3% 1.7x10-3 Integrin 3 active 83.5.0% 1.31x10-5 Integrin 3 inactive 63.7.7% 1.2x10-4 IR 34.0.3% 8.36x10-7 IGF-1R 42.0.1% 1.66x10-8 IR + IGF-1R (Combo) 36.6.0% 1.27x10-4 Promoter Development The expression of ribozymes cloned into the p21NewHp vector was driven by the CMV enhancer/chicken -actin promoter. This promis cuous enhancer/promoter was active in numerous cell lines a nd under a variety of physiologi cal states. Thus, using this promoter could be a problem since it will result in the expression of the ribozymes in multiple cell types and tissues. The targets of our ribozyme were physiologically required for normal retinal development and function. We only wanted to inhibit the abnormal expression of the target proteins while l eaving normal expression alone. But ubiquitous ribozyme expression could lead to the oblat ion of all normal and abnormal expression and lead to adverse effects. This was observed with the integrin ribozymes.

PAGE 145

129 Integrin Ribozyme Expression in vivo with the CMV/ -actin Enhancer Promoter As detailed in the introdu ction the various integrin dimers are very important for cell adhesion and migration and they play essential roles in normal eye development. We observed severe structural abnormalities in the eye in the OIR mouse model after injection of plasmids expressing ribozymes to the 1 and the 3 subunits of integrin. Figure 3.16. Reduction of pre-retin al neovascularization in the OIR mouse model with expression of the 1 or 3 integrin ribozymes. The 1 and 3 ribozymes significantly reduced the amount of pre-retinal neovascularization in the OIR m ouse model (Figure 3.16). Active 1 and 3 ribozymes resulted in 88.8.4% ( P =5.44x10-7) and 83.5.0% ( P =1.31x10-5) reduction in preretinal neovascularization, respectively. Their inactive forms resulted in a less signification reduction, 46.2.3% ( P =1.7x10-3) for inactive 1 ribozyme and 63.7.7% ( P =1.2x10-4) for inactive 3 ribozyme. These reductions with active or inactive ribozymes were much greater than a ny other ribozymes we tested in the same Integrin Ribzoymes Tested in OIR Mouse Model 0 20 40 60 80 100 120% Average Nuclei per Section Control 1 Rz Active 1 Rz Inactive 3 Rz Active 3 Rz Inactive

PAGE 146

130 mouse model. We also found that even the inactive versions of these two ribozymes could cause structural abnormalities in the developing eye due to antisense inhibition. Figure 3.17. Expression of 1 ribozyme in OIR model results in severe deformations of the eye. Figure 3.17 shows cross sections of a normal uninjected eye and of eyes expressing the integrin ribozymes. Overall the injected eyes showed a number of abnormalities, such as lens separation, retina detachment and closed iris. In addition the injected eyes were smaller than the unin jected eyes (although it is not unusual for intravitreal injection to affect eye size). While integrin ribozyme reduced neovascularization, these integrin ribozyme s also significantly inhibited the normal

PAGE 147

131 development of the eye. It is so interesti ng that the antisense e ffect of the inactive ribozymes was powerful enough to interfere with the developmental process. Due to the severity of deformation found with the CMV expression of the integrin ribozymes, we decided to use a proliferati ng endothelial cell specific pr omoter for integrin ribozyme expression. The Proliferating Endothelial Cell-Specific Promoter To overcome potential expre ssion problems with the CMV/ -actin promoter, we cloned and tested our ribozyme s in a vector that had a pr oliferating endothelial cellspecific promoter. Dr. Sullivan designed and constructed this promoter. Using this promoter, we were hoping to only target prolif erating endothe lial cells while not affecting the quiescent endothelial cells in the developed vasculature and any other cells in the retina. Figure 3.18. pLUC1297/1298 vectors and pLUC1297HHHP/1298HHHP clones Dr. Sullivan tested a number of enhan cers and promoters and eventually found that the combination of endothelin enhancer/c dc6 promoter provided the best specificity

PAGE 148

132 to endothelial cells. Endothelin (ET or ET-1), which is exclusively synthesized by vascular endothelium, is one of the most powerful vasoconstrictors known. Cdc6 is a 30,000-dalton protein essential for the init iation of DNA replication. This protein functions as a regulator at th e early steps of DNA replication. It is thought to be involved in the assembly of minichromosome main tenance proteins onto replicating DNA. It localizes in the cell nuc leus during cell cycle G1, but translocates to the cytoplasm at the start of S phase. Quiescent cells in G0 do not express this protein. Therefore in this specific vector, ET enhancer determined the expression specificity in endothelial cells and cdc6 promoter further narrowed the specif icity into proliferating endothelial cells. Dr. Sullivan produced two specific promoters, both of which had the cdc6 promoter. One vector had a 4X multimer of the ET enhancer, designated pLUC1297, and the other vector had 7X multimer of the ET enhancer, designated pLUC1298. The structure of these promoters is shown in Fi gure 3.18. Downstream of cdc6 promoter was a luciferase reporter gene followed by Poly A signal. The IGF-1R ribozyme was cloned and these two vectors followed by a self-cle aving hairpin ribozyme that generated a discrete 3-end to the run-off transc ript. (pLUC1297HHHP and pLUC1298HHHP). The four plasmids, pLUC1297, pLUC1298, pLUC1297HHHP and pLUC1298HHHP, were transfected into HRECs and fibroblast cells to examine the cellspecific expression of luciferase. We tested the expression in two fibroblast cell lines, shown as F1 and F2 in Figure 3.19, and in HR ECs from two different donors, shown as HREC 10 (ten-year-old donor) and HREC 14 (fourteen-year-old donor) in Figure 3.19. pLUC1297 and pLUC1298 showed high levels of luciferase expression in HRECs compared with fibroblasts (about 200 times higher in HREC 10). Similarly

PAGE 149

133 pLUC1297HHHP and pLUC1298HHHP also showed higher levels of luciferase in HRECs compared with fibroblasts (about 20 times higher in HREC 10). Luciferase expression was higher in HREC 10 than in HREC 14, which probably resulted from a difference in the donor cells. When co mparing pLUC1297HHHP with pLUC1297, or pLUC1298HHHP with pLUC1298, we found that th e ribozyme-inserted constructs had much lower luciferase expression level than th eir parent constructs. This is probably due to the loss of the PolyA signal in the plasmids. 0.0E+00 5.0E+05 1.0E+06 1.5E+06 2.0E+06NTpLUC1297pLUC1298pLUC1297 HHHPpLUC1 298HHHpLuciferase Activity RFU (% relative to NT) F1 F2 HREC 10 HREC 14 Figure 3.19. Verification of the cell specificity of the pro liferating endo thelial cellspecific enhancer/promoter. The New Promoter Tested in vivo We tested the pLUC1298HHHP construc t in the OIR mouse model. Figure 3.20 shows confocal images of eyes from these e xperiments. OS (left ey es) were un-injected eyes and OD (right eyes) were injected eyes. The vessels were labeled with endothelial cell specific agglutinin conjugated with r hodamine, and luciferase expression was

PAGE 150

134 immunofluorescently shown in green by secondary antibody. Un injected eyes (panels A and C) showed background green fluorescence, while the injected eyes (panels B, D, E and F) showed expression of luciferase onl y on the vasculature (green and yellow ). The magnifications of panel E and F were 400x, and panel F was showing greater detail of the boxed part in panel D. When comparing pa nel C and D (magnificat ion 200x), there was a greater density of abnormal, small, blood vesse ls evident in the uninjected eye (panel C) while the injected eye (panel D) showed a lower density of blood vessels on the retina. This suggests that the IGF-1R ribozyme wa s actively expressed and was reducing preretinal neovascularization. This was confirme d by examining H&E stained cross sections that quantitatively showed a reduction in pre-re tinal neovascularization in injected eye as detailed below. Figure 3.21 shows the results of the OI R mouse model injections with the proliferating endothelial cell-sp ecific constructs. Blue bars are injected eyes and brown bars are uninjected eyes. There were five groups of mice, injected with pLUC1297, pLUC1298, pLUC1297HHHP, pLUC1298HHHP pGE1298HHHP respectively. pGE1298HHHP had the same structure as pLUC1298HHHP except that the luciferase reporter gene was eliminated. Uninjected eyes from all groups were averaged together and the average nuclei number was set as 100%. The eyes injected with pLUC1297 and pLUC1298, compared with uninjected eyes, show ed no significant difference as expected ( P =0.47, 0.37, respectively). The eyes injected with the ribozyme-expressing constructs showed significant reduction in pre-retinal neovascularization. pLUC1297HHHP showed 48% ( P =3.58x10-7) reduction; pLC1298HHHP showed 54% ( P =1.46x10-4) reduction; and pGE1298HHH P showed 59% (P=2.89x10-10) reduction.

PAGE 151

135 Figure 3.20. The proliferating e ndothelial cell -specific promoter limits expression of luciferase to the actively prolifera ting blood vessels in the OIR model.

PAGE 152

136 Figure 3.21. Quantitative assessment of the IGF1R ribozymes ability to inhibit preretinal neovascularization when expressed from the promoter. Figure 3.22. New promoter tested in adult mouse model of laser-induced neovascularization.

PAGE 153

137 The proliferating endothelial cell specific constructs were also tested in an adult mouse model of laser-induced neovasculariza tion. The eyes from adult mouse were also stained with endothelial cell specific agglutinin conjugated with rhodamine (red) and a secondary antibody bound to luciferase (gr een). A video clip was made from the animation of a stack of pictur es (400x) focused at same hor izontal position but different vertical levels, about 1m apart between leve ls, taken using a confo cal microscope. Panel A in Figure 3.22 is a snapshot of the clip. A z-projection view of the same vessel was made using imageJ and Panel B is a snapshot of the z-projection view. The green staining was not seen in the interstitia l space outside vasculature, but was colocalized with blood vessels (yellow color for colocalization). Also the green staining was in cell-like shape aligned on vessel walls, which indicated the lu ciferase expression o ccurred in endothelial cells in these small vessels. For a better view of the colocalizations, please see the supplementary movie clips. Object 3.1. A blood vessel from the adult mouse model shows the luciferase expression is specific for prolifera ting endothelial cells. Object 3.2. The 3-D view of the blood vessel from the adult mouse model. Figure 3.23. The expression of the IGF-1R ri bozyme from the new promoter reduced aberrant blood vessel formati on in the adult laser model.

PAGE 154

138 We also perfused retina from the adult mouse model with rhodamine-labeled dextran to examine the state of the vasculat ure on the retina (Figure 3.23). The left panel showed the normal vasculature with no abnor mal (leaky) neovascularization. The middle panel is the eye that had neovascularization i nduced by laser treatment, also injected with empty vector pLUC1298. The hazy areas indica ted the leaky small vessel resulting from the abnormal neovascularization. The right pane l is the eye treated by laser but injected with pLUC1298HHHP, the IGF-1R ribozyme expressing construct. Compared with normal retina we can still see some hazy ar eas but there were much less of them in quantity and the size of leaky areas than pLUC 1298-injected eye. Th is indicated the IGF1R ribozyme inhibited la ser-induced neovasculari zation to some extent. The New Promoter Tested with Integrin Ribozyme The proliferating endothelial cell specif ic promoter was used to express the integrin ribozymes in vivo Five mouse pups were injected with the plasmid in one eye on day 1 of the OIR mouse model as usual. The eye sections are shown in Figure 3.24. Eye A in Figure 3.24 is the section of an uninjected eye. Eyes B, C, D, E and F are sections from injected eyes. Many abnorma lities still existed, such as smaller size in some injected eyes (especially eyes C a nd D), detached retina from choroid, unusual folding in the retina. Noneth eless, compared with CMV/ -actin-driven expression of the integrin ribozymes, much less eye deformati on resulted. The lens in the injected eyes were normal. With the CMV/ -actin promoter no open iris was found on any cross sections. Now, most eyes have an open iris Eye D also had an open iris but the section shown was too close to the cornea/iris boundary to see the pupil. Eye E looked no difference with a normal eye.

PAGE 155

139 Figure 3.24. Expression of integrin ribozyme driven by proliferat ing endothelial cellspecific promoter resulted in less eye deformation.

PAGE 156

140 Figure 3.25. Proliferating endot helial cell specific promoter with integrin ribozyme tested in OIR model The neovascularization quantification re sults showed different levels of reductions in the injected eyes (24.3 17.1% to 91.4.9% as shown in Figure 3.25). Interesting, the most deformed eye (eye C) showed greatest reduction, while the most normal eye (eye E) showed least reduction.

PAGE 157

141 CHAPTER 4 DISCUSSION This project involved developing multiple ribozymes and testing them in vitro and/or in vivo Our primary focus was to use these ri bozymes to inhibit the expression of proteins that play important roles in the a bnormal retinal neovascularization. A number of proteins were chosen as our targets in this study, in cluding IGF-1R, IR, VEGFR-1, VEGFR-2, and integrins. They have differe nt functions but are all important for endothelial cell physiology such as proliferation and migrati on, which are essential in the development of neovascularization. The developing the testing steps were sim ilar to all these ribozymes. We selected target sites in protein gene sequence, designe d the ribozyme accordingl y, tested cleavage reactivity in vitro transfected and functionally analy zed ribozymes in HRECs regarding mRNA levels, protein levels, physiological func tions of HRECs, and eventually tested in mouse models in vivo Ribozyme Testing Results and Antisense Effect All of these ribozymes have been sh own to have cleavage reactivity and can reduce the expression of target proteins. All th e testing results have been summarized in tables in the previous chapter. Taking IGF-1R ribozyme 1 as an example, more than 90% of the target RNA oligos were cleaved within the first 2 minutes in the cleavage time course study, indicating the ri bozyme was highly catalytically active. After HRECs were transfected with the IGF-1R ribozyme, mRNA levels for IGF1-R were reduced by about 40% compared with vector-transfected cells, and the inactive version did not result in any

PAGE 158

142 significant reduction. The IGF-1R ribozyme decr eased protein expression by 48% and the inactive version also decreased the protein expression about 31%. This decrease with inactive ribozyme treatment may have been caused by its antisense effect, in which the catalytic-deficient ribozyme cannot cleave th e mRNA but still complimentarily binds to target site and physically blocks translation. This effect was not unexpected to exist in studies that involve pr otein expression and function. For example, in the migration study, we observed 91% reduction in the cells ability to migrate with active IGF-1R transfection and 58% reduction with inactive IGF-1R transf ection, not surprising for the active version. IGF-1R ribozymes 1 and 2 induced 65% and 52% reductions in the pre-retinal neovascularization levels. Th eir inactive forms also resulted in reductions; these reductions are minimal or close to significant ( P =0.03, P =0.09, respectively), so there was minimal antisense effect in the in vivo test. This is not consistent with in vitro studies in HRECs but our explanations are: 1) A threshold must be achieved in the reduction of protein expression to see a re duction in functional analysis. The threshold may differ in different species and vary with methods of assay. 2) The number of IGF-1R proteins differs substantially in different situati ons including but not limited to species, cell phases, study conditions ( in vitro vs. in vivo ), and so on. According to Rubini et al. [355], taking mouse fibroblasts as a reference, cells in proliferation have more than 30,000 IGF1R proteins per cell while quiescent cells only have 15,000 to 20,000 IGF-1R proteins. In our studies, most in vitro experiments were done with confluent cultured cells, but in vivo experiments were done in developing retinas. So the antisense effect might have been diluted in the mouse model. However th e active ribozymes still resulted in

PAGE 159

143 neovascularization reductions in that they can completely cleave and oblate the protein functions versus possibly partial function i nhibition with inactive ribozymes blockade. More importantly, the active ribozyme has cataly tic ability so it can process much more proteins than inactive version to phy sically block in a 1:1 molar ratio. VEGFR-1 and VEGFR-2 Interactions In the traditional view, VEGFR-1 functions as a decoy receptor and negatively regulates VEGFR-2 signaling. This is basi cally accomplished by VEGFR-1 acting as a sink, binding VEGF ligand, and preventing activation of VEGFR-2. However, some recent data indicate that the kinase activity of VEGFR-1 plays an essential role during pathological angiogenesis and in wound healing, by potentiati ng VEGFR-2 signaling [100, 356, 357]. It has been accepted that th ere is cross talk between VEGFR-1 and VEGF-R2 signaling. PI3 kinase [356] and nitr ic oxide [358]have been proposed to be involved in VEGFR-1 regulat ion of VEGFR-2. In our studies, we showed that transfection of HRECs with the VEGFR1 ribozyme down-regulated theVEGFR-2 mRNA. On the other hand, th e transfection of HRECs with VEGFR-2 ribozyme also downregulated VEGFR-1 mRNA. This is a fu rther support of intraand inter-molecular cross talk between the two receptors. In the OIR mouse model, the VEGFR-1 and VEGFR-2 ribozymes both significantly reduced pre-retinal neovascularization, by 47% and 75%, respectively. The inhibition on VEGF R-2 inhibited neovascularization to a greater extent, which is consistent with the major role of VEGFR-2 in promoting endothelial cell proliferati on, migration and therefore a ngiogenesis, even though the interactions between VEGFR-1 and VEGFR2 exist. Another recognized role for VEGFR-1 kinase activity is its capability of recruiting hematopoietic stem cells from bone marrow precursors [105, 359]. It is possibl e that the decrease in neovascularization

PAGE 160

144 found from blocking of VEGFR1 signaling may affect stem cell involvement. In one study, a chimeric protein containing both the VEGF binding domains of VEGFR-1 and VEGFR-2 was constructed and expressed in a murine model of ischemic retinopathy. A single intravitreal injection the chimeric protein resulted in a >90% reduction of retinal neovascularization compared w ith control eyes [360]. This suggested that a combined targeting of VEGFR-1 and VE GFR-2 may bring about a deep er reduction in retinal neovascularization than either receptor alone A ribozyme that can target both receptors or an administration of both ribo zymes together could be tested. Besides the interactions inside VEGF system, IGF-1 may also crosstalk with VEGF signaling. We have shown that intravitrea l injections of IGF -I result in an acute increase in vascular permeability and vascul ar engorgement, followed by development of pre-retinal angiogenesis in rabbit eyes [361] In addition, IGF-1 production by HRECs, in turn, stimulates increased VE GF production [217] and visa versa [362]. It was also reported that elevated IGF-1 levels in vivo resulted in an increase in VEGF gene expression. Considering that VEGF and IGF-1 and their receptors can all be expressed by HRECs and both these growth factors have autocrine and paracrine function, it is reasonably to propose that the interaction be tween these signaling systems do exist and a better picture of their involvement in pre-retinal neovasc ularization should include their interactions. Apart from endothelial cells VEGF and its receptors ar e expressed in many other cell types, such as inflammatory cells [363] VEGF may function in an autocrine fashion on these cells. Unlike diabetic retinopathy, the ischemia-induced retinopathy was thought to be inflammatory-free; however, it is now known that inflammatory cells are involved

PAGE 161

145 [364]. These inflammatory cells may participat e in the processes of blood-retinal barrier breakdown and neovascularization [115]. Th e plasmid expressing VEGF receptor ribozymes has a CMV/ -actin promoter and does not excl usively target endothelial cells, so it is possible that the inhibition of VEGF signaling on other cell t ypes also contributed to the overall outcome of neovascularization reduction. The Proliferating Endothelial Cell Specific Promoters Expression of the IGF-1 ribozyme by the promiscuous CMV/ -actin promoter did not result in eye deformation like the inte grin ribozymes. However, IGF-1 and its receptor still play a major role in vascular development of both mouse and human eyes, thus the indiscriminant loss of IGF-IR coul d result in altered vascular development and propagate the ischemia observed in ROP infa nts or OIR mice [365]. Therefore, we also used the cell specific promoter to express the IGF-1R ribozyme in vitro and in vivo Our results demonstrated that the cell specific promoter limited expr ession to HRECs and no expression in fibroblasts. Our in vivo results also showed that the new cell specific promoter is active in the rapidly dividing vasculature of the eye. During construction of the new IGF-1R ribozyme plasmid, it was observed that ligation of the IGF-IR hammerhead ribozyme and the hairpin ribozyme caused the endothelin enhancers to be dele ted in several strains of bact eria, DH5, Stable 2s and Top Tens [365]. However, when placed at th e 3 end of the luciferase gene, the hammerhead/hairpin ribozyme insertion did not cause the deleti on of the endothelin enhancer. Thus the luciferase gene was also playing a role in stabilizing the construct. The insertion of the ribozymes made the vector lose the PolyA tail, which affected the stability of the transcript and reduced the expression levels of lu ciferase and IGF-1R

PAGE 162

146 ribozyme. This was confirmed in vitro in HRECs (Figure 3.17). The luciferase expression from these two plasmids was lower than their parent vectors (pLUC1297, pLUC1298). The in vivo study showed that the lucifera se expression fr om pLUC1298HHHP was exclusive limited to proliferating endotheli al cells. The colocaliz ation of luciferase and proliferating endothelia l cells was observed both in the OIR mouse model and the adult mouse model of laser-induced neovascul arization. However, the reduction of preretinal neovascularization from pLUC1298HHHP (54%) we re comparable with the reduction found with IGF-1R ribozyme driven by CMV/ -actin promoter (65%). In addition, expression of the IGF-1R riboz yme from the plasmid pGE1298HHHP, a modified version of pLUC1298HHHP with the de letion of luciferase gene, resulted in 59% reduction in pre-retina l neovascularization. Therefore the luciferase gene did not affect the expression of ribozymes. It is known that systemic administration of plasmid DNA alone by hydrodynamic administration results in initial high levels of expression 24hrs after injection and decreases to 7% of the peak value by day 10 [366, 367]. In our experiments, the luciferase expression and ribozyme activity wa s observed 17 days after administration in OIR model and 21 days after administration in adult mouse model. It is worth mentioning that the expression of luciferase, and by exte nsion of the IGF-1R ribozyme, from naked plasmid in OIR model or formulated plasmid in the adult mouse model exhibited significant expression th rough the time courses of the experiments. The idea of introducing a promoter that is specific for prolif erating endothelial cells originated from the integrin ribozyme in vivo study, since the ubiquitous knockdown of integrin resulted in severe eye deforma tion in the OIR mouse model. The expression of

PAGE 163

147 the same integrin ribozyme (against integrin 1 subunit) driven by the specific promoter showed much fewer problems. However, the de formations were still significant. But due to the specificity of the promoter, these deformations must result from affecting endothelial cells at the rapidl y proliferating vasculature of the eye. Even though decreases in abnormal neovascularization were found w ith the cell specific promoter, further refinement of the promoter, if possible, is required when e xpressing the integrin ribozymes. These problems, found with the integrin ribozymes result from the roles of the 1 integrin subunit in numerous processe s including their direct involvement in angiogenesis. These problems were not f ound with the IGF-1R ri bozyme with either promoter type due to the limited function of th is receptor in the developing vasculature of the eye. Therefore, while study of the integr in ribozyme will be useful from functional and developmental points of view, the use of integrins as therapeutic targets probably has limited or little value in the developing eye. Other Voices on Neovascularization in Diabetic Retinopathy As summarized earlier in the introducti on chapter, there is a tendency to propose that the abnormal neovascularization in diab etic retinopathy is the chronic pathological consequence of hypoxia in the retina. Howe ver, oxygen is the not the only nutrient supplied through blood vessels. Clinically the retinal angiography of diabetic re tinopathy patients shows nonperfused capillaries [368], whic h is indicating that hypoxia taki ng effect. It has also been reported that hyperoxia improved contrast sensitivity in early diabetic retinopathy [369] and that the supplemental oxygen improved di abetic macular edema [370]. However no studies have directly demonstrated reducti on of retinal oxygen levels in humans with diabetes compared with cont rols [368]. In the animal st udies, there was no significant

PAGE 164

148 difference in the pre-retina l oxygenation found between the cats [371] and dogs [372] within 1 year of diabetic ons et and the controls. But in th e long-term study, one group has reported that the retinal oxygen partial pressu re was reduced in cats with 6-8 years of diabetes [373]. Despite the direct eviden ce in the long-term cat study, most other supporting data of hypoxia are based on the overexpression of gr owth factors that are regulated by HIF, such as VEGF and PDGF. Even HIF activity was increased in diabetic rats [374], it is not necessary that the increases in growth factor levels are di rectly linked to HIF. In the diabetic retinas, besides the vascular cells, many other cell types are affected or harmed, including neurons, glial cells and microglial ce lls [375]. It is possible that the need to maintain neuron-dependent vision motivates an giogenesis to compensate for the nutrient deficiency in the neural retina In detail, VEGF could increas e initially to provide trophic support to neurons through VEGF receptors, but at the cost of increased vascular permeability [368]. The physiological compen sation response could convert into a pathological one in a chronic stress situation, and eventually lead to neovascularization and edema due to vascular leakage. One common problem in hypoxic and nut rition deficient cel ls is endoplasmic reticulum (ER) stress. ER stress can influe nce VEGF and PEGF expression levels [376] and thus affect the balance between cell su rvival and death signals. So it is not unreasonable to hypothesize that ER stress could be a potential target in the treatment of neovascular diabetic retinopathy. Final Words on RNA Silencing We used hammerhead ribozymes as a tool to inhibit gene expression. Ribozymes are only one category of RNA silencing tec hnologies. Gene silencing with antisense

PAGE 165

149 oligonucleotides is the earliest discovered and utilized, the easiest to design and has no target sequence requirement other than ta rget accessibility. Bu t, one significant disadvantage is that antisense oligonucleotides function in a 1:1 molar ratio with target mRNAs. This means a significant reduction in translation level may not be achieved without a fairly large amount of antisense oligonucleotides. Ribozymes, however, can catalyze the cleavage of target mRNAs and w ill be recycled and reused again, thus the dose of the RNA silencing agents can be si gnificantly reduced. However, the sequence requirement for ribozymes is the major obsta cle in the developmen t into convenient RNA silencing tools. RNAi, an endogenous and ubiquitous pathway, doesnt have much sequence requirement on targets. RNAi has othe r advantages such as ease of design, ease of synthesis and high specificity. Silencing w ith RNAi has been reported to exceed what can be achieved by antisense oligonucleotide s or ribozymes [377, 378]. In one head-tohead comparison, it has been shown th at siRNAs knocked down gene expression hundreds of time more efficiently than antis ense oligonucleotides [379]. RNAi is an attractive alternative as the ge ne silencing tool in my study. The antisense oligonucleotides have been studied intensively for the longest time and the first antisense DNA agent is now on th e market in the USA and Europe. However its mechanism is still not without contr oversy and it has been proposed that the therapeutic outcome could be a result of the CpG presence and the consequent immune stimulation in some cell types [380]. Two clinic al trials using ribozymes in gene therapy are in progress, in which retr oviral vectors, which express ribozymes targeting sequences in human HIV-1 RNA, are transduced in CD4 lymphocytes or CD34 hematopoietic precursors [381, 382]. Just three years after RNAi was shown to work in mammalian

PAGE 166

150 cells, the first Phase I clinical trials using R NAi have started in which RNAi is used to target the VEGF angiogenic pathway in ARMD patients. No evidence for clinical toxicity or disease progression has been shown in th ese studies conducted by Sirna Therapeutics [292]. These studies indicate that RNA silenc ing tools, especially ribozymes and RNAi, have a great potential to be used as therapeutic agents. New ribozyme types have been discove red. One recent repo rt indicated the existence of a metalolite-respons ive ribozyme in the mRNA of glm S, the Bacillus subtilis gene that encodes flucosamine frustose -6-phosphate aminotransferase [383]. The cleavage product is terminated by a cyclic 2-3 phosphate, very similar to the products of other self cleaving ribozymes, s uggesting that the transesterif ication reaction involves the nucleophilic attack from the 2 -oxygen. In another report, an element in the 3-flanking region of human -globin mRNA has been found that self cleaves [384]. The cleavage site is contained within a re gion that shows some similarity to the hammerhead ribozyme; however the 3-hydroxyl and 5-phosphate te rmini generated in the cleavage reaction imply a different mechanism of cleaving from other self cleaving ribozymes. As mentioned above, an HIV-directed hammerhead ribozyme has been tested in patients to exploit its ability to inhibit HIV replication [385]. The clinical trial is performed in ex vivo in which the peripheral blood T lymphocytes obtained from the HIV-infected patients are transduced with a retroviral vector coding a hammerhead ribozyme against HIV RNA expression. The tran sduced lymphocytes are injected back into HIV patients. The results showed that th e infusion of gene-altered, activated T-cells is safe, that the transduced cells persist fo r long intervals and the possible patient longterm survival resulting from the transduced cells [385]. In another study targeted against

PAGE 167

151 hepatitis C virus (HCV) replication, six hamme rhead ribozymes were designed that are targeted a conserved region of the plus and minus strands of the HCV genome and were expressed using recombinant adenovirus vector s. Testing in primary hepatocytes obtained from HCV-infected patients showed a benefici al antiviral effect of the ribozymes, and when used together with type 1 interferon, th e replication of HCV-po livirus (PV) chimera was inhibited up to 98% [386]. In anothe r study, synthetic and modified hammerhead ribozymes targeting 15 conserved sites at th e 5-untranslated region of HCV RNA were also tested for knock-down efficiency a nd stability [387], and a significant reduction (40%-90%) in gene expression of a reporter gene following the 5 untranslated region was observed [387]. Ribozymes targeted agains t hepatitis B virus (HBV) has also been proposed [388]. Modified hepatitis delta virus has been used to target HBV virus through its natural tropism to hepatocytes and the result of transgene de livery was positive [389, 390]. In cancer therapy, fusion proteins have been suggested as a target for ribozymes. The fusion proteins are expressed from ch imeric genes resulting from abnormal chromosomal translocations, which shuffle th at translocated exons and produce chimeric mRNAs [388]. These are tumor-specific chro mosomal abnormalities and only exist in the tumor cells [391], thus they provide a tumor cell-specific target and the normal cells are not targeted. These type of stra tegies could help to increase the effectiveness of current cancer treatments. In this study the in vivo application of the ribozymes required sufficient expression and stability of the ribozymes to survive the time course of the two animal models. The CMV/ -actin promoter produced qualitat ive expression of a GFP reporter gene beginning on day P11 of the mouse OIR model and extended beyond day P17 (data

PAGE 168

152 not shown). Thus it is expected that the hammerhead ribozy mes are also expressed in a similar manner and can meet the timeframe demand for expression in OIR mouse model. It has been reported that synt hetic siRNA, transfected into human cells, show an optimal effect around 24 hours and the RNAi starts to diminish in 4-7 days [392]. So if we use synthetic RNAs (either ribozymes or siRNAs ) as the gene silencing tool in the OIR mouse model, the synthetic siRNA will probably not last through the 17-day time course of the experiment. However, a vector e xpressing shRNA, similar to our ribozyme expression vectors, could be used. In one study, the AAVcloned shRNA introduced in mouse brain started to silence its target in 4-6 days and the silencing phenotype (Parkinsons disease) reached its peak ar ound two weeks and persisted for nearly two months [393]. Therefore it is probable that vector-expressed shRNAs could be successfully used in the OI R mouse model or even the adult mouse model of laserinduced retinal neovascularization, where the experiment termination is 3-4 weeks after injection. In our in vitro tests of the ribozymes, the ribozyme effects were determined by assaying both mRNA and protein levels us ing relative quantitative RT-PCR, real-time RT-PCR, western analysis and flow cytometry. We assume the transfection efficiency is consistent throughout all expe riments thus did not measur e it every time. However, variations may occur. Cell death resulting fr om transfections could also happen, which may not be the same for the transfection of empty vector, mock transfection, or ribozyme transfection. The normalization with living cell numbers may be useful in a fined measurement of reduced mRNA or proteins levels. In another aspect, the targeted protein could have a relative long half -life thus a modest drop in pr otein expression might not be

PAGE 169

153 seen in a short-period of time post-transf ection. Cullen suggested approach of the introduction of an expression plasmid encodi ng an epitope-tagged form of the target protein [303]. A western blotting using the an tibody against the epitope will be performed to measure the gene silencing effectiveness. In this way, the co-transfection efficiency is technically 100% same for silencing agents and the targets and protein half-life is not an issue any more. The construction of the epitope -tagged form of the targeted protein is laborious and time-consuming; however, this approach could give us an accurate measurement on the effectiveness of the silencin g agents. This is especially important in selecting a best-effective silencing agent for therapeutic purpose. The eye is an ideal target organ for gene therapy, in that it has relatively isolated compartment so that the local delivery of exogenous genes to the eye limits exposure to the rest of the body and reduces the dose. Sim ilar to our injection of ribozymes into the eye, siRNAs have also been injected into th e eye intravitreally and were readily diffused throughout the eye and detectable for at leas t five days [394]. VEGF and its receptors have been attractive targets in RNAi in vivo studies in the eye so far. In one study, hVEGF cDNA, expressed by an adenoviral vector, was subretinally inje cted in both eyes of mice. This was coupled with an siRNA ta rgeting against hVEGF mRNA in one eye, or siRNA targeting against GFP mRNA in the other eye. It was showed that eye injected with hVEGF siRNA had significantly less expr ession of hVEGF comp ared with the GFP siRNA control [395]. In the mouse model of CNV induced by laser photocoagulation, the area of CNV at sites of rupture of Bruchs membrane was significantly less in the eyes that were subretinally inject ed with mVEGF siRNA, compared with GFP siRNA controls [168]. These results directly lead to phase I clinical trials of si RNA against VEGF mRNA

PAGE 170

154 in ARMD patients with subfoveal CNV. In th e studies for corneal neovascularization, a systemic administration of siRNAs agai nst VEGF-A, VEGFR-1, or VEGFR-2 using a polymer delivery system was conducted [396] The polymer was composed of branched polyethylenimine (PEI) as one end, polyethylene glycol (PEG) in the middle and an RGD peptide motif at the other end. This tri-f unctional polymer can self assemble with negatively charged siRNA into a nanopartic le and RGD peptide will be exposed on the surface. The RGD peptide, a specific ligand for v 3 and 5 1 integrins on activated endothelial cells, can introduce the expression of siRNA to the neovasculature. PEF helps to prevent nonspecific binding to other tissu es. Thus the siRNA is delivered via ligandmediated endocytosis. The level of corneal neovascularization was significantly reduced with the administration of VEGF-A, VEGFR-1, or VEGFR-2 siRNAs, and the combination of all the three resulted in further reduction. In a nother study, an siRNA against VEGFR-1 called Sirna-27, was tested and found to maximally reduce VEGFR-1 levels in cultured endothelia l cells compared w ith other siRNA candidates [394]. This siRNA was further examined in mous e models of retinal and choroidal neovascularization. Sirna-027 significantly reduced VEGFR-1 mRNA levels by 57% or 40% after intravitr eal or periocular inj ection, respectively, as m easured by quantitative RT-PCR. In the CNV mouse model, the ar ea of neovascularization was decreased by 45% to 66% after the periocular or intrav itreous injection of Sirna-27. And in the ischemic retinopathy mouse model, the intrav itreous injection of 1.0 g of Sirna-027 significantly reduced retinal neovascularization [397]. All these studies used VEGFR-1 siRNAs and demonstrated that VEGFR-1 has an important role in stimulating ocular

PAGE 171

155 neovascularizations, which further argues ag ainst the hypothesis that VEGFR-1 is only a decoy receptor that negatively regulates the activity of VEGFR-2. Apart from eye diseases, siRNAs ta rgeting against the VEGF pathway, or angiogenesis, have been studied in the tr eatment of cancer, inflammation, and so on [398]. The growth hormones, their receptors signaling transduction factors, matrix metalloproteases and adhesion molecules have al l been used as the RNAi targets. These studies provide strong support that RNAi can be used in novel anti-angiogenesis therapies, from bench to bed.

PAGE 172

156 APPENDIX A LIST OF ABBREVIATIONS AAV Adeno-associated virus ABAM Antibiotic/antimycotic mix Ad Adenovirus AGE Advanced glycation end-products ALS Acid labile acid Ang-1 Angiopoeitin 1 Ang-2 Angiopoeitin 2 ARMD Age Related Macular Degeneration ARVO Association for Research in Vision and Ophthalmology ATP Adenosine triphosphate bFGF Basic fibroblast growth factor BFP Blue fluorescence protein BRB Blood-retinal barrier BrdU Bromo-uridine BSA Bovine serum albumin CAR Coxsackievirus-adenovirus receptor CFP Cyan fluorescence protein cGMP Cyclic guanosine 3,5-monophosphate CHO Chinese hamster ovary cells CMCT 1-cyclohexyl-(2-morpho linoethylo)cabodiimide methop -toluene sulfonate CMV Cytomegalovirus CNV Choroidal neovascularization DAG Diacylglycerol DMEM Dubellcos modifeid eagle medium DMSO Dimethyl sulfoxide DNA Deoxyribonucleic acid DR Diabetic retinopahty DTT Dithiothreitol EC Endothelial cells ECL Enhanced chemiluminescence ECM Extracellular matrix EDTA Ethylenediamine tetraacetic acid EGFR Epidermal growth factor receptor EGS External guide sequence eNOS Endothelial NO synthase ER Endoplasmic reticulum ET Endothelin PAF Platelet-derived factor

PAGE 173

157 FAK Focal adhesion kinase FBS Fetal bovine serum FGF Fibroblast growth factor GAGs Glycosaminoglycans GAS -interferon activated sequence GC Guanosine cytosine content GCL Ganglion cell layer GFP Green fluorescent protein GH Growth hormone GLUT1 Glucose transporter 1 HBSS Hanks balanced salt solution HBV Hepatitis B virus HCV Hepatitis C virus HDV Hepatitis delta virus H&E Hematoxylin-eosin HEK 293 Human embryonic kidney cells HIF Hypoxia inducible factor HIV Human immunodeficiency virus HRE Hypoxia response element HRECs Human retinal endothelial cells HSPGs Heparin sulfate proteoglycans HSV-1 Herpes simplex virus type 1 IACUC Institution Animal Care and Use Committee. IAP Integrin-associated protein ICAM-1 Intercellular adhesion molecule-1 Ig Immnoglobin IGF-1 Insulin-like growth factor 1 IGF-1R Insulin-like grow th factor 1 receptor IGF-2 Insulin-like growth factor 2 IGFBP Insulin-like growth factor binding protein ILM Inner limiting membrane INL Inner nuclear layer IP3 Inositol 1,4,5-triphosphate IPL Inner plexiform layer IR Insulin receptor IRS Insulin receptor substrate ITR Inverted terminal repeats LTR Long terminal repeats miRNA microRNA NAD Nicotinamide adenine dinucleotide NADPH Nicotinamide-adenine dinucleotide phosphate NFB Nuclear factorB NFL Nerve fibre layer NO Nitric oxide NOS Nitric oxide synthase NPDR Non proliferative diabetic retinopathy

PAGE 174

158 NRP Neurophilins OIR Oxygen-induced retinoppathy OLM Outer limiting membrane ONL Outer nuclear layer OPL Outer plexiform layer. ORF Open-reading frame PBS Phosphate buffered saline PDGF Platelet-derived growth factor PDR Proliferative diabetic retinopathy PEDF Pigment epithelium-derived factor PEG Polyethylene glycol PEI Polyethylenimine PIP2 Phosphatidylinositol 4,5-bisphosphate PKB Protein kinase B PKC Protein kinase C PLC Phospholipase C PlGF Placental growth factor PS Phosphorothioate PTB Phosphotyrosine binding PTGS Post-transcriptional gene silencing ptRNA Precursor tRNA rAAV Recombinant ade no associated virus RA Retinoic acid RAGE Advanced glycation end-products receptor RAR Retinoic acid receptor RFP Red fluorescence protein RGD Arginine-glyci ne-asparginine RISC RNA-inducing silencing complex ROP Retinopathy of prematurity ROS Reactive oxygen species RNA Ribonucleic acid RNAi RNA interference rRNA Ribosomal RNA RNasin Ribonuclease inhibitor RPE Retinal pigment epithelium RT Reverse transcription RXR Retinoid X receptor Rz Ribozyme scAAV Self complementary AAV SHP-2 Src homology 2 contai ning tyrosine phosphatase shRNA Short hairpin RNA siRNA Small interfering RNA SMCs Smooth muscle cells SnRNA Small nuclear RNA STAT Signal transducer and activator of transcription TBS Tris buffered saline

PAGE 175

159 T RV Type V receptor for tr ansforming growth factorTGF Transforming growth factor Tie 1 and 2 Angiopoeitin receptors 1 and 2 TNF Tumor necrosis factor tPA Tissue type plasminogen activator tRNA Transfer RNA TR Inverted terminal repeats TRS Terminal resolution site uPA Urokinase type plasminogen inhibitor VE cadherin Vascular endothelial cadherins VEGF Vascular endothe lial growth factor VEGFR-1 Vascular endothelial growth factor-receptor 1 VEGFR-2 Vascular endothe lial growth-receptor 2 VPF Vascular permeability factor VS Varkud satellite VSMCs Vascular smooth muscle cells YFP Yellow fluorescence protein

PAGE 176

160 LIST OF REFERENCES 1. Evans, J, Causes of blindness and partial sight in England and Wales 1990 1995, London: Her's Majesty's Stationery Office. 2. Klein, R, Klein, B E K, and Linton, K L P, Prevalence of Age-Related Maculopathy the Beaver Dam Eye Study. Ophthalmology, 1992. 99 (6): p. 933943. 3. Mitchell, P, Smith, W, Attebo, K, and Wang, J J, Prevalence of Cage-Related Maculopathy in Aust ralia the Blue M ountains Eye Study. Ophthalmology, 1995. 102 (10): p. 1450-1460. 4. Ambati, J, Ambati, B K, Yoo, S H, Ianchulev, S, and Adamis, A P, Age-related macular degeneration: Etiology, pathoge nesis, and therapeutic strategies. Survey of Ophthalmology, 2003. 48 (3): p. 257-293. 5. W.R. Green, S N K, 3rd, Senile macular degenerati on: a histopathologic study. Trans Am Ophthalmol Soc, 1977. 75 : p. 180-254. 6. Sunness, J S, Massof, R W, Johnson, M A, Finkelstein, D, and Fine, S L, Peripheral Retinal Function in Ag e-Related Macular Degeneration. Archives of Ophthalmology, 1985. 103 (6): p. 811-816. 7. Sunness, J S, Rubin, G S, Applegate, C A, Bressler, N M, Marsh, M J, Hawkins, B S, and Haselwood, D, Visual function abnormalities and prognosis in eyes with age-related geographic atrophy of the macula and good visual acuity. Ophthalmology, 1997. 104 (10): p. 1677-1691. 8. Dunaief, J L, Dentchev, T, Ying, G S, and Milam, A H, The role of apoptosis in age-related macular degeneration. Archives of Ophthalmology, 2002. 120 (11): p. 1435-1442. 9. Miller, H, Miller, B, and Ryan, S J, The Role of Retinal-Pigment Epithelium in the Involution of Subretinal Neovascularization. Investigative Ophthalmology & Visual Science, 1986. 27 (11): p. 1644-1652. 10. Green, W R, and Enger, C, Age-Related Macular Degeneration Histopathologic Studies the 1992 Zimmerman,Lorenz,E Lecture. Ophthalmology, 1993. 100 (10): p. 1519-1535.

PAGE 177

161 11. Witmer, A N, Vrensen, G F J M, Van Noorden, C J F, and Schlingemann, R O, Vascular endothelial growth fact ors and angiogenesis in eye disease. Progress in Retinal and Eye Research, 2003. 22 (1): p. 1-29. 12. Campochiaro, P A, Retinal and choroidal neovascularization. Journal of Cellular Physiology, 2000. 184 (3): p. 301-310. 13. Kliffen, M, Sharma, H S, Mooy, C M, Kerkvliet, S, and deJong, P T V M, Increased expression of angiogenic growth factors in age-related maculopathy. British Journal of Ophthalmology, 1997. 81 (2): p. 154-162. 14. Schwesinger, C, Yee, C, Rohan, R M, J oussen, A M, Fernandez, A, Meyer, T N, Poulaki, V, Ma, J J K, Redmond, T M, Liu, S Y, Adamis, A P, and D'Amato, R J, Intrachoroidal neovascularization in transgenic mice overexpressing vascular endothelial growth factor in the retinal pigment epithelium. American Journal of Pathology, 2001. 158 (3): p. 1161-1172. 15. Wells, J A, Murthy, R, Chibber, R, Nunn, A, Molinatti, P A, Kohner, E M, and Gregor, Z J, Levels of vascular endothelial growth factor are elevated in the vitreous of patients with s ubretinal neovascularisation. British Journal of Ophthalmology, 1996. 80 (4): p. 363-366. 16. Anderson, D H, Mullins, R F, Hageman, G S, and Johnson, L V, Perspective A role for local inflammation in the fo rmation of drusen in the aging eye. American Journal of Ophthalmology, 2002. 134 (3): p. 411-431. 17. Hutcheson, K, Retinopathy of prematurity. Current Opinion Ophthalmology, 2003. 14 (5): p. 286-290. 18. Zhang, Y F, and Stone, J, Role of astrocytes in the control of developing retinal vessels. Investigative Ophthalmol ogy & Visual Science, 1997. 38 (9): p. 1653-1666. 19. Aiello, L M, Perspectives on diabetic retinopathy. American Journal of Ophthalmology, 2003. 136 (1): p. 122-135. 20. Caldwelll, R B, Bartoli, M, Behzadian, M A, El-Remessy, A E B, Al-Shabrawey, M, Platt, D H, and Caldwell, R W, Vascular endothelial gr owth factor and diabetic retinopathy: pathophysiological mechani sms and treatmen t perspectives. DiabetesMetabolism Research and Reviews, 2003. 19 (6): p. 442-455. 21. Hinz, B J, de Juan, E, and Repka, M X, Scleral buckling surgery for active stage 4A retinopathy of prematurity. Ophthalmology, 1998. 105 (10): p. 1827-1830. 22. Capone, A, and Trese, M T, Lens-sparing vitreous surger y for tractional stage 4A retinopathy of prematurit y retinal detachments. Ophthalmology, 2001. 108 (11): p. 2068-2070.

PAGE 178

162 23. Group, T D C a C T R, The Effect of Intensive Treatment of Diabetes on the Development and Progression of Long-Term Complications in Insulin-Dependent Diabetes-Mellitus. New England Journal of Medicine, 1993. 329 (14): p. 977-986. 24. Grp, U P D S, Tight blood pressure control and risk of macrovascular and microvascular complications in type 2 diabetes: UKPDS 38. British Medical Journal, 1998. 317 (7160): p. 703-713. 25. Alder, V A, Su, E N, Yu, D Y, Cringle, S J, and Yu, P K, Diabetic retinopathy: Early functional changes. Clinical and Experimental Pharmacology and Physiology, 1997. 24 (9-10): p. 785-788. 26. Ban, Y, and Rizzolo, L J, Regulation of glucose transpor ters during development of the retinal pigment epithelium. Developmental Brain Research, 2000. 121 (1): p. 89-95. 27. Badr, G A, Tang, J, Ismail -Beigi, F, and Kern, T S, Diabetes downregulates GLUT1 expression in the retina and its microv essels but not in the cerebral cortex or its microvessels. Diabetes, 2000. 49 (6): p. 1016-1021. 28. Sone, H, Deo, B K, and Kumagai, A K, Enhancement of glucose transport by vascular endothelial growth fact or in retinal endothelial cells. Investigative Ophthalmology & Visual Science, 2000. 41 (7): p. 1876-1884. 29. Brownlee, M, Biochemistry and molecular cell bi ology of diabetic complications. Nature, 2001. 414 (6865): p. 813-820. 30. Williamson, J R, Chang, K, Frangos, M, Hasan, K S, Ido, Y, Kawamura, T, Nyengaard, J R, Vandenenden, M, Kilo, C, and Tilton, R G, Hyperglycemic Pseudohypoxia and Diabetic Complications. Diabetes, 1993. 42 (6): p. 801-813. 31. Setter, S M, Campbell, R K, and Cahoon, C J, Biochemical pathways for microvascular complications of diabetes mellitus. Annals of Pharmacotherapy, 2003. 37 (12): p. 1858-1866. 32. Swidan, S Z, and Montgomery, P A, Effect of blood glucose concentrations on the development of chronic complications of diabetes mellitus. Pharmacotherapy, 1998. 18 (5): p. 961-972. 33. Dyck, P J, and Giannini, C, Pathologic alterations in the diabetic neuropathies of humans: A review. Journal of Neuropathology an d Experimental Neurology, 1996. 55 (12): p. 1181-1193. 34. Brownlee, M, Nonenzymatic Glycosylation of Ma cromolecules Prospects of Pharmacological Modulation. Diabetes, 1992. 41 : p. 57-60.

PAGE 179

163 35. Haitoglou, C S, Tsilibary, E C, Brownlee, M, and Charonis, A S, Altered Cellular Interactions between Endothelial-Cells and Nonenzymatically Glucosylated Laminin Type-Iv Collagen. Journal of Biological Chemistry, 1992. 267 (18): p. 12404-12407. 36. Cai, H, and Harrison, D G, Endothelial dysfunction in cardiovascular diseases The role of oxidant stress. Circulation Research, 2000. 87 (10): p. 840-844. 37. Laight, D W, Carrier, M J, and Anggard, E E, Antioxidants, diabetes and endothelial dysfunction. Cardiovascular Research, 2000. 47 (3): p. 457-464. 38. Baynes, J W, Role of Oxidative Stress in Development of Complications in Diabetes. Diabetes, 1991. 40 (4): p. 405-412. 39. Taylor, A A, Pathophysiology of hypertension an d endothelial dysfunction in patients with diabetes mellitus. Endocrinology and Metabo lism Clinics of North America, 2001. 30 (4): p. 983-+. 40. SiflingerBirnboim, A, Lum, H, DelVecchio, P J, and Malik, A B, Involvement of Ca2+ in the H2O2-induced increase in endothelial permeability. American Journal of Physiology-Lung Cellular and Molecular Physiology, 1996. 14 (6): p. L973L978. 41. Way, K J, Katai, N, and King, G L, Protein kinase C and the development of diabetic vascular complications. Diabetic Medicine, 2001. 18 (12): p. 945-959. 42. Ishii, H, Koya, D, and King, G L, Protein kinase C activation and its role in the development of vascular complic ations in diabetes mellitus. Journal of Molecular Medicine-Jmm, 1998. 76 (1): p. 21-31. 43. Xia, P, Aiello, L P, Ishii, H, Jiang, Z Y, Park, D J, Robinson, G S, Takagi, H, Newsome, W P, Jirousek, M R, and King, G L, Characterization of vascular endothelial growth factor's ef fect on the activation of pr otein kinase C, its isoforms, and endothelial cell growth. Journal of Clinical Investigation, 1996. 98 (9): p. 20182026. 44. Bloomgarden, Z T, The epidemiology of complications. Diabetes Care, 2002. 25 (5): p. 924-932. 45. Imai, T, Morita, T, Shindo, T, Nagai, R, Yazaki, Y, Kurihara, H, Suematsu, M, and Katayama, S, Vascular smooth muscle cell-directed overexpression of heme oxygenase-1 elevates blood pressure thr ough attenuation of nitric oxide-induced vasodilation in mice. Circulation Research, 2001. 89 (1): p. 55-62. 46. Chakrabarti, S, Cukiernik, M, Hileeto, D, Evans, T, and Chen, S, Role of vasoactive factors in the pathogene sis of early changes in diabetic retinopathy. DiabetesMetabolism Research and Reviews, 2000. 16 (6): p. 393-407.

PAGE 180

164 47. Stevens, M J, Henry, D N, Thomas, T P, Killen, P D, and Greene, D A, Aldose Reductase Gene-Expression and Osmotic Dysregulation in Cultured Human Retinal-Pigment Epithelial-Cells. American Journal of Physiology, 1993. 265 (3): p. E428-E438. 48. Stitt, A W, Advanced glycation: an important pa thological event in diabetic and age related ocular disease. British Journal of Ophthalmology, 2001. 85 (6): p. 746753. 49. Moller, P, Loft, S, Lundby, C, and Olsen, N V, Acute hypoxia and hypoxic exercise induce DNA strand breaks and oxidative DNA damage in humans. Faseb Journal, 2001. 15 (7): p. 1181-1186. 50. Jerdan, J A, Michels, R G, and Glaser, B M, Diabetic Preretinal Membranes an Immunohistochemical Study. Archives of Ophthalmology, 1986. 104 (2): p. 286290. 51. Bek, T, and Ledet, T, Glycoprotein deposition in va scular walls of diabetic retinopathy A histopa thological and immunohi stochemical study. Acta Ophthalmologica Scandinavica, 1996. 74 (4): p. 385-390. 52. Nishikawa, T, Giardino, I, Ed elstein, D, and Brownlee, M, Changes in diabetic retinal matrix protein mRNA levels in a common transgenic mouse strain. Current Eye Research, 2000. 21 (1): p. 581-587. 53. Jian, B, Jones, P L, Li, Q Y, M ohler, E R, Schoen, F J, and Levy, R J, Matrix metalloproteinase-2 is associated with te nascin-C in calcific aortic stenosis. American Journal of Pathology, 2001. 159 (1): p. 321-327. 54. Cai, J, and Boulton, M, The pathogenesis of diabetic retinopathy: old concepts and new questions. Eye, 2002. 16 (3): p. 242-260. 55. Carmeliet, P, Angiogenesis in health and disease. Nature Medicine, 2003. 9 (6): p. 653-660. 56. Gariano, R F, Cellular mechanisms in retinal vascular development. Progress in Retinal and Eye Research, 2003. 22 (3): p. 295-306. 57. Luttun, A, Carmeliet, G, and Carmeliet, P, Vascular progenitors: From biology to treatment. Trends in Cardiova scular Medicine, 2002. 12 (2): p. 88-96. 58. Rafii, S, Lyden, D, Benezra, R, Hattori, K, and Heissig, B, Vascular and haematopoietic stem cells: Novel ta rgets for anti-angiogenesis therapy? Nature Reviews Cancer, 2002. 2 (11): p. 826-835. 59. Asahara, T, and Isner, J M, Endothelial progenitor cells for vascular regeneration. Journal of Hematotherapy & Stem Cell Research, 2002. 11 (2): p. 171-178.

PAGE 181

165 60. Grant, M B, May, W S, Caballero, S, Br own, G A J, Guthrie, S M, Mames, R N, Byrne, B J, Vaught, T, Spoerri, P E, Peck, A B, and Scott, E W, Adult hematopoietic stem cells provide functi onal hemangioblast activ ity during retinal neovascularization. Nature Medicine, 2002. 8 (6): p. 607-612. 61. Otani, A, Kinder, K, Ewalt, K, Otero, F J, Schimmel, P, and Friedlander, M, Bone marrow-derived stem cells target retinal astrocytes and can promote or inhibit retinal angiogenesis. Nature Medicine, 2002. 8 (9): p. 1004-1010. 62. Folkman, J, Anti-Angiogenesis New Concept for Therapy of Solid Tumors. Annals of Surgery, 1972. 175 (3): p. 409-&. 63. Liotta, L A, Steeg, P S, and Stetlerstevenson, W G, Cancer Metastasis and Angiogenesis an Imbalance of Positive and Negative Regulation. Cell, 1991. 64 (2): p. 327-336. 64. Pugh, C W, and Ratcliffe, P J, Regulation of angiogenesis by hypoxia: role of the HIF system. Nature Medicine, 2003. 9 (6): p. 677-684. 65. Kourembanas, S, Hannan, R L, and Faller, D V, Oxygen-Tension Regulates the Expression of the Platelet-Derived Gr owth Factor-B Chain Gene in Human Endothelial-Cells. Journal of Clinical Investigation, 1990. 86 (2): p. 670-674. 66. Shweiki, D, Itin, A, Soffer, D, and Keshet, E, Vascular Endothel ial Growth-Factor Induced by Hypoxia May Mediate Hypoxia-Initiated Angiogenesis. Nature, 1992. 359 (6398): p. 843-845. 67. Forsythe, J A, Jiang, B H, Iyer, N V, Agani, F, Leung, S W, Koos, R D, and Semenza, G L, Activation of vascular endothelial gr owth factor gene transcription by hypoxia-inducible factor 1. Molecular and Cellular Biology, 1996. 16 (9): p. 4604-4613. 68. Gleadle, J M, Ebert, B L, Fi rth, J D, and Ratcliffe, P J, Regulation of Angiogenic Growth-Factor Expression by Hypoxia, Tr ansition-Metals, and Chelating-Agents. American Journal of Phys iology-Cell Physiology, 1995. 37 (6): p. C1362-C1368. 69. Liu, Y X, Cox, S R, Morita, T, and Kourembanas, S, Hypoxia Regulates Vascular Endothelial Growth-Factor Gene-Expression in Endothelial-Cells Identification of a 5'-Enhancer. Circulation Research, 1995. 77 (3): p. 638-643. 70. Krishnamachary, B, Berg-Dixon, S, Kelly, B, Agani, F, Feldser, D, Ferreira, G, Iyer, N, LaRusch, J, Pak, B, Taghavi, P, and Semenza, G L, Regulation of colon carcinoma cell invasion by hypoxia-inducible factor 1. Cancer Research, 2003. 63 (5): p. 1138-1143. 71. Goldberg, M A, and Schneider, T J, Similarities between the Oxygen-Sensing Mechanisms Regulating the Expression of Vascular Endothelial Growth-Factor and Erythropoietin. Journal of Biological Chemistry, 1994. 269 (6): p. 4355-4359.

PAGE 182

166 72. Tian, H, McKnight, S L, and Russell, D W, Endothelial PAS domain protein 1 (EPAS1), a transcription factor selectiv ely expressed in endothelial cells. Genes & Development, 1997. 11 (1): p. 72-82. 73. Wiesener, M S, Turley, H, Allen, W E, Willam, C, Eckardt, K U, Talks, K L, Wood, S M, Gatter, K C, Harris, A L, Pugh, C W, Ratcliffe, P J, and Maxwell, P H, Induction of endotheli al PAS domain protein-1 by hypoxia: Characterization and comparison with hypoxia-inducible factor-1 alpha. Blood, 1998. 92 (7): p. 22602268. 74. Makino, Y, Cao, R H, Svensson, K, Bertil sson, G R, Asman, M, Tanaka, H, Cao, Y H, Berkenstam, A, and Poellinger, L, Inhibitory PAS domain protein is a negative regulator of hypoxia-ind ucible gene expression. Nature, 2001. 414 (6863): p. 550554. 75. Maxwell, P H, and Ratcliffe, P J, Oxygen sensors and angiogenesis. Seminars in Cell & Developmental Biology, 2002. 13 (1): p. 29-37. 76. Senger, D R, Galli, S J, Dvorak, A M, Perruzzi, C A, Harvey, V S, and Dvorak, H F, Tumor-Cells Secrete a Vascular-Per meability Factor That Promotes Accumulation of Ascites-Fluid. Science, 1983. 219 (4587): p. 983-985. 77. Ferrara, N, and Henzel, W J, Pituitary Follicular Cells Secrete a Novel HeparinBinding Growth-Factor Specific fo r Vascular Endothelial-Cells. Biochemical and Biophysical Research Communications, 1989. 161 (2): p. 851-858. 78. Leung, D W, Cachianes, G, Kuang, W J, Goeddel, D V, and Ferrara, N, Vascular Endothelial Growth-Factor Is a Secreted Angiogenic Mitogen. Science, 1989. 246 (4935): p. 1306-1309. 79. Keck, P J, Hauser, S D, Krivi, G, Sanzo, K, Warren, T, Feder, J, and Connolly, D T, Vascular-Permeability Factor, an Endot helial-Cell Mitogen Related to Pdgf. Science, 1989. 246 (4935): p. 1309-1312. 80. Ciulla TA, D R, Criswell M, Pratt LM., Changing therapeutic paradigms for exudative age-related macular dege neration: antiangiogenic agents and photodynamic therapy. Expert Opin Investig Drugs., 1999. 8 (12): p. 2173-2182. 81. Muller, Y A, Christinger, H W, Keyt, B A, and deVos, A M, The crystal structure of vascular endothelial grow th factor (VEGF) refined to 1.93 angstrom resolution: multiple copy flexibility and receptor binding. Structure, 1997. 5 (10): p. 1325-1338. 82. Olofsson, B, Pajusola, K, Kaipainen, A, vonEuler, G, Joukov, V, Saksela, O, Orpana, A, Petersson, R F, Alitalo, K, and Eriksson, U, Vascular endothelial growth factor B, a novel growth factor for endothelial cells. Proceedings of the National Academy of Sciences of the United States of America, 1996. 93 (6): p. 2576-2581.

PAGE 183

167 83. Lee, J, Gray, A, Yuan, J, Luoh, S M, Avraham, H, and Wood, W I, Vascular endothelial growth factor-r elated protein: A ligand and specific activator of the tyrosine kinase receptor Flt4. Proceedings of the Nationa l Academy of Sciences of the United States of America, 1996. 93 (5): p. 1988-1992. 84. Makinen, T, Veikkola, T, Mustjoki, S, Ka rpanen, T, Catimel, B, Nice, E C, Wise, L, Mercer, A, Kowalski, H, Kerjaschki, D, Stacker, S A, Achen, M G, and Alitalo, K, Isolated lymphatic endothelial cells transduce growth, survival and migratory signals via the VEGF-C/D receptor VEGFR-3. Embo Journal, 2001. 20 (17): p. 4762-4773. 85. Achen, M G, Jeltsch, M, Kukk, E, Makinen, T, Vitali, A, Wilks, A F, Alitalo, K, and Stacker, S A, Vascular endothelial growth fa ctor D (VEGF-D) is a ligand for the tyrosine kinases VEGF receptor 2 (Flk1) and VEGF receptor 3 (Flt4). Proceedings of the National Academy of Scie nces of the United States of America, 1998. 95 (2): p. 548-553. 86. Meyer, M, Clauss, M, Lepple-Wienhues, A, Waltenberger, J, Augustin, H G, Ziche, M, Lanz, C, Buttner, M, Rziha, H J, and Dehio, C, A novel vascular endothelial growth factor encoded by Orf virus, VEGF-E, mediates angiogenesis via signalling through VEGFR-2 (KDR) but not VEGFR-1 (Flt-1) receptor tyrosine kinases. Embo Journal, 1999. 18 (2): p. 363-374. 87. Junqueira-de-Azevedo, I D L, da Silva, M B, Chudzinski-Tavassi, A M, and Ho, P L, Identification and cloning of snake venom vascular endothelia l growth factor (svVEGF) from Bothrops erythromelas pitviper. Toxicon, 2004. 44 (5): p. 571-575. 88. Ferrara, N, Houck, K, Jakeman, L, and Leung, D W, Molecular and Biological Properties of the Vascular Endothelial Growth-Factor Family of Proteins. Endocrine Reviews, 1992. 13 (1): p. 18-32. 89. Tischer, E, Mitchell, R, Hartman, T, Silv a, M, Gospodarowicz, D, Fiddes, J C, and Abraham, J A, The Human Gene for Vascular Endothelial Growth-Factor Multiple Protein Forms Are Encode d through Alternative Exon Splicing. Journal of Biological Chemistry, 1991. 266 (18): p. 11947-11954. 90. Shima, D T, Kuroki, M, Deutsch, U, N g, Y S, Adamis, A P, and DAmore, P A, The mouse gene for vascular endot helial growth factor Ge nomic structure, definition of the transcriptional unit, and charac terization of transcriptional and posttranscriptional regulatory sequences. Journal of Biological Chemistry, 1996. 271 (7): p. 3877-3883. 91. Robinson, C J, and Stringer, S E, The splice variants of vascular endothelial growth factor (VEGF) and their receptors. Journal of Cell Science, 2001. 114 (5): p. 853-865.

PAGE 184

168 92. Byrne, A M, Bouchier-Hayes, D J, and Harmey, J H, Angiogenic and cell survival functions of Vascular Endothe lial Growth Factor (VEGF). Journal of Cellular and Molecular Medicine, 2005. 9 (4): p. 777-794. 93. Houck, K A, Leung, D W, Rowland, A M, Winer, J, and Ferrara, N, Dual Regulation of Vascular Endothelial Growth -Factor Bioavailability by Genetic and Proteolytic Mechanisms. Journal of Biological Chemistry, 1992. 267 (36): p. 2603126037. 94. Hoeben A, L B, Highley MS, Wildiers H, Van Oosterom AT, De Bruijn EA., Vascular endothelial grow th factor and angiogenesis. Pharmacol Rev., 2004. 56 : p. 549-580. 95. Bates, D O, Cui, T G, Doughty, J M, Wi nkler, M, Sugiono, M, Shields, J D, Peat, D, Gillatt, D, and Harper, S J, VEGF(165)b, an inhibitory splice variant of vascular endothelial growth factor, is downregulated in renal cell carcinoma. Cancer Research, 2002. 62 (14): p. 4123-4131. 96. Woolard, J, Wang, W Y, Beva n, H S, Qiu, Y, Morbidelli, L, Pritchard-Jones, R O, Cui, T G, Sugiono, M, Waine, E, Perrin, R, Foster, R, Digby-Bell, J, Shields, J D, Whittles, C E, Mushens, R E, Gillatt, D A, Ziche, M, Harper, S J, and Bates, D O, VEGF(165)b, an inhibitory vascular endot helial growth factor splice variant: Mechanism of action, in vivo effect on angiogenesis and endogenous protein expression. Cancer Research, 2004. 64 (21): p. 7822-7835. 97. Davis-Smyth, T, Presta, L G, and Ferrara, N, Mapping the charged residues in the second immunoglobulin-like domain of the va scular endothelial growth factor placenta growth factor receptor Flt-1 re quired for binding and structural stability. Journal of Biological Chemistry, 1998. 273 (6): p. 3216-3222. 98. Fuh, G, Li, B, Crowley, C, Cunningham, B, and Wells, J A, Requirements for binding and signaling of the kinase domain receptor for vascular endothelial growth factor. Journal of Biological Chemistry, 1998. 273 (18): p. 11197-11204. 99. Shinkai, A, Ito, M, Anazawa, H, Ya maguchi, S, Shitara, K, and Shibuya, M, Mapping of the sites involved in liga nd association and dissociation at the extracellular domain of the ki nase insert domain-contai ning receptor for vascular endothelial growth factor. Journal of Biological Chemistry, 1998. 273 (47): p. 31283-31288. 100. Park, J E, Chen, H H, Winer, J, Houck, K A, and Ferrara, N, Placenta GrowthFactor Potentiation of Vascular Endothe lial Growth-Factor Bioactivity, in-Vitro and in-Vivo, and High-Affinity Binding to Flt-1 but Not to Flk-1/Kdr. Journal of Biological Chemistry, 1994. 269 (41): p. 25646-25654. 101. Fong, G H, Zhang, L Y, Bryce, D M, and Peng, J, Increased hemangioblast commitment, not vascular disorganization, is the primary defect in flt-1 knock-out mice. Development, 1999. 126 (13): p. 3015-3025.

PAGE 185

169 102. Hiratsuka, S, Minowa, O, Kuno, J, Noda, T, and Shibuya, M, Flt-1 lacking the tyrosine kinase domain is sufficient for normal development and angiogenesis in mice. Proceedings of the National Academy of Sciences of the United States of America, 1998. 95 (16): p. 9349-9354. 103. Kendall, R L, and Thomas, K A, Inhibition of Vascular Endothelial-Cell GrowthFactor Activity by an Endogenously Encoded Soluble Receptor. Proceedings of the National Academy of Sciences of the United States of America, 1993. 90 (22): p. 10705-10709. 104. Kendall, R L, Wang, G, and Thomas, K A, Identification of a nat ural soluble form of the vascular endothelial grow th factor receptor, FLT-1, and its heterodimerization with KDR. Biochemical and Biophysical Research Communications, 1996. 226 (2): p. 324-328. 105. Barleon, B, Sozzani, S, Zhou, D, Weic h, H A, Mantovani, A, and Marme, D, Migration of human monocytes in response to vascular endothe lial growth factor (VEGF) is mediated via the VEGF receptor flt-1. Blood, 1996. 87 (8): p. 3336-3343. 106. Clauss, M, Weich, H, Breier, G, Knies, U, Rockl, W, Waltenberger, J, and Risau, W, The vascular endothelial growth factor receptor Flt-1 mediates biological activities Implications for a functional role of placenta growth factor in monocyte activation and chemotaxis. Journal of Biological Chemistry, 1996. 271 (30): p. 17629-17634. 107. Selvaraj, S K, Giri, R K, Perelman, N, Johnson, C, Malik, P, and Kalra, V K, Mechanism of monocyte activation a nd expression of proinflammatory cytochemokines by placenta growth factor. Blood, 2003. 102 (4): p. 1515-1524. 108. Shalaby, F, Ho, J, Stanford, W L, Fi scher, K D, Schuh, A C, Schwartz, L, Bernstein, A, and Rossant, J, A requirement for Flk1 in primitive and definitive hematopoiesis and vasculogenesis. Cell, 1997. 89 (6): p. 981-990. 109. Kaipainen, A, Korhonen, J, Mustone n, T, Vanhinsbergh, V W M, Fang, G H, Dumont, D, Breitman, M, and Alitalo, K, Expression of the Fms-Like Tyrosine Kinase-4 Gene Becomes Restricted to Lymphatic Endothelium during Development. Proceedings of the National Acad emy of Sciences of the United States of America, 1995. 92 (8): p. 3566-3570. 110. Fujisawa, H, and Kitsukawa, T, Receptors for collapsin/semaphorins. Current Opinion in Neurobiology, 1998. 8 (5): p. 587-592. 111. Fuh, G, Garcia, K C, and de Vos, A M, The interaction of neuropilin-1 with vascular endothelial growth factor and its receptor Flt-1. Journal of Biological Chemistry, 2000. 275 (35): p. 26690-26695.

PAGE 186

170 112. Whitaker, G B, Limberg, B J, and Rosenbaum, J S, Vascular endothelial growth factor receptor-2 and neuropilin-1 form a r eceptor complex that is responsible for the differential signaling pot ency of VEGF(165) and VEGF(121). Journal of Biological Chemistry, 2001. 276 (27): p. 25520-25531. 113. Soker, S, Miao, H Q, Nomi, M, Takashima, S, and Klagsbrun, M, VEGF(165) mediates formation of complexes containing VEGFR-2 and neuropilin-1 that enhance VEGF(165)-receptor binding. Journal of Cellular Biochemistry, 2002. 85 (2): p. 357-368. 114. Fakhari, M, Pullirsch, D, Abraham, D, Paya, K, Hofbauer, R, Holzfeind, P, Hofmann, M, and Aharinejad, S, Selective upregulation of vascular endothelial growth factor receptors neuropilin-1 and-2 in human neuroblastoma. Cancer, 2002. 94 (1): p. 258-263. 115. Ng YS, K D, Shima DT., VEGF function in vascular pathogenesis. Exp Cell Res., 2006. 312 : p. 527-537. 116. Cebe-Suarez S, Z-F A, Ballmer-Hofer K., The role of VEGF receptors in angiogenesis; complex partnerships. Cell Mol Life Sci., 2006: p. Epub ahead of print. 117. Meadows, K N, Bryant, P, and Pumiglia, K, Vascular endothelia l growth factor induction of the angiogenic phenot ype requires Ras activation. Journal of Biological Chemistry, 2001. 276 (52): p. 49289-49298. 118. Takahashi, T, Yamaguchi, S, Chida, K, and Shibuya, M, A single autophosphorylation site on KDR/Flk-1 is essential for VEGF-A-dependent activation of PLC-gamma and DNA synthe sis in vascular endothelial cells. Embo Journal, 2001. 20 (11): p. 2768-2778. 119. Takahashi, T, Ueno, H, and Shibuya, M, VEGF activates protein kinase Cdependent, but Ras-independent Raf-MEK -MAP kinase pathway for DNA synthesis in primary endothelial cells. Oncogene, 1999. 18 (13): p. 2221-2230. 120. Yu, Y, and Sato, J D, MAP kinases, phosphatidylinositol 3-kinase, and p70 S6 kinase mediate the mitogenic response of human endothelial cells to vascular endothelial growth factor. Journal of Cellular Physiology, 1999. 178 (2): p. 235246. 121. Kroll, J, and Waltenberger, J, The vascular endothelial gr owth factor receptor KDR activates multiple signal transduction pathways in porcine aortic endothelial cells. Journal of Biological Chemistry, 1997. 272 (51): p. 32521-32527. 122. Abedi, H, and Zachary, I, Vascular endothelial growth factor stimulates tyrosine phosphorylation and recruitment to new foca l adhesions of focal adhesion kinase and paxillin in endothelial cells. Journal of Biological Chemistry, 1997. 272 (24): p. 15442-15451.

PAGE 187

171 123. Kanno, S, Oda, N, Abe, M, Terai, Y, Ito, M, Shitara, K, Tabayashi, K, Shibuya, M, and Sato, Y, Roles of two VEGF receptors, Flt-1 and KDR, in the signal transduction of VEGF effects in human vascular endothelial calls. Oncogene, 2000. 19 (17): p. 2138-2146. 124. Rousseau, S, Houle, F, Landry, J, and Huot, J, p38 MAP kinase activation by vascular endothelial growth factor mediates actin reorganization and cell migration in human endothelial cells. Oncogene, 1997. 15 (18): p. 2169-2177. 125. Matsumoto, T, Bohman, S, Dixelius, J, Berge, T, Dimberg, A, Magnusson, P, Wang, L, Wikner, C, Qi, J H, Wernstedt, C, Wu, J, Bruheim, S, Mugishima, H, Mukhopadhyay, D, Spurkland, A, and Claesson-Welsh, L, VEGF receptor-2 Y951 signaling and a role for the adapter mo lecule TSAd in tumor angiogenesis. Embo Journal, 2005. 24 (13): p. 2342-2353. 126. Gille, H, Kowalski, J, Li, B, LeCouter, J, Moffat, B, Zioncheck, T F, Pelletier, N, and Ferrara, N, Analysis of biological effect s and signaling properties of Flt-1 (VEGFR-1) and KDR (VEGFR-2) A reasse ssment using novel receptor-specific vascular endothelial gr owth factor mutants. Journal of Biological Chemistry, 2001. 276 (5): p. 3222-3230. 127. Gerber, H P, McMurtrey, A, Kowalski, J, Yan, M H, Keyt, B A, Dixit, V, and Ferrara, N, Vascular endothelial growth factor regulates endothelial cell survival through the phosphatidylinositol 3 '-kinas e Akt signal transduction pathway Requirement for Flk-1/KDR activation. Journal of Biological Chemistry, 1998. 273 (46): p. 30336-30343. 128. Carmeliet, P, Lampugnani, M G, Moons, L, Breviario, F, Compernolle, V, Bono, F, Balconi, G, Spagnuolo, R, Oosthuyse, B, De werchin, M, Zanetti, A, Angellilo, A, Mattot, V, Nuyens, D, Lutgens, E, Clot man, F, de Ruiter, M C, Gittenberger-de Groot, A, Poelmann, R, Lupu, F, Herbert, J M, Collen, D, and Dejana, E, Targeted deficiency or cytosolic truncation of th e VE-cadherin gene in mice impairs VEGFmediated endothelial su rvival and angiogenesis. Cell, 1999. 98 (2): p. 147-157. 129. Gerber, H P, Dixit, V, and Ferrara, N, Vascular endothelial growth factor induces expression of the antiapoptotic proteins Bc l-2 and A1 in vascular endothelial cells. Journal of Biological Chemistry, 1998. 273 (21): p. 13313-13316. 130. Tran, J, Rak, J, Sheehan, C, Saibil, S D, LaCasse, E, Korneluk, R G, and Kerbel, R S, Marked induction of the IAP family anti apoptotic proteins su rvivin and XIAP by VEGF in vascular endothelial cells. Biochemical and Biophysical Research Communications, 1999. 264 (3): p. 781-788. 131. Dvorak, A M, and Feng, D, The vesiculo-vacuolar organelle (VVO): A new endothelial cell permeability organelle. Journal of Histochemistry & Cytochemistry, 2001. 49 (4): p. 419-431.

PAGE 188

172 132. Strickland, L A, Jubb, A M, Hongo, L A, Z hong, F, Burwick, J, Fu, L, Frantz, G D, and Koeppen, H, Plasmalemmal vesicle-associated protein (PLVAP) is expressed by tumour endothelium and is upregulated by vascular endothelia l growth factor-A (VEGF). Journal of Pathology, 2005. 206 (4): p. 466-475. 133. Fulton, D, Gratton, J P, McCabe, T J, Fontana, J, Fujio, Y, Walsh, K, Franke, T F, Papapetropoulos, A, and Sessa, W C, Regulation of endothelium-derived nitric oxide production by the protein kinase Akt. Nature, 1999. 399 (6736): p. 597-601. 134. Michell, B J, Griffiths, J E, Mitchelhi ll, K I, Rodriguez-Crespo, I, Tiganis, T, Bozinovski, S, de Montellano, P R O, Kemp, B E, and Pearson, R B, The Akt kinase signals directly to endot helial nitric oxide synthase. Current Biology, 1999. 9 (15): p. 845-848. 135. Miller, J W, Adamis, A P, Shima, D T, Damore, P A, Moulton, R S, Oreilly, M S, Folkman, J, Dvorak, H F, Brown, L F, Berse, B, Yeo, T K, and Yeo, K T, Vascular Endothelial Growth-Factor Vascular-Per meability Factor Is Temporally and Spatially Correlated with Ocular Angiogenesis in a Primate Model. American Journal of Pathology, 1994. 145 (3): p. 574-584. 136. Pierce, E A, Avery, R L, Foley, E D, Aiello, L P, and Smith, L E H, Vascular Endothelial Growth-Factor Vascular-Perme ability Factor Expression in a Mouse Model of Retinal Neovascularization. Proceedings of the National Academy of Sciences of the United States of America, 1995. 92 (3): p. 905-909. 137. Donahue, M L, Phelps, D L, Watkins, R H, LoMonaco, M B, and Horowitz, S, Retinal vascular endothelial growth factor (VEGF) mRNA expression is altered in relation to neovascularization in oxygen induced retinopathy. Current Eye Research, 1996. 15 (2): p. 175-184. 138. Dorey, C K, Aouididi, S, Reynaud, X, Dvorak, H F, and Brown, L F, Correlation of vascular permeability factor/vascular e ndothelial growth factor with extraretinal neovascularization in the rat. Archives of Ophthalmology, 1996. 114 (10): p. 12101217. 139. Adamis, A P, Miller, J W, Bernal, M T, Damico, D J, Folkman, J, Yeo, T K, and Yeo, K T, Increased Vascular Endothelial Growth -Factor Levels in the Vitreous of Eyes with Proliferat ive Diabetic-Retinopathy. American Journal of Ophthalmology, 1994. 118 (4): p. 445-450. 140. Aiello, L P, Avery, R L, Arrigg, P G, Keyt B A, Jampel, H D, Shah, S T, Pasquale, L R, Thieme, H, Iwamoto, M A, Park, J E, Nguyen, H V, Aiello, L M, Ferrara, N, and King, G L, Vascular Endothelial Growth-Facto r in Ocular Fluid of Patients with Diabetic-Retinopathy and Other Retinal Disorders. New England Journal of Medicine, 1994. 331 (22): p. 1480-1487.

PAGE 189

173 141. Malecaze, F, Clamens, S, Simorrepinatel, V, Mathis, A, Chollet, P, Favard, C, Bayard, F, and Plouet, J, Detection of Vascular Endothelial Growth-Factor Messenger-Rna and Vascular Endothelial Growth Factor-Like Activity in Proliferative Diabetic-Retinopathy. Archives of Ophthalmology, 1994. 112 (11): p. 1476-1482. 142. Peer, J, Folberg, R, Itin, A, Gn essin, H, Hemo, I, and Keshet, E, Upregulated expression of vascular endothelial grow th factor in proliferative diabetic retinopathy. British Journal of Ophthalmology, 1996. 80 (3): p. 241-245. 143. Ambati, J, Chalam, K V, Chawla, D K, DAngio, C T, Guillet, E G, Rose, S J, Vanderlinde, R E, and Ambati, B K, Elevated gamma-aminobutyric acid, glutamate, and vascular endothe lial growth factor levels in the vitreous of patients with proliferative diabetic retinopathy. Archives of Ophthalmology, 1997. 115 (9): p. 1161-1166. 144. Burgos, R, Simo, R, Audi, L, Mateo, C, Mesa, J, GarciaRamirez, M, and Carrascosa, A, Vitreous levels of vascular e ndothelial growth factor are not influenced by its serum concentr ations in diabetic retinopathy. Diabetologia, 1997. 40 (9): p. 1107-1109. 145. Hattenbach, L O, Allers, A, Gumbel, H O C, Scharrer, I, and Koch, F H J, Vitreous concentrations of TPA and plasminogen ac tivator inhibitor are associated with VEGF in proliferative diabetic vitreoretinopathy. Retina-the Journal of Retinal and Vitreous Diseases, 1999. 19 (5): p. 383-389. 146. Frank, R N, Amin, R H, Eliott, D, Puklin, J E, and Abrams, G W, Basic fibroblast growth factor and vascular endothelial grow th factor are present in epiretinal and choroidal neovascular membranes. American Journal of Ophthalmology, 1996. 122 (3): p. 393-403. 147. Spirin, K S, Saghizadeh, M, Lewin, S L, Zardi, L, Kenney, M C, and Ljubimov, A V, Basement membrane and growth factor gene expression in normal and diabetic human retinas. Current Eye Research, 1999. 18 (6): p. 490-499. 148. Aiello, L P, Northrup, J M, Keyt, B A, Takagi, H, and Iwamoto, M A, Hypoxic Regulation of Vascular Endothelial Growth-Factor in Retinal Cells. Archives of Ophthalmology, 1995. 113 (12): p. 1538-1544. 149. Adamis, A P, Shima, D T, Tolentino, M J, Gragoudas, E S, Ferrara, N, Folkman, J, DAmore, P A, and Miller, J W, Inhibition of vascular endothelial growth factor prevents retinal ischemia-associated iris neovascularization in a nonhuman primate. Archives of Ophthalmology, 1996. 114 (1): p. 66-71. 150. Robinson, G S, Pierce, E A, Rook, S L, Foley, E, Webb, R, and Smith, L E H, Oligodeoxynucleotides inhibit retinal ne ovascularization in a murine model of proliferative retinopathy. Proceedings of the National Academy of Sciences of the United States of America, 1996. 93 (10): p. 4851-4856.

PAGE 190

174 151. McLeod, D S, Taomoto, M, Cao, J T, Zhu, Z P, Witte, L, and Lutty, G A, Localization of VEGF receptor-2 (KDR/Flk-1 ) and effects of blocking it in oxygeninduced retinopathy. Investigative Ophthalmol ogy & Visual Science, 2002. 43 (2): p. 474-482. 152. Tolentino, M J, Miller, J W, Gra goudas, E S, Jakobiec, F A, Flynn, E, Chatzistefanou, K, Ferrara, N, and Adamis, A P, Intravitreous injections of vascular endothelial growth factor produce retinal ischemia and microangiopathy in an adult primate. Ophthalmology, 1996. 103 (11): p. 1820-1828. 153. Ozaki, H, Hayashi, H, Vinores, S A, Moromizato, Y, Campochiaro, P A, and Oshima, K, Intravitreal sustained releas e of VEGF causes retinal neovascularization in rabbits and breakdown of the blood-retinal barrier in rabbits and primates. Experimental Eye Research, 1997. 64 (4): p. 505-517. 154. Ohno-Matsui, K, Hirose, A, Yamamoto, S, Saikia, J, Okamoto, N, Gehlbach, P, Duh, E J, Hackett, S, Chang, M, Bok, D, Zack, D J, and Campochiaro, P A, Inducible expression of vascular endothel ial growth factor in adult mice causes severe proliferative retinopathy and retinal detachment. American Journal of Pathology, 2002. 160 (2): p. 711-719. 155. Das, A, and McGuire, P G, Retinal and choroidal angi ogenesis: pathophysiology and strategies for inhibition. Progress in Retinal and Eye Research, 2003. 22 (6): p. 721-748. 156. Amin, R, Puklin, J E, and Frank, R N, Growth-Factor Localization in Choroidal Neovascular Membranes of AgeRelated Macular Degeneration. Investigative Ophthalmology & Visual Science, 1994. 35 (8): p. 3178-3188. 157. Lopez, P F, Sippy, B D, Lambert, H M, Thach, A B, and Hinton, D R, Transdifferentiated retinal pigment ep ithelial cells are immunoreactive for vascular endothelial growth factor in surgically excised age-related macular degeneration-related choroi dal neovascular membranes. Investigative Ophthalmology & Visual Science, 1996. 37 (5): p. 855-868. 158. Yi, X J, Ogata, N, Komada, M, Yamamoto, C, Takahashi, K, Omori, K, and Uyama, M, Vascular endothelial growth fa ctor expression in choroidal neovascularization in rats. Graefes Archive for Clinical and Experimental Ophthalmology, 1997. 235 (5): p. 313-319. 159. Seo, M S, Kwak, N, Ozaki, H, Yamada, H, Okamoto, N, Yamada, E, Fabbro, D, Hofmann, F, Wood, J M, and Campochiaro, P A, Dramatic inhibition of retinal and choroidal neovascularization by oral administration of a kinase inhibitor. American Journal of Pathology, 1999. 154 (6): p. 1743-1753. 160. Kwak, N, Okamoto, N, Wood, J M, and Campochiaro, P A, VEGF is major stimulator in model of choroidal neovascularization. Investigative Ophthalmology & Visual Science, 2000. 41 (10): p. 3158-3164.

PAGE 191

175 161. Saishin, Y, Saishin, Y, Ta kahashi, K, Silva, R L E, Hylton, D, Rudge, J S, Wiegand, S J, and Campochiaro, P A, VEGF-TRAP(R1R2) suppresses choroidal neovascularization and VEGF-induced br eakdown of the blood-retinal barrier. Journal of Cellular Physiology, 2003. 195 (2): p. 241-248. 162. Okamoto, N, Tobe, T, Hackett, S F, Ozak i, H, Vinores, M A, LaRochelle, W, Zack, D J, and Campochiaro, P A, Transgenic mice with increased expression of vascular endothelial growth factor in the retina A new model of intraretinal and subretinal neovascularization. American Journal of Pathology, 1997. 151 (1): p. 281-291. 163. Nyberg, F, Hahnenberger, R, Jakobson, A M, and Terenius, L, Enhancement of Fgf-Like Polypeptides in the Retinae of Newborn Mice Exposed to Hyperoxia. Febs Letters, 1990. 267 (1): p. 75-77. 164. Zhang, N L, Samadani, E E, and Frank, R N, Mitogenesis and Retinal-Pigment Epithelial-Cell Antigen Expression in the Rat after Krypton Laser Photocoagulation. Investigative Ophthalmology & Visual Science, 1993. 34 (8): p. 2412-2424. 165. Sivalingam, A, Kenney, J, Brown, G C, Benson, W E, and Donoso, L, Basic Fibroblast Growth-Factor Leve ls in the Vitreous of Patients with Proliferative Diabetic-Retinopathy. Archives of Ophthalmology, 1990. 108 (6): p. 869-872. 166. Damore, P A, Mechanisms of Retinal and C horoidal Neovascularization. Investigative Ophthalmol ogy & Visual Science, 1994. 35 (12): p. 3974-3979. 167. Ozaki, H, Okamoto, N, Ortega, S, Chang, M, Ozaki, K, Sadda, S, Vinores, M A, Derevjanik, N, Zack, D J, Bas ilico, C, and Campochiaro, P A, Basic fibroblast growth factor is neither necessary nor sufficient for the d evelopment of retinal neovascularization. American Journal of Pathology, 1998. 153 (3): p. 757-765. 168. Tobe, T, Ortega, S, Luna, J D, Ozaki, H, Okamoto, N, Derevjanik, N L, Vinores, S A, Basilico, C, and Campochiaro, P A, Targeted disruption of the FGF2 gene does not prevent choroidal neovascularization in a murine model. American Journal of Pathology, 1998. 153 (5): p. 1641-1646. 169. Yamada, H, Yamada, E, Kwak, N, Ando, A, Suzuki, A, Esumi, N, Zack, D J, and Campochiaro, P A, Cell injury unmasks a latent proangiogenic phenotype in mice with increased expression of FGF2 in the retina. Journal of Cellular Physiology, 2000. 185 (1): p. 135-142. 170. Maisonpierre, P C, Suri, C, Jones, P F, Bartunkova, S, Wiegand, S, Radziejewski, C, Compton, D, McClain, J, Aldrich, T H, Papadopoulos, N, Daly, T J, Davis, S, Sato, T N, and Yancopoulos, G D, Angiopoietin-2, a natural antagonist for Tie2 that disrupts in vivo angiogenesis. Science, 1997. 277 (5322): p. 55-60.

PAGE 192

176 171. Oh, H, Takagi, H, Suzuma, K, Otan i, A, Matsumura, M, and Honda, Y, Hypoxia and vascular endothelial growth factor selectively up-regulate angiopoietin-2 in bovine microvascular endothelial cells. Journal of Biological Chemistry, 1999. 274 (22): p. 15732-15739. 172. Oh, H, Takagi, H, Takagi, C, Suzuma, K, Otani, A, Ishida, K, Matsumura, M, Ogura, Y, and Honda, Y, The potential angiogenic role of macrophages in the formation of choroidal neovascular membranes. Investigative Ophthalmology & Visual Science, 1999. 40 (9): p. 1891-1898. 173. Hackett, S F, Ozaki, H, Strauss, R W, Wahlin, K, Suri, C, Maisonpierre, P, Yancopoulos, G, and Campochiaro, P A, Angiopoietin 2 expression in the retina: Upregulation during physiologic and pathologic neovascularization. Journal of Cellular Physiology, 2000. 184 (3): p. 275-284. 174. Das, A, Talarico, N., Warren, E., McGuire, P.G., Inhibition of angiopoietin signaling suppresses retinal neovascularization in ARVO 2002: Ft. Lauderdale, FL. 175. Freyberger, H, Brocker, M, Yakut, H, Hammer, J, Effert, R, Schifferdecker, E, Schatz, N, and Derwahl, M, Increased levels of platelet-derived growth factor in vitreous fluid of patients with proliferative di abetic retinopathy. Experimental and Clinical Endocrinology & Diabetes, 2000. 108 (2): p. 106-109. 176. Mori, K, Gehlbach, P, Ando, A, McVe y, D, Wei, L, and Campochiaro, P A, Regression of ocular neovascularization in response to increased expression of pigment epithelium-derived factor. Investigative Ophthalmology & Visual Science, 2002. 43 (7): p. 2428-2434. 177. Mori, K, Gehlbach, P, Yamamoto, S, Dub, E, Zack, D J, Li, Q H, Berns, K I, Raisler, B J, Hauswirth, W W, and Campochiaro, P A, AAV-Mediated gene transfer of pigment epithelium-derived factor inhibits choroidal neovascularization. Investigative Ophthalmol ogy & Visual Science, 2002. 43 (6): p. 1994-2000. 178. Mori, K, Gehlbach, P, Ando, A, Dyer, G, Lipinsky, E, Chaudhry, A G, Hackett, S F, and Campochiaro, P A, Retina-specific expression of PDGF-B versus PDGF-A: Vascular versus nonvascular proliferative retinopathy. Investigative Ophthalmology & Visual Science, 2002. 43 (6): p. 2001-2006. 179. Vinores, S A, Seo, M S, Derevian ik, N L, and Campochiaro, P A, Photoreceptorspecific overexpression of platelet-derived growth factor induces proliferation of endothelial cells, pericytes, and glial cells and aberrant vascular development: an ultrastructural and immunocytochemical study. Developmental Brain Research, 2003. 140 (2): p. 169-183. 180. Stupack, D G, Integrins as a distinct subtype of dependence receptors. Cell Death and Differentiation, 2005. 12 (8): p. 1021-1030.

PAGE 193

177 181. Clemmons, D R, and Maile, L A, Interaction between insuli n-like growth factor-I receptor and alpha V beta 3 integrin linked signaling pathways: Cellular responses to changes in multiple signaling inputs. Molecular Endocrinology, 2005. 19 (1): p. 1-11. 182. ffrench-Constant, C, and Colognato, H, Integrins: versatile integrators of extracellular signals. Trends in Cell Biology, 2004. 14 (12): p. 678-686. 183. Brown, D A, and London, E, Structure and function of sphingolipidand cholesterol-rich membrane rafts. Journal of Biological Chemistry, 2000. 275 (23): p. 17221-17224. 184. Simons, K, and Toomre, D, Lipid rafts and signal transduction. Nature Reviews Molecular Cell Biology, 2000. 1 (1): p. 31-39. 185. Li, R H, Mitra, N, Gratkowski, H, Vilair e, G, Litvinov, R, Nagasami, C, Weisel, J W, Lear, J D, DeGrado, W F, and Bennett, J S, Activation of integrin alpha IIb beta 3 by modulation of transmemb rane helix associations. Science, 2003. 300 (5620): p. 795-798. 186. Klein, S, Giancotti, F G, Presta, M, Al belda, S M, Buck, C A, and Rifkin, D B, Basic Fibroblast Growth-Factor Modulates Integrin Expression in Microvascular Endothelial-Cells. Molecular Biology of the Cell, 1993. 4 (10): p. 973-982. 187. Mariotti, A, Kedeshian, P A, Dans, M, Curatola, A M, Gagnoux-Palacios, L, and Giancotti, F G, EGF-R signaling through Fyn kina se disrupts the function of integrin alpha 6 beta 4 at hemidesmosom es: role in epithelial cell migration and carcinoma invasion. Journal of Cell Biology, 2001. 155 (3): p. 447-457. 188. Byzova, T V, Goldman, C K, Pampori, N, Thomas, K A, Bett, A, Shattil, S J, and Plow, E F, A mechanism for modulation of cellu lar responses to VEGF: Activation of the integrins. Molecular Cell, 2000. 6 (4): p. 851-860. 189. Campochiaro, P A, and Hackett, S F, Ocular neovascularization: a valuable model system. Oncogene, 2003. 22 (42): p. 6537-6548. 190. Friedlander, M, Theesfeld, C L, Sugita, M, Fruttiger, M, Thomas, M A, Chang, S, and Cheresh, D A, Involvement of integrins alpha( v)beta(3) and alpha(v)beta(5) in ocular neovascular diseases. Proceedings of the Nationa l Academy of Sciences of the United States of America, 1996. 93 (18): p. 9764-9769. 191. Dwayne G. Stupack, D A C, ECM Remodeling Regulates Angiogenesis: Endothelial Integrins Look for New Ligands. Science's STKE, 2002: p. pe7. 192. Sahni, A, and Francis, C W, Stimulation of endothelial cel l proliferation by FGF-2 in the presence of fibrinogen requires alpha(v)beta(3). Blood, 2004. 104 (12): p. 3635-3641.

PAGE 194

178 193. Senger, D R, Perruzzi, C A, Streit, M, Koteliansky, V E, de Fougerolles, A R, and Detmar, M, The alpha(1)beta(1) and alpha(2)beta(1) Integrins provide critical support for vascular endothelial growth factor signaling, endothelial cell migration, and tumor angiogenesis. American Journal of Pathology, 2002. 160 (1): p. 195-204. 194. Stupack, D G, Puente, X S, Boutsaboual oy, S, Storgard, C M, and Cheresh, D A, Apoptosis of adherent cells by recruitmen t of caspase-8 to unligated integrins. Journal of Cell Biology, 2001. 155 (3): p. 459-470. 195. Soldi, R, Mitola, S, Strasly, M, Defilippi, P, Tarone, G, and Bussolino, F, Role of alpha(v)beta(3) integrin in the activation of vascular endothelial growth factor receptor-2. Embo Journal, 1999. 18 (4): p. 882-892. 196. De, S, Razorenova, O, McCabe, N P, O'Toole, T, Qin, J, and Byzova, T V, VEGFintegrin interplay controls tu mor growth and vascularization. Proceedings of the National Academy of Sciences of the United States of America, 2005. 102 (21): p. 7589-7594. 197. Maile, L A, and Clemmons, D R, Regulation of insulin-like growth factor I receptor dephosphorylation by SHPS-1 and the tyrosine phosphatase SHP-2. Journal of Biological Chemistry, 2002. 277 (11): p. 8955-8960. 198. Ling, Y, Maile, L A, Badley-Clarke, J, and Clemmons, D R, DOK1 mediates SHP2 binding to the alpha V beta 3 integrin and thereby regulates insulin-like growth factor I signaling in cultured vascular smooth muscle cells. Journal of Biological Chemistry, 2005. 280 (5): p. 3151-3158. 199. Calderwood, D A, Fujioka, Y, de Pereda, J M, Garcia-Alvarez, B, Nakamoto, T, Margolis, B, McGlade, C J, Liddi ngton, R C, and Ginsberg, M H, Integrin beta cytoplasmic domain interactions wi th phosphotyrosine-binding domains: A structural prototype for diver sity in integrin signaling. Proceedings of the National Academy of Sciences of the United States of America, 2003. 100 (5): p. 2272-2277. 200. Maile, L A, and Clemmons, D R, The alpha V beta 3 integrin regulates insulin-like growth factor I (IGF-I) receptor phosphorylation by altering the rate of recruitment of the Src-homology 2-c ontaining phosphotyrosine phosphatase-2 to the activated IGF-I receptor. Endocrinology, 2002. 143 (11): p. 4259-4264. 201. Sergent T, B W, Maile LA, Clemmons DR, The heparin binding domain fo vitronectin can enhance IGF1 signaling through activation of alpha(v)beta(3) in an RGD independent manner in Program of the Annual Me eting of The Endocrine Society 2004: New Orleans, LA. 202. Maile LA, C D, Activation of calpain in response to IGF-1 stimulation is required for release of SHP-2 from alpha(v)beta (3) integrin and thereby regulates the transfer of SHP-2 to the IGF-1 receptor. in Program of the Annual Meeting of The Endocrine Society 2004: New Orleans.

PAGE 195

179 203. Gao, A G, Lindberg, F P, Dimitry, J M, Brown, E J, and Frazier, W A, Thrombospondin modulates alpha(v)beta(3) function through integrin-associated protein. Journal of Cell Biology, 1996. 135 (2): p. 533-544. 204. Maile, L A, Badley-Clarke, J, and Clemmons, D R, The association between integrin-associated protein and SHPS-1 re gulates insulin-like growth factor-I receptor signaling in vascular smooth muscle cells. Molecular Biology of the Cell, 2003. 14 (9): p. 3519-3528. 205. Dawson, D W, Volpert, O V, Gillis, P, Cr awford, S E, Xu, H J, Benedict, W, and Bouck, N P, Pigment epithelium-derived factor: A potent inhibitor of angiogenesis. Science, 1999. 285 (5425): p. 245-248. 206. Stellmach, V, Crawford, S E, Zhou, W, and Bouck, N, Prevention of ischemiainduced retinopathy by the natural oc ular antiangiogenic agent pigment epithelium-derived factor. Proceedings of the National Academy of Sciences of the United States of America, 2001. 98 (5): p. 2593-2597. 207. Duh, E J, Yang, H S, Suzuma, I, Miyagi, M, Youngman, E, Mori, K, Katai, M, Yan, L, Suzuma, K, West, K, Davarya, S, Tong, P, Gehlbach, P, Pearlman, J, Crabb, J W, Aiello, L P, Campochiaro, P A, and Zack, D J, Pigment epitheliumderived factor suppresses ischemia-induced retinal ne ovascularization and VEGFinduced migration and growth. Investigative Ophthalmology & Visual Science, 2002. 43 (3): p. 821-829. 208. Spranger J, O M, Reimann M, Mohlig M, Ristow M, Francis MK, Cristofalo V, Hammes HP, Smith G, Boulton M, Pfeiffer AF, Loss of the antiangiogenic pigment epithelium-derived factor in patie nts with angiogen ic eye disease. Daibetes, 2001. 50 (12): p. 2641-2645. 209. Poulsen, J E, The Houssay Phenomenon in Man Recovery from Retinopathy in a Case of Diabetes with Simmonds Disease. Diabetes, 1953. 2 (1): p. 7-12. 210. Merimee, T J, Zapf, J, and Froesch, E R, Insulin-Like Growth-Factors Studies in Diabetics with and without Retinopathy. New England Journal of Medicine, 1983. 309 (9): p. 527-530. 211. Grant, M, Russell, B, Fitzgerald, C, and Merimee, T J, Insulin-Like GrowthFactors in Vitreous Studies in Control and Diabetic Subjects with Neovascularization. Diabetes, 1986. 35 (4): p. 416-420. 212. Dills, D G, Moss, S E, Klein, R, and Klein, B E K, Association of Elevated Igf-I Levels with Increased Retinopat hy in Late-Onset Diabetes. Diabetes, 1991. 40 (12): p. 1725-1730. 213. Smith, L E H, Kopchick, J J, Chen, W, Knapp, J, Kinose, F, Daley, D, Foley, E, Smith, R G, and Schaeffer, J M, Essential role of growth hormone in ischemiainduced retinal neovascularization. Science, 1997. 276 (5319): p. 1706-1709.

PAGE 196

180 214. Hellstrom, A, Perruzzi, C, Ju, M H, Engs trom, E, Hard, A L, Liu, J L, AlbertssonWikland, K, Carlsson, B, Niklasson, A, Sj odell, L, LeRoith, D, Senger, D R, and Smith, L E H, Low IGF-I suppresses VEGF-surv ival signaling in retinal endothelial cells: Direct correlation with clinical retinopathy of prematurity. Proceedings of the National Academy of Scie nces of the United States of America, 2001. 98 (10): p. 5804-5808. 215. Delafontaine, P, Insulin-like growth factor I and its binding proteins in the cardiovascular system. Cardiovascular Research, 1995. 30 : p. 825-834. 216. Delafontaine, P, Song, Y H, and Li, Y X, Expression, regulation, and function of IGF-1, IGF-1R, and IGF-1 binding proteins in blood vessels. Arteriosclerosis Thrombosis and Vascular Biology, 2004. 24 (3): p. 435-444. 217. Smith, L E H, Shen, W, Perruzzi, C, Soker, S, Kinose, F, Xu, X H, Robinson, G, Driver, S, Bischoff, J, Zhang, B, Schaeffer, J M, and Senger, D R, Regulation of vascular endothelial growth factor-dep endent retinal neovascularization by insulin-like growth factor-1 receptor. Nature Medicine, 1999. 5 (12): p. 1390-1395. 218. Liu, W L, Liu, Y Q, and Lowe, W L, The role of phosphatidylinositol 3-kinase and the mitogen-activated protein kinases in insulin-like growth factor-I-mediated effects in vascular endothelial cells. Endocrinology, 2001. 142 (5): p. 1710-1719. 219. Du, J, Meng, X P, and Delafontaine, P, Transcriptional regulation of the insulinlike growth factor-I receptor gene: Evid ence for protein kinase C-dependent and independent pathways. Endocrinology, 1996. 137 (4): p. 1378-1384. 220. Scheidegger, K J, Du, J, and Delafontaine, P, Distinct and common pathways in the regulation of insulin-like growth factor-1 receptor gene expression by angiotensin II and basic fibroblast growth factor. Journal of Biological Chemistry, 1999. 274 (6): p. 3522-3530. 221. Sakai, K, Busby, W H, Clar ke, J B, and Clemmons, D R, Tissue transglutaminase facilitates the polymerization of insulin -like growth factor-binding protein-1 (IGFBP-1) and leads to loss of IGFBP-1's abi lity to inhibit insulin-like growth factor-I-stimulated protein synthesis. Journal of Biological Chemistry, 2001. 276 (12): p. 8740-8745. 222. Moralez, A, Busby, W H, and Clemmons, D, Control of insulin-like growth factor binding protein-5 protease synthesis and secretion by human fibroblasts and porcine aortic smooth muscle cells. Endocrinology, 2003. 144 (6): p. 2489-2495. 223. Firth, S M, and Baxter, R C, Cellular actions of the insulin-like growth factor binding proteins. Endocrine Reviews, 2002. 23 (6): p. 824-854.

PAGE 197

181 224. Baxter, R C, Martin, J L, and Beniac, V A, High Molecular-Weight Insulin-Like Growth-Factor Binding-Protein Complex Purification and Properties of the AcidLabile Subunit from Human-Serum. Journal of Biological Chemistry, 1989. 264 (20): p. 11843-11848. 225. Jones, J I, Gockerman, A, Busby, W H, Wright, G, and Clemmons, D R, InsulinLike Growth-Factor Binding Protein-1 Stim ulates Cell-Migration and Binds to the Alpha-5-Beta-1 Integrin by Means of Its Arg-Gly-Asp Sequence. Proceedings of the National Academy of Sciences of the United States of America, 1993. 90 (22): p. 10553-10557. 226. Perks, C M, Newcomb, P V, Norman, M R, and Holly, J M P, Effect of insulin-like growth factor binding protein-1 on in tegrin signalling and the induction of apoptosis in human breast cancer cells. Journal of Molecu lar Endocrinology, 1999. 22 (2): p. 141-150. 227. Rajah, R, Valentinis, B, and Cohen, P, Insulin like growth factor (IGF)-binding protein-3 induces apoptosis and mediates the effects of transforming growth factorbeta 1 on programmed cell death through a p53and IGF-independent mechanism. Journal of Biological Chemistry, 1997. 272 (18): p. 12181-12188. 228. Firth, S M, Fanayan, S, Benn, D, and Baxter, R C, Development of resistance to insulin-like growth factor binding protein3 in transfected T47D breast cancer cells. Biochemical and Biophysical Research Communications, 1998. 246 (2): p. 325-329. 229. Miyake, H, Nelson, C, Rennie, P S, and Gleave, M E, Overexpression of insulinlike growth factor binding protein-5 helps accelerate progression to androgenindependence in the human prostate L NCaP tumor model through activation of phosphatidylinositol 3 '-kinase pathway. Endocrinology, 2000. 141 (6): p. 22572265. 230. Conover, C A, Glycosylation of Insulin-Like Gr owth-Factor Binding Protein-3 (Igfbp-3) Is Not Required for Potentiation of Igf-I Action Evidence for Processing of Cell-Bound Igfbp-3. Endocrinology, 1991. 129 (6): p. 3259-3268. 231. Karas, M, Danilenko, M, Fishman, D, LeRoith, D, Levy, J, and Sharoni, Y, Membrane-associated insulin-like growth fa ctor-binding protein-3 inhibits insulinlike growth factor-I-induced insulin-like growth factor-I receptor signaling in ishikawa endometrial cancer cells. Journal of Biological Chemistry, 1997. 272 (26): p. 16514-16520. 232. MohseniZadeh, S, and Binoux, M, Insulin-like growth factor (IGF) binding protein-3 interacts with the type 1 IG F receptor, reducing the affinity of the receptor for its ligand: an alternative m echanism in the regulation of IGF action. Endocrinology, 1997. 138 (12): p. 5645-5648.

PAGE 198

182 233. Williams, A C, Collard, T J, Perks, C M, Newcomb, P, Moorghen, M, Holly, J M P, and Paraskeva, C, Increased p53-dependent apoptosis by the insulin-like growth factor binding protein IGFBP-3 in human colonic adenoma-derived cells. Cancer Research, 2000. 60 (1): p. 22-27. 234. Butt, A J, Firth, S M, King, M A, and Baxter, R C, Insulin-like growth factorbinding protein-3 modulates expression of Bax and Bcl-2 and potentiates p53independent radiation-induced apoptosis in human breast cancer cells. Journal of Biological Chemistry, 2000. 275 (50): p. 39174-39181. 235. Hollowood, A D, Lai, T, Perks, C M, Newcomb, P V, Alderson, D, and Holly, J M P, IGFBP-3 prolongs the p53 response and enhances apoptosis following UV irradiation. International Jour nal of Cancer, 2000. 88 (3): p. 336-341. 236. Lee, D Y, Yi, H K, Hwang, P H, and Oh, Y, Enhanced expression of insulin-like growth factor binding protein-3 sensitizes th e growth inhibitory effect of anticancer drugs in gastric cancer cells. Biochemical and Biophysical Research Communications, 2002. 294 (2): p. 480-486. 237. Gucev, Z S, Oh, Y, Kelley, K M, and Rosenfeld, R G, Insulin-like growth factor binding protein 3 mediates retinoic acidand transforming growth factor beta 2induced growth inhibition in human breast cancer cells. Cancer Research, 1996. 56 (7): p. 1545-1550. 238. Huynh, H, Yang, X F, and Pollak, M, Estradiol and antiestrogens regulate a growth inhibitory insulinlike growth fa ctor binding protein 3 autocrine loop in human breast cancer cells. Journal of Biological Chemistry, 1996. 271 (2): p. 10161021. 239. Colston, K W, Perks, C M, Xie, S P, and Holly, J M P, Growth inhibition of both MCF-7 and Hs578T human breast cancer cell lines by vitamin D analogues is associated with increased expression of in sulin-like growth fa ctor binding protein3. Journal of Molecular Endocrinology, 1998. 20 (1): p. 157-162. 240. Rozen, F, Zhang, J C, and Pollak, M, Antiproliferative action of tumor necrosis factor-alpha on MCF-7 breast cancer cells is associated with increased insulin-like growth factor binding protein-3 accumulation. International J ournal of Oncology, 1998. 13 (4): p. 865-869. 241. Ogrady, P, Liu, Q J, Huang, S S, and Huang, J S, Transforming Growth-FactorBeta (Tgf-Beta) Type-V Receptor Has a Tgf-Beta-Stimulated Serine ThreonineSpecific Autophosphorylation Activity. Journal of Biological Chemistry, 1992. 267 (29): p. 21033-21037. 242. Fanayan, S, Firth, S M, and Baxter, R C, Signaling through the Smad pathway by insulin-like growth factor-binding protein-3 in breast cancer cel ls Relationship to transforming growth factor-beta 1 signaling. Journal of Biological Chemistry, 2002. 277 (9): p. 7255-7261.

PAGE 199

183 243. Conover, C A, Bale, L K, Durham, S K, and Powell, D R, Insulin-like growth factor (IGF) binding protein3 potentiation of IGF acti on is mediated through the phosphatidylinositol-3-kinase pathway and is associated with alteration in protein kinase B/AKT sensitivity. Endocrinology, 2000. 141 (9): p. 3098-3103. 244. Martin, J L, and Baxter, R C, Oncogenic ras causes resistance to the growth inhibitor insulin-like growth factor bindi ng protein-3 (IGFBP-3) in breast cancer cells. Journal of Biological Chemistry, 1999. 274 (23): p. 16407-16411. 245. Spagnoli, A, Torello, M, Nagalla, S R, Ho rton, W A, Pattee, P, Hwa, V, Chiarelli, F, Roberts, C T, and Rosenfeld, R G, Identification of STAT-1 as a molecular target of IGFBP-3 in the process of chondrogenesis. Journal of Biological Chemistry, 2002. 277 (21): p. 18860-18867. 246. Boisclair, Y R, Rhoads, R P, Ueki, I, Wang, J, and Ooi, G T, The acid-labile subunit (ALS) of the 150 kDa IGF-bindi ng protein complex: an important but forgotten component of th e circulating IGF system. Journal of Endocrinology, 2001. 170 (1): p. 63-70. 247. Holman, S R, and Baxter, R C, Insulin-like growth factor binding protein-3: Factors affecting binary and ternary complex formation. Growth Regulation, 1996. 6 (1): p. 42-47. 248. Chin, E, Zhou, J, Dai, J, Baxter, R C, and Bondy, C A, Cellular-Localization and Regulation of Gene-Expression for Compone nts of the Insulin-Like Growth-Factor Ternary Binding-Protein Complex. Endocrinology, 1994. 134 (6): p. 2498-2504. 249. Janosi JB, T S, Firth SM, Baxter RC & Delhanty PJD. Histochemical examination of the acid-labile subunit protein in human tissue. in Proceedings of the 5th International Symposium on Insulin-like Growth Factors, 1999. 250. Ooi, G T, Cohen, F J, Tseng, L Y H, Rechler, M M, and Boisclair, Y R, Growth hormone stimulates transcri ption of the gene encodi ng the acid-labile subunit (ALS) of the circulating insulin-like growth factor-binding protein complex and ALS promoter activity in rat liver. Molecular Endocrinology, 1997. 11 (7): p. 9971007. 251. Schindler, C, and Darnell, J E, Transcriptional Responses to Polypeptipe Ligands the Jak-Stat Pathway. Annual Review of Biochemistry, 1995. 64 : p. 621-651. 252. CarterSu, C, Schwartz, J, and Smit, L S, Molecular mechanism of growth hormone action. Annual Review of Physiology, 1996. 58 : p. 187-207. 253. Zapf, J, Hauri, C, Futo, E, Hussain, M, Rutishauser, J, Maack, C A, and Froesch, E R, Intravenously Injected Insulin-Like Grow th-Factor (Igf) I/Igf Binding Protein-3 Complex Exerts Insulin-Li ke Effects in Hypophysectomiz ed, but Not in Normal Rats. Journal of Clinic al Investigation, 1995. 95 (1): p. 179-186.

PAGE 200

184 254. Rother, K I, and Accili, D, Role of insulin receptors and IGF receptors in growth and development. Pediatric Nephrology, 2000. 14 (7): p. 558-561. 255. Kondo, T, Vicent, D, Suzuma, K, Yanagi sawa, M, King, G L, Holzenberger, M, and Kahn, C R, Knockout of insulin and IGF-1 receptors on vascular endothelial cells protects against re tinal neovascularization. Journal of Clini cal Investigation, 2003. 111 (12): p. 1835-1842. 256. Bronson, S K, Reiter, C E N, and Gardner, T W, An eye on insulin. Journal of Clinical Investigation, 2003. 111 (12): p. 1817-1819. 257. Fritz, J J, Lewin, A, Hauswirth, W, Agarwal, A, Grant, M, and Shaw, L, Development of hammerhead ribozymes to modulate endogenous gene expression for functional studies. Methods, 2002. 28 (2): p. 276-285. 258. Kurreck, J, Antisense technologies Improvement through novel chemical modifications. European Journal of Biochemistry, 2003. 270 (8): p. 1628-1644. 259. Zamecnik, P C, and Stephenson, M L, Inhibition of Rous-Sarcoma VirusReplication and Cell Transformation by a Specific Oligodeoxynucleotide. Proceedings of the National Academy of Scie nces of the United States of America, 1978. 75 (1): p. 280-284. 260. Shaw, L C, and Lewin, A S, Protein-Induced Folding of a Group-I Intron in Cytochrome-B Pre-Messenger-Rna. Journal of Biological Chemistry, 1995. 270 (37): p. 21552-21562. 261. Milner, N, Mir, K U, and Southern, E M, Selecting effective antisense reagents on combinatorial oligonucleotide arrays. Nature Biotechnology, 1997. 15 (6): p. 537541. 262. Brown, D A, Kang, S H, Gryaznov, S M, Dedionisio, L, Heidenreich, O, Sullivan, S, Xu, X, and Nerenberg, M I, Effect of Phosphorothi oate Modification of Oligodeoxynucleotides on Specific Protein-Binding. Journal of Biological Chemistry, 1994. 269 (43): p. 26801-26805. 263. Fedora, M J, and Williamson, J R, The catalytic diversity of RNAS. Nature Reviews Molecular Cell Biology, 2005. 6 (5): p. 399-412. 264. Tsui, L C, The Spectrum of Cystic-Fibrosis Mutations. Trends in Genetics, 1992. 8 (11): p. 392-398. 265. Michel, F, Hanna, M, Green, R, Bartel, D P, and Szostak, J W, The Guanosine Binding-Site of the Tetrahymena Ribozyme. Nature, 1989. 342 (6248): p. 391-395. 266. Frank, D N, and Pace, N R, Ribonuclease P: Unity and diversity in a tRNA processing ribozyme. Annual Review of Biochemistry, 1998. 67 : p. 153-180.

PAGE 201

185 267. Altman S, K L I G R, Cech TR, Atkins JF, editors., In the RNA World. 2nd ed. 1999, Cold Spring Harbor, NY.: Cold Spring Harbor Laboratory Press. 351-380. 268. Guerriertakada, C, Gardiner, K, Marsh, T, Pace, N, and Altman, S, The Rna Moiety of Ribonuclease-P Is the Cata lytic Subunit of the Enzyme. Cell, 1983. 35 (3): p. 849857. 269. Niranjanakumari, S, Stams, T, Crary, S M, Christianson, D W, and Fierke, C A, Protein component of the ribozyme ribonuc lease P alters substrate recognition by directly contacting precursor tRNA. Proceedings of the National Academy of Sciences of the United States of America, 1998. 95 (26): p. 15212-15217. 270. Reich, C, Olsen, G J, Pace, B, and Pace, N R, Role of the Protein Moiety of Ribonuclease-P, a Ribon ucleoprotein Enzyme. Science, 1988. 239 (4836): p. 178181. 271. Bothwell, A L M, Garber, R L, and Altman, S, Nucleotide-Sequence and Invitro Processing of a Precursor Molecu le to Escherichia-Coli 4.5 S Rna. Journal of Biological Chemistry, 1976. 251 (23): p. 7709-7716. 272. Alifano, P, Rivellini, F, Piscitelli, C, Arraiano, C M, Bruni, C B, and Carlomagno, M S, Ribonuclease-E Provides Substrates for Ribonuclease P-Dependent Processing of a Polycistronic Messenger-Rna. Genes & Development, 1994. 8 (24): p. 3021-3031. 273. Komine, Y, Kitabatake, M, Yokogawa, T, Nishikawa, K, and Inokuchi, H, A Transfer-Rna-Like Structure Is Present in 10sa Rna, a Small Stable Rna from Escherichia-Coli. Proceedings of the National Acad emy of Sciences of the United States of America, 1994. 91 (20): p. 9223-9227. 274. Forster, A C, and Altman, S, External Guide Sequences for an Rna Enzyme. Science, 1990. 249 (4970): p. 783-786. 275. Raj, S M L, and Liu, F Y, Engineering of RNase P ribozyme for gene-targeting applications. Gene, 2003. 313 : p. 59-69. 276. Hutchins, C J, Rathjen, P D, Fo rster, A C, and Symons, R H, Self-Cleavage of Plus and Minus Rna Transcripts of Avocado Sunblotch Viroid. Nucleic Acids Research, 1986. 14 (9): p. 3627-3640. 277. Buzayan, J M, Gerlach, W L, and Bruening, G, Nonenzymatic Cleavage and Ligation of Rnas Complementary to a Plant-Virus Satellite Rna. Nature, 1986. 323 (6086): p. 349-353. 278. Pley, H W, Flaherty, K M, and Mckay, D B, 3-Dimensional Structure of a Hammerhead Ribozyme. Nature, 1994. 372 (6501): p. 68-74.

PAGE 202

186 279. Scott, W G, Finch, J T, and Klug, A, The Crystal-Structure of an All-Rna Hammerhead Ribozyme a Proposed Mechanism for Rna Catalytic Cleavage. Cell, 1995. 81 (7): p. 991-1002. 280. Branch, A D, and Robertson, H D, A Replication Cycle for Viroids and Other Small Infectious Rnas. Science, 1984. 223 (4635): p. 450-455. 281. Shimayama, T, Nishikawa, S, and Taira, K, Generality of the Nux Rule KineticAnalysis of the Results of Systematic Muta tions in the Trinucleotide at the Cleavage Site of Hammerhead Ribozymes. Biochemistry, 1995. 34 (11): p. 3649-3654. 282. Murray, J B, Seyhan, A A, Walter, N G, Burke, J M, and Scott, W G, The hammerhead, hairpin and VS ribozymes are catalytically profic ient in monovalent cations alone. Chemistry & Biology, 1998. 5 (10): p. 587-595. 283. Joyce, G F, RNA cleavage by the 10-23 DNA enzyme. Ribonucleases, Pt A, 2001. 341 : p. 503-517. 284. Cedergren, R, Rna the Catalyst. Biochemistry and Cell Biology-Biochimie Et Biologie Cellulaire, 1990. 68 (6): p. 903-906. 285. Hampel, A, Tritz, R, Hicks, M, and Cruz, P, Hairpin Catalytic Rna Model Evidence for Helices and Sequen ce Requirement for Substrate Rna. Nucleic Acids Research, 1990. 18 (2): p. 299-304. 286. Earnshaw, D J, and Gait, M J, Hairpin ribozyme cleavage catalyzed by aminoglycoside antibiotics and the polyamin e spermine in the absence of metal ions. Nucleic Acids Research, 1998. 26 (24): p. 5551-5561. 287. Khan AU, L S, The white halo plaque phenotype of bacteriophage T4: Its uses and applications in screen ing and mapping of splici ngdefective mutants. J Biochem Mol Biol Biophys, 2001. 5 : p. 237-242. 288. Kuo, M Y P, Sharmeen, L, Dintergottlieb, G, and Taylor, J, Characterization of Self-Cleaving Rna Sequences on the Ge nome and Antigenome of Human Hepatitis Delta-Virus. Journal of Virology, 1988. 62 (12): p. 4439-4444. 289. Collins, R A, and Saville, B J, Independent Transfer of Mitochondrial Chromosomes and Plasmids during Unst able Vegetative Fusion in Neurospora. Nature, 1990. 345 (6271): p. 177-179. 290. Fire, A, Xu, S Q, Montgomery, M K, Ko stas, S A, Driver, S E, and Mello, C C, Potent and specific genetic interference by double-stranded RNA in Caenorhabditis elegans. Nature, 1998. 391 (6669): p. 806-811. 291. Lavorgna, G, Dahary, D, Lehner, B, So rek, R, Sanderson, C M, and Casari, G, In search of antisense. Trends in Biochemical Sciences, 2004. 29 (2): p. 88-94.

PAGE 203

187 292. Dykxhoorn DM, P D, Lieberman J., The silent treatment: siRNAs as small molecule drugs. Gene Therapy, 2006. 13 (6): p. 541-552. 293. Barik, S, Development of gene-specific double-stranded RNA drugs. Annals of Medicine, 2004. 36 (7): p. 540-551. 294. Schwarz, D S, Hutvagner, G, Haley, B, and Zamore, P D, Evidence that siRNAs function as guides, not primers, in the Drosophila and human RNAi pathways. Molecular Cell, 2002. 10 (3): p. 537-548. 295. Nykanen, A, Haley, B, and Zamore, P D, ATP requirements and small interfering RNA structure in the RNA interference pathway. Cell, 2001. 107 (3): p. 309-321. 296. Chiu, Y L, and Rana, T M, RNAi in human cells: Basic structural and functional features of small interfering RNA. Molecular Cell, 2002. 10 (3): p. 549-561. 297. Bartel, D P, MicroRNAs: Genomics, biogene sis, mechanism, and function. Cell, 2004. 116 (2): p. 281-297. 298. Lee, Y, Ahn, C, Han, J J, Choi, H, Kim, J, Yim, J, Lee, J, Provost, P, Radmark, O, Kim, S, and Kim, V N, The nuclear RNase III Drosha initiates microRNA processing. Nature, 2003. 425 (6956): p. 415-419. 299. Lund, E, Guttinger, S, Calado, A, Dahlberg, J E, and Kutay, U, Nuclear export of microRNA precursors. Science, 2004. 303 (5654): p. 95-98. 300. Jackson, A L, and Linsley, P S, Noise amidst the silence: off-target effects of siRNAs? Trends in Genetics, 2004. 20 (11): p. 521-524. 301. Paddison, P J, Caudy, A A, Bernstein, E, Hannon, G J, and Conklin, D S, Short hairpin RNAs (shRNAs) induce sequencespecific silencing in mammalian cells. Genes & Development, 2002. 16 (8): p. 948-958. 302. Brummelkamp, T R, Bernards, R, and Agami, R, A system for stable expression of short interfering RNAs in mammalian cells. Science, 2002. 296 (5567): p. 550-553. 303. BR., C, Induction of stable RNA interference in mammalian cells. Gene Therapy, 2006. 13 (6): p. 503-508. 304. Green, R, and Lorsch, J R, The path to perdition is paved with protons. Cell, 2002. 110 (6): p. 665-668. 305. Benkovic, S, Schray, K., Then Enzymes ed. Boyer. Vol. 8. 1973, New York: Academic Press. 201-238. 306. Young, K J, Gill, F, and Grasby, J A, Metal ions play a passi ve role in the hairpin ribozyme catalysed reaction. Nucleic Acids Research, 1997. 25 (19): p. 3760-3766.

PAGE 204

188 307. Zhang, H D, Kolb, F A, Brondani, V, Billy, E, and Filipowicz, W, Human Dicer preferentially cleaves dsRN As at their termini without a requirement for ATP. Embo Journal, 2002. 21 (21): p. 5875-5885. 308. Reynolds, A, Leake, D, Boese, Q, Scaringe, S, Marshall, W S, and Khvorova, A, Rational siRNA design for RNA interference. Nature Biotechnology, 2004. 22 (3): p. 326-330. 309. Kim, D H, Behlke, M A, Rose, S D, Chang, M S, Choi, S, and Rossi, J J, Synthetic dsRNA Dicer substrates enhan ce RNAi potency and efficacy. Nature Biotechnology, 2005. 23 (2): p. 222-226. 310. Bitko, V, Musiyenko, A, Shulyayeva, O, and Barik, S, Inhibition of respiratory viruses by nasally administered siRNA. Nature Medicine, 2005. 11 (1): p. 50-55. 311. Hutvagner, G, McLachlan, J, Pasquinelli, A E, Balint, E, Tuschl, T, and Zamore, P D, A cellular function for the RNA-interference enzyme Dicer in the maturation of the let-7 small temporal RNA. Science, 2001. 293 (5531): p. 834-838. 312. Persengiev, S P, Zhu, X C, and Green, M R, Nonspecific, concentration-dependent stimulation and repression of mammalian gene expression by small interfering RNAs (siRNAs). Rna-a Publication of the Rna Society, 2004. 10 (1): p. 12-18. 313. Sledz, C A, Holko, M, de Veer, M J, Silverman, R H, and Williams, B R G, Activation of the interferon system by short-interfering RNAs. Nature Cell Biology, 2003. 5 (9): p. 834-839. 314. Bridge, A J, Pebernard, S, Ducra ux, A, Nicoulaz, A L, and Iggo, R, Induction of an interferon response by RNAi vectors in mammalian cells. Nature Genetics, 2003. 34 (3): p. 263-264. 315. Soutschek, J, Akinc, A, Bramlage, B, Charisse, K, Constien, R, Donoghue, M, Elbashir, S, Geick, A, Hadwiger, P, Har borth, J, John, M, Kesavan, V, Lavine, G, Pandey, R K, Racie, T, Rajeev, K G, Rohl I, Toudjarska, I, Wang, G, Wuschko, S, Bumcrot, D, Koteliansky, V, Limmer, S, Manoharan, M, and Vornlocher, H P, Therapeutic silencing of an endogenous gene by syst emic administration of modified siRNAs. Nature, 2004. 432 (7014): p. 173-178. 316. Layzer, J M, McCaffrey, A P, Tanner, A K, Huang, Z, Kay, M A, and Sullenger, B A, In vivo activity of nucl ease-resistant siRNAs. Rna-a Publication of the Rna Society, 2004. 10 (5): p. 766-771. 317. Schiffelers, R M, Ansari, A, Xu, J, Zhou, Q, Tang, Q Q, Storm, G, Molema, G, Lu, P Y, Scaria, P V, and Woodle, M C, Cancer siRNA therapy by tumor selective delivery with ligand-targeted sterically stabilized nanoparticle. Nucleic Acids Research, 2004. 32 (19): p. -.

PAGE 205

189 318. Harborth, J, Elbashir, S M, Vandenburgh, K, Manninga, H, Scaringe, S A, Weber, K, and Tuschl, T, Sequence, chemical, and structural variation of small interfering RNAs and short hairpin RNAs and the effect on mammalian gene silencing. Antisense & Nucleic Acid Drug Development, 2003. 13 (2): p. 83-105. 319. Chiu, Y L, and Rana, T M, siRNA function in RNAi: A chemical modification analysis. Rna-a Publication of the Rna Society, 2003. 9 (9): p. 1034-1048. 320. Czauderna, F, Fechtner, M, Dames, S, Aygun, H, Klippel, A, Pronk, G J, Giese, K, and Kaufmann, J, Structural variations and stabili sing modifications of synthetic siRNAs in mammalian cells. Nucleic Acids Research, 2003. 31 (11): p. 2705-2716. 321. Amarzguioui, M, Holen, T, Babaie, E, and Prydz, H, Tolerance for mutations and chemical modifications in a siRNA. Nucleic Acids Research, 2003. 31 (2): p. 589595. 322. Snyder, R O, Miao, C H, Patijn, G A, Sp ratt, S K, Danos, O, Nagy, D, Gown, A M, Winther, B, Meuse, L, Cohen, L K, Thompson, A R, and Kay, M A, Persistent and therapeutic concentrations of human factor IX in mice after hepatic gene transfer of recombinant AAV vectors. Nature Genetics, 1997. 16 (3): p. 270-276. 323. Gerard, R D, and Collen, D, Adenovirus gene therapy for hypercholesterolemia, thrombosis and restenosis. Cardiovascular Research, 1997. 35 (3): p. 451-458. 324. Virus Vectors & Gene Therapy:Problems, Promises & Prospects 2004, Microbiology @ Leicester. 325. Robbins, P D, and Ghivizzani, S C, Viral vectors for gene therapy. Pharmacology & Therapeutics, 1998. 80 (1): p. 35-47. 326. Wang, C Y, and Huang, L, Ph-Sensitive Immunoliposomes Mediate Target-CellSpecific Delivery and Controlled Expre ssion of a Foreign Gene in Mouse. Proceedings of the National Academy of Scie nces of the United States of America, 1987. 84 (22): p. 7851-7855. 327. Stavridis, J C, Deliconstantinos, G, Psallidopoulos, M C, Armenakas, N A, Hadjiminas, D J, and Hadjiminas, J, Construction of Transfe rrin-Coated Liposomes for Invivo Transport of Exogenous DNA to Bone-Marrow Erythroblasts in Rabbits. Experimental Cell Research, 1986. 164 (2): p. 568-572. 328. Dzau, V J, Mann, M J, Morishita, R, and Kaneda, Y, Fusigenic viral liposome for gene therapy in cardiovascular diseases. Proceedings of the National Academy of Sciences of the United States of America, 1996. 93 (21): p. 11421-11425. 329. Kootstra, N A, and Verma, I M, Gene therapy with viral vectors. Annual Review of Pharmacology and Toxicology, 2003. 43 : p. 413-439.

PAGE 206

190 330. Lu, Y, Recombinant adeno-associated virus as delivery vector for gene therapy A review. Stem Cells and Development, 2004. 13 (1): p. 133-145. 331. Hallahan, D E, Mauceri, H J, Seung, L P, Dunphy, E J, Wayne, J D, Hanna, N N, Toledano, A, Hellman, S, Kufe, D W, and Weichselbaum, R R, Spatial and Temporal Control of Gene-The rapy Using Ionizing-Radiation. Nature Medicine, 1995. 1 (8): p. 786-791. 332. Rivera, V M, Clackson, T, Natesan, S, Pollo ck, R, Amara, J F, Keenan, T, Magari, S R, Phillips, T, Courage, N L, Cerasoli, F, Holt, D A, and Gilman, M, A humanized system for pharmacologic control of gene expression. Nature Medicine, 1996. 2 (9): p. 1028-1032. 333. Wang, Y, Xu, J, Pierson, T, OMalley, B W, and Tsai, S Y, Positive and negative regulation of gene expression in eukaryotic cells with an inducible transcriptional regulator. Gene Therapy, 1997. 4 (5): p. 432-441. 334. Merten, O W, Geny-Fiamma C, and Douar, A M, Current issues in adenoassociated viral vector production. Gene Therapy, 2005. 12 : p. S51-S61. 335. Li, C W, Bowles, D E, van Dyke, T, and Samulski, R J, Adeno-associated virus vectors: potential applications for cancer gene therapy. Cancer Gene Therapy, 2005. 12 (12): p. 913-925. 336. Summerford, C, and Samulski, R J, Membrane-associated heparan sulfate proteoglycan is a receptor for adeno-associated virus type 2 virions. Journal of Virology, 1998. 72 (2): p. 1438-1445. 337. Summerford, C, Bartlett, J S, and Samulski, R J, alpha V beta 5 integrin: a coreceptor for adeno-associated virus type 2 infection. Nature Medicine, 1999. 5 (1): p. 78-82. 338. Qing, K, Mah, C, Hansen, J, Zhou, S Z, Dwarki, V, and Srivastava, A, Human fibroblast growth factor receptor 1 is a co-receptor for infection by adenoassociated virus 2. Nature Medicine, 1999. 5 (1): p. 71-77. 339. Bartlett, J S, Wilcher, R, and Samulski, R J, Infectious entry pathway of adenoassociated virus and adeno-associated virus vectors. Journal of Virology, 2000. 74 (6): p. 2777-2785. 340. McCarty, D M, Monahan, P E, and Samulski, R J, Self-complementary recombinant adeno-associated virus (scAAV) vectors promote efficient transduction independently of DNA synthesis. Gene Therapy, 2001. 8 (16): p. 12481254. 341. Young, S M, McCarty, D M, Degtya reva, N, and Samulski, R J, Roles of adenoassociated virus Rep protein and hum an chromosome 19 in site-specific recombination. Journal of Virology, 2000. 74 (9): p. 3953-3966.

PAGE 207

191 342. Xiao, X, Li, J, and Samulski, R J, Production of high-tite r recombinant adenoassociated virus vectors in th e absence of helper adenovirus. Journal of Virology, 1998. 72 (3): p. 2224-2232. 343. Grimm, D, Kay, M A, and Kleinschmidt, J A, Helper virus-free, optically controllable, and two-plasmid-based produc tion of adeno-associated virus vectors of serotypes 1 to 6. Molecular Therapy, 2003. 7 (6): p. 839-850. 344. Zolotukhin, S, Byrne, B J, Mason, E, Zolotukhin, I, Potter, M, Chesnut, K, Summerford, C, Samulski, R J, and Muzyczka, N, Recombinant adeno-associated virus purification using no vel methods improves in fectious titer and yield. Gene Therapy, 1999. 6 (6): p. 973-985. 345. Clark, K R, Liu, X L, McGr ath, J P, and Johnson, P R, Highly purified recombinant adeno-associated virus vectors are biologica lly active and free of detectable helper and wild-type viruses. Human Gene Therapy, 1999. 10 (6): p. 1031-1039. 346. Bartlett, J S, Kleinschmidt, J, Boucher, R C, and Samulski, R J, Targeted adenoassociated virus vector transduction of nonpermissive cells mediated by a bispecific F(ab 'gamma)(2) antibody. Nature Biotechnology, 1999. 17 (2): p. 181186. 347. Wu, P, Xiao, W, Conlon, T, Hughes, J, Agbandje-McKenna, M, Ferkol, T, Flotte, T, and Muzyczka, N, Mutational analysis of the ade no-associated virus type 2 (AAV2) capsid gene and construction of AAV2 vectors with altered tropism. Journal of Virology, 2000. 74 (18): p. 8635-8647. 348. Worgall, S, Wolff, G, FalckP edersen, E, and Crystal, R G, Innate immune mechanisms dominate elimination of ade noviral vectors following in vivo administration. Human Gene Therapy, 1997. 8 (1): p. 37-44. 349. Hong, S S, Karayan, L, Tournier, J, Curiel, D T, and Boulanger, P A, Adenovirus type 5 fiber knob binds to MHC class I alpha 2 domain at the surface of human epithelial and B ly mphoblastoid cells. Embo Journal, 1997. 16 (9): p. 2294-2306. 350. Bergelson, J M, Cunningham, J A, Droguett G, KurtJones, E A, Krithivas, A, Hong, J S, Horwitz, M S, Crowell, R L, and Finberg, R W, Isolation of a common receptor for coxsackie B viruses and adenoviruses 2 and 5. Science, 1997. 275 (5304): p. 1320-1323. 351. Rother, R P, Fodor, W L, Springhorn, J P, Birks, C W, Setter, E, Sandrin, M S, Squinto, S P, and Rollins, S A, A Novel Mechanism of Retrovirus Inactivation in Human Serum Mediated by Anti-Al pha-Galactosyl Natural Antibody. Journal of Experimental Medicine, 1995. 182 (5): p. 1345-1355.

PAGE 208

192 352. Rollins, S A, Birks, C W, Setter, E, Squinto, S P, and Rother, R P, Retroviral vector producer cell killing in human serum is mediated by natural antibody and complement: Strategies for evadi ng the humoral immune response. Human Gene Therapy, 1996. 7 (5): p. 619-626. 353. Shaw, L C, Afzal, A, Lewin, A S, Timmers A M, Spoerri, P E, and Grant, M B, Decreased expression of the insulin-like growth factor 1 receptor by ribozyme cleavage. Investigative Ophthalmol ogy & Visual Science, 2003. 44 (9): p. 41054113. 354. Grant, M B, and Guay, C, Plasminogen-Activator Produc tion by Human Retinal Endothelial-Cells of Nondiabetic and Diabetic Origin. Investigative Ophthalmology & Visual Science, 1991. 32 (1): p. 53-64. 355. Rubini, M, Hongo, A, DAmbrosio, C, and Baserga, R, The IGF-I receptor in mitogenesis and transformation of mouse em bryo cells: Role of receptor number. Experimental Cell Research, 1997. 230 (2): p. 284-292. 356. Autiero, M, Waltenberger, J, Communi, D, Kranz, A, Moons, L, Lambrechts, D, Kroll, J, Plaisance, S, De Mol, M, Bono, F, Kliche, S, Fellbrich, G, Ballmer-Hofer, K, Maglione, D, Mayr-Beyrle, U, Dewerc hin, M, Dombrowski, S, Stanimirovic, D, Van Hummelen, P, Dehio, C, Hicklin, D J, Persico, G, Herbert, J M, Communi, D, Shibuya, M, Collen, D, Conway, E M, and Carmeliet, P, Role of PIGF in the intraand intermolecular cross talk betwee n the VEGF receptors Flt1 and Flk1. Nature Medicine, 2003. 9 (7): p. 936-943. 357. Hiratsuka, S, Maru, Y, Okada, A, Seiki, M, Noda, T, and Shibuya, M, Involvement of Flt-1 tyrosine kinase (vascular endot helial growth factor receptor-1) in pathological angiogenesis. Cancer Research, 2001. 61 (3): p. 1207-1213. 358. Bussolati, B, Dunk, C, Grohman, M, K ontos, C D, Mason, J, and Ahmed, A, Vascular endothelial growth factor r eceptor-1 modulates vascular endothelial growth factor-mediated angiogenesis via nitric oxide. American Journal of Pathology, 2001. 159 (3): p. 993-1008. 359. Hattori, K, Heissig, B, Wu, Y, Dias, S, Te jada, R, Ferris, B, Hicklin, D J, Zhu, Z P, Bohlen, P, Witte, L, Hendrikx, J, Hackett, N R, Crystal, R G, Moore, M A S, Werb, Z, Lyden, D, and Rafii, S, Placental growth factor reconstitutes hematopoiesis by recruiting VEGFR1(+) stem cells from bone-marrow microenvironment. Nature Medicine, 2002. 8 (8): p. 841-849. 360. Tripathi, R C, Li, J P, Tripathi, B J, Chalam, K V, and Adamis, A P, Increased level of vascular endothelial growth fact or in aqueous humor of patients with neovascular glaucoma. Ophthalmology, 1998. 105 (2): p. 232-237. 361. Spoerri PE, R W, Player D, Groome AB, Alexander T, Bodkin NL, Hansen BC, Grant MB., Diabetes related increase in plasminogen activator inhibitor-1 expression in monkey retinal capillaries. Int J Diabetes, 1998. 6

PAGE 209

193 362. Grant, M B, Ellis, E A, Caballero, S, and Mames, R N, Plasminogen activator inhibitor-1 overexpression in nonpr oliferative diabetic retinopathy. Experimental Eye Research, 1996. 63 (3): p. 233-244. 363. Gaudry, M, Brgerie, O, Andrieu, V, ElBe nna, J, Pocidalo, M A, and Hakim, J, Intracellular pool of vascular endothelia l growth factor in human neutrophils. Blood, 1997. 90 (10): p. 4153-4161. 364. Ishida, S, Usui, T, Yamashiro, K, Kaji, Y, Amano, S, Ogura, Y, Hida, T, Oguchi, Y, Ambati, J, Miller, J W, Gragoudas, E S, Ng, Y S, D'Amore, P A, Shima, D T, and Adamis, A P, VEGF(164)-mediated inflammation is required for pathological, but not physiological, ischemia-i nduced retinal neovascularization. Journal of Experimental Medicine, 2003. 198 (3): p. 483-489. 365. LC Shaw, H P, A Afzal, SL Calzi, PE Spoerri, SM Sullivan and MB Grant, Proliferating endothelial cell-specific expression of IGF-I receptor ribozyme inhibits retinal neovascularization. Gene Therapy, 2006: p. 1-9. 366. Anwer, K, Kao, G, Proctor, B, Rolland, A, and Sullivan, S, Optimization of cationic lipid/DNA complexes for system ic gene transfer to tumor lesions. Journal of Drug Targeting, 2000. 8 (2): p. 125-135. 367. Anwer, K, Meaney, C, Kao, G, Hussain, N, Shelvin, R, Earls, R M, Leonard, P, Quezada, A, Rolland, A P, and Sullivan, S M, Cationic lipid-based delivery system for systemic cancer gene therapy. Cancer Gene Therapy, 2000. 7 (8): p. 1156-1164. 368. Gariano, R F, and Gardner, T W, Retinal angiogenesis in development and disease. Nature, 2005. 438 (7070): p. 960-966. 369. Harris, A, Arend, O, Danis, R P, Ev ans, D, Wolf, S, and Martin, B J, Hyperoxia improves contrast sensitivity in early diabetic retinopathy. British Journal of Ophthalmology, 1996. 80 (3): p. 209-213. 370. Nguyen, Q D, Shah, S M, Van Anden, E, Sung, J U, Vitale, S, and Campochiaro, P A, Supplemental oxygen improves diabetic macular edema: A pilot study. Investigative Ophthalmol ogy & Visual Science, 2004. 45 (2): p. 617-624. 371. Stefansson, E, Hatchell, D L, Fisher, B L, Sutherland, F S, and Machemer, R, Panretinal Photocoagulation and Retinal Oxygenation in Normal and Diabetic Cats. American Journal of Ophthalmology, 1986. 101 (6): p. 657-664. 372. Stefansson, E, Peterson, J I, and Wang, Y H, Intraocular Oxygen-Tension Measured with a Fiber-Optic Sensor in Normal and Diabetic Dogs. American Journal of Physiology, 1989. 256 (4): p. H1127-H1133. 373. Linsenmeier, R A, Braun, R D, McRiple y, M A, Padnick, L B, Ahmed, J, Hatchell, D L, McLeod, D S, and Lutty, G A, Retinal hypoxia in l ong-term diabetic cats. Investigative Ophthalmol ogy & Visual Science, 1998. 39 (9): p. 1647-1657.

PAGE 210

194 374. Poulaki, V, Joussen, A M, Mitsiades, N, Mitsiades, C S, Iliaki, E F, and Adamis, A P, Insulin-like growth factor -I plays a pathogenetic role in diabetic retinopathy. American Journal of Pathology, 2004. 165 (2): p. 457-469. 375. Gardner, J L, and Lisberger, S G, Serial linkage of target selection for orienting and tracking eye movements. Nature Neuroscience, 2002. 5 (9): p. 892-899. 376. Holcik, M, Sonenberg, N, and Korneluk, R G, Internal ribosome initiation of translation and the control of cell death. Trends in Genetics, 2000. 16 (10): p. 469473. 377. Dykxhoorn, D M, Novina, C D, and Sharp, P A, Killing the messenger: Short RNAs that silence gene expression. Nature Reviews Molecular Cell Biology, 2003. 4 (6): p. 457-467. 378. Dykxhoorn, D M, and Lieberman, J, The silent revolution: RNA interference as basic biology, research tool, and therapeutic. Annual Review of Medicine, 2005. 56 : p. 401-423. 379. Bertrand, J R, Pottier, M, Vekris, A, Opolon, P, Maksimenko, A, and Malvy, C, Comparison of antisense oligonucleotides and siRNAs in cell culture and in vivo. Biochemical and Biophysical Re search Communications, 2002. 296 (4): p. 10001004. 380. Tassakka, A C M A R, Savan, R, Watanuki, H, and Sakai, M, The in vitro effects of CpG oligodeoxynucleotides on the expre ssion of cytokine genes in the common carp (Cyprinus carpio L.) head kidney cells. Veterinary Immunology and Immunopathology, 2006. 110 (1-2): p. 79-85. 381. Wong-Staal, F, Poeschla, E M, and Looney, D J, A controlled, phase 1 clinical trial to evaluate the safety and effects in HIV-1 infected humans of autologous lymphocytes transduced with a ri bozyme that cleaves HIV-1 RNA. Human Gene Therapy, 1998. 9 (16): p. 2407-2425. 382. Amado, R G, Mitsuyasu, R T, Symonds, G, Rosenblatt, J D, Zack, J, Sun, L O, Miller, M, Ely, J, and Gerlach, W, A phase I trial of autologous CD34(+) hematopoietic progenitor cells trans duced with an anti-HIV ribozyme. Human Gene Therapy, 1999. 10 (13): p. 2255-+. 383. Winkler, W C, Nahvi, A, Roth, A, Collins, J A, and Breaker, R R, Control of gene expression by a natural me tabolite-responsive ribozyme. Nature, 2004. 428 (6980): p. 281-286. 384. Teixeira, A, Tahiri-Alaoui, A, West, S, Thomas, B, Ramadass, A, Martianov, I, Dye, M, James, W, Proudfoot, N J, and Akoulitchev, A, Autocatalytic RNA cleavage in the human beta-globin pre-mR NA promotes transcription termination. Nature, 2004. 432 (7016): p. 526-530.

PAGE 211

195 385. Sarver, N, Cantin, E M, Chang, P S, Za ia, J A, Ladne, P A, Stephens, D A, and Rossi, J J, Ribozymes as Potential Anti-Hiv-1 Therapeutic Agents. Science, 1990. 247 (4947): p. 1222-1225. 386. Macejak, D G, Jensen, K L, Pavco, P A, Phipps, K M, Heinz, B A, Colacino, J M, and Blatt, L M, Enhanced antiviral effect in cell culture of type 1 interferon and ribozymes targeting HCV RNA. Journal of Viral Hepatitis, 2001. 8 (6): p. 400-405. 387. Macejak, D G, Jensen, K L, Jamison, S F, Domenico, K, Roberts, E C, Chaudhary, N, von Carlowitz, I, Bellon, L, Tong, M J, Conrad, A, Pavco, P A, and Blatt, L M, Inhibition of hepatitis C vi rus (HCV)-RNA-dependent translation and replication of a chimeric HCV poliovirus using synthetic stabilized ribozymes. Hepatology, 2000. 31 (3): p. 769-776. 388. Khan, A U, and Lal, S K, Ribozymes: A modern tool in medicine. Journal of Biomedical Science, 2003. 10 (5): p. 457-467. 389. Hsieh SY, T J, Delta virus as a vector for the delivery of biologically-active RNAs: possibly a ribozyme specific for chro nic hepatitis B virus infection. Adv Exp Med Biol., 1992. 312 : p. 125-128. 390. Netter HJ, H S, Lazinski D, Taylor J., Modified HDV as a vector for the delivery of biologically-active RNAs. Prog Clin Biol Res., 1993. 382 : p. 373-376. 391. Cobaleda, C, Perez-Losada, J, and Sanchez-Garcia, I, Chromosomal abnormalities and tumor development: from genes to therapeutic mechanisms. Bioessays, 1998. 20 (11): p. 922-930. 392. Bitko V, B S, Phenotypic silencing of cytoplasm ic genes using sequence-specific double-stranded short interfering RNA and its application in the re verse genetics of wild type negative-strand RNA viruses. BMC Microbiol., 2001. 1 (34). 393. Hommel, J D, Sears, R M, Georgesc u, D, Simmons, D L, and DiLeone, R J, Local gene knockdown in the brain using viral-mediated RNA interference. Nature Medicine, 2003. 9 (12): p. 1539-1544. 394. Shen, J, Samul, R, Silva, R L, Akiyama, H, Liu, H, Saishin, Y, Hackett, S F, Zinnen, S, Kossen, K, Fosnaugh, K, Vargeese, C, Gomez, A, Bouhana, K, Aitchison, R, Pavco, P, and Campochiaro, P A, Suppression of ocular neovascularization with siRNA targeting VEGF receptor 1. Gene Therapy, 2006. 13 (3): p. 225-234. 395. Reich, S, Fosnot, J, Kuroki, A, Tang, W X, Yang, X Y, Maguire, A, Bennett, J, and Tolentino, M, Small interfering RNA (siRNA) targ eting VEGF effectively inhibits ocular neovascularization in a mouse model. Molecular Vision, 2003. 9 (31-32): p. 210-216.

PAGE 212

196 396. Kim, B, Tang, Q Q, Biswas, P S, Xu, J, Schiffelers, R M, Xie, F Y, Ansari, A M, Scaria, P V, Woodle, M C, Lu, P, and Rouse, B T, Inhibition of ocular angiogenesis by siRNA targeting vascular e ndothelial growth fa ctor pathway genes Therapeutic strategy for herpetic stromal keratitis. American Journal of Pathology, 2004. 165 (6): p. 2177-2185. 397. PA., C, Potential applications for RNAi to probe pathogenesis and develop new treatments for ocular disorders. Gene Therapy, 2006. 13 (6): p. 559-562. 398. Lu, P Y, Xie, F Y, and Woodle, M C, Modulation of angiogenesis with siRNA inhibitors for no vel therapeutics. Trends in Molecular Medicine, 2005. 11 (3): p. 104-113.

PAGE 213

197 BIOGRAPHICAL SKETCH Hao Pan was born in Nanjing, China, in Dec 1978 and completed his B.S. in Nanjing University, China, in 2001, majoring in pharmaceutical biotechnology. He was enrolled in the Biomedical Sciences program at College of Medicine, University of Florida, in 2001, and began his research wo rk the following year under the guidance of Dr. Maria B. Grant.