THE ROLE OF SPHINGOLIPIDS IN RETINAL VASCULAR INTEGRITY
By
Nermin Mohammad Mohammad Kady

A DISSERTATION
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
Genetics- Doctor of Philosophy
2017

ABSTRACT
THE ROLE OF SPHINGOLIPIDS IN RETINAL VASCULAR INTEGRITY
By
Nermin Mohammad Mohammad Kady
Diabetic retinopathy is a vision-threatening microvascular complication of diabetes mellitus. The
increase in the prevalence of diabetes in the population will most likely lead to a higher incidence
of diabetic retinopathy (DR), a devastating complication with limited treatment options. Years of
investigations have yet to yield a fully understood mechanism for the causes of vascular
dysfunction in DR. Dyslipidemia, as well as hyperglycemia, is a major metabolic insult of
diabetes and is positively correlated with the development of DR. The objective of this study is to
provide a mechanistic link between retinal dyslipidemia in diabetes and retinal vascular pathology
in DR. Our studies revealed two main pathways of sphingolipid metabolism involved in the
development of DR. The first pathway leads to disruption of blood retinal barrier (BRB) and thus
increased vascular permeability-- an early sign of DR. The second pathway plays a role in retinal
endothelial damage and its defective repair.
First, we addressed the link between dyslipidemia and BRB using both in vitro and in vivo
experiments. We demonstrated that diabetes induces down regulation of an essential retinal fatty
acid elongase enzyme, Elongation of very long chain fatty acids-4 (ELOVL4), in the diabetic
retina. Down regulation of ELOVL4 plays a crucial role in diabetes-induced blood retinal barrier
dysfunction. Overexpression of ELOVL4 in retinal endothelial cells enhances barrier properties
of retinal vasculature. Ceramides were found to co-localize with tight junction complexes.
Lipidomic analysis of tight junction isolates revealed the presence of ELOVL4-produced VLC
ceramides, notably, omega-linked acyl-ceramides that were previously identified only in the skin

permeability barrier.
Second, we investigated the effect of dyslipidemia on the function of both cell types involved in
revascularization; human retinal endothelial cells (HRECs) and circulating angiogenic cells
(CACs). We demonstrated that diabetes induces activation of acid sphingomyelinase (ASM), a
key enzyme of sphingolipid metabolism in human retinal endothelial cells (HRECs) and human
CD34+ CACs. Diabetes induced ASM upregulation has deleterious effect on both retinal
endothelial cell and CD34+ CAC function. Diabetic HRECs with high ASM showed defective
ability to form blood-vessel-like tubular structure and diabetic CD34+ CACs with high ASM
showed defective incorporation into endothelial tubes formed by HRECs.
Taken together, these findings indicate that modulation of sphingolipid metabolism to normalize
retinal vascular dysfunction and improve retinal vascular repair can represent a novel therapeutic
strategy for treating DR.

To the soul of my mother

	
  
	
  

iv

ACKNOWLEDGEMENTS

I would like to express my deep appreciation to those who have contributed to my work and
supported me in one way or the other during this exciting journey.
First of all, I would like to express my sincere gratitude to my mentor, Dr. Julia Busik, for her
continuous support, motivation and wealth of knowledge. Dr. Busik has supported me not only
by growing as a research scientist, but also academically and emotionally along the path to finish
my doctorate degree. I greatly appreciate all of her contributions towards helping me find
solutions in the lab and towards composing my doctorate literature. Outside of the lab, Dr. Busik
and her husband, Dr. Denis, were truly kind and hospitable with their invitation to the whole
Busik lab to spend time at their house up North. We all enjoyed their generosity and kindness. I
do not exaggerate in the slightest when I say I was incredibly lucky to have the chance to join the
Busik lab and have Dr. Busik as a mentor. The pleasure was truly mine.
I would like to thank my committee members Dr. Brian Schutte, Dr. Laura McCabe, Dr. Andras
Komaromy and Dr. Simon Petersen-Jones for their support, valuable discussions and
constructive criticism throughout all the committee meetings.
I would like to specially thank Dr. Brian Schutte for his tremendous support since day one when
Dr. Sainan Wei introduced me to him. I am especially grateful to Dr. Schutte and his team
consisting of Dr. Walid Fakhouri, Dr. Youssef Koussa, Dr. Tamer Mansour and Dr. Arianna
Smith for their support, help and constructive feedback even during the harder times of my
doctorate degree. Discussion with Dr. Schutte at committee meetings pushed me to approach all
of my data more critically and with healthy skepticism. I am very grateful to Dr. Laura McCabe
for her thorough questioning and gentle but firm encouragement during committee meetings. I

	
  
	
  

v

am very appreciative of Dr. Andras Komaromy and his expertise in all things retina related, he
was invaluable to my doctorate study. Another special thanks must be made to Dr. Simon
Petersen- Jones who not only served as my Genetics Program representative but also helped
greatly with his famed experience in gene therapy. The success of my virus experiments is due in
no small part to his consultation.
I would like to thank Dr. Barbara Sears, the former Genetics Graduate Program Director. When I
applied to the PhD program, Dr. Sears was greatly concerned with the well-being of the graduate
students and continued to be concerned throughout my years here at Michigan State University.
My first year was quite a trial and Dr. Sears was fabulously supportive in helping me adjust.
I would like to thank Dr. Sainan Wei, the Director of DNA Diagnostics and Cytogenetics Lab in
the Department of Pediatrics and Human Development. I first came to the United States as a
visiting scholar to be trained by Dr. Wei on different molecular genetic techniques. Dr. Wei
encouraged me to apply to the PhD program and introduced me to my future mentor, Dr. Busik.
She opened my eyes and played an important role in changing my career. Thank you Dr. Wei.
I am deeply thankful to Dr. David Antonetti of the Kellogg Eye Center at the University of
Michigan. His expertise in vascular permeability as well as his collaboration and invaluable
scientific input were instrumental to my research. His generosity extended so far as to inviting
me to do my in vitro experiments at his lab. The environment and guidance he provided helped
me greatly in my research.
I would like to thank Dr. Xuwen Liu from Dr. Antonetti’s lab. Dr. Liu taught me how to perform
tricky permeability experiments. Dr. Liu also had a great impact on my research skills. He taught
me how to stay organized under the culture hood, how to be prepared for my experiments more
efficiently as well as teaching me better time management skills for lab. Thank you Xuwen!

	
  
	
  

vi

I was blessed during this exciting journey to be surrounded by great people; Svetlana
Navitskaya, Dr. Qi Wang, Dr. Sandra Hammer, Dr. Harshini Chakravarthy, Chao Huang, Yan
Levitsky, Kiera Fisher, Maseray, Cassie Larrivee and Abid Ahmad. I would like to thank
Svetlana, for working with me patiently and teaching me different lab techniques especially
isolating retinal tissues and culturing cells. I would also like to thank her for her support during
hard times. I would like to thank Qi, Harshini, Chao and Sandra for their help during my
experiments, for their constructive discussions at lab meetings and for their friendship when we
made trips to the MSU Dairy Store for ice cream.
I would like to thank Dr. Todd Lydic, our collaborator in the Department of Chemistry, for his
expertise in mass spectrometry and for analyzing my lipid samples. I would also like to thank Dr.
Maria Grant and Dr. Eleni Beli, our collaborators from Indiana University, for providing blood
samples for my experiments.
I am deeply grateful to Dr. Sandra O’Reilly for teaching me how to handle mice with great
confidence and for showing me different in vivo techniques. Thank you, Sandra, for taking time
out of your busy schedule to teach me. I would also like to thank Dr. Melinda Frame for her
assistance in confocal microscopy. I really enjoyed your class, Dr. Frame, which was of great
help when scanning my images. I wish to express my deepest appreciation to Alicia Withrow for
her skills with cutting, embedding and staining my retina samples for electron microscopy. I
would also like to thank Amy Porter and Kathy Campbell for their help with my
immunocytochemistry slides.
I would like to thank my undergraduate assistants Meesum Syed, John Jarad and Kent Gamber.
They worked tirelessly on the project during the late part of my work, spending hours during

	
  
	
  

vii

their weekends for tight junction isolations. Good luck on your next steps in your careers and
educations!
I am forever grateful to my professors and friends back home at Mansoura University for their
faith in my abilities and for their support.
I would like to thank my mother and father in law for their support and unconditional love and
especially their encouragement in my doctoral pursuit.
I am deeply grateful to my father, Dr. Mohamed Kady, for shaping my interest in science—you
are the greatest inspiration for me. Your emotional support, sacrifice, patience, encouragement
and faith in my abilities have been a great pillar of strength for me throughout my PhD journey. I
am truly thankful to my sisters, Eman and Aya, for always being there for me and showering me
with support. To my father and sisters, thank you for helping fortify my career. And special
thanks for your unconditional love.
Finally, I would like to thank my beloved husband, Amir Abougabal, for his constant, steady
support and for his love.
I would not have made it this far without all of you. Thank you so much, I hope to make you all
proud.

	
  
	
  

viii

TABLE OF CONTENTS

LIST OF TABLES ...................................................................................................................... xii
LIST OF FIGURES ................................................................................................................... xiii
KEY TO ABBREVIATIONS..................................................................................................... xv
Chapter 1. Introduction................................................................................................................ 1
1.1 Background ............................................................................................................................. 1
1.2 Diabetic retinopathy ............................................................................................................... 2
1.2.1 Introduction .................................................................................................................... 2
1.2.2 Clinical classification of diabetic retinopathy ............................................................. 4
1.2.2.1 Pre-clinical diabetic retinopathy ......................................................................... 5
1.2.2.2 Non-proliferative diabetic retinopathy ............................................................... 5
1.2.2.3 Diabetic macular edema ....................................................................................... 6
1.2.2.4 Pre-proliferative diabetic retinopathy ................................................................ 6
1.2.2.5 Proliferative diabetic retinopathy ....................................................................... 7
1.2.3 Current treatment options ............................................................................................ 7
1.3 Diabetes induced blood retinal barrier breakdown ............................................................. 9
1.3.1 Retinal blood supply ..................................................................................................... 9
1.3.2 Blood retinal barrier and tight junction .................................................................... 10
1.3.2.1 Proteins forming tight junction ......................................................................... 14
1.3.2.2 A lipid-protein hybrid model of tight junction................................................. 17
1.3.3 Blood retinal barrier breakdown in diabetic retinopathy........................................ 18
1.3.3.1 Tight junction modulation ................................................................................. 18
1.3.3.2 Endothelial cell loss ............................................................................................. 19
1.3.3.3 Pericyte loss ......................................................................................................... 20
1.3.3.4 Leukostasis........................................................................................................... 21
1.3.3.5 Glial cell dysfunction .......................................................................................... 22
1.4 Diabetic dyslipidemia............................................................................................................ 23
1.4.1 Fatty acids ..................................................................................................................... 24
1.4.2 Dyslipidemia and diabetic retinopathy ...................................................................... 26
1.4.3 Retinal lipid composition in diabetes ......................................................................... 27
1.4.4 Fatty acid elongase, ELOVL4, and very long chain ceramides ............................... 28
1.4.5 ASM and short chain ceramides ................................................................................. 33
1.5 Reparative circulating angiogenic cells .............................................................................. 34
1.6 Objectives of the dissertation ............................................................................................... 36
1.7 Overview of chapters ............................................................................................................ 38
REFERENCES ............................................................................................................................ 39

	
  
	
  

ix

Chapter 2. Overexpression of ELOVL4 stabilizes tight junctions and prevents VEGFinduced vascular permeability in diabetic retina and retinal cells......................................... 58
2.1 Abstract .................................................................................................................................. 58
2.2 Introduction ........................................................................................................................... 60
2.3 Materials and methods ......................................................................................................... 63
2.3.1 Cell culture ................................................................................................................... 63
2.3.2 Virus-mediated hELOVL4 overexpression ............................................................... 63
2.3.3 siRNA transfection ....................................................................................................... 64
2.3.4 Permeability assay in vitro .......................................................................................... 64
2.3.5 Western blotting ........................................................................................................... 65
2.3.6 Quantitative real-time PCR ........................................................................................ 66
2.3.7 Immunocytochemistry ................................................................................................. 67
2.3.8 Immunogold electron microscopy .............................................................................. 69
2.3.9 TJ isolation and mass spectrometry ........................................................................... 69
2.3.10 Animal model ............................................................................................................. 70
2.3.11 Retinal vascular permeability ................................................................................... 71
2.3.12 Immunohistochemistry .............................................................................................. 72
2.3.13 Statistical analysis ...................................................................................................... 72
2.4 Results .................................................................................................................................... 73
2.4.1 Expression level of ELOVL4 controls retinal endothelial permeability ................. 73
2.4.2 ELOVL4 overexpression prevents VEGF- and IL-1β induce increase in
permeability .......................................................................................................................... 76
2.4.3 ELOVL4 overexpression increases the barrier properties of tight junction
complex .................................................................................................................................. 78
2.4.4 Ceramide co-localizes with tight junction complex .................................................. 81
2.4.5 Identification of VLC omega-linked acyl-ceramide species in the tight
junction complex ................................................................................................................... 85
2.4.6 Verification of AAV2 mediated expression of hELOVL4 in mouse
retina ...................................................................................................................................... 90
2.4.7 AAV2-hELOVL4 overexpression prevents diabetes-induced increase in
retinal vascular permeability in vivo .................................................................................. 94
2.5 Discussion............................................................................................................................... 96
REFERENCES .......................................................................................................................... 101
Chapter 3. Increase in acid sphingomyelinase level in human retinal endothelial cells
and CD34+ circulating angiogenic cells isolated from diabetic individuals is associated with
dysfunctional retinal vasculature and vascular repair process in diabetes ......................... 109
3.1 Abstract ................................................................................................................................ 109
3.2 Introduction ......................................................................................................................... 111
3.3 Materials and methods ....................................................................................................... 114
3.3.1 Circadian study of human CD34+ CACs ................................................................. 114
3.3.2 Postmortem imaging of human retina and cell culture .......................................... 114
3.3.3 Human CD34+ CACs isolation .................................................................................. 116
3.3.4 Tube formation assay ................................................................................................ 116
3.3.5 Quantitative real-time PCR ..................................................................................... 117

	
  
	
  

x

3.3.6 ELISA assay ............................................................................................................... 117
3.3.7 Statistical analysis ...................................................................................................... 117
3.4 Results .................................................................................................................................. 122
3.4.1 ASM expression level is increased in HREC, CD34+ CACs and blood
plasma of diabetic donors ................................................................................................... 122
3.4.2 Diabetes induces decrease in HREC tube formation .............................................. 123
3.4.3 Diabetes induced increase in ASM is associated with CD34+ CACs
dysfunction........................................................................................................................... 125
3.4.4 Diabetes induced increase in ASM is associated with loss of circadian release
of CD34+ CACs .................................................................................................................. 128
3.5 Discussion............................................................................................................................. 130
REFERENCES .......................................................................................................................... 134
Chapter 4. Summary and future perspectives ....................................................................... 141
4.1 Summary .............................................................................................................................. 141
4.2 Future perspectives ............................................................................................................. 143

	
  
	
  

xi

LIST OF TABLES

Table 1. Average body weights and non-fasting blood glucose level of control and diabetic
mice............................................................................................................................................... 91
Table 2. Clinical characteristics of control and diabetic subjects involved in the circadian
study .......................................................................................................................................... 118
Table 3. Clinical characteristics of control and diabetic HRECs donors and CD34+ CACs
subjects involved in the study ................................................................................................. 120

	
  
	
  

xii

LIST OF FIGURES

Figure 1. Frequency of retinopathy by duration of diabetes in years ...................................... 3
Figure 2. The anatomy of the retina .......................................................................................... 10
Figure 3. Intercellular Junctions ............................................................................................... 12
Figure 4. Morphology of the tight junction .............................................................................. 13
Figure 5. Schematic representation of different constituents of tight junction..................... 14
Figure 6. Schematic diagram of VLC-FA and VLC-PUFA biosynthetic pathways mediated
a series of elongation and desaturation reactions .................................................................... 30
Figure 7. Synthesis of VLC ceramides ...................................................................................... 32
Figure 8. Acid sphingomyelinase (ASM) is a central enzyme of sphingolipid
metabolism ................................................................................................................................... 34
Figure 9. Expression level of ELOVL4 controls retinal endothelial permeability ............... 74
Figure 10. ELOVL4 overexpression block cytokine-induced retinal endothelial cell
permeability ................................................................................................................................. 77
Figure 11. ELOVL4 overexpression increases the barrier properties of the tight junction
complex ........................................................................................................................................ 79
Figure 12. Ceramide co-localizes with tight junction complex ............................................... 82
Figure 13. Tight junction ceramide content ............................................................................. 87
Figure 14. Verification of AVV2 mediated expression of hELOVL4 in mouse retina ......... 92
Figure 15. ELOVL4 overexpression prevents diabetes-induced vascular leakage in
mice............................................................................................................................................... 95
Figure 16. Diabetes induced increase in ASM expression ..................................................... 122
Figure 17. Diabetes impairs tube formation capacity of HRECs ......................................... 124
Figure 18. Reduced incorporation of diabetic CD34+ CACs into diabetic HRECs
tubes ........................................................................................................................................... 126

	
  
	
  

xiii

Figure 19. Loss of circadian release of diabetic CD34+ CACs .............................................. 129

	
  
	
  

xiv

KEY TO ABBREVIATIONS

AAV2 ......................................................................................Adeno-associated virus serotype 2
ACCORD ................................................... Action to Control Cardiovascular Risk in Diabetes
adMD ........................................................................... Autosomal dominant macular dystrophy
AGEs ...................................................................................... Advanced glycation end products
AJ ....................................................................................................................... Adherens junction
ALA ................................................................................................................ Alpha-linolenic acid
AP .......................................................................................................................................... Apical
ASM ........................................................................................................... Acid sphingomyelinase
Bl .................................................................................................................................... Basolateral
BRB .............................................................................................................. Blood retinal barrier
BRECs .......................................................................................... Bovine retinal endothelial cells
CACs ................................................................................................. Circulating angiogenic cells
cDNA .............................................................................. Complementary deoxyribonucleic acid
Cer .................................................................................................................................. Ceramides
CETP ....................................................................................... Cholesteryl ester transfer protein
ChREBP ..................................................... Carbohydrate-regulatory element binding protein
CMV .................................................................................................................... Cytomegalovirus
CSME ................................................................................. Clinically significant macular edema
CVD ........................................................................................................... Cardiovascular disease
CXCR-4 ................................................................................... C-X-C chemokine receptor type 4

	
  
	
  

xv

DAPI .............................................................. 4’, 6-Diamidino-2-phenylindole, Dihydrochloride
DCCT ........................................................................ Diabetes Control and Complications Trial
DCCT/EDIC ................................... Diabetes Control and Complications Trial/ Epidemiology
of Diabetes Interventions and Complications
DHA ............................................................................................................. Docosahexaenoic acid
DM ....................................................................................................................... Diabetes mellitus
DME ........................................................................................................ Diabetic macular edema
DPBS ................................................................................ Dulbecco’s phosphate-buffered saline
DR .................................................................................................................. Diabetic retinopathy
DS ................................................................................................................................. Desmosome
ELISA............................................................................... Enzyme-linked immunosorbent assay
ELOVL.........................................................................Elongation of very long chain fatty acids
EPA.............................................................................................................. Eicosapentaenoic acid
ETDRS ............................................................ Early Treatment of Diabetic Retinopathy Study
FA .................................................................................................................................... Fatty acid
FADS .......................................................................................................... Fatty acid desaturases
FAS ...................................................................................................................Fatty acid synthase
FBS ................................................................................................................... Fetal bovine serum
FCS ........................................................................................................................ Fetal calf serum
FIELD ............................................ Fenofibrate Intervention and Event Lowering in Diabetes
FITC-albumin ..................................................................... Fluorescein isothiocyanate-albumin
GAPDH ................................................................. Glyceraldehyde-3-phosphate dehydrogenase
GlcCer .............................................................................................................. Glucosylceramides

	
  
	
  

xvi

GUK ...................................................................................................... Guanylate kinase domain
HCD ...................................................................................... High energy collisional dissociation
HDL ......................................................................................................... High density lipoprotein
hELOVL4 ..................................................... Human elongation of very long chain fatty acid 4
HRECs ........................................................................................ Human retinal endothelial cells
HSCs ....................................................................................................... Hematopoietic stem cells
HSL........................................................................................................ Hormone -sensitive lipase
ICAM-1 ................................................................................... Intercellular adhesion molecule-1
IDF .......................................................................................... International Diabetes Federation
IL-1β .................................................................................................................. Interleukin-1 beta
iNOS ............................................................................................. Inducible nitric oxide synthase
IRMA ......................................................................... Intraretinal microvascular abnormalities
JAMs .............................................................................................. Junctional adhesion molecule
LA ................................................................................................................................Linoleic acid
LDL .......................................................................................................... Low density lipoprotein
LPL ..................................................................................................... Lipoprotein lipase enzymes
LXR ...................................................................................................................... Liver X receptor
MA ......................................................................................................................... Microaneurysm
MAGUK ....................................................................... Membrane –associated guanylate kinase
MDCK ..................................................................................... Madin- Darby canine kidney cells
MMPs ................................................................................................. Matrix metalloproteinases
MNCs ................................................................................................................ Mononuclear cells

	
  
	
  

xvii

mRNA................................................................................................. messenger ribonucleic acid
MUFAs ............................................................................................ Monounsaturated fatty acids
Mv .....................................................................................................................................Microvilli
NA ....................................................................................................................Numerical aperture
NADPH ......................................Nicotinamide adenine dinucleotide phosphate, reduced form
NEFA ......................................................................................................Non-esterified fatty acids
NO................................................................................................................................. Nitric oxide
NPDR .............................................................................. Non-proliferative diabetic retinopathy
OCT .............................................................................................. Optical coherence tomography
OsOS4................................................................................................................ Osmium tetroxide
PBS ........................................................................................................... Phosphate buffer saline
PCs ................................................................................................................ Phosphatidylcholines
PDGF-β .................................................................................... Platelet-derived growth factor- β
PDGFRβ ..................................................................... Platelet-derived growth factor receptor-β
PDR ......................................................................................... Proliferative diabetic retinopathy
PDZ.................................................................. Postsynaptic density protein-95/Drosophila disc
large tumor suppressor Dlg1/Zonula occludens-1
PKC ...................................................................................................................... Protein kinase C
PPAR-α .........................................................Peroxisome proliferator-activated receptor alpha
qPCR .............................................................................. quantitative polymerase chain reaction

	
  
	
  

xviii

PUFA ................................................................................................... Polyunsaturated fatty acid
RBC .......................................................................................................................... Red blood cell
RITC................................................................................. Rhodamine B isothiocyanate-Dextran
RPE................................................................................................ Retinal pigment epithelia cells
S.E.M. ................................................................................................ Standard error of the mean
SCDs ...................................................................................................... Stearoyl-coA desaturases
SDF-1 ............................................................................................... Stromal cell-derived factor-1
SDS ......................................................................................................... Sodium dodecyl sulphate
SH3 ........................................................................................................... Src homology 3 domain
siRNA .........................................................................................................Small interfering RNA
SM............................................................................................................................ Sphingomyelin
smCBA .......................................................................................................CMV-chicken - β actin
SREBP-1c............................................................. Sterol regulatory element binding protein-1c
STGD3 ..................................................................................... Stargardt-like macular dystrophy
STZ ...........................................................................................................................Streptozotocin
TBS ..................................................................................................................Tris-buffered saline
TER ................................................................................... Transendothelial electrical resistance
TGF-β .......................................................................................... Transforming growth factor –β
TJ .............................................................................................................................. Tight junction
TNF-α ...................................................................................................... Tumor necrosis factor-α
VEGF .................................................................................... Vascular endothelial growth factor

	
  
	
  

xix

VLC-PUFA ............................................................. Very long chain polyunsaturated fatty acid
VLC ....................................................................................................................... Very long chain
VLCFA ................................................................................................. Very long chain fatty acid
VLDL ...............................................................................................Very low-density lipoprotein
WESDR ............................................. Wisconsin Epidemiologic Study of Diabetic Retinopathy
WHO ................................................................................................. World Health Organization
ZO-1 ................................................................................................................ Zonula occludens-1

	
  
	
  

xx

Chapter 1. Introduction
1.1 Background
Diabetes mellitus (DM) is a serious endocrine disease characterized by chronic
hyperglycemia--which results from defective insulin production or action. Defective insulin
action leads to impaired carbohydrate, protein and fat metabolism. The American Diabetes
Association diagnoses diabetes as fasting plasma glucose ≥126 mg/dl (7.0 mmol/l) or A1C ≥
6.5% (1).
Diabetes mellitus is classified into two major etiological categories; type 1 or type 2. Type 1
DM, previously defined as insulin–dependent DM, is characterized by absolute insulin
deficiency due to autoimmune destruction of insulin secreting pancreatic β-cells of the islets of
Langerhans. Insulin-dependent DM, generally occurring in children and adolescents, accounts
for approximately 10 % of diabetic patients (1,2). Type 2 DM, previously designated as noninsulin –dependent DM, is due to insulin resistance in peripheral tissue and defective
compensatory insulin secretion from β cells. Type 2 DM accounts for 90-95% of all cases of
diabetes and occurs mostly in adults. Type 2 DM is normally correlated with obesity and a
sedentary lifestyle (3–5).
Approximately 415 million people are estimated to currently have diabetes worldwide, this
number is estimated to increase to approximately 642 million (or 1 in 10 adults) by 2040 (6).
The 2015 International Diabetes Federation (IDF) estimates 5 million deaths caused by diabetes
every year (1 person every 6 seconds) (6–8). In addition, The World Health Organization (WHO)
projects that diabetes will be the 7th main cause of death in 2030 (9). The estimated total medical
cost of diagnosed diabetes is more than 300 billion dollars per year (10).

	
  
	
  

1

Complications of diabetes can be divided into macrovascular and microvascular disorders. The
macrovascular complications include coronary heart disease, cerebrovascular disease and
peripheral vascular disease. The microvascular complications include diabetic retinopathy (DR),
diabetic nephropathy, and diabetic neuropathy. These complications highlight the importance of
diabetes as a major public health problem.
1.2 Diabetic retinopathy
1.2.1 Introduction
Diabetic retinopathy (DR) is one of the most common and feared complications of diabetes,
affecting approximately 93 million people worldwide. Of the 93 million, 28 million suffer from
vision-threatening complications such as proliferative diabetic retinopathy (PDR) and /or
diabetic macular edema (DME) (11). In both type 1 and type 2 DM, the prevalence of DR is
strongly associated with the duration of diabetes. Nearly all patients with type 1 diabetes and >
60% of patients with type 2 diabetes have some degree of retinopathy after 20 years of the
disease (12). In the Wisconsin Epidemiologic Study of Diabetic Retinopathy (WESDR), it was
found that duration of diabetes is a strong predictor for development and severity of diabetic
retinopathy. WESDR found that the frequency of any retinopathy among type 1 diabetes was 2%
in those with < 2 years duration of diabetes, 25% after 5 years and 97.5% at 15 or more years of
diabetes. Furthermore, frequency of PDR was very rare before 10 years of diabetes, nonexistent
at 3 years, 4% after 10 years and 25% after 15 years of diabetes (Figure 1) (13). DR is the
leading cause of visual impairment in the world; 3.6% of type 1 diabetics and 1.6% of type 2
diabetic individuals suffered from vision loss. In type 1 diabetes, 86% of the vision loss resulted
in blindness from DR, mostly due to neovascularization of the proliferative stage (12,14). 33% of

	
  
	
  

2

vision loss in type 2 diabetic individuals resulted in complete vision loss from DR, mostly due to

Frequency)of)re,nopathy)%)

macular edema and very rarely PDR (12,14).

120"
100"
80"
60"
40"
20"
0"
0"

2"

4"

6"

8"

10"

12"

14"

16"

Dura,on)of)DM)in)years))
Figure 1. Frequency of retinopathy by duration of diabetes in years. The development and
the severity of diabetic retinopathy is directly correlated with duration of diabetes (Modified
form Klein et al., 1984) (13).

	
  
	
  

3

1.2.2 Clinical classification of diabetic retinopathy
Diabetic retinopathy is a chronic, progressive sight-threatening disease, with multiple retinal
lesions accumulating along the course of the pathology. It is essential to classify and stage the
severity of the disease in order to provide more effective therapeutic strategies to prevent the
progression of DR. Retinal lesions seen in the fundus can be classified into 3 categories.
Category 1 retinal lesions result from the breakdown of the blood retinal barrier and leaky
vasculature (observed as hemorrhage, hard exudate, retinal edema). Category 2 retinal lesions
result from structural changes in vascular wall (observed as microaneurysms). Lastly, category 3
retinal lesions result from ischemic changes (observed as soft exudate or cotton wool spots,
intraretinal microvascular abnormalities (IRMA), neovascularization, vitreous hemorrhage and
fibrous tissue formation) (15). Many attempts were made to develop standardized classification
of DR. From 1968, there was the Airlie House Symposium classification that was modified for
use in the Early Treatment of Diabetic Retinopathy Study (ETDRS). ETDRS was the gold
standard classification system for DR. It graded retinopathy using 7 standard stereophotographic
fields, classifying DR retinal lesions into 13 levels ranging from level 10 (no retinopathy) to
level

85

(advanced

proliferative

diabetic

retinopathy

with

vitreous

hemorrhage,

neovascularization or retinal detachment at the center of the macula). The term “clinically
significant macular edema (CSME)” was first introduced by ETDRS, see below for details (16).
The complexity of the ETDRS grading system rendered the scale to be clinically inapplicable.
Another classification has been proposed by Wilkinson et al. 2003 in an attempt to simplify the
classification of DR-- the International Clinical Disease Severity Scale of DR. It is based on
WESDR and ETDRS classification, the retinopathy pathology is graded into 5 stages; (1) no
apparent retinopathy; no diabetic fundus lesions, (2) mild non-proliferative diabetic retinopathy;
	
  
	
  

4

few microaneurysms, (3) moderate non-proliferative diabetic retinopathy; microaneurysm,
hemorrhage or venous beading, (4) severe non-proliferative diabetic retinopathy; 4:2:1 rule,
hemorrhages exist in the 4 quadrants, venous beading in 2 quadrants or more and IRMA in 1
quadrant or more, and finally, (5) proliferative diabetic retinopathy; neovascularization, vitreous
hemorrhage or tractional retinal detachment (17). Stages of DR are listed below:
1.2.2.1 Pre-clinical diabetic retinopathy is characterized by subtle alterations that occur before
ophthalmoscopic signs of DR. These alterations are not visible on fundus examination but they
were demonstrated by histology and fluorescein angiography. Early alterations include pericyte
loss, thickening of the blood vessel basement membrane, capillary occlusion and endothelial cell
degeneration and loss (18,19).
Another prominent feature of this stage is diffuse mild BRB loss with increased vascular
permeability (20,21). In addition, neurosensory dysfunction and altered color vision were also
detected in this stage (22).
1.2.2.2 Non-proliferative diabetic retinopathy
This stage is characterized by visible retinal lesions upon fundus examination. Lesions include
microaneurysms (MAs), intraretinal hemorrhage, hard exudates and altered blood retinal barrier
(14).
MA is one of the earliest ophthalmoscopic signs, it appears as red dot with a well-defined border.
With fluorescein angiography, MAs appear hyperfluorescent and leaky (23). MA is usually
associated with loss of pericytes, thickening of the blood vessel basement membrane, capillary
closure affecting both the arterial and venous side of the circulation at the posterior pole of

	
  
	
  

5

the retina (24,25). An interesting feature of MAs is that counting rate of MA formation and
disappearance has been demonstrated as a good indicator of retinopathy progression (14,26).
Intraretinal hemorrhage results from breakdown of MAs or small vessels. Hard exudates result
from extravascular leakage of lipids or lipoproteins into retinal tissue appearing as yellow
glistening deposits close to the macula. Abnormal blood retinal barrier leads to accumulation of
fluid in retinal tissue resulting in retinal edema.
1.2.2.3 Diabetic macular edema is the main cause of vision loss, particularly, in type 2 diabetes
(27,28). Altered blood retinal barrier results in accumulation of fluid in the retinal tissue
especially in macular area and is detected by optical coherence tomography (OCT). Progression
of edema depends on Starling law, where the gradient induced between plasma blood pressure,
tissue pressure and tissue osmotic pressure play important roles in controlling movement of
fluids (29).
ETDRS introduced a definition for clinically significant macular edema, where the diagnosis
mainly depends on the presence of retinal thickening or hard exudates at or within 500 microns
of the center of the macula (16).
1.2.2.4 Pre-proliferative diabetic retinopathy Its name indicates increased risk of progression
to proliferative retinopathy. This stage is characterized by signs of retinal ischemia: soft exudates
or cotton wool spots, venous beading, IRMAs and several areas of capillary non-perfusion.
Cotton wool spots are grayish white areas due to nerve fiber infarction which result from
occlusion of pre-capillary arterioles (30). The presence of a high number of cotton wool spots
(more than eight in one eye) is indicative of a severe case that may progress into the proliferative

	
  
	
  

6

stage (31). IRMAs represent non-leaky collaterals next to non-perfused areas (32).
1.2.2.5 Proliferative diabetic retinopathy is a vision threatening complication particularly in
type 1 diabetes (14). It is characterized by neovascularization, that arise either from the optic
disc, retina or iris. These new blood vessels are fragile and leaky with an abnormal blood retinal
barrier. They pierce the internal limiting membrane causing vitreous hemorrhage, fibrosis and
retinal detachment. These vessels usually arise secondary to massive areas of capillary occlusion
and retinal ischemia, which in turn stimulates the production of angiogenic factors (33).
Capillary occlusion usually results from inflammatory changes, microthrombosis and apoptosis
(19).
In summary, impaired vision or vision loss usually takes place in the late stage of the disease,
due to either macular edema or neovascularization.
1.2.3 Current treatment options
Diagnosis of patients in the very early stages of the disease (preclinical stage) before developing
vision-threatening lesions provide the best window of opportunity for therapy, especially since
this category of patients represent a majority and respond better to intensive diabetic therapy
(34). Controlling diabetes-induced metabolic dysregualtion is a milestone of DR treatment and
can be achieved through optimizing blood-glucose levels (35). Tight glycemic control (HbA1C <
7 %) has become a standard intervention to slow down the development of DR. Several clinical
trials (DCCT, UKPDS) have shown the benefit of intensive glycemic control in both type 1 and
type 2 diabetic patients (36,37). However, maintaining the recommended blood-glucose level
was found to be difficult in many patients. Moreover, tight glycemic control was associated with

	
  
	
  

7

critical complications such as hypoglycemic episodes and diabetic ketoacidosis. In the ACCORD
clinical trial, intensive glycemic control was associated with increased mortality rate in type 2
diabetics especially those at high risk of cardiovascular events (38). Thus glycemic control
should be aimed with great caution and each patient should be treated individually. Until recent
years, laser photocoagulation was the only available therapeutic option for treating either
proliferative diabetic retinopathy (PDR) or diabetic macular edema (DME) (39). Although
proven beneficial, laser photocoagulation does not result in vision improvement but rather results
in slowing the vision loss. It is also associated with some adverse effects: night blindness, color
vision changes, reduced contrast sensitivity, acute glaucoma and worsening of macular edema
(35).
These adverse effects of laser photocoagulation brought the need for new approaches in treating
DR. Recently introduced intravitreal injection of anti-VEGF agents is considered a new modality
in treating both PDR and diabetic macular edema. This class of drugs include: Pegaptanib
(Macugen, RNA aptamer), Ranibizumab (Lucentis, recombinant antibody fragment (Fab)),
Bevacizumab (Avastin, monoclonal antibody), and Aflibercept (Regeneron, fusion protein of
VEGF receptors) (35,39,40). All these agents act by neutralizing the proangiogenic factor VEGF
and inhibiting VEGF-induced blood retinal barrier breakdown and retinal neovascularization.
However, this is an invasive approach of short duration, requiring repeated intravitreal injections
with increasing incidence of local complications; endophthalmitis, vitreous hemorrhage and
retinal detachment (40). Although, VEGF has been identified as playing an important role in the
development of DR, endogenous VEGF is important for the survival and maintenance of retinal
neurons (41) and photoreceptors (42), so great caution is required during long-term therapy.
Other treatment options are also available as intravitreal injection of corticosteroids and
	
  
	
  

8

vitrectomy both are invasive with local complications. The currently available treatment options
are highly invasive and only improve vision but do not restore it completely. Identification of
novel therapeutic strategies aiming to prevent vision-threatening lesions of DR is crucial.
1.3 Diabetes induced blood retinal barrier breakdown
1.3.1 Retinal blood supply
The retina is a transparent multi-layer neural tissue (Figure 2). It is responsible for converting
visible light into electrochemical impulses that are transmitted via the optic nerve to the cerebral
visual cortex for interpretation of the image. This requires a protective environment with
minimal interference to the passage of light which is supported by a unique vascular structure
with restrictive or selective control of permeability for proper vision (34). Furthermore, the retina
is a highly metabolically active tissue with the highest oxygen consumption per unit weight (43).
In order to achieve this metabolic support with minimal interference to passage of light, the
retina receives dual blood supply from the choriocapillaries and the retinal capillaries. The outer
retina (photoreceptors) is supplied by the choriocapillaries while the inner retina is supplied by
retinal capillaries originating from central retinal artery (44). The endothelium of
choriocapillaries is highly fenestrated while that of retinal capillaries lack fenestration and posses
selective barrier function.

	
  
	
  

9

Choroid)
RPE)(outer)BRB))
Rods)and)
Cones))

OLM)

Outer)nuclear)
layer)
Inner)nuclear)layer)
Ganglion)cell)layer)
Endothelial)cell)(inner)BRB)))

Figure 2. The anatomy of the retina. Image of rat retina showing the retinal layers. Antiglucose transporter 1 staining (brick color) was used and nuclei were detected with DAPI (blue
color). The blood retinal barrier (BRB) is formed of tight junctions, between retinal capillary
endothelial cells (inner BRB) and retinal pigment epithelial cells (RPE, outer BRB). OLM: outer
limiting membrane.
1.3.2 Blood retinal barrier and tight junction
The regulation of the flux of blood-born solutes and fluids into the retinal tissue is controlled by
the outer and inner blood retinal barriers (BRB) (Figure 2). The outer BRB is formed by the tight
junctions between retinal pigment epithelial cells (RPE) and the inner BRB is formed by the
retinal neurovascular unit composed of tightly adherent retinal capillary endothelial cells
surrounded by pericytes and glial cells (Muller and astrocytes) (45) . Pericytes and glial cells
play an important role in the development and stabilization of the inner BRB through their direct

	
  
	
  

10

physical contact and intercellular signaling with the capillary endothelial cells. Pericyte
recruitment to endothelial cells through PDGF-β/PDGFRβ was found to be essential for tight
junction formation (46). Another well-known pathway is Angiopoietin-1/Tie2, where pericytes
secret angiopoietin-1 inducing phosphorylation of Tie-2 on endothelial cells and promoting
endothelial cell maturation (47), vessel integrity (48) and inducing tight junction occludin
expression (49). The physical integrity of these cells is essential for regulating BRB; any
disturbances by pathological condition, as in DR, will cause barrier integrity dysfunction.
Retinal capillary endothelial cells regulate the transport of the solutes through 2 pathways; the
paracellular pathway and the transcellular pathway. Retinal endothelial cells are characterized by
low rate of transcellular pathways (50) through membrane transporter, vesicle- (caveola),
receptor- or carrier- mediated transport (51). The paracellular pathway is mostly regulated by the
tight junction, the most apical part of inter-endothelial junctional complex, and also regulated by
adherens junctions, desmosomes and gap junction (Figure 3).

	
  
	
  

11

Figure 3. Intercellular Junctions. A. Schematic representation of different structures within
the intercellular junctional complex (adapted form Matter and Balda, 2003) (52). B. Electron
micrograph showing main structures of the intercellular junctional complex from apical to
basolateral surface (adapted from Tsukita et al., 2001) (53). Mv: microvilli, TJ: tight junction,
AJ: adherens junction, DS: desmosome.
The tight junction is a multifunctional structure with (1) barrier functions regulating paracellualr
passage of water and water-soluble ions, (2) fence functions, preventing diffusion of lipid and
proteins between the apical and basolateral membrane (54), (3) cell signaling functions,
controlling cell proliferation and gene transcription (55). By electron microscopy, tight junctions
appear as an area where the plasma membranes of two adjacent cells come together and fuse,
obliterating the intercellular space (56). While by freeze –fracture microscopy they appear as
anastomosing band-like mesh work (57) ( Figure 4).

	
  
	
  

12

A

B

C

Figure 4. Morphology of the tight junction. A. Freeze-fracture replica electron microscope of
the tight junction. Tight junction appears as continuous, anastomosing fibrils (arrowheads) on the
P face with grooves on the E face (arrows). (Mv: microvilli, AP: apical membrane, Bl:
basolateral membrane). B. Ultrathin sectional view of tight junctions, showing kissing points of
tight junctions (arrowheads), where the intercellular spaces are obliterated. C. Schematic
structure of tight junction strands (adapted from Tsukita et al., 2001) (53).
Tight junctions are composed of distinct transmembrane proteins including claudin, occludin,
and junctional adhesion molecule (JAMs) that are coupled to the intracellular anchoring or
scaffolding protein zonula occludens (ZO) which is linked to the actin cytoskeleton (Figure 5).
Some of these proteins are stable, while others cycle continuously between the plasma membrane
and intracellular compartments and also along or within the plasma membrane. This cycling
provides the dynamic nature of the junction; opening and closing in response to variable stimuli
and regulating the inter-endothelial passage of different molecules (58,59).

	
  
	
  

13

Figure 5. Schematic representation of different constituents of tight junction. Tight
junctions are composed of several transmembrane proteins including, claudin, occludin,
junctional adhesion molecule (JAMs) that are coupled to the intracellular scaffolding protein
zonula occludens (ZO), which is linked to the actin cytoskeleton (adapted from Runkle and
Antonetti, 2011) (45).
1.3.2.1 Proteins forming tight junction
•

Occludin. Occludin, a 65 KDa transmembrane protein, was the first integral membrane protein
identified as part of the tight junction of chicken liver by Furuse et al. (1993) (60). It is tetraspan
transmembrane protein with two extracellular loops, one intracellular loop and cytoplasmic
amino- and carboxy-termini. Occludin physically interacts with a number of proteins through its
carboxy- and amino- termini. The carboxy terminus of occludin interacts with ZO-1 (61), ZO-2
(62), gap junction protein connexin 32 (63), or occludin itself (64). ZO-1 interaction with

	
  
	
  

14

occludin plays an important role in (1) connecting occludin with actin cytoskeleton (65) and (2)
localization of occludin at the tight junction (61). Its amino- terminus was found to interact with
Itch and E3 ubiquitin-protein ligase (59,66).
Occludin is expressed mainly at the tight junctions of endothelial and epithelial cells (60).
Occludin is a functional constituent of the paracellular barrier, where its expression correlates
with barrier properties of different tissues. Occludin is expressed at higher levels in endothelial
cells with tight barrier property as in arterial, brain, and retinal endothelium compared to those
with other barriers as in venous endothelium or endothelial cells of non-neuronal tissue (67,68).
Moreover, increased expression of occludin in bovine retinal endothelial cells was associated
with decreased paracellular permeability (69), while silencing of occludin decreased barrier
property and increased permeability (67,70,71). Although occludin is an essential determinant of
tight junction permeability, it is not required for tight junction formation (72,73).
•

Claudins. Claudins, a large family of 24 transmembrane proteins (20-27 KDa) (53) , are
tetraspan transmembrane proteins with two extracellular loops and cytoplasmic amino- and
carboxy-termini. The carboxy termini interacts with ZO family protein (74) and was found to be
essential in targeting claudin proteins into tight junctions (75). Caludin-1, -5,-15 are specifically
expressed in endothelial cells and constitute blood retinal barrier (76). Unlike occludin, claudin
was found to be an important regulator of tight junction formation. A number of studies showed
expression of either claudin-1 and -2 or claudin-5 in fibroblast reconstituted tight junction
strands (76,77).
In addition, claudin regulates barrier function of the tight junction, where overexpression of
claudin in MDCK cells results in increased TER
	
  
	
  

and decreased paracellular permeability (78).

15

In claudin-5 deficient mice paracellular permeability to small molecules was severely impaired
(79). In retinal tissue, hypoxia–induced decrease in claudin-5 was associated with increased
vascular permeability and impaired barrier function (80).
Collectively, Claudins play a critical role in tight junction formation and regulation of barrier
function.
•

The Zonula occludens family. ZO-1 (210-225 KDa) was the first protein to be discovered in
association with tight junction (81). Later other members of the family were discovered, ZO-2
(160KDa) (82) and ZO-3 ( 130 KDa) (83), they were found to co-immunoprecipitate with ZO-1.
ZO proteins belong to the membrane –associated guanylate kinase (MAGUK) homologue family
and are present within tight junctions of endothelial cells and epithelial cells. In addition, they
exist in adherens junction in those cells that lack tight junction formation as in fibroblast and
cardiac myocytes (62). This family characterized by the presence of three PDZ domain, one Src
homology 3(SH3) domain, guanylate kinase (GUK) domain and a proline-rich C-terminus (84).
Through these multiple domains, ZO proteins bind to diverse set of junction proteins. ZO-1
interacts with other ZO proteins via the PDZ domain, forming heterodimer complexes (65,85),
with claudin -1 through PDZ-1 domain (74).

ZO-1 interacts with occludin via GUK domain

(61) and cytoskeleton F-actin via proline-rich C-terminus (65). ZO proteins may serve as link
connecting occludin and claudin 5 to the cytoskeleton. The ZO proteins family was demonstrated
to have a functional role in tight junction organization and assembly. Silencing of either ZO-1 or
ZO-2 in MDCK cells delayed tight junction assembly and formation (86,87). Other studies
showed that ZO-1 and ZO-2 deficient mouse epithelial cells lacks tight junction completely and
claudin fails to polymerize at tight junction, this was associated with decreased TER and increase

	
  
	
  

16

paracellualr permeability (88). In vivo studies demonstrated that ZO-1 knockout mice possess a
lethal phenotype due to developmental defects in the neural tube and yolk sac angiogenesis (89).
In addition to their role in tight junction, ZO- proteins have been reported to play a role in gene
transcription and cell differentiation (90).
1.3.2.2 A lipid-protein hybrid model of tight junction
Several studies have demonstrated the lateral diffusion of different molecules from one cell to its
neighboring cell (91,92). Lee et al. (2008) explained that if the tight junction is formed of only
transinteracting proteins across the intercellular space then such a structure would not allow
these type of lateral movements (93). The tight junction represents fusion points between
apposing plasma membranes, however there is no evidence for adhesion or fusion between
extracellular components or membrane-bridging of any of the tight junction proteins (93).
Therefore, the protein cannot be the only constituent of the tight junction.
The idea of the lipid-protein hybrid model is not recent. Lipidic nature of tight junction itself was
first proposed more than 30 years ago (1980s) by Kachar et al. (1982). They showed that tight
junction strands are formed of inverted lipid cylinders, where apposing cells get in contact with
each other through fusion points between their exoplasmic leaflets (94). This was further
supported by other studies - freeze fracture preparations demonstrated the presence of
phospholipids and cholesterol particles in tight junction strands (95,96). The model of lipid
bridges at tight junctions has been postulated as a route for lateral transfer of phospholipids and
lipid-protein complex between neighboring cells (91,92). Collectively, these studies postulate
that the tight junction is an assembly of lipid-protein interactions.

	
  
	
  

17

Although there is a good basis supporting the presence of lipids in the tight junction, the
structural relationship between the proteins and the lipids in the junctional complex is not well
understood.
1.3.3 Blood retinal barrier breakdown in diabetic retinopathy
Blood retinal barrier dysfunction and increased permeability is one of the earliest signs of DR
even before any ophthalmoscopic signs are visible (20). Retinal vasculature forming the inner
blood retinal barrier was demonstrated to be the primary site of leakage in diabetic patients
(21,97). As mentioned above, the inner BRB is formed by what is called the retinal
neurovascular unit; tightly adherent retinal capillary endothelial cells surrounded by pericytes
and glial cells (Muller and astrocyte). The physical integrity of these cells is essential for
regulating the BRB, any disturbances by pathological condition, as in DR, will cause barrier
integrity dysfunction.
1.3.3.1 Tight junction modulation
Tight junctions (TJ) form a physical barrier regulating the flux of ions and molecules through the
paracellular pathway into retinal tissue. Any changes in the content or localization of the tight
junction proteins results in tight junction disruption and increased paracellular permeability.
Among the different tight junction proteins, downregulation of occludin has been reported as a
leading cause of the BRB breakdown particularly in association with VEGF. Induction of
experimental diabetes and intravitreal injection of VEGF (71,98) were associated with either
decreased occludin content or its dislocalization from the tight junction. This decreased occludin
expression or dislocalization was associated with increased vascular permeability - an

	
  
	
  

18

observation mirrored by treatment of BRECs with VEGF (71). Molecular analysis suggests that
VEGF activates downstream proteins, including protein kinase C(PKC) β, that induce Ser490
phosphorylation, ubiquitination, internalization and degradation of occludin. PKC β inhibitors
were found to block VEGF-induced occludin phosphorylation and decrease paracellular
permeability (59,99–101).
Alterations in other tight junction proteins have been associated with blood retinal barrier break
down and increased vascular permeability. ZO-1 protein content has been shown to be reduced
in the diabetic mouse retina, with concomitant increase in retinal vascular permeability (102). In
addition, VEGF treatment of endothelial cells in vitro, results in significant decrease of ZO-1
content and immunoreactivity at the endothelial cell border (103,104). Intravitreous injection of
VEGF caused tyrosine phosphorylation of ZO-1 in the rat retina (99). Claudin 5 transcript was
reduced in the diabetic rat retina and in VEGF treated BRECs (105).
Taken together, altered expression, phosphorylation and localization of tight junction proteins
affect tight junction function and the paracellular permeability.
1.3.3.2 Endothelial cell loss
Endothelial cell death and formation of acellular capillaries is one of the earliest signs of diabetic
retinopathy (106).

A chronic hyperglycemic environment, that results in accumulation of

advanced glycation end products (AGEs) and oxidative stress, has been shown to be involved in
the intracellular glucose toxicity and retinal endothelial cell injury (107,108). Moreover,
leukostasis, an early feature of DR, may be a direct cause of retinal endothelial cell death.
Inhibition of leukostasis by using intercellular adhesion molecule-1 (ICAM-1) and CD18

	
  
	
  

19

neutralizing antibodies prevent retinal endothelial cell loss (109). In combination with defective
repair of impaired vessels due to diabetes induced endothelial progenitor dysfunction, leukostasis
and chronic hyperglycemic environment (110,111) may result in endothelial cell damage.
Endothelial cell death may be a direct cause of BRB disruption.
1.3.3.3 Pericyte loss
The blood retinal barrier demands proper pericyte function, and loss of pericytes may lead to
increased vascular permeability.
Several studies have shown that endothelial cell barrier property and permeability were directly
correlated to the density of pericytes (112). In vitro studies have demonstrated that co-culturing
of pericytes with endothelial cells enhances barrier property and decreases permeability
(113,114). While endothelium-specific knockout of PDGF-B, an essential factor for pericyte
survival, resulted in decreased pericyte density and increased retinal vascular regression (115).
Pericyte death has been observed as an early feature of DR in both diabetic humans and rats
(106,116).
Diabetes induced pericyte loss may be attributed to many factors; hyperglycemia induced
inhibition of PDGF signaling (117), increased expression of angipiotein 2 (118,119), or
excessive production of AGEs and connective tissue growth factor (CTGF) (120,121). All these
factors induce pericyte detachment from the endothelial cells and destabilization of the junctional
complex (119,120,121).	
  
	
  

	
  
	
  

20

Pericytes play an important role in providing physical support for endothelial cells and
maintaining barrier properties.
1.3.3.4 Leukostasis
Leukostasis is increased adhesion of leukocytes to retinal vasculature and is an early event in the
pathogenesis of diabetic retinopathy. Leukostasis correlates with the development of retinal
capillary occlusion, death of retinal endothelial cells and breakdown of blood retinal barrier
(122).
This process is mediated by expression of intercellular adhesion molecule-1 (ICAM-1) on
endothelial cells and β2 integrin CD18 on leukocytes (123,124). As diabetes progresses,
expression of ICAM-1 increases which in turn correlates with retinal leukostasis (122).
Inhibition of ICAM-1 and CD18 expression, by either using neutralizing antibodies
(109,122,124) or by knockdown of ICAM-1 and CD-18 in mice (125), prevent diabetes induced
leukostasis and retinal vascular leakage .
Many factors have been demonstrated to be involved in diabetes induced leukostasis; VEGF,
PKC activation, and inducible nitric oxide synthase (iNOS). Lu et al. (1999) showed that
intraocular injection of VEGF results in increased expression of ICAM-1 in retinal vasculature
(126). While, inhibition of VEGF with neutralizing antibodies results in significant decrease in
ICAM-1 expression and leukocyte adhesion (127). It was found that PKC activation is also
involved in retinal leukostasis; oral administration of PKC β inhibitor, LY333531, decreases
leukocyte adhesion to retinal circulation (128). iNOS has also been considered to have role in
leukostasis, mice treated with iNOS inhibitor and iNOS knock out mice prevented diabetes

	
  
	
  

21

induced increase in leukostasis and BRB permeability (102). Leukocyte adhesion has been
correlated with blood brain barrier dysfunction through disorganization of tight junctions and
adherens junction (129).
Taken together, these studies show that leukostasis has a role in retinal vascular damage and
BRB dysfunction.
1.3.3.5 Glial cell dysfunction
As mentioned above, retinal glial cells (Muller and astrocytes) participate in the establishment of
BRB. Astrocytes have been shown to be restricted to the vascularized area of the retina,
coexisting with the retinal blood vessels (130–132). More interestingly, they have been found to
enhance the formation of the blood retinal barrier (130,133).
Gardner et al. (1997) demonstrated that astrocyte conditioned media increased expression of
tight junction protein ZO-1 and the barrier property of BRECs (134). In vivo studies also showed
that astrocytes play a critical role in maintaining tight junction complex, where astrocyte loss
coincides with the loss of occludin, claudin-5 and Zo-1 localization in brain tight junction (135).
In the in vitro co-culture model, Muller cells were also shown to enhance barrier properties of
BRECs and its dysfunction lead to break down of the blood retinal barrier (136). Muller
dysfunction by glial toxin or siRNA results in excessive production of VEGF and reduced
expression of claudin-5, which was associated with increased vascular leakage (137).
It has also been suggested that Muller cell activation under pathological conditions result in
overexpression of a number of cytokines e.g. TGF-β and VEGF which may induce breakdown of

	
  
	
  

22

the BRB via proteolytic degradation of occludin by matrix metalloproteinases (MMPs)
(138,139). Of particular relevance to diabetes, diabetic rats show evidence of reactive gliosis of
muller cells with excessive expression of glial fibrillary acidic protein (GFAP), this was
accompanied by reduced and redistributed occludin in retinal endothelial cells (140).
Retinal microglia have also been reported to be activated in early diabetes with excessive
production of proinflammatory cytokines IL-1β and TNF-α (141) which are important mediators
of leukostasis and increased permeability (142,143).
All together, this demonstrates the importance of glial cells in maintenance of BRB.
1.4 Diabetic dyslipidemia
Dyslipidemia can be defined as abnormal serum level of lipids due to disordered or disrupted
lipid metabolism; release or uptake by either adipose or hepatic tissue. In adipose tissue, during
the postprandial period, insulin inhibits fat mobilization and stimulates fat deposition through
suppression of hormone -sensitive lipase (HSL) and activation of lipoprotein lipase enzymes
(LPL) (144–146). In hepatic tissue, following food intake and increase in the carbohydrate flux
into the liver, insulin stimulates fatty acid synthesis (de novo lipogenesis) which in turn get
esterified into triglycerides (147). Therefore, abnormalities in insulin, as in in type 1 diabetes
(dependence) and in type 2 diabetes (resistance), are expected to have a profound effect on lipid
profile. In poorly controlled type 1 diabetes, elevated triglycerides and LDL cholesterol
commonly occur (148–150). Type 2 diabetes have high levels of free fatty acids, triglycerides,
atherogenic lipoproteins (small dense LDL) and reduced high density lipoprotein (HDL)
cholesterol (148,151–153). Type 1 diabetes differs from type 2 diabetes in that lipid

	
  
	
  

23

abnormalities in the former are reversible and can be corrected by insulin replacement while in
the latter these abnormities cannot be fully corrected (148).
Dyslipidemia is a disorder of lipid metabolism with altered production of different lipid classes.
Diabetes associated alteration in fatty acid class is discussed below.
1.4.1 Fatty acids
In type 2 diabetes, Insulin resistance coupled with increased release of non-esterified fatty acids
(NEFA) from adipose tissue and decreased clearance of circulating triacylglycerol results in
elevated blood level of both esterified and non-esterified fatty acids (NEFA) (146). Moreover,
the uncontrolled release of NEFA leads to increased fatty acid flux to the liver, which in turn
stimulate hepatic synthesis of triglycerides and thus VLDL secretion (147,148,151). The
accumulated VLDL leads to the generation of atherogenic lipoprotein (small dense LDL) and the
reduction of HDL cholesterol through the action of cholesteryl ester transfer protein (CETP)
(147,148).
Several studies have demonstrated the effect of fatty acids content on serum low density
lipoprotein (LDL), where saturated fatty acids have been associated with increased concentration
of small, dense LDL in plasma and unsaturated fatty acids decreased the concentration of these
atherogenic lipoprotein particles (154,155).
Diabetes is characterized by altered saturation and elongation of fatty acids with increased level
of saturated, short chain fatty acids and decreased polyunsaturated, long chain fatty acids
(156,157). Fatty acids that are either synthesized by the cytosolic enzyme fatty acid synthase
(FAS) or derived from the diet undergo series of

	
  
	
  

desaturation and elongation reactions by

24

physically associated, membrane-bound enzymes localized in the endoplasmic reticulum called
desaturases and elongases enzymes (158–160). Fatty acid desaturases and elongases play an
essential role in hepatic and whole body lipid composition.
Desaturases are enzymes that introduce double bonds at certain positions on the acyl chain of
fatty acids. According to the position at which the double bond is introduced, desaturases are
divided into 2 groups; stearoyl-coA desaturases (SCDs) including the Δ9 desaturases and fatty
acid desaturases (FADS) including Δ6, 5 desaturases (161). Previous studies have revealed
defective desaturase activity in type 1 diabetes, which can be reverted by insulin treatment (162–
165).
Elongase enzymes are believed to perform the first condensation (regulatory) step in the fourstep fatty acid (FA) elongation cycle, which takes place in the endoplasmic reticulum with the
help of 3-ketoacyl-CoA reductase, 3-hydroxyacyl-CoA dehydratase, and 2,3- trans-enoyl-CoA
reductase. Fatty acid elongation is a complex reaction involving the addition of	
  two	
  carbons to
the fatty acyl-CoA using malonyl-CoA as the carbon donor. To date, seven elongases, termed
ELOVL 1-7 (elongation-of-very-long-chain-fatty acids) have been identified in mammals.
ELOVLs exhibit tissue-specific distribution and have distinct substrate specificities. ELOVL1,
ELOVL5 and ELOVL6 are ubiquitously expressed among tissues, while ELOVL2, ELOVL3
and ELOVL4 are highly expressed in liver/testis (166), skin/liver (159,167) and retina/brain/skin
(159,167–169), respectively. ELOVL1, ELOVL3 and ELOVL6 are responsible for the
elongation of both saturated and monosaturated FAs, while ELOVL2 and ELOVL5 elongate
polyunsaturated FAs (PUFAs). ELOVL4 utilizes both saturated and polyunsaturated long fatty
acids (C≥C24) to synthesize very long chain fatty acids (C≥26) (158–160). Elongases are well

	
  
	
  

25

controlled enzymes, they are regulated at a transcriptional level where different cellular stimuli
have been shown to alter their expression (158). Wang et al. (2006) have demonstrated that
ELOVLs are regulated in hepatic tissue by hormones (insulin), transcriptional factors (SREBP1c, LXR and PPAR-α), and nutrients (glucose, fat). Insulin, liver X receptor (LXR) and glucose
induce ELOVL-6 expression through the induction of the transcription factors: sterol regulatory
element binding protein-1c (SERBP-1c) and carbohydrate-regulatory element binding protein
(ChREBP), respectively (165). Lipolytic transcription factor PPAR-α induces both ELOVL5
and ELOVL6. Interestingly, Jakobsson et al. (2005) revealed differential regulation of ELOVL3
in brown adipocyte where ELOVL3 expression was induced in activated adipocyte cells by
glucocorticoids, norepinephrine and PPAR-α treatment but was inhibited by LXR (170).
Although activated brown adipose tissue is associated with increased lipolysis, fatty acid –β
oxidation and turn over, ELOVL3 was still induced to replenish tissue pools of specific fatty
acids. Similar to desaturases, elongases are suppressed in diabetes due to reduced insulin and
insulin-induced SREBP (165,171).
1.4.2 Dyslipidemia and diabetic retinopathy
Clinical data support a possible role of dyslipidemia in the development of diabetic retinopathy
in both type 1 and type 2 diabetes. Small early clinical trials revealed that a lipid lowering diet
could reduce the amount of retina hard exudates (172) and a diet rich in unsaturated fatty acids
(linoleic acid) delay the development and the progression of DR (173). Augustin et al. (1993)
demonstrated that lipid peroxide levels were significantly elevated in the vitreous of patients
with proliferative diabetic retinopathy (174). Data from the Early Treatment Diabetic
Retinopathy Study (ETDRS) showed that severity of retinal hard exudate was positively

	
  
	
  

26

associated with serum level of total cholesterol and LDL cholesterol (175–177). Moreover, the
DCCT/EDIC cohort study revealed a positive correlation between severity of DR in type 1
diabetes with the small size of lipoprotein subclasses (VLDL, LDL, HDL) and with increased
LDL concentration (178). A prospective analysis of the DCCT data indicates strong association
between clinically significant macular edema (CSME) and elevated serum lipids in particular
high triglycerides and total -to -HDL cholesterol ratio (179). Recent intervention studies
performed by ACCORD Eye study group in type 2 diabetes have determined that intensive
dyslipidemia therapy with combination of PPARα activator fenofibrate and HMG-CoA reductase
inhibitor simvastatin, significantly slowed the progression of diabetic retinopathy (180).
Taken together, these studies revealed that dyslipidemia is highly associated with the
development of diabetic retinopathy.
1.4.3 Retinal lipid composition in diabetes
Fatty acid composition of any particular tissue is determined by complex interplay among two
metabolic pathways; de novo lipogenesis using glucose as substrate and polyunsaturated fatty
acid (PUFA) biosynthesis (Sprecher) pathway using essential fatty acids, i.e., linoleic acid (LA,
C18:2n-6) or α-linolenic acid (ALA, C18:3n-3) (Figure 6). (181,186). Excessive carbohydrate,
glucose, or dietary fatty acid precursors undergo a series of desaturation and elongation reactions
resulting in the production of either saturated, monounsaturated (MUFAs) or polyunsaturated
fatty acids (PUFAs).
The retina possesses a unique lipid profile, it is highly enriched in n-3 PUFA, docosahexaenoic
acid (DHA, 22:6n-3) (182,183) and very long chain polyunsaturated fatty acids (VLC-PUFA,

	
  
	
  

27

C28-38) (184,185). This is concomitant with the high expression of Elongases 4, 2, 1 and 6
(186). ELOVL4 shows the greatest expression level among retinal elongases. ELOVL4 is known
for its substrate specificity for long chain fatty acids (≥C24) with synthesis of very long chain
fatty acids (VLCFA), which may explain the high content (or abundance) of VLC-PUFA in
retinal tissue. Our lab has previously demonstrated that diabetes-induced down regulation of
retinal elongases resulted in significantly different retinal lipid profiles, including higher
saturation index, shorter fatty acid chains, and decreased long-chain-to-short-chain
polyunsaturated fatty acids (PUFA) ratio and n3-to-n6 PUFA ratio, with a significant decreases
in DHA and VLC-PUFA (186).
Apart from diabetes induced retinal fatty acid remodeling, we have previously demonstrated
dysregulation of sphingolipid metabolism in the diabetic retina with pathological activation of
the acid sphingomyelinase (ASM) enzyme and increased acid sphingomyelinase-mediated
production of pro-inflammatory and pro-apoptotic ceramides (187).
These changes in retina lipid metabolism were associated with higher expressions of proinflammatory cytokines, adhesion molecules and retinal vascular degeneration (186,188,189).
This evidence suggests that lipid metabolism may play an important role in pathogenesis of DR.
1.4.4 Fatty acid elongase, ELOVL4, and very long chain ceramides
Elongation of very long chain fatty acids-4 (ELOVL4) belongs to a seven-member elongase
(ELO) family that is involved in the synthesis of very long chain fatty acids (VLCFA). ELOVL4
was identified by genetic linkage and haplotype analysis in families with two dominant forms of
macular dystrophy; Stargardt-like macular dystrophy (STGD3) and autosomal dominant macular

	
  
	
  

28

dystrophy (adMD). Genetic mapping revealed that a 5-base pair deletion mutation in the proteincoding region of a newly discovered retinal gene called ELOVL4 was found in all affected
members. The new protein, ELOVL4, had a homology with a group of yeast proteins involved in
the synthesis of VLCFA, Zhang et al. (2001) suggested that this enzyme may be involved in the
elongation of long chain fatty acids (168).
ELOVL4 is restricted to the tissues where extremely long fatty acids (≥C24) exist. With the
highest expression of ELOVL4 in the retinal tissue (168,169,190–192), followed by the skin,
brain and testis (167,169). In the human retina, ElOVL4 is mainly expressed in the inner segment
of photoreceptors and their cell bodies in the outer nuclear layer, with some expression in the
ganglionic cell layer (192). ELOVL4 seems to be the only elongase that is involved in the
biosynthesis of both saturated and polyunsaturated very long chain fatty acids (C≥26). Agbaga et
al. (2008) revealed the metabolic function of ELOVL4, demonstrating that ELOVL4 is involved
in the synthesis of both saturated (VLC-FA, C≥26) and polyunsaturated (VLC-PUFA, C28-36)
very long chain fatty acids in cultured cells that have been transfected with mouse ELOVL4
(Figure 6) (193).

	
  
	
  

29

Elovl1& Elovl1& Elovl1& Elovl4& Elovl4&

20:0& 22:0& 24:0& 26:0& 28:0&
Elovl5/6!

Plant&oil&

Fish&oil&

Fatty!acid!metabolism!in!mammals!
Elovl2/5( Elovl2/5(

Plant&oil&

Fish&oil&

Elovl1/3&

26:5&
28:5&
30:5&
32:5&
34:5&
36:5&
38:5&

Elovl4&

26:6&

Fish&oil&

28:6&
Elovl4& 30:6&
Elovl4&
Elovl4&
Elovl4&
Elovl4&

32:6&
34:6&
36:6&
38:6&

Figure 6. Schematic diagram of VLC-FA and VLC-PUFA biosynthetic pathways mediated
a series of elongation and desaturation reactions. Fatty acids are synthesized through either de
novo lipogenesis pathway using glucose as substrate or polyunsaturated fatty acid (PUFA)
remodeling pathway (Sprecher) using essential fatty acids, i.e., linoleic acid (LA, C18:2n-6) or
α-linolenic acid (ALA, C18:3n-3). ELOVL4 is necessary for elongating long chain fatty acids
generating VLC-FA and VLC-PUFA (C≥26). Fatty acids in solid boxes, those that accumulate in
animal and human tissues (modified from Tikhonenko et al., 2010) (186).

	
  
	
  

30

Moreover, retinas of heterozygous Stgd3 mice carrying the pathogenic 5 base pair deletion in
exon 6 of the ELOVL4 gene (194) and photoreceptor-specific homozygous ELOVL4 knock-out
mice (195) showed a significant decrease in VLC-PUFA-containing phosphatidylcholines (PCs,
C28-C36) that have been shown to interact with rhodopsin (184). This decrease was associated
with retinal manifestations similar to those seen in STGD3 patients. STGD3 is an autosomal
dominant Stargardt – like macular degeneration characterized by the accumulation of lipofuscin,
RPE atrophy, degeneration of macular photoreceptor and reduced vision (196,197). The diseases
is caused by 5 base pair deletion in exon 6 of ELOVL4 gene, resulting in C- terminal truncated
ELOVL4 with its mislocalization and aggregation outside the endoplasmic reticulum. Moreover
this mutant ELOVL4 exerts a dominant negative effect on the wild type ELOVL4, changing its
subcellular distribution. The presence of the misfolded mutant protein leads to photoreceptor
degeneration and vision loss (192). As this misfolding is not a problem in DR, the role or the
function of ELOVL4 in BRB maintenance rather than photoreceptor damage is observed in DR.
While, mice homozygous for either Stgd3 (167,197) or ELOVL4 null mutations (198) exhibit
neonatal lethal phenotype due to defective skin permeability barrier, excessive water loss and
dehydration. The homozygous and null mutation mice were found to have complete absence of
epidermal VLC-FA- containing acylceramides (VLCFA, C≥28), these phenotypes were rescued
by epidermal –specific expression of wild-type (WT) ELOVL4 transgene (199).
We can conclude that ELOVL4 plays a major role in the biosynthesis of VLC-PUFAs which are
esterified into PCs in the retina, and in the biosynthesis of saturated or monounsaturated VLCFAs which are incorporated into skin ceramides (Figure 7).
The Function of VLC-PUFAs in the retina is not known but they are mainly present in the retinal
outer segment in PCs and are able to span both
	
  
	
  

leaflets of the lipid bilayer. VLC-PUFAs may

31

play a role in the stability of highly curved cell membranes, such as the rims of photoreceptors
disk membranes (193). Karan et al. (2005) has found that photoreceptor-specific overexpression
of mutant ELOVL4 was associated with irregularities in disk membrane organization (196).
Most of the identified retinal fatty acids are VLC-PUFAs (C28-C36) with six double bonds.
However, Brush et al. (2010) has demonstrated the presence of saturated and monounsaturated
VLC-FAs (C 26-30), with C26:0 is the most abundant VLC-FA incorporated into retinal
ceramides (185). The role of these VLC-FAs in the retina is largely unknown.
ELOVL4 was initially discovered as the etiological gene of STGD3 (168). It has also been
shown to be associated with other retinal diseases. Abcouwer et al. (2010) demonstrated that
ELOVL4 is one of the enzymes that was found to be down regulated by ischemia-reperfusion in
the retina (200).
Sphingosine*

Sphingosine*

OH*

CH2OH*

NH2*

VLC"ceramide"
Sphingosin
OH"
Sphingosin
e"
CH2OH"

Fa-y*acid*

e"

+"
OH"

VLC9FA"(C"≥"26)"

Fa.y"acid"

NH
2"

ELOVL49"produced"VL9CFA"(C"≥"26)""

"""

Figure 7. Synthesis of VLC ceramides. ELOVL4-produced VLC-FA incorporates into the
structure of VLC-ceramides.
We
	
  
	
  

have

previously

demonstrated

that

expression level of elongases and desaturases

32

in the liver and retina is affected by diabetes with dramatic downregulation of elongases,
including ELOVL4 in the retina (165,171,186), this may be attributed to the reduced level (type
1 DM) or reduced action (type 2 DM) of insulin, the well-known regulator of the elongases.
Diabetes-induced down regulation of elongase in the retina was associated with significant
decrease in total retinal DHA and VLC-PUFA incorporated into retinal phosphatidylcholine.
These changes of retinal lipid metabolism were associated with higher expressions of proinflammatory cytokines, adhesion molecules and retinal vascular injury (186,188,189).
According to these data, it is conceivable that ELOVL4 plays a critical role in maintaining retinal
homeostasis.
1.4.5 ASM and short chain ceramides
Acid sphingomyelinase (ASM) is the key enzyme of sphingolipid metabolism. ASM exists in
two cellular components; the intracellular lysosome and the outer leaflet of the cell membrane.
Activation of acid sphingomyelinase by either receptor (CD95) (201) or stress stimuli (UV
radiation) (202) results in its translocation from the lysosome to the outer leaflet of the cell
membrane catalyzing the hydrolysis of membrane sphingomyelin (SM) into ceramides (Figure
8). The produced ceramide molecules self-associate into ceramide-enriched membrane domains
and result in reorganization of membrane rafts, aggregation of signaling proteins, receptor
clustering and enhanced cellular signal transduction (203,204). We have previously
demonstrated that ASM exhibits distinct substrate specificity towards short saturated acyl chains
(SM) with the production of pro-inflammatory and pro-apoptotic short chain ceramides (C≤16)
when it gets activated in different pathological conditions such as diabetes (187). ASM was
found to be highly activated in diabetic retinal vasculature (189), inhibition of ASM exhibits a

	
  
	
  

33

protective effect-- preventing diabetes induced retinal inflammation and vascular degeneration
(187–189).
Phosphorylcholine)

ASM)
Short)chain)
sphingomyelin)))

Short)chain)
ceramide))

Figure 8. Acid sphingomyelinase (ASM) is a central enzyme of sphingolipid metabolism.
ASM hydrolyzes SM to pro-apoptotic and pro-inflammatory short chain ceramides.

1.5 Reparative circulating angiogenic cells
Circulating angiogenic cells (CACs), a population of vascular progenitors originated from
hematopoietic stem cells (HSCs) (205) in the bone marrow, are now recognized as key regulators
for vascular repair. HSCs isolated from bone marrow or CACs from peripheral blood of control
animals have been shown to repair ischemic damage and aid in reperfusion of ischemic tissues
(205–207). This reparative process takes place through a series of orchestrated steps:
mobilization of CACs from bone marrow, their recruitment and extravasation to the sites of
damaged vessels.

	
  
	
  

34

CACs mobilize in response to a number of physiological or pathological stimuli to the site of
injury (208) promoting the repair of the damaged endothelium, either by incorporation and
differentiation into mature endothelial cells (205,209) or by secretion of paracrine factors that
enhance the integrity of the resident endothelium (210–212) . CACs mobilization is regulated by
sympathetic signaling to bone marrow stromal cells, which results in stromal cell-derived factor1 (SDF-1) down regulation and progenitor egress from the bone marrow (213,214).

In diabetes bone marrow-derived CACs are dysfunctional, with defective proliferation,
migration, adhesion and differentiation potential (110,215,216). Several studies have shown an
association between DR risk and both reduced number (217–219) and function of CACs
(187,220,221). We have previously demonstrated that the defective release of BM-derived CACs
in T2D rat model precedes the development of DR (217). The impaired release of CACs to the
site of injury is due to defective migration of these cells in response to ischemic induced factors;
SDF-1 and VEGF (222). The blunted response of CACs is due to reduced bioavailability of nitric
oxide (NO) in diabetes, essential for CACs migration (223–225) and survival (226).
Our laboratory identified marked increase in ASM expression in diabetic CACs (187,227).
Increase in ASM activity in CACs in diabetic animal models was associated with ceramide
accumulation in the cell membrane, and thereby increasing membrane rigidity and impairing
migration (227). Interestingly, we demonstrated that inhibition of ASM in diabetic CACs
decreased membrane rigidity and improved the migration of these cells to damaged retinal
vasculature, preventing diabetes induced increase in retinal vascular permeability and retinal
capillary loss (227).

	
  
	
  

35

1.6 Objectives of the dissertation
Diabetic retinopathy (DR) is a microvascular disease that results from diabetes-induced retinal
damage. The underlying mechanisms of vascular dysfunction in diabetes are yet to be fully
understood. Dyslipidemia, as well as hyperglycemia, is a major metabolic insult of diabetes and
is positively correlated with the development of DR. We previously explored the novel concept
that altered retinal lipid metabolism and dyslipidemia leads to retinal inflammation and vascular
degeneration (186–189). We found that early - stage diabetes induced marked decrease in the
expression of many retinal fatty acid elongases, in particular severe downregulation of the most
abundant elongase; elongation of very long chain fatty acids 4 (ELOVL4). Downregulation of
elongases in the diabetic retina lead to significant remodeling of retinal fatty acids with
decreased long-chain-to-short-chain fatty acids ratio and n3-to-n6 PUFA ratio, and higher
saturation index with significant decrease in n3-DHA and n3 VLCPUFA. These changes were
associated with increased expression of adhesion molecules, pro-inflammatory cytokines, and
retinal vascular pathology (186).
ELOVL4 was initially discovered as the etiological gene of STGD3 retinal disease with
association to a significant decrease in VLC-PUFA-containing phosphatidylcholines. It has also
been shown to be essential in maintenance of the water permeability barrier in skin (197,198). To
date, no studies have approached the direct role of ELOVL4 in retinal vascular permeability.
In addition to ELOVL4, we have previously demonstrated activation of the central enzyme of
sphingolipid metabolism, acid sphingomyelinase (ASM), as the key metabolic abnormality in
diabetic retinal vasculature and CACs (188,189).

	
  
	
  

36

Thus, the central hypothesis of this dissertation is as follows:
Diabetes-induced disruption of retinal lipid metabolism plays a critical role in retinal
vascular dysfunction through two pathways
•

Down regulation of ELOVL4-produced very long chain ceramides, contribute to
diabetes induced retinal vascular permeability.

•

Up regulation of ASM-produced short chain ceramides, contribute to retinal
endothelial damage and defective function of the reparative progenitor cells.

Thus normalizing the ratio between pro-barrier long chain ceramides and proinflammatory and pro-apoptotic short chain ceramides could prevent diabetes-induced
retinal damage and progression of DR.

	
  
	
  

37

1.7 Overview of chapters
Chapter I is a review of the relevant literature addressing diabetic retinopathy disease, Blood
retinal barrier break down in DR, diabetic dyslipidemia and its contribution to retinal vascular
degeneration.
Chapter II investigates the role of ELOVL4 and VLC ceramides in diabetes-induced increase in
retinal vascular permeability and to identify if lipids in particular VLC ceramides localize in the
tight junction and play a role in controlling the paracellular permeability of retinal endothelial
cells. Experiments described in this chapter demonstrate that ELOVL4-mediated production of
VLC-ceramides is necessary for retinal endothelial barrier function and normalization of retinal
ELOVL4 expression may provide a potential therapeutic strategy for the prevention of diabetesinduced early breakdown of the blood-retina barrier by modulating retinal sphingolipid
metabolism.
Chapter III investigates the effect of diabetes-induced ASM activity on the function of both cell
types involved in revascularization; retinal endothelial cells and circulating angiogenic cells
(CACs). Experiments described in this chapter demonstrate that ASM has deleterious effect on
the vascular repair potential of both mature retinal endothelial cells and circulating angiogenic
cells in diabetes and correcting this defect could treat vasodegeneration, enhancing vessel repair
and thus preventing progression into PDR stage. This study is published in Journal of Clinical
Lipidology.
Chapter IV summarizes the obtained data, the conclusions that can be reached from the studies
included in the thesis and suggests directions for

	
  
	
  

future research related to the present study.

38

REFERENCES

	
  
	
  

39

REFERENCES

1.

American Diabetes Association. Diagnosis and classification of diabetes mellitus.
Diabetes Care. 2014;37(SUPPL.1):81–90.

2.

Williams R, Airey M, Baxter H, Forrester J, Kennedy-Martin T, Girach a. Epidemiology
of diabetic retinopathy and macular oedema: a systematic review. Eye (Lond).
2004;18(10):963–83.

3.

Stumvoll M, Goldstein B, van Haeften T. Type 2 diabetes: principles of pathogenesis and
therapy. Lancet. 2005;365:1333–46.

4.

Olokoba AB, Obateru OA, Olokoba LB. Type 2 diabetes mellitus: A review of current
trends. Oman Med J. 2012;27(4):269–73.

5.

Lin Y, Sun Z. Current views on type 2 diabetes. J Endocrinol. 2010;204(1):1–11.

6.

International
Diabetes
Federation.
http://www.diabetesatlas.org (2016).

7.

Zimmet P, Alberti KG, Magliano DJ, Bennett PH. Diabetes mellitus statistics on
prevalence and mortality: facts and fallacies. Nat Rev Endocrinol. 2016;12:616–22.

8.

Kanguru L, Bezawada N, Hussein J, Bell J. The burden of diabetes mellitus during
pregnancy in low- and middle-income countries: a systematic review. Glob Health Action.
2014;7:23987.

9.

Mathers CD, Loncar D. Projections of global mortality and burden of disease from 2002
to 2030. PLoS Med. 2006;3(11):2011–30.

10.

American Diabetes Association. Economic costs of diabetes in the U.S. in 2012. Diabetes
Care. 2013;36(4):1033–46.

11.

Yau JWY, Rogers SL, Kawasaki R, Lamoureux EL, Kowalski JW, Bek T, et al. Global
prevalence and major risk factors of diabetic retinopathy. Diabetes Care. 2012;35(3):556–
64.

12.

Fong DS, Aiello L, Gardner TW, King GL, Blankenship G, Cavallerano JD, et al.
Retinopathy in Diabetes. Diabetes Care. 2004;27(SUPPL. 1): s84-s87

13.

Klein R, Klein BEK, Moss SE, Davis MD, Demets DL. The Wisconsin Epidemiologic
Study of Diabetic Retinopathy. Arch Ophthalmol. 1984;102:520–6.

14.

Cunha-Vaz J, Bernardes R. Nonproliferative retinopathy in diabetes type 2. Initial stages
and characterization of phenotypes. Prog Retin Eye Res. 2005;24(3):355–77.

	
  
	
  

40

IDF

Diabetes

Atlas

7th

edn

15.

Wong K. Defining diabetic retinopathy severity. Diabet Retin Evidence-Based Manag.
2010;(10):105–20.

16.

Early Treatment Diabetic Retinopathy Study Research Group. Grading Diabetic
Retinopathy from Stereoscopic Color Fundus Photographs-an Extension of the Modified
Airlie House Classification. Ophthalmology [Internet]. 1991;98(5):786–806.

17.

Wilkinson CP, Ferris FL, Klein RE, Lee PP, Agardh CD, Davis M, et al. Proposed
international clinical diabetic retinopathy and diabetic macular edema disease severity
scales. Ophthalmology. 2003;110(9):1677–82.

18.

Hammes HP, Feng Y, Pfister F, Brownlee M. Diabetic retinopathy: Targeting
vasoregression. Diabetes. 2011;60(1):9–16.

19.

Lorenzi M, Gerhardinger C. Early cellular and molecular changes induced by diabetes in
the retina. Diabetologia. 2001;44(7):791–804.

20.

Cunha-Vaz J, Faria de Abreu JR, Campos AJ. Early breakdown of the blood-retinal
barrier in diabetes. Br J Ophthalmol. 1975;59(11):649–56.

21.

Vinores S a, McGehee R, Lee a, Gadegbeku C, Campochiaro P a. Ultrastructural
localization of blood-retinal barrier breakdown in diabetic and galactosemic rats. J
Histochem Cytochem. 1990;38(9):1341–52.

22.

Daley ML, Watzke RC, Riddle MC. Early loss of blue-sensitive color vision in patients
with type I diabetes. Diabetes Care. 1987;10(6):777–81.

23.

Kohner EM, Henkind P. Correlation of fluorescein angiogram and retinal digest in
diabetic. Am J Opthalmol. 1970;69(3):403-14.

24.

Ashton N. Studies of the Retinal Capillaries in Relation To Diabetic and Other
Retinopathies. Br J Ophthalmol. 1963;47:521–38.

25.

Cunha-Vaz JG. Pathophysiology of Diabetic Retinopathy. Br J Ophthalmol.1978;62:35155

26.

Klein R, Moss SE, Klein BEK. Retinal Microaneurysm counts and 10-year progression of
diabetic retinopathy. Arch Ophthalmol.1995; 113:1386-91.

27.

Klein R, Klein BE, Moss SE, Cruickshanks KJ. The Wisconsin Epidemiologic Study of
Diabetic Retinopathy. XV. The long-term incidence of macular edema. Ophthalmology
[Internet]. 1995;102(1):7–16.

28.

Aiello LP, Gardner TW, King GL, Blankenship G, Cavallerano JD, et al. Diabetic
retinopathy. Diabetes Care. 1998;21(1):143-56.

29.

Cunha-Vaz J, Travassos A. Breakdown of the blood-retinal barriers and cystoid macular

	
  
	
  

41

edema. Surv Ophthalmol. 1984;28(SUPPL. 2):485–92.
30.

Kohner EM, Dollery CT, Bulpitt CJ. Cotton-wool spots in diabetic retinopathy. Diabetes.
1969;18(10):691–704.

31.

Kohner EM, oakley NW. Diabetic Retinopathy.Metabolism. 1975; 24(9):1085-102.

32.

Frank RN. Diabetic retinopathy. Prog Retin Eye Res. 1995;14(2):361–92.

33.

Aiello LP, Avery RL, Arrigg PG, Keyt BA, Jampel HD et al. Vascular endothelial growth
factor in ocular fluid of patients with diabetic retinopathy and other retinal disoredrs. N
Engl J Med. 1994; 331: 1480-7.

34.

Antonetti DA, Barber AJ, Bronson SK, Freeman WM, Gardner TW, Jefferson LS, et al.
Diabetic retinopathy: Seeing beyond glucose-induced microvascular disease. Diabetes.
2006;55(9):2401–11.

35.

Simó R, Hernández C. Advances in the medical treatment of diabetic retinopathy.
Diabetes Care. 2009;32(8):1556–62.

36.

The Diabetes Control and Complications Trial Research Group. The effect of intensive
treatment of diabetes on the development and progression of long-term complications in
Insulin- dependent diabetes mellitus. N Engl J Med. 1993; 329:977-86.

37.

UK Prospective Diabetes Study (UKPDS) Group. Intensive blood-glucose control with
sulphonylureas or insulin compared with conventional treatment and risk of complications
in patients with type 2 diabetes (UKPDS 33). Lancet. 1998; 352 (9131): 837–53.

38.

The Action of Control Cardiovascular Risk in Diabetes Study Group. Effect of intensive
glucose lowering in type 2 diabetes. N Engl J Med. 2008; 358: 2545-59.

39.

Mohamed Q, Gillies MC, Womg TY. Management of diabetic retinopathy. JAMA.
2007;298(8):902-16.

40.

Osaadon P, Fagan XJ, Lifshitz T, Levy J. A review of anti-VEGF agents for proliferative
diabetic retinopathy. Eye (Lond). 2014;28(5):510–20.

41.

Nishijima K, Ng Y-S, Zhong L, Bradley J, Schubert W, Jo N, et al. Vascular Endothelial
Growth Factor-A Is a Survival Factor for Retinal Neurons and a Critical Neuroprotectant
during the Adaptive Response to Ischemic Injury. Am J Pathol. 2007;171(1):53–67.

42.

Saint-Geniez M, Maharaj ASR, Walshe TE, Tucker BA, Sekiyama E, Kurihara T, et al.
Endogenous VEGF is required for visual function: Evidence for a survival role on Muller
cells and photoreceptors. PLoS One. 2008;3(11):e3554.

43.

Arden GB, Arden GB, Sidman RL, Sidman RL, Arap W, Arap W, et al. Spare the rod and
spoil the eye. Br J Ophthalmol. 2005;89(August):764–9.

	
  
	
  

42

44.

Klaassen I, Van Noorden CJF, Schlingemann RO. Molecular basis of the inner bloodretinal barrier and its breakdown in diabetic macular edema and other pathological
conditions. Progress in Retinal and Eye Research. 2013;34: 19–48.

45.

Runkle EA, Antonetti DA. The blood-retinal barrier: structure and functional significance.
Methods Mol Biol. 2011;686: 133-48.

46.

Kim JH, Kim JH, Yu YS, Kim DH, Kim KW. Recruitment of pericytes and astrocytes is
closely related to the formation of tight junction in developing retinal vessels. J Neurosci
Res. 2009;87(3):653–9.

47.

Davis S, Aldrich TH, Jones PF, Acheson A, Compton DL, Jain V, et al. Isolation of
angiopoietin-1, a ligand for the TIE2 receptor, by secretion-trap expression cloning. Cell.
1996;87(7):1161–9.

48.

Sato TN, Tozawa , Deutsch U, Wolburg-Buchholz K, Fujiwara Y, et al. Distinct roles of
the receptor tyrosine kinases Tie-1 and Tie-2 in blood vessel formation. Nature.
1995;376(6535):70-4.

49.

Hori S, Ohtsuki S, Hosoya KI, Nakashima E, Terasaki T. A pericyte-derived angiopoietin1 multimeric complex induces occludin gene expression in brain capillary endothelial
cells through Tie-2 activation in vitro. J Neurochem. 2004;89(2):503–13.

50.

Sagaties MJ, Raviola G, Schaeffer S, Miller C. The structural basis of the inner bloodretina barrier in the eye of Macaca mulatta. Investig Ophthalmol Vis Sci.
1987;28(12):2000–14.

51.

Zlokovic B V. The blood-brain barrier in health and chronic neurodegenerative disorders.
Neuron. 2008;57(2):178–201.

52.

Matter K and Balda MS. Siganaling to and from tight junctions. Nature Reviews
Molecular Cell Biology. 2003;4: 225-37.
Tsukita S, Furuse M, Itoh M. Multifunctional strands in tight junctions. Nat Rev Mol Cell
Biol. 2001;2(4):285-93.

53.
54.

Van Meer G, Simons K, Simons K. The function of tight junctions in maintaining
differences in lipid composition between the apical and the basolateral cell surface
domains of MDCK cells. EMBO J. 1986;5(7):1455–64.

55.

González-Mariscal L, Lechuga S, Garay E. Role of tight junctions in cell proliferation and
cancer. Prog Histochem Cytochem. 2007;42(1):1–57.

56.

Farquhar MG, Palade GE. Junctional complexes in various epithelia. J cell
Biol.1963;17(2):375–412.

57.

Staehelin L.A. Further observations on the fine structure of freeze-cleaved tight junctions.
J Cell Sci. 1973;13:763–86.

	
  
	
  

43

58.

Shen L, Weber CR, Turner JR. The tight junction protein complex undergoes rapid and
continuous molecular remodeling at steady state. J Cell Biol. 2008;181(4):683–95.

59.

Murakami T, Felinski EA, Antonetti DA. Occludin phosphorylation and ubiquitination
regulate tight junction trafficking and vascular endothelial growth factor-induced
permeability. J Biol Chem. 2009;284(31):21036–46.

60.

Furuse M, Hirase T, Itoh M, Nagafuchi A, Yonemura S, Tsukita S, et al. Occludin: A
Novel Integral Membrane Protein Localizing at Tight Junctions. J cell
Biol.1993;123(6):1777–88.

61.

Furuse M, Itoh M, Hirase T, Nagaftichi A, Yonemura S, Tsukita S, et al. Direct
association of occludin with ZO-1 and its possible involvement in the localization of
occludin at tight junctions. J Cell Biol. 1994;127(6 I):1617–26.

62.

Itoh M, Morita K, Tsukita S. Characterization of ZO-2 as a MAGUK Family Member
Associated with Tight as well as Adherens Junctions with a Binding Affinity to Occludin
and α Catenin. J Biol Chem.1999;274(9):5981–6.

63.

Kojima T, Sawada N, Chiba H, Kokai , Yamamoto M, . Induction o f tight Junctions in
human connexin 32 (hCx32) -transfected mouse hepatocytes  : Connexin 32 interacts wi th
occludin. Biochem Biophys Res Commun.1999;266(1):222–9.

64.

Nusrat A, Chen JA, Foley CS, Liang TW, Tom J, Cromwell M, et al. The coiled-coil
domain of occludin can act to organize structural and functional elements of the epithelial
tight junction. J Biol Chem. 2000;275(38):29816–22.

65.

Fanning AS, Jameson BJ, Lynne a, Anderson JM, Jesaitis L a, Melvin J. The Tight
Junction Protein ZO-1 Establishes a Link between the Transmembrane Protein Occludin
and the Actin Cytoskeleton. J Biol Chem.1998;273(45):29745–53.

66.

Traweger A, Fang D, Liu YC, Stelzhammer W, Krizbai IA, Fresser F, et al. The tight
junction-specific protein occludin is a functional target of the E3 ubiquitin-protein ligase
itch. J Biol Chem. 2002;277(12):10201–8.

67.

Kevil CG, Okayama N, Trocha SD, Kalogeris TJ, Coe LL, et al. Expression of zonula
occludens and adherens junctiona proteins in human venous and and arterial endothelial
cells: role of occludin in endothelial solute barriers. Microcirculation.1998;5(2-3)197–
210.

68.

Hirase T, Staddon JM, Saitou M, Ando-Akatsuka Y, Itoh M, Furuse M, et al. Occludin as
a possible determinant of tight junction permeability in endothelial cells. J Cell Sci.
1997;110 (Pt 14):1603–13.

69.

Antonetti DA, Wolpert EB, DeMaio L, Harhaj NS, Scaduto RC. Hydrocortisone decreases
retinal endothelial cell water and solute flux coincident with increased content and
decreased phosphorylation of occludin. J Neurochem. 2002;80(4):667–77.

	
  
	
  

44

70.

Phillips BE, Cancel L, Tarbell JM, Antonetti DA. Occludin independently regulates
permeability under hydrostatic pressure and cell division in retinal pigment epithelial
cells. Invet Ophthalmol Vis Sci. 2008;49(6)2568-76.

71.

Antonetti DA, Barber AJ, Khin S, Lieth E, Tarbell JM, Gardner T. Vascular peremability
in experimental diabetes is associated with reduced endothelial occludin content: vascular
endothelial growth factor decreases occludin in retinal endothelial cells.
Diabetes.1998;47(12):1953-9.

72.

Saitou M, Fujimoto K, Doi Y, Itoh M, Fujimoto T, Furuse M, et al. Occludin-deficient
embryonic stem cells can differentiate into polarized epithelial cells bearing tight
junctions. J Cell Biol. 1998;141(2):397–408.

73.

Saitou M, Furuse M, Sasaki H, Schulzke JD, Fromm M, Takano H, et al. Complex
phenotype of mice lacking occludin, a component of tight junction strands. Mol Biol Cell.
2000;11(12):4131–42.

74.

Itoh M, Furuse M, Morita K, Kubota K, Saitou M, Tsukita S. Direct binding of three tight
junction-associated MAGUKs, ZO-1, ZO-2 and ZO-3, with COOH termini of claudins. J
Cell Biol. 1999;147(6):1351–63.

75.

Rüffer C, Gerke V. The C-terminal cytoplasmic tail of claudins 1 and 5 but not its PDZbinding motif is required for apical localization at epithelial and endothelial tight
junctions. Eur J Cell Biol. 2004;83(4):135–44.

76.

Morita K, Sasaki H, Furuse M, Tsukita S. Endothelial Claudin: Claudin-5/TMVCF
constitutes tight junction starnds in endothelial cells. J Cell Biol. 1999;147(1):185–94.

77.

Furuse M, Sasaki H, Fujimoto K, Tsukita S. A single gene product, claudin-1 or -2,
reconstitutes tight junction strands and recruits occludin in fibroblasts. J Cell Biol.
1998;143(2):391–401.

78.

Inai T, Kobayashi J, Shibata Y. Claudin-1 contributes to the epithelial barrier function in
MDCK cells. Eur J Cell Biol. 1999;78(12):849–55.

79.

Nitta T, Hata M, Gotoh S, Seo Y, Sasaki H, Hashimoto N, et al. Size-selective loosening
of the blood-brain barrier in claudin-5-deficient mice. J Cell Biol. 2003;161(3):653–60.

80.

Koto T, Takubo K, Ishida S, Shinoda H, Inoue M, Tsubota K, et al. Hypoxia disrupts the
barrier function of neural blood vessels through changes in the expression of claudin-5 in
endothelial cells. Am J Pathol. 2007;170(4):1389–97.

81.

Stevenson BR, Siliciano JD, Mooseker MS, Goodenough DA. Identification of Z0-1: a
high molecular weight polypeptide associated with tight junction (zonula occludens) in a
variety of epithelia. J Cell Biol. 1986;103(September):755–66.

82.

Gumbiner B, Lowenkopf T, Apatira D. Identification of a 160-kDa polypeptide that binds
to the tight junction protein ZO-1. Proc Natl Acad Sci U S A [Internet].

	
  
	
  

45

1991;88(8):3460–4.
83.

Haskins J, Gu L, Wittchen ES, Hibbard J, Stevenson BR. ZO-3, a novel member of the
MAGUK protein family found at the tight junction, interacts with ZO-1 and occludin. J
Cell Biol. 1998;141(1):199–208.

84.

Woods DF, Bryant PJ. ZO-1, DlgA and PSD-95/SAP90: homologous proteins in tight,
septate and synaptic cell junctions. Mech Dev. 1993;44(2–3):85–9.

85.

Wittchen ES, Haskins J, Stevenson BR. Protein Interactions at the Tight Junction. Actin
has multiple bending partners, and ZO-1 forms independent complexes with ZO-2 and
ZO-3. J Biol Chem.1999;274(49):35179–85.

86.

McNeil E, Capaldo CT, Macara IG. Zonula occludens-1 function in the assembly of tight
junctions in Madin-Darby canine kidney epithelial cells. Mol Biol Cell.2006;17(4):192232.

87.

Hernandez S, Chavez Munguia B, Gonzalez-Mariscal L. ZO-2 silencing in epithelial cells
perturbs the gate and fence function of tight junctions and leads to an atypical monolayer
architecture. Exp Cell Res. 2007;313(8):1533–47.

88.

Umeda K, Ikenouchi J, Katahira-Tayama S, Furuse K, Sasaki H, Nakayama M, et al. ZO1 and ZO-2 Independently Determine Where Claudins Are Polymerized in Tight-Junction
Strand Formation. Cell. 2006;126(4):741–54.

89.

Katsuno T, Umeda K, Matsui T, Hata M, Tamura A, Itoh M, et al. Deficiency of zonula
occludens-1 causes embryonic lethal phenotype associated with defected yolk sac
angiogenesis and apoptosis of embryonic cells. Mol Biol Cell. 2008;19(6):2465-75.

90.

Balda MS, Matter K. The tight junction protein ZO-1 and an interacting transcription
factor regulate ErbB-2 expression. EMBO J. 2000;19(9):2024–33.

91.

Laffafian I, Hallett MB. Lipid-Protein Cargo Transfer  : A Mode of Direct Cell-to-Cell
Communication for Lipids and Their Associated Proteins. J cell Physiol. 2007
;210(2):336–42.

92.

Grebenkfimper K. Translational diffusion measurements of a fluorescent phospholipid
between MDCK-I cells support the lipid model of the tight junctions. Chem Phys Lipids.
1994;71(94):133–43.

93.

Lee DBN, Jamgotchian N, Allen SG, Abeles MB, Ward HJ. A lipid-protein hybrid model
for tight junction. Am J Physiol Renal Physiol. 2008 ;295(6):F1601-12.

94.
Kachar B, Reese TS. Evidence for the lipidic nature of tight junction strands.1982
Nature.1982;296:464-6.
95.

	
  
	
  

Kan FW. Cytochemical Evidence for the Presence of Phospholipids in Epithelial Tight

46

Junction Strands. J Histochem Cytochem.1993;41(5):649-56.
96.

Feltkamp CA, Van der waerden AW. Junction formation between cultured normal rat
hepatocytes. An ultrastructural study on the presence of cholesterol and the structure of
developing tight- junction strands. J Cell Sci.1983;63:271–86.

97.

Vinores SA, Gadegbeku C, Campochiaro PA, Greent WR. Immunohistochemical
localization of blood-retinal barrier breakdown in human diabetics. Am J Pathol
1989;134(2):231–5.

98.

Barber AJ, Antonetti DA. Mapping the Blood Vessels with Paracellular Permeability in
the Retinas of Diabetic Rats. Investig Ophthalmol Vis Sci. 2003;44(12):5410–6.

99.

Antonetti DA, Barber AJ, Hollinger LA, Wolpert EB, Gardner TW. Vascular Endothelial
Growth Factor Induces Rapid Phosphorylation of Tight Junction Proteins Occludin and
Zonula Occluden 1. J Biol Chem. 1999;274(33):23463–7.

100. Harhaj NS, Felinski EA, Wolpert EB, Sundstrom JM, Gardner TW, Antonetti DA. VEGF
activation of protein kinase C stimulates occludin phosphorylation and contributes to
endothelial permeability. Investig Ophthalmol Vis Sci. 2006;47(11):5106–15.
101. Murakami T, Frey T, Lin C, Antonetti DA. Protein kinase Cβ phosphorylates occludin
regulating tight junction trafficking in vascular endothelial growth factor - Induced
permeability in vivo. Diabetes. 2012;61(6):1573–83.
102. Leal EC, Manivannan A, Hosoya K, Terasaki T, Cunha-Vaz J, Ambrósio AF, et al.
Inducible nitric oxide synthase isoform is a key mediator of leukostasis and blood-retinal
barrier breakdown in diabetic retinopathy. Invest Ophthalmol Vis Sci. 2007;48(11):5257–
65.
103. Ghassemifar R, Lai CM, Rakoczy PE. VEGF differentially regulates transcription and
translation of ZO-1α+ and ZO-α1- and mediates trans-epithelial resistance in cultured
endothelial and epithelial cells. Cell Tissue Res. 2006;323(1):117–25.
104. Peters S, Cree IA, Alexander R, Turowski P, Ockrim Z, Patel J, et al. Angiopoietin
modulation of vascular endothelial growth factor: Effects on retinal endothelial cell
permeability. Cytokine. 2007;40(2):144–50.
105. Klaassen I, Hughes JM, Vogels IMC, Schalkwijk CG, Van Noorden CJF, Schlingemann
RO. Altered expression of genes related to blood-retina barrier disruption in
streptozotocin-induced diabetes. Exp Eye Res. 2009;89(1):4–15.
106. Mizutani M, Kern TS, Lorenzi M. Accelerated death of retinal microvascular cells in
human and experimental diabetic retinopathy. J Clin Invest. 1996;97(12):2883–90.
107. Hammes HP, Du X, Edelstein D, Taguchi T, Matsumura T, Ju Q, et al. Benfotiamine
blocks three major pathways of hyperglycemic damage and prevents experimental diabetic
	
  
	
  

47

retinopathy. Nat Med. 2003;9(3):294-9.
108. Brownlee Michael. The pathobiology of diabetic complications. A unifying mechanism.
Diabetes. 2005;54(6):1615-25.
109. Joussen AM, Murata T, Tsujikawa A, Kirchhof B, Bursell S-E, Adamis AP. LeukocyteMediated Endothelial Cell Injury and Death in the Diabetic Retina. Am J Pathol [Internet].
2001;158(1):147–52.
110. Chen Y, Lin S, Lin F, Wu T, Tsao C. High Glucose Impairs Early and Late Endothelial
Progenitor Cells by Modifying Nitric Oxide – Related but Not Oxidative Stress –
Mediated Mechanisms. Diabetes. 2007;56(6):1559–68.
111. Churdchomjan W, Kheolamai P, Manochantr S, Tapanadechopone P, Tantrawatpan C, UPratya Y, et al. Comparison of endothelial progenitor cell function in type 2 diabetes with
good and poor glycemic control. BMC Endocr Disord. 2010;10:5.
112. Stewart PA, Tuor UI. Blood-eye barriers in the rat: Correlation of ultrastructure with
function. J Comp Neurol. 1994;340(4):566–76.
113. Al Ahmad A, Gassmann M, Ogunshola OO. Maintaining blood-brain barrier integrity:
Pericytes perform better than astrocytes during prolonged oxygen deprivation. J Cell
Physiol. 2009;218(3):612–22.
114. Wisniewska-Kruk J, Hoeben KA, Vogels IMC, Gaillard PJ, Van Noorden CJF,
Schlingemann RO, et al. A novel co-culture model of the blood-retinal barrier based on
primary retinal endothelial cells, pericytes and astrocytes. Exp Eye Res. 2012;96(1):181–
90.
115. Enge M, Bjarnega M, Gerhardt H, Gustafsson E, Asker N, Hammes H, et al.
Endothelium-specific platelet-derived growth factor-B ablation mimics diabetic
retinopathy. EMBO J. 2002;21(16):4307–16.
116. Kuwabara T, Cogan DG. Retinal vascular patterns. VI. Mural cells of the retinal
capillaries. Arch Opthalmol.1963;69:492-502.
117. Geraldes P, Hiraoka-yamamoto J, Matsumoto M, Clermont A. Activation of PKCδ and
SHP1 by hyperglycemia causes vascular cell apoptosis and diabetic retinopathy. Nat Med.
2009;15(11):1298–306.
118. Patel JI, Hykin PG, Gregor ZJ, Boulton M, Cree IA. Angiopoietin concentrations in
diabetic retinopathy. Br J Ophthalmol. 2005;89(4):480–3.
119. Rangasamy S, Srinivasan R, Maestas J, McGuire PG, Das A. A potential role for
angiopoietin 2 in the regulation of the blood-retinal barrier in diabetic retinopathy.
Investig Ophthalmol Vis Sci. 2011;52(6):3784–91.
120. Beltramo E, Pomero F, Allione A, D’Alù
	
  
	
  

F, Ponte E, Porta M. Pericyte adhesion is

48

impaired on extracellular matrix produced by endothelial cells in high hexose
concentrations. Diabetologia. 2002;45(3):416–9.
121. Kuiper EJ, Witmer AN, Klaassen I, Oliver N, Goldschmeding R, Schlingemann RO.
Differential expression of connective tissue growth factor in microglia and pericytes in the
human diabetic retina. Br J Ophthalmol. 2004;88(8):1082–7.
122. Miyamoto K, Khosrof S, Bursell SE, Rohan R, Murata T, Clermont AC, et al. Prevention
of leukostasis and vascular leakage in streptozotocin-induced diabetic retinopathy via
intercellular adhesion molecule-1 inhibition. Proc Natl Acad Sci U S A.
1999;96(19):10836–41.
123. McLeod DS, Lefer DJ, Merges C, Lutty G a. Enhanced expression of intracellular
adhesion molecule-1 and P-selectin in the diabetic human retina and choroid. Am J Pathol.
1995;147(3):642–53.
124. Barouch FC, Miyamoto K, Allport JR, Fujita K, Bursell SE, Aiello LP, et al. Integrinmediated neutrophil adhesion and retinal leukostasis in diabetes. Investig Ophthalmol Vis
Sci. 2000;41(5):1153–8.
125. Joussen AM, Poulaki V, Le ML, Koizumi K, Esser C, Janicki H, et al. A central role for
inflammation in the pathogenesis of diabetic retinopathy. FASEB J. 2004;18(12):1450–2.
126.

Lu M, Perez VL, Ma N, Miyamoto K, Peng H, Liao JK, Adamis AP. VEGF Increases
Retinal Vascular ICAM-1 Expression In Vivo. IOVS.1999;40(8):1808-12

127. Joussen AM, Poulaki V, Qin W, Kirchhof B, Mitsiades N, Wiegand SJ, et al. Retinal
vascular endothelial growth factor induces intercellular adhesion molecule-1 and
endothelial nitric oxide synthase expression and initiates early diabetic retinal leukocyte
adhesion in vivo. Am J Pathol. 2002;160(2):501–9.
128. Nonaka A, Kiryu J, Tsujikawa A, Yamashiro K, Miyamoto K, Nishiwaki H, et al. PKC-β
inhibitor (LY333531) attenuates leukocyte entrapment in retinal microcirculation of
diabetic rats. Investig Ophthalmol Vis Sci. 2000;41(9):2702–6.
129. Bolton SJ, Anthony DC, Perry VH. Loss o f the tight junction proteins occludin and
zonula occludens-1 from cerebral vascular endothelium during neutrophil - induced blood
– brain barrier breakdown in vivo. Neuroscience.1998;86(4):1245–57.
130. Reese TS, Karnovsky MJ. Fine structural localizatin of blood-brain barrier to exogenous
peroxidase. J Cell Biol. 1967;34(1):207–17.
131. Schnitzer J. Retinal astrocytes: their restriction to vascularized parts of the mammalian
retina. Neurosci Lett. 1987;78(1):29–34.
132. Schnitzer J. Astrocytes in the guinea pig, horse, and monkey retina: Their occurrence
coincides with the presence of blood vessels. Glia. 1988;1(1):74–89.
	
  
	
  

49

133. Tout S, Chan-Ling T, Hollander H, Stone J. The role of Muller cells in the formation of
the blood-retinal barrier. Neuroscience. 1993;55(1):291–301.
134. Gardner TW, Lieth E, Khin SA, Barber AJ, Bonsall DJ, Lesher T, et al. Astrocytes
increase barrier properties and ZO-1 expression in retinal vascular endothelial cells.
Investig Ophthalmol Vis Sci. 1997;38(11):2423–7.
135. Willis CL, Leach L, Clarke GJ, Nolan CC, Ray DE. Reversible disruption of tight junction
complexes in the rat blood-brain barrier, following transitory focal astrocyte loss. Glia.
2004;48(1):1–13.
136. Tretiach M, Madigan MC, Wen L, Gillies MC. Effect of Muller cell co-culture on in vitro
permeability of bovine retinal vascular endothelium in normoxic and hypoxic conditions.
Neurosci Lett. 2005;378(3):160–5.
137.

Shen W, Li S, Chung SH, Gillies MC. Retinal vascular changes after glial disruption in
rats. J Neurosci Res. 2010;88(7):1485–99.

138. Behzadian M a, Wang XL, Windsor LJ, Ghaly N, Caldwell RB. TGF-beta increases
retinal endothelial cell permeability by increasing MMP-9: possible role of glial cells in
endothelial barrier function. Invest Ophthalmol Vis Sci. 2001;42(3):853–9.
139. Giebel SJ, Menicucci G, McGuire PG, Das A. Matrix metalloproteinases in early diabetic
retinopathy and their role in alteration of the blood-retinal barrier. Lab Invest.
2005;85(5):597–607.
140. Barber AJ, Antonetti DA, Gardner TW. Altered expression of retinal occludin and glial
fibrillary acidic protein in experimental diabetes. Penn State Retin Res Group Invest
Ophthalmol Vis Sci. 2000;41(11):3561–8.
141. Krady JK, Basu A, Allen CM, Xu Y, LaNoue KF, Gardner TW, et al. Minocycline
reduces proinflammatory cytokine expression, microglial activation, and caspase-3
activation in a rodent model of diabetic retinopathy. Diabetes. 2005;54(5):1559–65.
142. Martiney JA, Litwak M, Berman JW, Arezzo JC, Brosnan CF. Pathophysiologic effect of
interleukin-1b in the rabbit retina. Am J Pathol [Internet]. 1990;137(6):1411–23.
143. Vinores SA, Xiao WH, Shen J, Campochiaro. TNF-alpha is critical for ischemia-induced
leukostasis, but not retinal neovascularization nor VEGF- induced leakage. J
Neuroimmunol. 2007;182(1-2):73-9.
144. Coppack SW, Evans RD, Fisher RM, Fryan KN, Gibbons GF, et al. Adipose tissue
metabolism in obesity Lipase action in vivo before and after a mixed meal. Metabolism.
1992; 41(3):264-72.
145. Dimitriadis G, Mitrou P, Lambadiari V, Maratou E, Raptis SA. Insulin effects in muscle
and adipose tissue. Diabets Res Clin Pract. 2011;93:S52-9.
	
  
	
  

50

146. Frayn K. Adipose tissue as a buffer for daily lipid flux. Diabetologia. 2002;45(9):1201–
10.
147. Leavens KF, Birnbaum MJ. Insulin signaling to hepatic lipid metabolism in health and
disease. Crit Rev Biochem Mol Biol. 2011;46(3):200–15.
148. Goldberg IJ. Clinical review 124: Diabetic dyslipidemia - Causes and consequences. J
Clin Endocrinol Metab. 2001;86(3):965–71.
149. Guy J, Ogden L, Wadwa RP, Hamman RF, Mayer-Davis EJ, et al. Lipid and Lipoprotein
Profiles in Youth With and Without Type 1 Diabetes: The SEARCH for Diabetes in
Youth Case-Control Study. Diabetes Care. 2009;32(3):416–20.
150. Petitti DB, Imperatore G, Palla SL, Daniels SR, Dolan LM, et al. Serum Lipids and
Glucose Control. The earch for diabets in youth study.Arch Pediatr Adolesc Med
[Internet]. 2007;161:159–65.
151. Ginsberg HN. Diabetic dyslipidemia: Basic mechanisms underlying the common
hypertriglyceridemia and low HDL cholesterol levels. Diabetes. 1996;45(3 SUPPL.):27–
30.
152. Krauss RM. Heterogeneity of plasma low-density lipoprotein and atheroscelrosis risk. curr
opin lipidol. 1994;5(5):339-49.
153. Rader DJ. Effect of Insulin Resistance, Dyslipidemia, and Intra-abdominal Adiposity on
the Development of Cardiovascular Disease and Diabetes Mellitus. Am J Med.
2007;120(3 SUPPL. 1):S12–8.
154. Adams TH, Walzem RL, Smith DR, Tseng S, Smith SB. Hamburger high in total,
saturated and trans-fatty acids decreases HDL cholesterol and LDL particle diameter, and
increases TAG, in mildly hypercholesterolaemic men. Br J Nutr. 2010;103(1):91–8.
155. Kelley DS, Siegel D, Vemuri M, Mackey BE. Docosahexaenoic acid supplementation
improves fasting and postprandial lipid profiles in hypertriglyceridemic men. Am J Clin
Nutr. 2007;86:324–33.
156. Tilvis RS, Taskinen MR, Miettinen TA. Effect of insulin treatment on fatty acids of
plasma and erythrocyte membrane lipids in type 2 diabetes. Clin Chim Acta. 1988;171(2–
3):293–303.
157. Seigneur M, Freyburger G, Gin H, Claverie M. Serum fatty acid profiles in type I and type
II diabetes: metabolic alterations of fatty acids of the main serum lipids. Diabetes Res.
1994;23:169–77.
158. Jakobsson A, Westerberg R, Jacobsson A. Fatty acid elongases in mammals: Their
regulation and roles in metabolism. Prog Lipid Res. 2006;45(3):237–49.
159. Uchida Y. The role of fatty acid
	
  
	
  

elongation

51

in

epidermal

structure

and

function. Dermatoendocrinol. 2011;3(2):65–9.
160. Kihara A. Very long-chain fatty acids: Elongation, physiology and related disorders. J
Biochem. 2012;152(5):387–95.
161. Guillou H, Zadravec D, Martin PGP, Jacobsson A. The key roles of elongases and
desaturases in mammalian fatty acid metabolism: Insights from transgenic mice. Prog
Lipid Res. 2010;49(2):186–99.
162. Brenner RR. Hormonal modulation of   Δ6 and  Δ5 desaturases: Case of diabetes.
Prostaglandins Leukot Essent Fat Acids. 2003;68(2):151–62.
163. Rimoldi OJ, Finarelli GS, Brenner R. Effects of Diabetes and Insulin on Hepatic Δ 6
Desaturase Gene Expression. Biochem Biophys Res Commun. 2001;283(2):323–6.
164. Waters K, Ntambi JM. Insulin and dietary fructose induce stearoyl-coA desaturase 1 gene
expression in liver of diabetic mice. J Biol Chem. 1994;269(44):27773-7
165. Wang Y, Botolin D, Xu J, Christian B, Mitchell E, Jayaprakasam B, et al. Regulation of
hepatic fatty acid elongase and desaturase expression in diabetes and obesity. J Lipid Res.
2006;47(9):2028–41.
166. Ohno Y, Suto S, Yamanaka M, Mizutani Y, Mitsutake S, Igarashi Y, et al. ELOVL1
production of C24 acyl-CoAs is linked to C24 sphingolipid synthesis. Proc Natl Acad Sci
U S A. 2010;107(43):18439–44.
167. Vasireddy V, Uchida Y, Salem N, Kim SY, Mandal NA, et al. Loss of functional
ELOVL4 depletes very long-chain fatty acids (≥28) and the unique ω-o-acylceramides in
skin leading to neonatal death. Hum Mol Genet. 2007;16(5):471-82.
168. Zhang K, Kniazeva M, Han M, Li W, Yu Z, yang Z, et al. A 5-bp deletion in ELOVL4 is
associated with two related forms of autosomal dominant macular dystrophy. Nat Genet.
2001;27(1):89–93.
169. Mandal MNA, Ambasudhan R, Wong PW, Gage PJ, Sieving PA, Ayyagari R.
Characterization of mouse orthologue of ELOVL4: Genomic organization and spatial and
temporal expression. Genomics. 2004;83(4):626–35.
170. Jakobsson A, Jörgensen JA, Jacobsson A. Differential regulation of fatty acid elongation
enzymes in brown adipocytes implies a unique role for Elovl3 during increased fatty acid
oxidation. Am J Physiol Endocrinol Metab. 2005;289(4):E517–26.
171. Green CD, Olson LK. Modulation of palmitate-induced endoplasmic reticulum stress and
apoptosis in pancreatic β-cells by stearoyl-CoA desaturase and Elovl6. Am J Physiol
Endocrinol Metab.2011;300(4):E640–9.
172. Van Eck WF. The effect of a low fat diet on the serum lipids in diabetes and its
	
  
	
  

52

significance in diabetic retinopathy. Am J Med [Internet]. 1959;27:196–211.
173. Houtsmuller AJ, Zahn KJ, Henkes HE. Unsaturated fats and progression of diabetic
retinopathy. Doc Opthalmol. 1980;48(2):363–71.
174. Augustin AJ, Breipohl W, Böker T, Lutz J, Spitznas M. Increased lipid peroxide levels
and myeloperoxidase activity in the vitreous of patients suffering from proliferative
diabetic retinopathy. Graefe’s Arch Clin Exp Ophthalmol. 1993;231(11):647–50.
175. Chew EY, Klein ML, Iii FLF, Remaley N a, Murphy RP, Chantry K, et al. Association of
elevated serum lipid levels with retinal hard exudate in diabetic retinopathy. Arch
Opthalmol.1996;114:1079-84.
176. Ferris FL, Chew EY, Hoogwerf B. Serum lipids and diabetic retinopathy. Diabetes care.
1996;19(11):1291-93.
177. Davis MD, Fisher MR, Gangnon RE, Barton F, Aiello LM, Chew EY, et al. Risk factors
for high-risk proliferative diabetic retinopathy and severe visual loss: Early Treatment
Diabetic Retinopathy Study Report #18. Invest Ophthalmol Vis Sci. 1998;39(2):233–52.
178. Lyons TJ, Jenkins AJ, Zheng D, Lackland DT, McGee D, Garvey WT, et al. Diabetic
retinopathy and serum lipoprotein subclasses in the DCCT/EDIC cohort. Investig
Ophthalmol Vis Sci. 2004;45(3):910–8.
179. Miljanovic B, Glynn RJ, Nathan DM, Manson JE, Schaumberg DA. A prospective study
of serum lipids and risk of diabetic macular edema in type 1 diabetes. Diabetes.
2004;53(11):2883–92.
180. The Action of Control Cardiovascular Risk in Diabetes Study Group. Effects of Medical
Therapies on Retinopathy Progression in Type 2 Diabetes. N Engl J Med.2010;363:23344.
181. Sprecher H, Chen Q. Polyunsaturated fatty acid biosynthesis: a microsomal-peroxisomal
process. Prostaglandins Leukot Essent Fat Acids. 1999;60:317-21.
182. Lecomte M, Paget C, Ruggiero D, Wiernsperger N, Lagarde M. Docosahexaenoic acid is
a major n-3 polyunsaturated fatty acid in bovine retinal microvessels. J Neurochem.
1996;66(5):2160–7.
183. Fliesler SJ, Anderson RE. Chemistry and metabolism of lipids in the vertebrate retina.
prog. lipid. Res.1983;22:79-131.
184. Aveldaño MI. Phospholipid species containing long and very long polyenoic fatty acids
remain with rhodopsin after hexane extraction of photoreceptor membranes.
Biochemistry. 1988;27(4):1229–39.
185. Brush RS, Tran JTA, Henry KR, McClellan ME, Elliott MH, Mandal MNA. Retinal
sphingolipids and their very-long-chain fatty acid-containing species. Investig
	
  
	
  

53

Ophthalmol Vis Sci. 2010;51(9):4422–31.
186. Tikhonenko M, Lydic T a, Wang Y, Chen W, Opreanu M, Sochacki A, et al. , Remodeling
of retinal fatty acids in animals model of diabetes: a decrease in long-chain
polyunsaturated fatty acids is associated with a decrease in fatty acid elongases ELOVL2
and ELOVL4. Diabetes 2010;59(1):219–27.
187. Tikhonenko M, Lydic TA, Opreanu M, Li Calzi S, Bozack S, McSorley KM, et al. N-3
Polyunsaturated Fatty Acids Prevent Diabetic Retinopathy by Inhibition of Retinal
Vascular Damage and Enhanced Endothelial Progenitor Cell Reparative Function. PLoS
One. 2013;8(1):1–10.
188. Opreanu M, Lydic T a, Reid GE, McSorley KM, Esselman WJ, Busik J V. Inhibition of
cytokine signaling in human retinal endothelial cells through downregulation of
sphingomyelinases by docosahexaenoic acid. Invest Ophthalmol Vis Sci.
2010;51(6):3253–63.
189. Opreanu M, Tikhonenko M, Bozack S, Lydic TA, Reid GE, McSorley KM, et al. The
unconventional role of acid sphingomyelinase in regulation of retinal microangiopathy in
diabetic human and animal models. Diabetes. 2011;60(9):2370–8.
190. Zhang XM, Yang Z, Karan G, Hashimoto T, Baehr W, Yang XJ, et al. Elovl4 mRNA
distribution in the developing mouse retina and phylogenetic conservation of Elovl4
genes. Mol Vis [Internet]. 2003;9(April):301–7.
191. Lagali PS, Liu J, Ambasudhan R, Kakuk LE, Bernstein SL, Seigel GM, et al.
Evolutionarily conserved ELOVL4 gene expression in the vertebrate retina. Investig
Ophthalmol Vis Sci. 2003;44(7):2841–50.
192. Grayson C, Molday RS. Dominant negative mechanism underlies autosomal dominant
Stargardt-like macular dystrophy linked to mutations in ELOVL4. J Biol Chem.
2005;280(37):32521–30.
193. Agbaga M-P, Brush RS, Mandal MNA, Henry K, Elliott MH, Anderson RE. Role of
Stargardt-3 macular dystrophy protein (ELOVL4) in the biosynthesis of very long chain
fatty acids. Proc Natl Acad Sci U S A. 2008;105(35):12843–8.
194. McMahon A, Jackson SN, Woods AS, Kedzierski W. A stargardt disease-3 mutation in
the mouse Elovl4 gene causes retinal deficiency of C32-C36 acyl
phosphatidylcholines.FEBS lett. 2007;581(28):5459-63.
195. Harkewicz R, Du H, Tong Z, Alkuraya H, Bedell M, Sun W, et al. Essential role of
ELOVL4 protein in very long chain fatty acid synthesis and retinal function. J Biol Chem.
2012;287(14):11469–80.
196. Karan G, Lillo C, Yang Z, Cameron DJ, Locke KG, Zhao Y, et al. Lipofuscin
accumulation, abnormal electrophysiology, and photoreceptor degeneration in mutant
ELOVL4 transgenic mice: a model for macular degeneration. Proc Natl Acad Sci U
	
  
	
  

54

S A. 2005;102(11):4164–9.
197.

McMahon A, Butovich IA, Mata NL, Klein M, Ritter R, Richardson J, et al. Retinal
pathology and skin barrier defect in mice carrying a Stargardt disease-3 mutation in
elongase of very long chain fatty acids-4. Mol Vis. 2007;13(October 2006):258–72.

198. Li W, Sandhoff R, Kono M, Zerfas P, Hoffmann V, Ding BCH, et al. Depletion of
ceramides with very long chain fatty acids causes defective skin permeability barrier
function, and neonatal lethality in ELOVL4 deficient mice. Int J Biol Sci. 2007;3(2):120–
8.
199. McMahon A, Butovich I a, Kedzierski W. Epidermal expression of an Elovl4 transgene
rescues neonatal lethality of homozygous Stargardt disease-3 mice. J Lipid Res.
2011;52(6):1128–38.
200. Abcouwer SF, Lin C, Wolpert EB, Shanmugam S, Schaefer EW, Freeman WM, et al.
Effects of ischemic preconditioning and bevacizumab on apoptosis and vascular
permeability following retinal ischemia-reperfusion injury. Invest Ophthalmol Vis Sci.
2010 Nov;51(11):5920–33.
201. Grassmé H, Jekle A, Riehle A, Schwarz H, Berger J, Sandhoff K, et al. CD95 Signaling
via Ceramide-rich Membrane Rafts. J Biol Chem. 2001;276(23):20589–96.
202. Zeidan YH, Hannun YA. Activation of acid sphingomyelinase by protein kinase Cςmediated phosphorylation. J Biol Chem. 2007;282(15):11549–61.
203. Grassmé H, Riethmüller J, Gulbins E. Biological aspects of ceramide-enriched membrane
domains. Prog Lipid Res. 2007;46(3–4):161–70.
204. Kolesnick RN, Goñi FM, Alonso A. Compartmentalization of Ceramide Signaling:
Physical Foundations and Biological Effects. J Cell Physiol. 2000;184(3):285–300.
205. Grant MB, May WS, Caballero S, Brown G a J, Guthrie SM, Mames RN, et al. Adult
hematopoietic stem cells provide functional hemangioblast activity during retinal
neovascularization. Nat Med. 2002;8(6):607–12.
206. Kocher AA, Schuster MD, Szabolcs MJ, Takuma S, Burkhoff D, et al. Neovascularization
of ischemic myocardium by human bone - marrow – derived angioblasts prevents
cardiomyocyte apoptosis, reduces remodeling and improves cardiac function. Nat
Med.2001;7(4):430-6.
207. Kalka C, Masuda H, Takahashi T, Kalka-Moll WM, Silver M, Kearney M, et al.
Transplantation of ex vivo expanded endothelial progenitor cells for therapeutic
neovascularization. Proc Natl Acad Sci U S A. 2000;97(7):3422–7.
208. Hristov M, Erl W, Weber PC. Endothelial progenitor cells: Mobilization, differentiation,
and homing. Arterioscler Thromb Vasc Biol. 2003;23(7):1185–9.
	
  
	
  

55

209. Hirschi KK, Ingram DA, Yoder MC. Assessing identity, phenotype, and fate of
endothelial progenitor cells. Arterioscler Thromb Vasc Biol. 2008;28(9):1584–95.
210. Kamihata H, Matsubara H, Nishiue T, Fujiyama S, Tsutsumi Y, Ozono R, et al.
Implantation of Bone Marrow Mononuclear Cells Into Ischemic Myocardium Enhances
Collateral Perfusion and Regional Function via Side Supply of Angioblasts, Angiogenic
ligands, and Cytokines. Circulation. 2001;104(9):1046–52.
211. Kinnaird T, Stabile E, Burnett ES, Shou M, Lee CW, Barr S, Fuchs S, et al. Local
Delivery of Marrow-Derived Stromal Cells Augments Collateral Perfusion Through
Paracrine Mechanisms. Circulation. 2004;109(12):1543–9.
212. Tongers J, Roncalli JG, Losordo DW. Role of endothelial progenitor cells during
ischemia-induced vasculogenesis and collateral formation. Microvasc Res.
2010;79(3):200–6.
213. Méndez-Ferrer S, Lucas D, Battista M, Frenette PS. Haematopoietic stem cell release is
regulated by circadian oscillations. Nature. 2008;452(7186):442–7.
214. Katayama Y, Battista M, Kao WM, Hidalgo A, Peired AJ, Thomas SA, et al. Signals from
the sympathetic nervous system regulate hematopoietic stem cell egress from bone
marrow. Cell. 2006;124(2):407–21.
215. Schatteman GC, Hanlon HD, Jiao C, Dodds SG, Christy BA. Blood-derived angioblasts
accelerate blood-flow restoration in diabetic mice. J Clin Invest. 2000;106(4):571–8.
216. Tepper OM, Galiano RD, Capla JM, Kalka C, Gagne PJ, Jacobowitz GR, et al. Human
endothelial progenitor cells from type II diabetics exhibit impaired proliferation, adhesion,
and incorporation into vascular structures. Circulation. 2002;106(22):2781–6.
217. Busik J V, Tikhonenko M, Bhatwadekar A, Opreanu M, Yakubova N, Caballero S, et al.
Diabetic retinopathy is associated with bone marrow neuropathy and a depressed
peripheral clock. J Exp Med. 2009;206(13):2897–906.
218. Kusuyama T, Omura T, Nishiya D, Enomoto S, Matsumoto R, Takeuchi K, et al. Effects
of treatment for diabetes mellitus on circulating vascular progenitor cells. J Pharmacol Sci.
2006;102:96–102.
219. Fadini GP, Pucci L, Vanacore R, Baesso I, Penno G, Balbarini A, et al. Glucose tolerance
is negatively associated with circulating progenitor cell levels. Diabetologia.
2007;50(10):2156–63.
220. Loomans CJ, de Koning EJ, Staal FJ, Rookmaaker MB, Verseyden C, de Boer HC, et al.
Endothelial progenitor cell dysfunction: a novel concept in the pathogenesis of vascular
complications of type 1 diabetes. Diabetes. 2004;53(1):195–9.
	
  
	
  

56

221. Caballero S, Sengupta N, Afzal A, Chang KH, Li Calzi S, Guberski DL, et al. Ischemic
vascular damage can be repaired by healthy, but not diabetic, endothelial progenitor cells.
Diabetes. 2007;56(4):960–7.
222. Jarajapu YP, Thampi P, Caballero S, Boulton ME, Grant MB. Differential paracrine
dysfunction and hypoxic desensitization in diabetic CD34+ and CD14+ bone marrow
derived (BMDC). Circulation. 2009;120:S1124 .
223. Segal MS, Shah R, Afzal A, Perrault CM, Chang K, et al. Nitric oxide cytoskeletalinduced alterations reverse the endothelail progenitor cell migratory defect associated with
diabetes. Diabetes.2006;55(1):102-9.
224. Aicher A, Zeiher AM, Dimmeler S. Mobilizing endothelial progenitor cells. Hypertension.
2005;45(3):321–5.
225. Thum T, Fraccarollo D, Schultheiss M, Froese S, Galuppo P, Widder JD, et al.
Endothelial nitric oxide synthase uncoupling impairs endothelial progenitor cell
mobilization and function in diabetes. Diabetes. 2007;56(3):666–74.
226. Vasa M, Breitschopf K, Zeiher AM, Dimmeler S. Nitric oxide activates telomerase and
delays endothelial cell senescence. Circ Res. 2000;87(7):540–2.
227. Chakravarthy H, Navitskaya S, O'Reilly S, Gallimore J, Mize H, et al. Role of Acid
Sphingomyelinase in Shifting the Balance Between Proinflammatory and Reparative Bone
Marrow Cells in Diabetic Retinopathy. Stem cells. 2016;34(4):972-83.

	
  
	
  

57

Chapter 2. Overexpression of ELOVL4 stabilizes tight junctions and prevents VEGFinduced vascular permeability in diabetic retina and retinal cells
Authors: Nermin Kady, Xuwen Liu, Todd Lydic, Meesum Syed, Svetlana Navitskaya, Qi Wang,
Sandra Hammer, Sandra O’Reilly, Chao Huang, Sergey S. Seregin, Andrea Amalfitano, Vince A.
Chiodo, Sanford L. Boye, William W. Hauswirth, David A. Antonetti and Julia V. Busik
2.1 Abstract
Tight junctions contribute to the control of permeability in the blood brain and blood
retinal barrier and loss of the junctional complex contributes to the pathology of a host of
diseases of these neural tissues including brain tumors and dementia and blinding eye diseases
such as diabetic retinopathy. Tight junctions involve close apposition of transmembrane proteins
between cells. While, tight junction proteins have been studied in detail, the role of lipids in tight
junction structure and permeability has received much less attention. In this study, using diabetes
induced loss of blood-retinal barrier as a model, we addressed the role of very long chain (VLC
≥26) ceramides in tight junction integrity and function. VLC fatty acids that incorporate into
VLC ceramides are produced by elongation of very long chain fatty acids 4 (ELOVL4).
ELOVL4 is significantly reduced in the diabetic retina leading to decreased levels of VLC fatty
acids. Overexpression of ELOVL4 in bovine retinal endothelial cells (BRECs) significantly
decreased basal permeability, inhibited VEGF and IL-1β induced permeability and
simultaneously prevented VEGF-induced decrease in occludin expression and border staining of
tight junction proteins; ZO-1 and claudin-5. Knockdown of ELOVL4 was associated with
increased paracellular permeability. Moreover, intravitreal delivery of hELOVL4-AAV2 reduced
diabetes-induced increase in vascular permeability in mice. Importantly, ultrastructure analysis

	
  
	
  

58

reveals that ceramides co-localize with tight junction complexes. Lipidomic analysis of tight
junction isolates revealed the presence of omega-linked acyl-VLC ceramides. These very
unusual structures are known to be important for the maintenance of skin water permeability
barrier, but they have never been shown in tight junctions. The data presented here demonstrate
that ELOVL4 is required for maintaining healthy blood retinal barrier and normalization of
retinal ELOVL4 expression could prevent blood-retinal barrier dysregulation in DR through
increase in VLC ceramides and stabilization of tight junctions.

	
  
	
  

59

2.2 Introduction
Diabetic Retinopathy (DR) remains a major cause of loss of vision despite recent advances in
therapeutic options. Loss of the blood-retinal barrier (BRB) and increased vascular permeability
marks early stages of DR and contributes to disease progression. Several factors have been
shown to contribute to increased vascular permeability. Prominent among them are diabetesinduced increase in the level of vascular endothelial growth factor (VEGF) (1–3), interleukin-1
beta (IL-1β) (4,5), tumor necrosis factor α (TNF α) (5), extracellular proteases; matrix
metalloproteases 2 and 9 (6–8), and pericyte loss (9–11). Importantly, antibodies targeting VEGF
demonstrate prevention of visual loss and even restoration of vision in many patients with
diabetic macular edema (12). Hyperglycemia contributes an important role in the expression of
these pro-inflammatory and pro-angiogenic factors (11,13–16). However, the etiology of DR
progression remain incompletely understood and recent assessment of the Diabetes
Complications and Control Trial suggest that only 11% of blindness may be accounted for by
hyperglycemia (17).

Several recent clinical trials, including the ACCORD Eye Trial, have demonstrated that in
addition to hyperglycemia, diabetic dyslipidemia may contribute a critical role in the
development of DR (18). In agreement with the results of clinical trials, recent studies of diabetic
animal models have demonstrated that altered retinal lipid metabolism and dyslipidemia leads to
retinal inflammation and vascular degeneration (19–21). In particular, our laboratory identified a
marked decrease in the expression of several retinal fatty acid elongases. Elongation is a
complex reaction involving four enzymatic steps that result in addition of	
   two	
   carbons to the
carboxyl end of fatty acids. Elongases perform the first rate-limiting condensation step in the

	
  
	
  

60

fatty acid elongation cycle in a substrate specific manner. Seven elongation of very long chain
fatty acids (ELOVL) 1-7 have been identified in mammals with each elongase exhibiting a
characteristic substrate specificity and tissue distribution (22,23). Elongases are highly expressed
in normal retina where they actively participate in de novo lipogenesis, as well as mono and
polyunsaturated fatty acid (PUFA) synthesis. Downregulation of elongases in diabetic retina
results in decreased long-chain-to-short-chain fatty acids ratio. Moreover, synthesis of the major
n3 PUFA that accumulate in the retina, docosahexaenoic acid (DHA; 22:6, n3), requires two
extra elongation steps compared to the n6 PUFA, arachidonic acid (20:4,n6), thus decrease in
elongation contributes to decreased n3-to-n6 PUFA ratio in the retina. This decrease in n3-to-n6
ratio was shown to be associated with increased expression of adhesion molecules, proinflammatory cytokines, and retinal vascular pathology (19).

ELOVL4 is the highest expressed elongase in the retina (24). ELOVL4 elongates extremely long
fatty acids ≥C24 and its expression is restricted to the tissues with extremely long fatty acids
such as the retina (24–27), thymus and skin (26), and at a lower level in the brain and testis(27).
Fatty acids with ≥C24 chain are used as precursors by ELOVL4 for synthesis of ≥ C26 very long
chain PUFA (VLCPUFA) and saturated VLCFAs (23,28–30). Saturated VLCFAs are primarily
incorporated into ceramides (Cer) and glucosylceramides (GlcCer) that are known to be highly
expressed in epidermis (22,23). These ceramides with extremely long FA are major lipid
components of stratum corneum and they have been shown to be essential in maintenance of the
water permeability barrier in skin (22,31,32). Genetic studies with ELOVL4 mutant mice
revealed that ELOVl4 plays an essential role in skin barrier function as these mice exhibit skin

	
  
	
  

61

barrier defect and neonatal death phenotype (33–35). Besides the role of ELOVL4 in the skin,
ELOVL4 dysfunction has been associated with Stargardt-like macular dystrophy (STGD3) (36–
38). Interestingly, ELOLV4 is one of the enzymes that was found to be downregulated in the
retina by both diabetes (19) and retinal ischemia-reperfusion (39).

We hypothesized that VLC ceramides support BRB function in an ELOVL4 dependent manner.
This study was designed to test the role of ELOVL4 and VLC ceramides in diabetes-induced
increase in retinal vascular permeability and to identify if lipids, in particular VLC ceramides,
localize in the tight junction and play a role in controlling the paracellular permeability of retinal
endothelial cells.

	
  
	
  

62

2.3 Materials and methods
2.3.1 Cell culture
Bovine retinal endothelial cells (BREC) were isolated and grown in culture as previously
described (2) and cultured on 1 ug/cm2 fibronectin coated plates (Sigma-Aldrich) in MCDB-131
(Sigma-Aldrich) growth media supplemented with 10% FCS (Hyclone, Logan, UT), 10 ng/mL
epidermal growth factor (Sigma-Aldrich), 0.2 mg/mL endothelial cell growth medium additive
(EndoGro; Vec Technologies, Rensselaer, NY), 0.09 mg/mL heparin (Fisher Scientific,
Pittsburgh,

PA),

0.01

mL/mL

antibiotic-antimycotic

(GIBCO

100x,

Invitrogen-Life

Technologies, Carlsbad, CA), and 8 ug/ml tylosin (Sigma-Aldrich), and maintained at 37°C in
95% air and 5% CO2 in a humidified cell culture incubator. All experiments were performed
with cells at passage 4-6. For biochemical purification of tight junction, Human retinal pigment
epithelial cell line ARPE-19 (HRPE) cells were grown in Dulbecco’s modified Eagle’s
medium/F12 (1:1 ratio, 5 mmol/l glucose) supplemented with 10% fetal calf serum, 1%
penicillin/streptomycin, at 37°C in humidified 95% air and 5% CO2.
2.3.2 Virus-mediated hELOVL4 overexpression
Human ELOVL4 was expressed using an E-1 and E-3 deleted adenoviral vector system
containing the human ELOVL4 cDNA (AdhELOVL4) under control of the CMV promoter.
Where subconfluent BRECs were transduced with either AdhELOVL4 or AdEmpty as a
negative control at a multiplicity of infection (MOI) 20,000 overnight in MCDB media
supplemented with 2 % serum. Successful overexpression was confirmed by quantitative realtime PCR (qPCR) and immunoblotting for hELOVL4 at the time of permeability studies (60 h
post viral transduction).
For in vivo experiments, human ELOVL4 was

	
  
	
  

engineered

63

into

adeno-associated

virus

serotype 2 vectors containing four capsid tyrosine to phenylalanine (Y – F) mutations (AAV2
mut quad) (40) under control of the ubiquitous truncated chimeric CMV-chicken β-actin
(smCBA) promoter. Empty AAV2 mut quad construct was used as a negative control. The virus
was produced at the Department of Ophthalmology, University of Florida, Gainesville, FL.
2.3.3 siRNA transfection
BREC in culture were resuspended in nucleofector solution (Amaxa Biosystems, Gaithersburg,
MD) to a final concentration 5x105 cells/100 ul. Cell suspension was mixed with 100nM control
or ELOVL4 StealthTM siRNA from Invitrogen Life Technologies (Carlsbad, CA, USA) into the
electroporation cuvette, and were electroporated

(Nucleofector program S-005; Amaxa

Biosystems) and maintained in culture as above. After either 48h or 72 h, cells were collected for
total mRNA and protein expression, respectively.
2.3.4 Permeability assay in vitro
Ad-Empty and Ad-hELOVL4 BRECs were grown to confluence on fibronectin-coated, 0.4-um
pores transwell filters (Transwell; Corning Costar, Acton, MA, USA) at 75 x 103 cells/well,
pretreated with 1% FCS, 0.01 ml/ml antibiotic/antimycotic and 100 nM hydrocortisone for 36 h.
Apical and basolateral chambers were treated with recombinant human VEGF

165

(50 ng/ml for

30 min), or IL-1β (10ng/ml for 15 min) where indicated. After treatments, 10 uM of 70 kDa
RITC-dextran was added to the apical chamber. BREC paracellualr permeability to 70 KDa
rhodamine isothiocyanate dextran (Sigma-Aldrich) was measured as previously described (2,3).
Aliquots were collected from the basolateral chamber every 30 min for 4 h and placed in a 96well black/clear bottom plate. At the last time point, aliquots were collected from the apical

	
  
	
  

64

chamber and placed in the 96-well plate. The RITC-dextran fluorescence was quantified with
fluorescence imager (FluorImager 560/590; TECAN infinite 1000), and the rate of diffusive
permeability (Po) was calculated by the following formula:
Po= [(FA / Δt) VA]/(FLA)
Where, Po = diffusive flux (cm / s); FA is basolateral fluorescence; FL is apical fluorescence; Δt
is change in time; A is the surface area of the filter (in cm2); and VA is the volume of the
basolateral chamber (in cm3).
The average Po for control conditions (AdEmpty) used in all experiments ranged from 2.5 to 2.8
x 10-6 cm/s.
2.3.5 Western blotting
Protein extraction and Western blot was performed using NuPAGE system (Invitrogen) as
described previously (3,21). Briefly, AdEmpty and AdhELOVL4 BRECs were lysed in TritonX-100-deoxycolate-SDS-based lysis buffer. Protease inhibitor cocktail tablet (Roche,
Indianapolis, IN, USA) and phosphatase inhibitor (1 mM Na4PPi, 10 mM NaF, 100 µM
glycerophosphate, 1 mM Na3PO4) were freshly added. Protein concentration was determined by
the BioRad DC protein colorimetric assay (BioRad, Hercules, CA, USA). 30 µg of protein were
loaded on 4-12 % NuPAGE Bis -Tris gradient gels (Life Technologies, Carlsbad, CA) and ran at
200 V for 1 h, followed by transfer to nitrocellulose membrane, and immunoblotting with the
following primary antibodies; rabbit polyclonal anti-ELOVl4 (Sigma, Cat. No. SAB4502884),
rat monoclonal anti- ZO-1 (Millipore, Cat. No. MABT11), mouse monoclonal anti-occludin
(Invitrogen, Cat. No. 33-1500) and rabbit polyclonal anti-claudin-5 (Invitrogen, Cat. No. 34-

	
  
	
  

65

1600), all used in dilution 1:1000. As control for protein loading, membranes were blotted with
either mouse monoclonal anti- α tubulin (Sigma, Cat. No. T5168) or mouse monoclonal anti-β
actin antibody (Cell Signaling, Cat. No. 4967) (1:4000). Primary antibodies were detected by
horseradish peroxidase –conjugated or infrared secondary antibody (IRDye; LI-COR, USA).
Immunoreactive bands were detected with either chemiluminescence with horseradish
peroxidase substrate Lumigen TMA-6 (Lumigen, Southfield, MI) and FluorChem E system (Cell
Biosciences, Santa Clara, CA) or Odyssey scanning system (LI-COR Odyssey Imaging System,
USA).
2.3.6 Quantitative real-time PCR
Total RNA was extracted from BRECs or mouse retinas using Qiagen RNeasy (Qiagen Inc.,
Valencia, CA, USA) or mirVana isolation kit (Life Technologies, Carlsbad, CA) according to the
manufacturer’s instructions. NanoDrop 2000 (Thermo Scientific, IL, USA) was used to
determine total RNA concentration. Total RNA was reverse transcribed into cDNA using
superscript III first-strand synthesis system (Invitrogen, Carlsbad, CA). Human and mouse genespecific primers for ELOVL4 were used. Expression levels were normalized to either bovine
GAPDH or mouse cyclophilin. A no reverse transcriptase control was done as a negative control.
Human ELOVL4: GCACTCAACGACACGGTAGA and GGACCCAGCCACACAAAC
Bovine GAPDH: CATTGACCTTCACT CATGGT and ACCCTTCAAGTGAGCCCCAG
Mouse ELOVL4: CCATAGGCCACCACGATT and TCAAAGGCTACCTTCCTCATTAC
Mouse Cyclophilin: ATTCATGTGCCAGGGTGGTGA and CCGTTTGTGGGTCCAGCA

	
  
	
  

66

2.3.7 Immunocytochemistry
Cellular localization of the tight junction complexes in BRECs and mouse retina was evaluated
by immunocytochemistry as described previously (5). Ad-Empty and Ad-hELOVL4 transduced
cells were seeded on 12-mm diameter polystyrene coverslips (Nunc, Naperville, IL) in a 24 well
tissue culture plate at density of 8 x 104 cells in 1 ml MCDB-131complete media/ well for 1day.
For experiments the cells were pre-treated with 1% FBS, 0.01 ml/ml antibiotic/antimycotic and
100 nM hydrocortisone for 36 h and then treated with VEGF (50 ng/ml) for 30 min. After
treatment cells were washed twice with DPBS (with Ca2+, Invitrogen), fixed with 1%
paraformaldehyde for 10 min at room temperature, washed with TBS, permeabilized with 0.2%
TritonX-100/ TBS for 10 min at room temperature and blocked with 10 % goat serum in 0.1%
triton X -100/TBS for 1 h at room temperature. Cells were incubated for 2 days at 4 °C with the
following primary antibodies in the blocking solution (2% goat serum in 0.1% TritonX100/TBS): rat monoclonal anti-ZO-1 (Millipore, Cat. No. MABT11), mouse monoclonal anticeramide (Sigma, Cat. No. C8104) or rabbit polyclonal anti-claudin-5 (Invitrogen, Cat. No. 341600) used in concentration 1:300. After 3 washes with TBS, primary antibodies were detected
with Alexa Fluor 647 goat anti-rat IgG (650/668), Alexa Fluor 488 goat anti-mouse F (ab’) 2 IgG
(496/519) or Alexa Fluor 594 goat anti-rabbit IgG (590/617) (Molecular Probes, Invitrogen). The
nuclei were stained with Hoechst (361/497) during secondary antibody incubation. Coverslips
were mounted on slides using ProLong® Gold Antifade Mountan (Life Technologies,
CARLSBAD, CA). XY images were sequentially scanned using confocal laser scanning
microscope Nikon A1 Plus (Nikon Instruments, Inc., Melville, NY) with an Apo 60x/ NA 1.4,
oil immersion objective. The near-IR, green, red and blue channels were acquired sequentially
using 647-nm, 488.5-nm, 560.9-nm and 402.6-

	
  
	
  

nm excitation, respectively. Near-IR channel

67

was acquired using 660-nm long pass emission filter while 500-550 band pass, 575-625 band
pass and 428-463 band pass emission filters were used for green, red and blue channels,
respectively.
Tight junction protein staining at the cell border was quantified by semi-quantitative ranking
score system as previously described (5). With scale 1-5: 1 for 0% disruption of cell border
staining where borders are straight and clear, 2 for 25% disruption with slight border disruption,
3 for 50% disruption with about 50% or more of cells border show disruption at many points, 4
for 75% disruption where cell shape is disrupted, 5 for 100% disruption where no clear cell
shape and complete loss of staining between cell borders. Two observers examined ranking score
in a blinded manner. For each condition, four images were examined. For each image four fields
were analyzed and ranking scores were calculated. One-way ANOVA was used to determine the
statistical difference between different conditions.
For co-localization, confocal images were analyzed using Nikon NIS Elements software to
determine Pearson’s coefficient of co-localization.
To evaluate TJ protein organization in retinal vasculature, whole retina flat mounts were
immunolabeled for occludin using mouse monoclonal anti-occludin Alexa Fluor 488 conjugate
(1:75) (Thermo Fisher Scientific, Cat. No. OC-3F10) for 3 days at 4 °C. Following incubation,
retinas were washed and flat mounted on slides with Fluoromount mounting medium (Sigma
Aldrich). Images were acquired using Nikon C2 (Nikon Instruments, Inc., Melville, NY).

	
  
	
  

68

2.3.8 Immunogold electron microscopy
Bovine retinas were fixed with 4% paraformaldehyde/0.5% glutaraldehyde for 24 h at 4 °C, cut
into 5x5 mm pieces and analyzed using osmium tetroxide (41) or double immunogold labeling
(42) as previously described. Retinal pieces were washed with PBS, dehydrated in a graded
series of ethanol then infiltrated on LR-white resin and polymerized at 60 °C for 24 h in gelatin
capsule. Ultrathin sections (70nm) were obtained by PowerTome Ultramicrotome (RMC
Boeckeler Instruments, Tucson, AZ) and collected on formvar coated gold grids. Grids were
incubated in blocking solution (0.01M PBS + 5 %BSA + 5 % normal goat serum) for 10-30
min., then rinsed 3 times. Simultaneous double immunogold labeling was done using the
following primary antibodies; rabbit polyclonal anti-occludin (Invitrogen, Cat. No. 71-1500) and
mouse monoclonal anti-ceramide (Sigma, Cat. No. C8104) used in dilution 1:1. Primary
antibodies were detected with mixture of goat anti-rabbit IgG-15 nm gold conjugate and goat
anti-mouse IgG-6 nm gold conjugate diluted (1:20). Finally, grids were stained with uranyl
acetate and lead citrate and images were examined with JEOL 100 CX TEM (Japan Electron
Optical laboratory, Tokyo, Japan) operated at 100 KV and equipped with Gatan ORIUS CCD
camera.
2.3.9 TJ isolation and mass spectrometry
In order to determine molecular species of ceramides associated with the tight junction
complexes, tight junctions were isolated as previously described (43). Briefly, ARPE-19 cells
(15 x 150 mm culture plates) were lysed with hypo-osmotic buffer, collected, homogenized and
centrifuged at 100,000 xg for 60 minutes at 4°C against a saturated sucrose cushion. The
collected membrane was subsequently cryo	
  
	
  

lysed using liquid nitrogen, then incubated

69

with either rabbit polyclonal anti-PKC zeta (Sigma, Cat. No. P 0713) or rat monoclonal anti-ZO1 ((Millipore, Cat. No. MABT11) using protein A or G dynabeads, respectively according to
manufacturer’s instructions (Invitrogen, USA). Tight junction complexes bound to the beads
were subjected to lipid extraction as previously described (44). Lipids obtained from tight
junctions were analyzed by high resolution/accurate mass spectrometry and tandem mass
spectrometry in positive and negative ionization modes, using an LTQ-Orbitrap Velos mass
spectrometer (Thermo Scientific) operating at a resolving power of 100,000 defined at m/z 400.
Sphingolipids were identified as previously described in both positive ion mode as [M+H]+ and
[M-H2O+H]+ molecular ions (45) and in negative ion mode as [M+HCOO-H]- ions (46).
Ceramides of interest were further examined by higher-energy collisional dissociation MS/MS at
100,000 resolving power to confirm sphingolipid backbone and head group moieties. Mouse
epidermal lipid extracts were similarly prepared and analyzed for optimization of analytical
conditions used for detecting putative very long chain fatty acid-containing sphingolipids.
2.3.10 Animal model
All procedures involving animal models were performed according to the Association for
Research in Vision and Ophthalmology (ARVO) Statement for the Use of Animals in
Ophthalmic and Vision Research, and were approved and monitored by Institutional Animal
Care and Use Committee at MSU. Streptozotocin (STZ)-induced male C57BL/6J mice were
used as a model of type I of diabetes to access retinal vascular permeability and tight junction
protein organization. 13-week-old Male C57BL/6J mice (Jackson Laboratory) were made
diabetic by intraperitoneal injections of STZ (Sigma Aldrich) (65 mg/ kg) for five consecutive
days. Age matched nondiabetic control mice

	
  
	
  

received an equivalent amount of vehicle

70

(citrate buffer, pH 4.5). Diabetes was confirmed by blood glucose higher than 13.8 mmol/l (300
mg/dl) and decrease in the body weight gain. After two weeks of confirmed diabetes, animals
received intravitreal injection of either AAV2-hELOVL4 quad mutant (right eye) or AAV2Empty vector (left eye) as a control (2 x 109 viral genomes/eye in 1 ul). Animals were sacrified
6-8 weeks following viral injection (8-10 week of STZ induction) for retinal vascular
permeability and retinal analysis.
2.3.11 Retinal vascular permeability
Two months after induction of diabetes, retinal vascular permeability was assessed as initial
retinal vascular damage in DR. Briefly; Mice were injected with FITC-albumin (0.5 mg in 50 µl
PBS) (Sigma Aldrich). After 2 h, blood was collected, mice were perfused with 1%
formaldehyde and fixed eyes were enucleated. Obtained retinas were flat mounted with 4 radial
cuts and placed on glass slides with Fluoromount mounting medium (Sigma Aldrich). Images
were obtained by using confocal laser scanning microscope Nikon C2 (Nikon Instruments, Inc.,
Melville, NY) with 20x / NA 0.75 objective. Green channel was acquired using 488 nm
excitation wavelength and 500-550 band pass emission filters.
For each retina, at least 5 different fields were randomly acquired, measured using Angiogenesis
tube formation application-MetaMorph imaging software (Molecular Devices, Downingtown,
PA) by subtracting residual fluorescence inside the vessels from total fluorescence to determine
leakage fluorescence intensity. The fluorescence intensity of the leakage was averaged between
at least 5 fields for each retina and normalized to blood plasma fluorescence level.

	
  
	
  

71

2.3.12 Immunohistochemistry
Retinas were processed as previously described (47). Briefly, samples were fixed in zinc fixative,
paraffin embedded and 4-5 micron sections were obtained with Reichert Jung 2030 rotary
microtome. Sections were dried at a 56°C slide incubator for 2–24 h, deparaffinized and rinsed in
several changes of distilled water followed by Tris buffered saline pH 7.4. Sections were blocked
with 10% normal goat serum (Life Technologies, Carlsbad, CA) in 0.3%triton x-100 for 1 h at
room temperature and immunolabeled with Isolectin GS-IB4 with Alexa Fluor 594 conjugate
(1:50) (Molecular Probes, Invitrogen) and rabbit polyclonal anti-ELOVl4 (1:50) (Sigma) in
blocking solution, overnight at 4°C. ELOVL4 primary antibody was detected by Alexa Fluor 488
goat anti-rabbit IgG (496/519) (Molecular Probes, Invitrogen). Images were captured using
Nikon TE2000 fluorescence microscope equipped with Photometrics CoolSNAP HQ2 camera.
For each slide, 3 different fields were collected. Co-localization of ELOVL4 (green) with retinal
vasculature (red) was determined to evaluate transduction efficiency of AAV2 vector. Percentage
of co-localization (yellow) was quantified using MetaMorph imaging software (Molecular
Devices, Downingtown, PA).
2.3.13 Statistical analysis Data are expressed as mean ± S.E.M. Results were analyzed for
statistical significance by the Student’s t-test or one-way analysis of variance (ANOVA)
followed by post hoc Tukey test. (GraphPad Prism 7, GraphPad Software, San Diego, CA) was
used for all statistical analysis. Significance was established at P < 0.05.

	
  
	
  

72

2.4 Results
2.4.1 Expression level of ELOVL4 controls retinal endothelial permeability
The effect of ELOVL4 on retinal vascular endothelial cell permeability was evaluated in
monolayer of BRECs, a model of inner blood retinal barrier, after overexpression through the use
of adenovirus (AdhELOVL4) or short interfering RNA (siRNA) to inhibit expression of
ELOVL4.

To ensure efficient overexpression and inhibition of hELOVL4, expression of

ELOVL4 in BRECs was evaluated using qPCR and western blot. Human ELOVL4 was
significantly expressed in adhELOVL4-transduced BRECs compared with AdEmpty-transduced
cells (Fig. 9A and B), with human specific qPCR primers revealing expression only after
transduction of the human exogenous gene (9A) while Western blot reveals a 50% increase in
steady state ELOVL4 over endogenous protein (9B). To evaluate the effect of hELOVL4 on
endothelial permeability under basal conditions, paracellular permeability to 70 kDa RITCdextran was measured in AdhELOVL4 and AdEmpty transduced cells. Data from these
experiments indicate that under basal conditions, overexpression of hELOVL4 results in a
significant decrease in permeability to 70kDa RITC dextran (53.7 ± 5.7 % of control, P=0.0016;
Fig. 9C). ELOVL4 was knocked down using siRNA. ELOVL4 mRNA and protein levels were
significantly reduced in BRECs transfected with ELOVL4-siRNA in comparison to cells
transfected with control siRNA (Fig 9D and E). Moreover, knockdown of ELOVL4 was
associated with increased paracellualr permeability (528.2 ± 242.2 % of control, P= 0.0317; Fig
9F).

	
  
	
  

73

ELOVL4 overxpression

AdEmpty
AdhELOVL4
C

0.004
AdEmpty

0.003
0.002
0.001

ND

0.000

AdhELOVL4

ELOVL4

37.7

tubulin

55

0.020
0.015
0.010
0.005
0.000

Relative cell permeability
( % of control )

B

ELOVL4 /alpha tubulin

Relative expression of

hELOVL4 mRNA / GAPDH

A

**

150
**

100
50
0

Control-siRNA

ELOVL4 inhibition

ELOVL4-siRNA

0.00020

F

Control
siRNA

0.00010
0.00005

ELOVL4

0.00000

actin

-actin

****

0.00015

ELOVL4
siRNA
37.7
45

Relative cell permeability
( % of control )

E

ELOVL4 /

Relative expression of

Bovine ELOVL4/ GAPDH

D

**

1.0
0.8
0.6
0.4
0.2
0.0

*

800
600
400
200
0

Figure 9. Expression level of ELOVL4 controls retinal endothelial permeability.
IL-1

VEGF

(A-B) BRECs were transduced with either AdEmpty (white bars) or AdhELOVL4 (black bars)
G

H

200

***

***

AdEmpty

200

Relative cell permeability
( % of control )

Relative cell permeability
( % of control )

and collected 60 h after adenovirus transduction for (A) RNA analysis by RT qPCR or (B)
*

**

AdEmpty

AdEmpty+VEGF
protein by Western blot to confirm
adenovirus-mediated
overexpression of hELOVL4.
Human
AdEmpty+IL-1
150
150

AdhELOVL4+VEGF

AdhELOVL4+IL-1

ELOVL4
mRNA expression was normalized to100
GAPDH (A) and protein expression was
100
50
normalized
to α-tubulin loading control (B).

0

0

	
  
	
  

50

74

Figure 9 (cont’d)
(C) Paracellular permeability to 70kDa RITC-dextran was determined in BRECs transduced with
either AdEmpty (white bar) or AdhELOVL4 (black bar) in untreated cells. (D-F) BRECs were
transfected with either control siRNA (white bars) or ElOVL4-siRNA (black bars) and collected
for analysis. Expression profile of ELOVL4 was determined by either (D) RT qPCR, 48 h after
transfection or (E) Western blot, 72 h after transfection to confirm silencing of ELOVL4. Bovine
ELOVL4 mRNA expression was normalized to GAPDH (D) and protein expression was
normalized to β-actin loading control (E). (F) Paracellular permeability to 70kDa RITC-dextran
was determined in BRECs transfected with either control siRNA (white bars) or ElOVL4-siRNA
(black bars). Results are shown as mean ± SEM, n =3-6. *p< 0.05, **p<0.01, ***p<0.001,
****p<0.0001.

	
  
	
  

75

2.4.2 ELOVL4 overexpression prevents VEGF- and IL-1 β- induced increase in
permeability
The effectiveness of anti-VEGF therapy for treating vision loss associated with diabetic macular
edema (12, 48–52) and growing evidence for a contribution of chronic inflammation with
elevated levels of pro-inflammatory factors, including interleukin-1β (IL-1β) (53–57) provide
compelling evidence these cytokines lead to increased BRB permeability during early stages of
DR. Our next goal was to evaluate the effect of overexpression of ELOVL4 on VEGF- and IL1β -induced endothelial cell permeability. AdhELOVL4 and AdEmpty transduced BRECs were
grown to confluence as mentioned above and treated with VEGF (50 ng/ml, 30 min) or IL-1β
(10 ng/ml, 15 min). Paracellular permeability to 70 kDa RITC-dextran was measured. VEGF
significantly increased cell permeability (169.4 ± 13.4% of control; Fig. 10A). Similarly, IL-1β
increased cell permeability (152.8 ± 4.9 % of control, Fig. 10B). Human ELOVL4
overexpression prevented VEGF- and IL-1β -induced increase in permeability (88.5 ± 8.6 % of
control; Fig. 10A) and (97.5 ± 9.3% of control; Fig. 10B), respectively.

	
  
	
  

76

IL-1

VEGF

B

200

***

150

***

AdEmpty
AdEmpty+VEGF
AdhELOVL4+VEGF

100
50

200

Relative cell permeability
( % of control )

Relative cell permeability
( % of control )

A

150
100

*

**

AdEmpty
AdEmpty+IL-1
AdhELOVL4+IL-1

50
0

0

Figure 10. ELOVL4 overexpression block cytokine-induced retinal endothelial cell
permeability.
(A-B) Paracellular permeability to 70kDa RITC-dextran was determined in BRECs transduced
with either AdEmpty (white bar) or AdhELOVL4 (black bar), and either treated with VEGF (50
ng/ml) for 30 min (gray and black bars) (A); or treated with IL-1β (10 ng/ml) for 15 min (gray
and black bars) (B). Results are shown as mean ± SEM, n =3-6. *p< 0.05, **p<0.01,
***p<0.001.

	
  
	
  

77

2.4.3 ELOVL4 overexpression increases the barrier properties of tight junction complex
To explore the mechanism by which overexpression of ELOVL4 reduces retinal vascular
permeability, we examined the effect of ELOVL4 on TJ protein content and localization. Under
basal condition, overexpression of hELOVL4 significantly increased occludin content in
AdhELOVL4 transduced BRECs compared to AdEmpty transduced cells and prevented the
VEGF – induced decrease in occludin (Fig. 11A). For ZO-1 and Claudin-5 no significant effects
of hELOVL4 on total protein expression were observed (Fig. 11 B and C). Although hELOVL4
had no effect on ZO-1 and Claudin-5 total protein levels, immunostaining revealed that
overexpression of hELOVL4 both increased protein staining at cell border, and prevented
VEGF-induced disruption of continuous immunostaining of ZO-1 and claudin-5 at cell border
(Fig. 11 D and E). These data suggest that overexpression of ELOVL4 may stabilize the TJ
complex and suppress the VEGF-induced disruption of the complex.

.

	
  
	
  

78

A

AdEmpty
Basal

VEGF

AdhELOVL4

B

Basal VEGF

AdEmpty

AdhELOVL4

Basal VEGF

Basal VEGF

C

Occludin

Zo-1

Claudin-5

actin

actin

actin

*

0.2
0.1
0.0

Claudin-5 / -actin

0.3

1.0

*
0.06

Zo-1 / -actin

Occludin / -actin

0.4

AdhELOVL4+VEGF

0.08

**

AdhELOVL4
Basal VEGF

AdhELOVL4

AdEmpty+VEGF

AdEmpty

AdEmpty
Basal VEGF

0.04
0.02

*

0.8
0.6
0.4
0.2

0.00

0.0

Figure 11. ELOVL4 overexpression increases the barrier properties of the tight junction
complex.
(A-C) Confluent AdEmpty (white bars) and AdhELOVL4 (light gray bars) transduced BRECs
were treated with VEGF (dark gray and black bars) (50 ng/ml) for 30 min. Total cell lysates were
immunoblotted for (A) Occludin, (B) ZO-1, and (C) Claudin-5. β-actin served as loading control.
Representative Western blots from three independent experiments are shown on top, with
densitometry quantification below. Results are shown as mean ± SEM (n=3). *p< 0.05,
**p<0.01, ***p<0.001.

	
  
	
  

79

Figure 11 (cont’d)
D

AdEmpty
+ VEGF

AdhELOVL4

AdhELOVL4
+ VEGF

ZO-1

AdEmpty

Overlay

Ceramide

Cell border disruption
Frequency
(% Total)

Zo-1

E

AdEmpty
+ VEGF

AdhELOVL4

***

100.0%
80.0%

100%
75%

60.0%

50%
25%

40.0%

0%

20.0%
0.0%

AdEmpty

AdEmpty +VEGF

AdhELOVL4

AdhELOVL4 + VEGF

AdhELOVL4
+ VEGF

Claudin-5

AdEmpty

*

120.0%

Overlay

Ceramide

Cell border disruption
Frequency
(% Total)

Claudin-5
**

120%

***

100%
80%

100%

75%
60%

50%
25%

40%

0%

20%
0%

AdEmpty

AdEmpty +VEGF

AdhELOVL4 AdhELOVL4 + VEGF

(D-E) Confluent monolayers of BRECs treated as in A-C were immunolabeled for (D) ZO-1 (Far
red) and Ceramide (green), or (E) Claudin-5 (red) and ceramide (green) and confocal images
were taken. The results are from 2 independent experiments, for each condition, four images
were examined. For each image four fields were analyzed. Scale bar, 10 µm. Quantification of
each protein staining at the cell border, is shown on the far right, the results represent frequency
of each score as described in methodology. *p< 0.05, **p<0.01, ***p<0.001.

	
  
	
  

80

2.4.4 Ceramide co-localizes with tight junction complex
Changes in tight junction protein content and localization after ELOVL4 overexpression
prompted us to determine if hELOVL4-produced lipids co-localize with TJ complexes. As
mentioned above ELOVL4 is responsible for production of VLC poly, mono unsaturated and
saturated fatty acids. In the skin VLC fatty acids incorporated into VLC ceramides play an
important role in the maintenance of barrier function (22,31). We tested the hypothesis that VLC
ceramides incorporated into TJ complexes in the retinal endothelial cells. Immunocytochemistry
for TJ proteins and ceramide was performed in BRECs. Interestingly, ceramide (green) was
found to co-localize with the tight junction complexes and follow the same distribution as ZO-1
and claudin-5 at the cell border (Fig. 11 D and E).Quantification of co-localization of ceramide
with either ZO-1 or claudin-5 in confocal images reveals an average Pearson coefficient of colocalization of 0.77 and 0.65, respectively.

In addition to cell border distribution, typical

punctate green cytoplasmic staining of ceramides was observed. To further verify the specificity
of the immunofluorescence co-localization, we performed immunogold labeling and
transmission electron microscopy using retinal vasculature in bovine retina. The tight junctions
were first visualized by Osmium tetroxide (OsO4) Staining (Fig. 12A). Neighboring sections
were then used for immunogold labeling of occludin and ceramide. As shown in Fig. 12B,
ceramide (smaller, 6 nm gold particles) is localized along the tight junction in close proximity
with occludin protein (larger, 15 nm gold particles), as well as in the intercellular space between
two neighboring endothelial cells (Fig. 12B, right, upper panel).

	
  
	
  

81

A

Osmiun tetroxide staining

RB
C

s

x10K

x67K

x200K

Figure 12. Ceramide co-localizes with tight junction complex.
(A) Tight junctions stained by Osmium tetroxide (OsO4) staining (black arrows).
(left) Lower magnification x10K, red blood cells (RBCs) in the intravascular compartment and
the tight junction between two neighboring retinal endothelial cells (black arrow). Bar 1 µm.
(middle) Tight junction is enlarged for better visualization x67K (black arrow). Bar 0.1 µm.
(right) Kissing point, hallmark of tight junction (black arrowhead). Bar 0.5 µm.

	
  
	
  

82

Figure 12 (cont’d)
B

Immunogold labeling

RBCs

x27K

x100k

RBCs

x40K

x140K

(B) Immunogold localization of ceramide and TJ protein occludin in ultrathin sections of bovine
retina. (B, left panel) Lower magnification x27K (upper) and x40K (lower), of the tight junction
(black arrow), with RBC in the intravascular compartment. Bar 0.5 µm and 0.2 µm, respectively.
(B, right panel) higher magnification x100k (upper), x140k (lower), Ceramide (smaller, 6 nm
gold particles) (white arrowhead) localized along the tight junction in juxtaposition to the tight
junction protein, occludin (larger, 15 nm gold

	
  
	
  

particles) (black arrow). Bar 50 nm.

83

Figure 12 (cont’d)
Although most of ceramide labeling localized along the cytoplasmic side of the plasma
membrane, it can also be visualized in the intercellular space between two neighboring
endothelial cells (black arrowhead).

	
  
	
  

84

2.4.5 Identification of VLC omega-linked acyl-ceramide species in the tight junction
complex
In order to determine which ceramide species interact with the tight junction complex, tight
junction was isolated as previously described (43). ARPE-19 cells were lysed with hypo-osmotic
buffer, collected, homogenized and centrifuged at 100,000xg for 60 minutes at 4°C against a
saturated sucrose cushion. The collected membrane was subsequently cryo-lysed using liquid
nitrogen, and tight junctions were immuno-isolated using either anti-PKC zeta or anti-ZO-1.
Lipid extracts of three independent tight junction preparations were subjected to high
resolution/accurate mass direct infusion-based lipidomic analysis. In order to guide detection and
analysis of sphingolipids containing very long chain fatty acids, we concurrently analyzed lipid
extracts of mouse epidermis, which is known to contain high levels of ceramides and omegalinked acyl ceramides containing very long chain fatty acids.
Interestingly, lipidomic analysis revealed that several ceramide molecular species co-localized
with isolated tight junction complexes (Fig. 13A). While the majority of observed ceramide
(Cer) species were detected between m/z 500-m/z 650 in both positive and negative ionization
modes, corresponding to ceramides containing N-linked fatty acids ranging from C16-C24 in
length, a small number of Cer species were detected at higher m/z values which coincided with
C26:4 and C26:3 very long chain fatty acid-containing ceramides, which were observed above
m/z 650 (Fig. 13A, Inset). Moreover, several ions were observed in the mass range of 900-1100
m/z, which were initially identified as putative omega-linked acyl-ceramide species based on
agreement of the observed accurate mass measurements and isotope distributions with the
theoretical values for each putative Cer species. The assignment of these ions was further
supported by Orbitrap high resolution/accurate

	
  
	
  

mass higher-energy collisional dissociation

85

MS/MS to evaluate the presence of the sphingosine backbone, with identical analysis also
performed on mouse epidermal lipid extracts. . The top panel of Fig. 13B shows the epidermis
tandem mass spectrum from HCD-MS/MS of m/z 943.8562, initially identified as a putative
acyl-ceramide h42:1 (total acyl carbons: total double bonds). The product ion at m/z 264.2682 in
this spectrum is indicative of a d18:1 sphingosine backbone. The bottom panel of Fig. 13B
shows the HCD-MS/MS analysis of ARPE19 tight junctions purified by anti-PKC zeta IP with a
lipid of m/z 943.8562. A product ion at m/z 264.2682 was also observed in the tight junction
MS/MS spectrum, although at lower abundance than that observed for the epidermis lipid
extract. The tight junction isolation was repeated in ARPE-19 homogenates using anti ZO-1
antibody for immunoprecipitation of tight junctions, to assess the potential antibody-specific
dependence of the obtained tight junction lipid profiles. Mass spectrometry analysis of lipid
extracts from anti ZO-1 precipitated tight junctions was performed as described above. The
detected ceramide molecular species were comparable to those from tight junctions obtained
from anti-PKZ immunoprecipitation, including the presence of the acyl-ceramide (d18:1/h24:0)
at m/z 943.8562. These data combined with accurate mass-based formula prediction and
available published data suggest an assignment of acyl-cer d18:1/h24:0, with an omega O-linked
linoleic acid (33–35). Additional criteria used for assignment of putative acyl-ceramide species
included absence of these ions in sample blanks and Protein-A bead negative controls, and a lack
of other detected known sphingolipids which could theoretically be co-isolated during MS/MS
analysis in the mass range in which these ions were observed in the isolated tight junction
complexes. Relative quantitation of all epidermal ceramide species obtained by the same
analytical methods, including Cer and acyl-cer species containing very long chain fatty acids, is
shown for comparison in Fig. 13C.

	
  
	
  

86

3
2.5

A

2
1.5

35

1

AcylCer(d18:1/30:6)

10

AcylCer(d18:1/28:6)

AcylCer(d18:1/h24:0)

15

Cer(d18:1/22:1)

% Total Cer

20

Cer(d18:1/26:3)

0
25

Cer(d18:0/20:0)

30

Cer(d18:1/26:4)

0.5

5

Cer(d18:0/24:0)

Cer(d18:1/24:0)

Cer(d18:1/24:1)

Cer(d18:1/24:2)

Cer(d18:1/23:1)

Cer(d18:0/22:0)

Cer(d18:1/22:0)

Cer(d18:1/20:0)

Cer(d18:1/18:0)

Cer(d18:1/18:1)

Cer(d18:0/17:0)

Cer(d18:0/16:0)

Cer(d18:1/16:0)

Cer(d18:1/16:1)

Cer(d18:1/15:0)

Cer(d18:1/14:0)

0

Figure 13. Tight junction ceramide content.
Lipid extracts of isolated tight junction complexes were subjected to high resolution/accurate
mass spectrometry and tandem mass spectrometry to determine sphingolipid content. (A) Data
are expressed as the average +/- standard deviation of n=3 independent tight junction
preparations. Individual ceramide molecular species are presented as the percent of total
ceramide abundance. Ceramide species were quantitated in negative ion MS mode as
[M+HCOO-H]- ions. The sphingosine backbone and amide-linked fatty acid for the most
abundant isomer at each m/z are given. Inset: Ceramide species containing >C24 fatty acids.

	
  
	
  

87

Figure 13 (cont’d)

B
OH
100

264.2682

HO
OH

Relative Abundance

HN

O

O

O

50

Mouse epidermis
*
668.4731
438.7692

250

100

300

350

400

450

500

550

600 650
m/z

700

750

800

850

900

950

1000

800

850

900

950

1000

255.2315

Relative Abundance

283.2627
ARPE-19 Tight junction
50

264.2682
*
668.4731

250

300

350

400

450

500

550

600 650
m/z

700

750

(B) High resolution tandem mass spectrometry of epidermal and tight junction putative acyl
ceramide. Positive ion mode Orbitrap Higher Energy Collisional Dissociation (HCD) MS/MS at
100,000 resolving power of m/z 943.8562, corresponding to putative protonated acyl-ceramide
(d18:1/h24:0) in mouse epidermis (top panel) and ARPE-19 tight junctions (bottom panel) lipid
extracts. * Matrix ion.

	
  
	
  

88

Figure 13 (cont’d)

AcylCer(d18:0/34:0)

AcylCer(d18:1/h36:0)

AcylCer(d18:1/34:0)

10

AcylCer(d18:1/34:1)

15

AcylCer(d18:0/31:0)

AcylCer(d18:1/h24:0)

% Total Cer

20

AcylCer(d18:1/31:0)

25

AcylCer(d18:1/30:0)

1.4
1.2
1
0.8
0.6
0.4
0.2
0

C

5

0

(C) Mouse epidermis ceramide content. Lipid extracts of mouse epidermis were subjected to
high resolution/accurate mass spectrometry and tandem mass spectrometry to determine
sphingolipid content. Data are expressed as the average +/- standard deviation of n=3 replicate
analyses. Individual ceramide molecular species are presented as the percent of total ceramide
abundance. Ceramide species were quantitated in negative ion MS mode as [M+HCOO-H]- ions.
The sphingosine backbone and amide-linked fatty acid for the most abundant isomer at each m/z
are given. Inset: Epidermal acyl-ceramide species.

	
  
	
  

89

2.4.6 Verification of AAV2 mediated expression of hELOVL4 in mouse retina
In order to determine the effect of ELOVL4 on diabetes-induced blood retinal barrier
dysfunction in vivo, 16 week old diabetic and non-diabetic mice received single intravitreal
injection of 2 x 109 viral genomes of either AAV2-hELOVL4 mut quad (Rt eye) or AAV2Empty vector (left eye) as a control two weeks after STZ induction. Animals were sacrified 6-8
weeks following viral injection (8-10 week of STZ induction). Body weight gain and blood
glucose concentration of diabetic mice and their age matched non-diabetic controls used in this
study are described in Table 1.
In agreement with our previous study (19), ELOVL4 mRNA expression was reduced in diabetic
retinas as compared to control (Fig. 14A).

Transduction with AAV2- hELOVL4 led to

hELOVL4 overexpression in both diabetic and non-diabetic retina as determined using human
transgene-specific primers. As expected, hELOVL4 transcript was undetectable in AAV2-Empty
injected contralateral retinas (Fig. 14B).
To determine the effect of diabetes with and without AAV2-hELOVL4 overexpression on
vascular levels of ELOVL4, retinal cross sections were immunolabeled for ELOVl4 (green).
Isolectin-B4 (red) was used to counterstain retinal vasculature. AAV2-Empty injected retinas
were used as control. Diabetes induced downregulation of endothelial expression of ELOVL4,
which was reversed by AAV2-hELOVL4 treatment (Fig. 14C).

	
  
	
  

90

Table 1. Average body weights and non-fasting blood glucose level of control and diabetic
mice.

Group

	
  
	
  

Weight (g)

Control (n=17)

29.6 ± 0.95

Diabetic (n=25)

24.6 ± 0.33

P value

<0.0001

Blood Glucose
(mg/dL)

181.6 ± 4.62

479.4 ± 21.26

91

P value

<0.0001

Control

0.25

Diabetic

B

Control AAV2-hELOVL4
Diabetic AAV2-hELOVL4

Control AAV2-Empty
Diabetic AAV2-Empty
0.004

**

Relative expression of
hELOVL4 / cyclophilin

Relative expression of
mouse ELOVL4 / cyclophilin

A

0.20
0.15
0.10
0.05
0.00

0.003
0.002
0.001
ND

ND

0.000

C
Control
AAV2-hELOVL4

Diabetic
AAV2-Empty

Diabetic
AAV2-hELOVL4

ELOVL4

Control
AAV2-Empty

Control AAV2-Empty
Control AAV2-hELOVL4
Diabetic AAV2-Empty
Diabetic AAV2-hELOVL4

Co-localization %

Isolectin B4

20

**

****

15
10
5

Overlay

0

Figure 14. Verification of AAV2 mediated expression of hELOVL4 in mouse retina.
STZ-induced diabetic mice were intravitrealy injected with either AAV2-hELOVL4 mut quad
(right eye) or control AAV2-Empty vector (left eye) 2 weeks after diabetes induction. Retinal
tissue was analyzed 6-8 weeks after viral injection. (A) qRT-PCR quantification of endogenous
mouse ELOVL4 mRNA levels in control (white bar) and diabetic (black bar) retinas (n=3-4 per
group). (B) qRT-PCR quantification of exogenous human ELOVL4 mRNA levels in control
(white and dark gray bars) and diabetic (black and light gray bars) retinas (n= 4-6 per group).

	
  
	
  

92

Figure 14 (cont’d)
(C) Immunohistochemical detection of ELOVL4 (green, upper panel) co-localization with retinal
vasculature (red, middle panel) with yellow color (lower panel) indicating vascular expression in
retinal cross sections. Quantification of co-localization (yellow) is shown on the far right (n=3
per group). Scale bar, 20 µm. *p< 0.05, **p<0.01, ***p<0.001, ****p<0.0001.

	
  
	
  

93

2.4.7 AAV2-hELOVl4 overexpression prevents diabetes-induced increase in retinal
vascular permeability in vivo
We next determined whether ELOVL4 overexpression in retinal vasculature could prevent
retinal vascular permeability at early stages of diabetes. Retinal vascular peremability was
assessed 6 weeks after vector administration (8 weeks of diabetes) by measuring leakage of
FITC-albumin. As expected, diabetes induced significant increase in vascular permeability (6.7
fold of control, P=0.0051; Fig. 15A) in retinas receiving AAV2-Empty vector (left eye) relative
to non-diabetic controls (left eye). Notably, treatment with AAV2-hELOLVL4 (right eye)
prevented the diabetes-induced increase in vascular permeability. FITC-albumin leakage was
significantly lower in diabetic mouse retinas receiving AAV2-hELOVL4 (right eye) than AAV2Empty diabetic mouse retina (left eye) (1.6 fold of control, P=0.1202; Fig. 15A). These findings
are consistent with the results obtained in BRECs.
To confirm the beneficial effect of ELOVL4 overexpression on tight junction distribution at cell
border, whole retina flat mounts were immunolabeled for occludin in AAV2-Empty and AAV2hELOVL4 transduced retinas. In AVV2-Empty diabetic retinas (left eye), occludin
immunoreactivity was evidently disrupted and intermittently lost in retinal blood vessels. These
changes were prevented in AAV2-hELOVL4 diabetic retinas (right eye), where occludin
distribution was similar to that in non-diabetic control mice (Fig. 15B)

	
  
	
  

94

Control AAV2-Empty
Control AAV2-hELOVL4
Diabetic AAV2-Empty
Diabetic AAV2-hELOVL4

A
Control AAV2-hELOVL4 Diabetic AAV2-Empty

Diabetic AAV2-hELOVL4

10
Vascular permeability
(Fold Change)

Control AAV2-Empty

**

6
4
2
0

B
Control AAV2-Empty

**

8

Control AAV2-hELOVL4 Diabetic AAV2-Empty

Diabetic AAV2-hELOVL4

Figure 15. ELOVL4 overexpression prevents diabetes-induced vascular leakage in mice.
(A) Representative fluorescent images of retinal vascular permeability. Retina permeability was
evaluated by measuring leakage fluorescence intensity as shown on the far right (n= 8-12 per
group). Scale bar, 50 µm. (B) Immunostaining of flat mounted whole retina for occludin.
Arrowheads show disruption of continuous staining in diabetic retina. Scale bar, 20 µm. *p<
0.05, **p<0.01, ***p<0.001.

	
  
	
  

95

2.5 Discussion	
  
Hyperglycemia and dyslipidemia are the major metabolic abnormalities in diabetes, however
historically diabetes was described as a problem of glucose metabolism and diabetic
complications, including diabetic retinopathy, were generally considered as a result of
hyperglycemia. The results of clinical trials, including DCCT, ACCORD and FIELD, however,
do not support this one-sided view as they clearly demonstrate a large contribution of
dyslipidemia to the pathogenesis of diabetic retinopathy (18,58–61). The mechanism(s) of
dyslipidemia-induced retinal damage are not yet well elucidated, however it became clear that
changes in retinal-specific lipid metabolism in diabetes play a major role in DR pathogenesis.

Elongases and desaturases are key enzymes in fatty acid metabolism; they regulate the length
and degree of saturation of fatty acids and thereby their function and metabolic fate. As we and
others have previously demonstrated, expression level of elongases and desaturases in the liver
and retina is affected by diabetes with dramatic downregulation of elongases, including the most
abundant retinal elongase ELOVL4 in the retina (19,62–64). ELOVL4 has substrate specificity
profile for fatty acids of C24 or longer and it can elongate both poly, mono-unsaturated and
saturated very long chain fatty acids (29). The role of ELOVL4 in the retina was mainly
described for VLCPUFA (38,65,66). VLCPUFAs are mainly present in the rod outer-segment
membrane, where they are suggested to contribute a role in photoreceptor function by stabilizing
the rims of photoreceptor disks (38). Decrease in ELOVL4 in diabetic retina was associated with
a decrease in VLCPUFA (i.e., 32:6n3) synthesis (19).

	
  
	
  

96

Although the highest retinal expression of ELOVL4 is found in photoreceptor inner segment
(37), it is also expressed in other retinal layers (67,68). Herein, we demonstrated expression of
ELOVL4 in retinal vasculature. The product of ELOVL4-mediated elongation will depend on
the precursor fatty acids. As the photoreceptors are rich in n3 PUFA, the main products of
ELOVL4 in photoreceptor cells are C32-36 PUFAs. Endothelial and epithelial cells have much
lower n3 PUFA and higher saturated fatty acids content leading to higher level of VLC saturated
fatty acid production that can be incorporated into ceramides. This is further supported by the
studies demonstrating that ELOVL4 works in concert with ceramide synthases leading to VLC
ceramide production (69). The role of VLC ceramides is best known in skin permeability barrier
where they represent a major component of stratum corneum (33–35). Whether decrease in
ELOVL4 and thus VLC ceramides production in the retina plays a role in BRB breakdown in
diabetes was not known and represented a major focus of this study.
The studies here reveal that overexpression of ELOVL4 by recombinant adenoviral system
normalized VEGF and IL-1β induced retinal vascular permeability in BRECs and increased
steady state content of tight junction protein occludin. Further, cell border organization was
normalized for both ZO-1 and claudin 5. Importantly, intravitreal delivery of hELOVL4 in
diabetic retina using AAV2 viral vector increased hELOVL4 expression in retinal endothelium
and prevented diabetes-induced increase in vascular permeability in diabetic mice. These data
strongly support the role of ELOVL4 in protective pro-barrier function.
Although this is the first study to demonstrate the role of ELOVL4 in retinal BRB, the essential
role of ELOVL4 in skin barrier function is well known. Homozygous ELOVL4 mutant mice die
within few hours after birth due defective skin barrier and severe dehydration (33–35), a

	
  
	
  

97

phenotype that can be rescued by skin specific expression of ELOVL4 to normalize skin
permeability (70). This defective water insulation function of the skin was attributed to ELOVL4
loss of function and subsequent decrease in production of VLCFA (C≥26) and thereby VLC
ceramides and acylceramides, critical lipid components of the skin barrier function, probably due
its packing property within stratum corneum that resists water loss. The VLC ceramides and
especially	
   omega-linked acyl-ceramide species have never before been shown outside of stratum
corneum.
The data presented in this study suggest that retinal pro-barrier effects of ELOVL4 are mediated
by VLC ceramide production and incorporation into tight junction structures. Ceramide was
found to co-localize with tight junction complexes and follow the same distribution as ZO-1 and
claudin 5. Moreover, mass spectrometry analysis revealed very long chain ceramides in tight
junction isolates. The presence of these VLC ceramides in BRB tight junctions suggests that
similar to skin, they might serve pro-barrier function in the retinal microvasculature.
The role of plasma membrane lipid composition in the maintenance and organization of tight
junction complexes has being postulated for a long time. In endothelial cells, gamma linolenic
acid 18:3n6 and EPA 20:5n3 increased occludin content with subsequent decrease in paracellular
permeability (71). Park et al. have demonstrated that gangliosides exert a protective effect on rat
intestinal permeability, preventing lipopolysaccharide induced occludin degradation (72). TNF-α
induced alteration in lipid composition of TJ microdomains causing occludin protein
redistribution and increased paracellular permeability (73). Palmitoylation- covalent attachment
of the palmitic fatty acid to cysteine residues of proteins- of tight junction protein claudin was
found to be essential for protein localization and

	
  
	
  

enhanced property of tight junctions (74).

98

In addition to structural role of lipids in surrounding plasma membrane, lipidic nature of tight
junction itself was first proposed more than 30 years ago by Kachar et al. They suggested that
tight junction strands are formed of inverted lipid cylinders where apposing cells contact each
other through fusion points between exoplasmic leaflet of the two neighboring cells (75). This
was further supported by freeze fracture preparations that demonstrate the presence of
phospholipids and cholesterol particles in tight junction strands (76,77). Model of lipid bridges at
tight junctions has been postulated as a route for lateral transfer of phospholipids and lipidprotein complexes between neighboring cells (78,79). Our study is the first to demonstrate that
ceramides co-localize within tight junction proteins demonstrated by immunofluorescent
staining, electron microscopy and tight junction purification followed by MS analysis.
Specifically we demonstrated the presence of VLC ceramides and, importantly, omega-linked
acyl-ceramide species by mass spectrometry analysis of tight junction structures. Until now VLC
ceramides were only found in water permeability barrier of the skin.
The exact confirmation of omega-linked acyl-VLC ceramides in the tight junction structure is
not known, and will require further investigation, however these very long highly saturated lipids
may have a role in stabilizing tight junction proteins and protecting against degradation. Indeed,
overexpression of hELOVL4 in BRECs, significantly increased occludin content and prevented
the VEGF – induced decrease in occludin. This effect was due to translational or
posttranslational regulation as occludin mRNA level was not increased after ELOVL4
overexpression (occludin mRNA/ GAPDH; in AdEmpty transduced cells = 0.005 ± 0.002, while
in hELOVL4 transduced cells = 0.004 ± 0.0002).

	
  
	
  

99

Based on the structure, omega-linked acyl-VLC ceramides could be incorporated in the plasma
membranes and aid in organizing junctional domain. The long fatty acid tail of omega-linked
acyl-ceramides could potentially allow for traversing the thickness of double layer, bridging the
space between tight junction proteins of apposing cells and thereby regulating barrier function.
More detailed structural studies would be required to assign the exact function to these lipids in
the tight junctions. In summary, we have shown that ELOVL4-mediated production of VLCceramides is necessary for retinal endothelial barrier function. Overexpression of ELOVL4
prevented diabetes-induced retinal permeability and BRB breakdown. VLC ceramides may be
necessary for tight junction stability and thereby enhancing barrier property. Normalization of
retinal ELOVL4 expression may provide a potential therapeutic strategy for the prevention of
diabetes-induced early breakdown of the blood-retina barrier by modulating retinal sphingolipid
metabolism.

	
  
	
  

100

REFERENCES

	
  
	
  

101

REFERENCES

1.

Antonetti DA, Barber AJ, Hollinger LA, Wolpert EB, Gardner TW. Vascular Endothelial
Growth Factor Induces Rapid Phosphorylation of Tight Junction Proteins Occludin and
Zonula Occluden 1. J Biol Chem. 1999;274(33):23463–7.

2.

Murakami T, Felinski EA, Antonetti DA. Occludin phosphorylation and ubiquitination
regulate tight junction trafficking and vascular endothelial growth factor-induced
permeability. J Biol Chem. 2009;284(31):21036–46.

3.

Harhaj NS, Felinski E a, Wolpert EB, Sundstrom JM, Gardner TW, Antonetti DA. VEGF
activation of protein kinase C stimulates occludin phosphorylation and contributes to
endothelial permeability. Invest Ophthalmol Vis Sci. 2006;47(11):5106–15.

4.

Zhu W, London NR, Gibson CC, Davis CT, Tong Z, Sorensen LK, et al. Interleukin
receptor activates a MYD88-ARNO-ARF6 cascade to disrupt vascular stability. Nature.
2012;492(7428):252–5.

5.

Aveleira C a, Lin C-M, Abcouwer SF, Ambrósio AF, Antonetti D a. TNF-α signals
through PKCζ/NF-κB to alter the tight junction complex and increase retinal endothelial
cell permeability. Diabetes. 2010;59(11):2872–82.

6.

Behzadian MA, Wang X, Windsor LJ, Ghaly N, Caldwell RB. TGF-beta
Increases Retinal Endothelial Cell Permeability by Increasing MMP-9  : Possible
Role of Glial Cells in AND. Invest Ophthalmol Vis Sci. 2001;42(3):853-9.

7.

Navaratna D, Mcguire PG, Menicucci G, Das A. Proteolytic degradation of VE-cadherin
alters the blood-retinal barrier in diabetes. Diabetes. 2007;56(9)2380-2387.

8.

Kim J, Kim C-S, Lee YM, Jo K, Shin SD, Kim JS. Methylglyoxal induces
hyperpermeability of the blood-retinal barrier via the loss of tight junction proteins and the
activation of matrix metalloproteinases. Graefes Arch Clin Exp Ophthalmol.
2012;250(5):691–7.

9.

Enge M, Bjarnega M, Gerhardt H, Gustafsson E, Asker N, Hammes H, et al.
Endothelium-specific platelet-derived growth factor-B ablation mimics diabetic
retinopathy. EMBO J. 2002;21(16):4307–16.

10.

Gavard J, Patel V, Gutkind JS. Angiopoietin-1 prevents VEGF-induced endothelial
permeability by sequestering Src through mDia. Dev Cell. 2008;14(1):25–36.

11.

Geraldes P, Hiraoka-yamamoto J, Matsumoto M, Clermont A. Activation of PKCδ 47and
SHP1 by hyperglycemia causes vascular cell apoptosis and diabetic retinopathy. Nat Med.
2009;15(11): 1298-306.

12.

Nguyen QD, Brown DM, Marcus DM,

	
  
	
  

102

Boyer DS, Patel S, Feiner L, et al.

Ranibizumab for diabetic macular edema: Results from 2 phase iii randomized trials:
RISE and RIDE. Ophthalmology. 2012;119(4):789–801.
13.

Cunha-Vaz J. Ribeiro L, Lobo C. Phenotypes and biomarkers of diabetic retinopathy. Prog
Retin Eye Res. 2014;41:90-111.

14.

Klaassen I, Van Noorden CJ, Schlingemann RO. Molecular basis of the inner bloodretinal barrier and its breakdown in diabetic macular edema and other pathological
conditions. Prog Retin Eye Res. 2013;34:19–48.

15.

Engerman RL, Kern TS. Experimental Galactosemia Produces Diabetic-like Retinopathy.
Diabetes.1984;33(1):97–100.

16.

Antonetti DA , Barber AJ, Khin S, Lieth E, Tarbell JM, Gardner TW. Vascular
Permeability in Experimental Diabetes Is associated with Reduced Endothelial Occludin
Content:vascular endothelial growth factor decreases occludin in retinal endothelial cells.
Diabetes. 1998;47(12):1953-9.

17.

Giacco F, Brownlee M. Oxidative stress and diabetic complications. Circ Res.
2010;107(9):1058–70.

18.

The Action of Control Cardiovascular Risk in Diabetes Study Group. Effects of Medical
Therapies on Retinopathy Progression in Type 2 Diabetes. N Engl J Med.2010;363:23344.

19.

Tikhonenko M, Lydic T a, Wang Y, Chen W, Opreanu M, Sochacki A, et al. , Remodeling
of retinal fatty acids in animals model of diabetes: a decrease in long-chain
polyunsaturated fatty acids is associated with a decrease in fatty acid elongases ELOVL2
and ELOVL4. Diabetes. 2010;59(1):219–27.

20.

Opreanu M, Tikhonenko M, Bozack S, Lydic T a, Reid GE, McSorley KM, et al. The
unconventional role of acid sphingomyelinase in regulation of retinal microangiopathy in
diabetic human and animal models. Diabetes. 2011;60(9):2370–8.

21.

Opreanu M, Lydic T a, Reid GE, McSorley KM, Esselman WJ, Busik J V. Inhibition of
cytokine signaling in human retinal endothelial cells through downregulation of
sphingomyelinases by docosahexaenoic acid. Invest Ophthalmol Vis Sci.
2010;51(6):3253–63.

22.

Uchida Y. The role of fatty acid elongation in epidermal structure and function.
Dermatoendocrinol. 2011;3(2):65–9.

23.

Kihara A. Very long-chain fatty acids: Elongation, physiology and related disorders. J
Biochem. 2012;152(5):387–95.

24.

Mandal MN a, Ambasudhan R, Wong PW, Gage PJ, Sieving P a, Ayyagari R.
Characterization of mouse orthologue of ELOVL4: genomic organization and spatial and

	
  
	
  

103

temporal expression. Genomics. 2004;83(4):626–35.
25.

Lagali PS, Liu J, Ambasudhan R, Kakuk LE, Bernstein S, Seigel GM, et al. Evolutionarily
Conserved ELOVL4 Gene Expression in the Vertebrate Retina. Invest Ophthalmol Vis
Sci. 2003;44(7):2841–50.

26.

Umeda S, Ayyagari R, Suzuki MT, Ono F. Molecular Cloning of ELOVL4 Gene from
Cynomolgus Monkey ( Macaca fascicularis ). Exp.Anim. 2003;52(2):129–35.

27.

Zhang X, Yang Z, Karan G, Hashimoto T, Baehr W, Yang X. Elovl4 mRNA distribution
in the developing mouse retina and phylogenetic conservation of Elovl4 genes. Molecular
Vision. 2003;9:301–7.

28.

Agbaga M-P, Mandal MN a, Anderson RE. Retinal very long-chain PUFAs: new insights
from studies on ELOVL4 protein. J Lipid Res. 2010;51(7):1624–42.

29.

Agbaga M-P, Brush RS, Mandal MNA, Henry K, Elliott MH, Anderson RE. Role of
Stargardt-3 macular dystrophy protein (ELOVL4) in the biosynthesis of very long chain
fatty acids. Proc Natl Acad Sci U S A. 2008;105(35):12843–8.

30.

Busik JV, Esselman WJ, Reid GE. Examining the role of lipid mediators in diabetic
retinopathy. Clin Lipidol. 2012;7(6):661–75.

31.

Wertz PW, Downing DT. Ceramides of pig epidermis  : structure determination. J lipid
Res.1983;24(6):759–65.

32.

Imokawa g, Abe A, Jin K, Higaki Y, Kawashima M, Hidano A. Decreased level of
ceramides in stratum corneum of atopic dermatitis: an etiologic factor in atopic dry skin. J
Invest Dermatol. 1991;96(4):523-6.

33.

Vasireddy V, Uchida Y, Salem N, Kim SY, Mandal NA, et al. Loss of functional
ELOVL4 depletes very long-chain fatty acids (≥28) and the unique ω-o-acylceramides in
skin leaidng to neonatal death. Hum Mol Genet. 2007;16(5):471-82.

34.

McMahon A, Butovich IA, Mata NL, Klein M, Ritter R, Richardson J, et al. Retinal
pathology and skin barrier defect in mice carrying a Stargardt disease-3 mutation in
elongase of very long chain fatty acids-4. Mol Vis. 2007;13:258–72.

35.

Li W, Sandhoff R, Kono M, Zerfas P, Hoffmann V, Ding BCH, et al. Depletion of
ceramides with very long chain fatty acids causes defective skin permeability barrier
function, and neonatal lethality in ELOVL4 deficient mice. Int J Biol Sci. 2007;3(2):120–
8.

36.

Okuda A, Naganuma T, Ohno Y, Abe K, Yamagata M, Igarashi Y, et al. Heterooligomeric interactions of an ELOVL4 mutant protein: implications in the molecular
mechanism of Stargardt-3 macular dystrophy. Mol Vis. 2010;16:2438–45.

37.

Zhang K, Kniazeva M, Han M, Li W, Yu

	
  
	
  

104

Z, Yang Z, et al. A 5-bp deletion in

ELOVL4 is associated with two related forms of autosomal dominant macular dystrophy.
Nat Genet. 2001;27(1):89–93.
38.

Karan G, Lillo C, Yang Z, Cameron DJ, Locke KG, Zhao Y, et al. Lipofuscin
accumulation, abnormal electrophysiology, and photoreceptor degeneration in mutant
ELOVL4 transgenic mice: a model for macular degeneration. Proc Natl Acad Sci U S A.
2005;102(11):4164–9.

39.

Abcouwer SF, Lin C, Wolpert EB, Shanmugam S, Schaefer EW, Freeman WM, et al.
Effects of ischemic preconditioning and bevacizumab on apoptosis and vascular
permeability following retinal ischemia-reperfusion injury. Invest Ophthalmol Vis Sci.
2010;51(11):5920–33.

40.

Ryals RC, Boye SL, Dinculescu A, Hauswirth WW, Boye SE. Quantifying transuction
efficiencies of unmodified and tyrosine capsid mutant AAV vectors in vitro using two
ocular cell lines. Mol Vis. 2011;17:1090-102.

41.

Wigglesworth VB. Lipid staining for the electron microscope: a new method. J Cell Sci.
1975;19(3):425–37.

42.

Mahendrasingam S, Wallam CA, Hackney CM. Two approaches to double postembedding immunogold labeling of freeze-substituted tissue embedded in low
temperature Lowicryl HM20 resin. Brain Res Brain Protoc. 2003;11(2):134–41.

43.

Tang VW. Proteomic and bioinformatic analysis of epithelial tight junction reveals an
unexpected cluster of synaptic molecules. Biol Direct. 2006;1:37.

44.

Busik J V, Reid GE, Lydic TA. Global analysis of retinal lipids by complementary
precursor ion and neutral loss mode tandem mass spectrometry. Methods Mol Biol.
2009;579:33-70.

45.

Lydic T a, Busik J V, Reid GE. A monophasic extraction strategy for the simultaneous
lipidome analysis of polar and nonpolar retina lipids. J Lipid Res. 2014;55(8):1797–809.

46.

Byeon SK, Lee JY, Lee J-S, Moon MH. Lipidomic profiling of plasma and urine from
patients with Gaucher disease during enzyme replacement therapy by nanoflow liquid
chromatography–tandem mass spectrometry. J Chromatogr A. 2015;1381:132–9.

47.

Chakravarthy H, Beli E, Navitskaya S, O’Reilly S, Wang Q, Kady N, et al. Imbalances in
mobilization and activation of pro-inflammatory and vascular reparative bone marrowderived cells in diabetic retinopathy. PLoS One. 2016;11(1):1–21.

48.

Aiello LP, Avery RL, Arrigg PG, Keyt BA, Jampel HD et al. Vascular endothelial growth
factor in ocular fluid of patients with diabetic retinopathy and other retinal disoredrs. N
Engl J Med. 1994;331:1480-7.

49.

Mathews MK, Merges C, Mcleod DS, Lutty GA. Vascular Endothelial Growth Factor and
Vascular Permeability Changes in Human
Diabetic Retinopathy. Invest Opthalmol Vis

	
  
	
  

105

Sci.1997;38:2729-41.
50.

Hammes HP, Lin J, Bretzel RG, Brownlee M, Breier G. Upregulation of the vascular
endothelial growth factor/vascular endothelial growth factor receptor system in
experimental background diabetic retinopathy of the rat. Diabetes. 1998;47(3):401–6.

51.

Qaum T, Xu Q, Joussen AM, Clemens MW, Qin W, Miyamoto K, et al. VEGF-initiated
Blood – Retinal Barrier Breakdown in Early Diabetes. Invest Opthalmol Vis Sci.
2001;42:2408–13.

52.

Caldwell RB, Bartoli M, Behzadian MA, El-Remessy AEB, Al-Shabrawey M, Platt DH,
et al. Vascular endothelial growth factor and diabetic retinopathy: pathophysiological
mechanisms and treatment perspectives. Diabetes Metab Res Rev. 2003;19(6):442–55.

53.

Koleva-Georgieva DN, Sivkova NP, Terzieva D. Serum inflammatory cytokines IL-1β,
IL-6, TNF-α and VEGF have influence on the development of diabetic retinopathy. Folia
Med. 2011;53(2):44–50.

54.

Martiney J a., Berman JW, Brosnan CF. Chronic inflammatory effects of interleukin-1 on
the blood-retina barrier. J Neuroimmunol. 1992;41(2):167–76.

55.

Claudio L, Martiney JA, Brosnan CF. Ultrastructural studies of the blood-retina barrier
after exposure to interleukin-1 beta or tumor necrosis factor-alpha . Lab
Invest.1994;70(6):850-61.

56.

Demircan N, Safran BG, Soylu M, Ozcan a a, Sizmaz S. Determination of vitreous
interleukin-1 (IL-1) and tumour necrosis factor (TNF) levels in proliferative diabetic
retinopathy. Eye (Lond). 2006 ;20(12):1366–9.

57.

Kowluru R a, Odenbach S. Role of interleukin-1beta in the development of retinopathy in
rats: effect of antioxidants. Invest Ophthalmol Vis Sci. 2004;45(11):4161–6.

58.

Lyons TJ, Jenkins AJ, Zheng D, Lackland DT, McGee D, Garvey WT, et al. Diabetic
retinopathy and serum lipoprotein subclasses in the DCCT/EDIC cohort. Investig
Ophthalmol Vis Sci. 2004;45(3):910–8.

59.

Klein RL, McHenry MB, Lok KH, Hunter SJ, Le N-A, Jenkins AJ, et al. Apolipoprotein
C-III protein concentrations and gene polymorphisms in Type 1 diabetes: associations
with microvascular disease complications in the DCCT/EDIC cohort. J Diabetes
Complications. 2005;19(1):18–25.

60.

Scott R, Best J, Forder P, Taskinen MR, Simes J, et al. Fenofibrate Intervention and Event
Lowering in Diabetes (FIELD) study: baseline characteristics and short-term effects of
fenofibrate. Cardiovasc Diabetol. 2005;4:13.

61.

Ferris FL, Chew EY, Hoogwerf BJ. Serum Lipids and Diabetic Retinopathy. Diabetes
care.1996;19(11):1291–3.

	
  
	
  

106

62.

Wang Y, Botolin D, Xu J, Christian B, Mitchell E, Jayaprakasam B, et al. Regulation of
hepatic fatty acid elongase and desaturase expression in diabetes and obesity. J Lipid Res.
2006;47(9):2028–41.

63.

Wang Q, Tikhonenko M, Bozack SN, Lydic TA, Yan L, Panchy NL, et al. Changes in the
Daily Rhythm of Lipid Metabolism in the Diabetic Retina. PLoS One. 2014;9(4):1-13.

64.

Green CD, Olson LK. Modulation of palmitate-induced endoplasmic reticulum stress and
apoptosis in pancreatic β-cells by stearoyl-CoA desaturase and Elovl6. Am J Physiol
Endocrinol Metab. 2011;300(4):E640-9.

65.

McMahon A, Jackson SN, Woods AS, Kedzierski W. A stargardt disease-3 mutation in
the mouse Elovl4 gene causes retinal deficiency of C32-C36 acyl phosphatidylcholines.
FEBS lett. 2007;581(28):5459-63.

66.

Harkewicz R, Du H, Tong Z, Alkuraya H, Bedell M, Sun W, et al. Essential role of
ELOVL4 protein in very long chain fatty acid synthesis and retinal function. J Biol Chem.
2012;287(14):11469–80.

67.

Grayson C, Molday RS. Dominant negative mechanism underlies autosomal dominant
Stargardt-like macular dystrophy linked to mutations in ELOVL4. J Biol Chem.
2005;280(37):32521–30.

68.

Sommer JR, Estrada JL, Collins EB, Bedell M, Alexander C a, Yang Z, et al. Production
of ELOVL4 transgenic pigs: a large animal model for Stargardt-like macular
degeneration. Br J Ophthalmol. 2011;95(12):1749–54.

69.

Sandhoff R. Very long chain sphingolipids  : Tissue expression , function and synthesis.
FEBS lett. 2010;584:1907–13.

70.

McMahon A, Butovich I a, Kedzierski W. Epidermal expression of an Elovl4 transgene
rescues neonatal lethality of homozygous Stargardt disease-3 mice. J Lipid Res.
2011;52(6):1128–38.

71.

Jiang WG, Bryce RP, Horrobin DF, Mansel RE. Regulation of Tight Junction
Permeability and Occludin Expression by Polyunsaturated Fatty Acids. Biochem Biophys
Res commun.1998;244(2):414–20.

72.

Park EJ, Thomson ABR, Clandinin MT. Protection of intestinal occludin tight junction
protein by dietary gangliosides in lipopolysaccharide-induced acute inflammation. J
Pediatr Gastroenterol Nutr. 2010;50(3):321–8.

73.

Li Q, Zhang Q, Wang M, Zhao S, Ma J, Luo N, et al. Interferon-gamma and tumor
necrosis factor-alpha disrupt epithelial barrier function by altering lipid composition in
membrane microdomains of tight junction. Clin Immunol. 2008 126(1):67–80.

74.

Van Itallie CM, Gambling TM, Carson JL, Anderson JM. Palmitoylation of claudins is
required for efficient tight-junction
localization. J Cell Sci. 2005;118(7):1427–

	
  
	
  

107

36.
75.

Kachar B, Reese TS. Evidence for the lipidic nature of tight junction strands.
Nature.1982;296:464-6.

76.

Kan FW. Cytochemical Evidence for the Presence of Phospholipids in Epithelial Tight
Junction Strands. J Histochem Cytochem.1993;41(5):649-56.

77.

Feltkamp CA, Van der waerden AW. Junction formation between cultured normal rat
hepatocytes. An ultrastructural study on the presence of cholesterol and the structure of
developing tight- junction strands. J Cell Sci.1983;63:271–86.

78.

Laffafian I, Hallett MB. Lipid-Protein Cargo Transfer  : A Mode of Direct Cell-to-Cell
Communication for Lipids and Their Associated Proteins. J cell Physiol.
2007;210(2):336–42.

79.

Grebenkfimper K. Translational diffusion measurements of a fluorescent phospholipid
between MDCK-I cells support the lipid model of the tight junctions. Chem Phys Lipids.
1994;71(94):133–43.

	
  

	
  
	
  

108

Chapter 3. Increase in acid sphingomyelinase level in human retinal endothelial cells and
CD34+ circulating angiogenic cells isolated from diabetic individuals is associated with
dysfunctional retinal vasculature and vascular repair process in diabetes
This chapter is a modified version of an article accepted in Clinical Lipidology.
Authors: Nermin Kady, Yuanqing Yan, Tatiana Salazar, Qi Wang, Harshini Chakravarthy, Chao
Huang, Eleni Beli, Svetlana Navitskaya, Maria Grant, Julia Busik.
3.1 Abstract
Diabetic retinopathy (DR) is a microvascular disease that results from retinal vascular
degeneration and its defective repair due to diabetes induced endothelial progenitor dysfunction.
Understanding the key molecular factors involved in vascular degeneration and repair is
paramount for developing effective DR treatment strategies. We propose that diabetes-induced
activation of acid sphingomyelinase (ASM) plays essential role in retinal endothelial and CD34+
circulating angiogenic cell (CAC) dysfunction in diabetes.
Primary cultures of human retinal endothelial cells (HRECs) isolated from control and diabetic
donor tissue and human CD34+ CACs from control and diabetic patients were used in this study.
ASM mRNA expression was assessed by quantitative PCR, ASM concentration was determined
by ELISA assay. To evaluate the effect of diabetes-induced ASM on HRECs and CD34+ CACs
function, tube formation, CAC incorporation into endothelial tubes, and diurnal release of CD34+
CACs in diabetic individuals was determined.
ASM expression level was significantly increased in HRECs isolated from diabetic compared to
control donor tissue, as well as CD34+CACs and plasma of diabetic patients. A significant
decrease in tube area was observed in HRECs from donors with signs of NPDR as compared to
control HRECs. The tube formation deficiency

	
  
	
  

was associated with increased expression of

109

ASM in diabetic HRECs. Moreover, diabetic CD34+ CACs with high ASM showed defective
incorporation into endothelial tubes. Diurnal release of CD34+ CACs was disrupted with the
rhythmicity lost in diabetic patients.
Collectively, these findings support that diabetes-induced ASM upregulation has a marked
detrimental effect on both retinal endothelial cells and CACs. ASM may provide novel
therapeutic target for treating DR.
	
  
	
  

	
  
	
  

110

3.2 Introduction
	
  
Diabetic retinopathy (DR) is a microvascular disease that results from diabetes-induced
retinal damage that is further exacerbated by bone marrow dysfunction. Bone marrow
dysfunction leads to decreased release of cells into the circulation and changes in hematopoiesis
resulting in increased circulating pro-inflammatory monocytes and diminished repair due to
defective progenitor cells. Although DR influence all retinal cells, clinical manifestations of DR
are mainly due to changes in retinal vessels, where early histological alterations include pericyte
loss, thickening of basement membrane, capillary occlusion and endothelial cell degeneration
(1,2) . These are followed by break down of blood retinal barrier (BRB) and leaky vasculature
leading to hemorrhages, hard exudates, and retinal edema; structural changes involving the
vascular wall leading to microaneurysms; and finally neovascularization, vitreous hemorrhage
and fibrous tissue formation (3). Impaired vision due to macular edema, or vision loss due to
neovascularization-induced vitreous hemorrhage or tractional retinal detachment usually takes
place in the later stages of the disease.

Circulating angiogenic cells (CACs), a population of vascular progenitors originated
from HSC (4), are considered as key regulators for healthy maintenance of retinal vasculature.
Diabetic metabolic abnormalities lead to defective vascular maintenance due, in part, to failed
attempts by dysfunctional CACs to repair damaged endothelium.

	
  
	
  

111

HSCs isolated from bone marrow or CACs from peripheral blood of control (healthy)
animals have been shown to repair ischemic damage and aid in reperfusion of ischemic tissues
(4–6). Several studies have shown an association between DR risk and both reduced number (7–
10) and function of CACs (11–16).

Several key hyperglycemia- and dyslipidemia-activated pathways leading to retinal
endothelial cell and CAC dysfunction have been identified. Prominent among these are pathways
that promote an increase of pro-inflammatory cytokines, pro-inflammatory lipids and proangiogenic factors (17–26). We have previously demonstrated activation of the central enzyme
of sphingolipid metabolism, acid sphingomyelinase (ASM), as a key metabolic abnormality in
diabetic retinal vasculature and CACs. ASM hydrolyzes sphingomyelin (SM) into proinflammatory and pro-apoptotic ceramide. Activation of ASM plays an important role in signal
transduction in response to various stimuli including IL-1β (27,28) and TNF-α (29). Endothelial
cells represent a major source of ASM (30–33). Inhibition of ASM exhibits protective effect in
diabetes preventing diabetes-induced retinal inflammation and vascular degeneration (15,33,34).

Previously, we have identified key defects in circadian regulation of CACs. We showed
that bone marrow denervation results in loss of circadian release of vascular reparative cells from
the bone marrow and generation of increased numbers of proinflammatory cells. Using a rat
model of T2D, we showed that the decrease in CACs release from diabetic bone marrow is
caused by bone marrow neuropathy and that these changes precede the development of diabetic
retinopathy. We observed a marked reduction in clock gene expression in the retina and in
CACs. Denervation of the bone marrow resulted

	
  
	
  

in progenitors being “trapped” within the

112

bone marrow and in loss of the circadian release of these cells into the circulation. This reduction
in the circadian peak of CAC release into the circulation led to diminished reparative capacity
and resulted in development of acellular retinal capillaries (7). We also showed that Per2 mutant
mice recapitulate key aspects of diabetes without the associated metabolic abnormalities. In Per2
mutant mice, we observed a threefold decrease in proliferation and 50% reduction in nitric oxide
levels in CACs. Tyrosine hydroxylase-positive nerve processes and neurofilament-200 staining
were reduced in Per2 mutant mice (suggestive of diabetic neuropathy) and increased acellular
capillaries were identified (35). We also showed that as CD34+ CACs acquired differentiation
markers (towards the endothelial lineage), robust oscillations of clock genes are observed (36).
It is well accepted in diabetic complications field that cells isolated from diabetic tissue
keep diabetic phenotype for several passages even when cultured in normal glucose. This is due
to “metabolic memory”, or "legacy effect" for vascular disease in diabetes - the prolonged
benefits of good glycemic control, as well as the prolonged harm poor control in diabetic
patients(11,37–39). In this study we used HREC cells isolated from control and diabetic donor
tissue as a model.
In the current study, we have focused exclusively on human CACs. We asked if the
defect in circadian release observed in rodents with diabetes occurred in humans. We examined
the effect of diabetes-induced ASM activity on the function of human CACs and retinal
endothelial cells comparing the angiogenic ability of control (low ASM) and diabetic (high
ASM) HRECs to form tube–like structures in vitro and determining the capacity of control (with
low ASM) and diabetic (with high ASM) CACs to support endothelial tube formation.

	
  
	
  

113

3.3 Materials and methods
3.3.1 Circadian study of human CD34+ CACs
The study was approved by the University of Florida IRB #411-2010. All study subjects
provided informed consent. Individuals were brought into the Clinical Research Center at the
University of Florida for 48 h. During the first 24 h, individuals were evaluated and on the
evening of the first day, a heparin lock was inserted into their forearm. During the second 24-h
period, the individuals had 1 mL of blood removed every 2 h for a total of 24 h and analyzed for
the number of CD34+ cells by flow cytometry. Clinical characteristics of the patients are
presented in Table 2.
3.3.2 Postmortem imaging of human retina and cell culture
Primary cultures of HRECs were prepared from postmortem tissue obtained from National
Disease Research Interchange, Philadelphia, PA and Midwest Eye-Banks, Ann Arbor, MI. The
tissue was received within 36 h after death. The donor characteristics are provided in Table 3.
Primary HRECs were isolated as previously described (40). As previously demonstrated the
cells isolated from control and diabetic donors keep their phenotypes for 4-6 passages due to
metabolic memory phenomenon (11,37–39). On arriving in the laboratory, the eyes were
placed on sterile gauze, and they were washed with povidone- iodine solution (Purdue Pharma
L.P. Stamford, CT). After 10 minutes, the globe was punctured approximately 2 mm from the
limbus, a circumferential incision was made and the anterior chamber was removed. A vitreous
spatula was used to loosen the vitreous adherent to the anterior retina. When all the vitreous
was removed, the retina was gently removed from the layer of retinal pigment epithelium and
cut at the optic nerve. Before proceeding into the isolation of HRECs, retinas were rinsed,

	
  
	
  

114

flatmounted and retinal imaging was taken using a Nikon SMZ-800 Stereo Microscope with
Prior Proscan 3 Motorized XY System with Z Drive and MetaMorph Modules to perform
image stitching, to properly determine the stage of retinopathy of the donors used for isolation
of HRECs. Retinas included in this study have at least three signs of non-proliferative diabetic
retinopathy (NPDR) such as microaneurysms, intraretinal haemorrhages and intraretinal
microvascular abnormalities (IRMA). Retina was then placed on a 53-µm Nylon mesh filter
(Sefar America, Buffalo, NY.), washed with a solution containing Glucose (5.5 mM), LGlutamine (2 mM) and 1x MEM Non-Essential Amino Acids Solution (GIBCO, Life
Technologies, Carlsbad, CA). The retinas were then placed into a 25-ml flask containing 100
U/ml of collagenase, Type 1 (Worthington Biochemical Co., Lakewood, NJ) in the abovementioned solution containing 22% Bovine serum albumin (Sigma) and 0.01% Soybean
Trypsin Inhibitor (Sigma). Retinas were then mechanically agitated using a shaker and allowed
to digest at 37ºC for approximately 60 minutes or until no tissue fragments could be seen.
After digestion, cells were centrifuged at 1000 rpm for 5 minutes. The supernatant was
removed and pellet was suspended in fresh media; 1:1 mix of low glucose Dulbecco’s
modified Eagle’s medium (DMEM 1g/L) / F-12 nutrient mix (GIBCO, Life Technologies,
Carlsbad, CA) supplemented with 10 % fetal calf serum (Hyclone, Logan, UT), 1% endothelial
cell growth supplement (Millipore, MA), 1% insulin-transferrin- sodium selenite media
supplement (Sigma) and 1% penicillin-streptomycin antimycotic (GIBCO, Invitrogen-Life
Technologies, Carlsbad, CA). Glucose concentration in the final media was adjusted to 5.5
mM. The cells were maintained at 37°C in 95% air and 5% CO2 in a humidified cell culture
incubator. Passages 3 to 5 were used in the experiments.

	
  
	
  

115

3.3.3 Human CD34+ CACs isolation
Human peripheral blood samples (150 ml) were collected into Sodium Citrate-containing
CPTTM glass vacuum tubes (BD, Franklin Lakes, NJ). Written informed consent was obtained
from each patient, and all procedures were approved by the Institutional Review Board at the
University of Florida (IRB # 408-2010). Peripheral blood mononuclear cells (MNCs) were
isolated from the blood by density gradient centrifugation using Lympholyte (Cedarlane
Laboratories Ltd., Ontario, Canada). The CD34+ cell fraction was then isolated from the MNCs
using the EasySepTM CD34+ positive selection system according to the manufacturer’s
instructions (Stem Cell Technologies, Vancouver, BC, Canada). Clinical characteristics of the
patients are presented in Table 3.
3.3.4 Tube formation assay
Tube formation assay was performed using BD BioCoat Angiogenesis System-Endothelial Cells
Tube Formation Matrigel Matrix 96-well plate (BD Biosciences Discovery Labware, Bedford,
MA) according to the manufacturer’s instructions. Briefly, isolated CD34+ CACs and HRECs
were labeled with Qtracker 655 and Qtracker 525 (Invitrogen); respectively. Control or diabetic
HRECs were mixed in a 4:1 ratio with either control or diabetic CD34+ cells, seeded into
Matrigel Matrix 96-well plate) and incubated for 16 to 18 h at 37°C (5% CO2). After incubation,
wells were assessed for the presence of tube-like structures and images were taken in 10x
magnifications using a Nikon TE2000 fluorescence microscope equipped with Photometrics
CoolSNAP HQ2 camera. At least three different fields were randomly selected and captured to
collect images for each well. Tube area and percentage of CD34+ incorporated into tubules were
calculated using MetaMorph software system (Molecular Devices, Downingtown, PA). Statistics

	
  
	
  

116

were performed on 3 independent wells per condition with minimum three images taken from
each well.
3.3.5 Quantitative real-time PCR
Total RNA was extracted from HRECs, human CD34+ cells using QuickGene RNA (Fujifilm,
Minato-Ku, Tokyo, Japan) or Qiagen RNeasy (Qiagen Inc., Valencia, CA, USA) according to
the manufacturer’s instructions. NanoDrop 2000 (Thermo Scientific, IL, USA) was used to
determine total RNA concentration. Total RNA was reverse transcribed into cDNA using
superscript III first-strand synthesis system (Invitrogen, Carlsbad, CA). Human gene-specific
primers for ASM were used. Expression levels were normalized to human cyclophilin. Sequence
of specific primers used is given below:
Human ASM:

caacctcgcgctgaagaa

Human Cyclophilin: aaggtcccaaagacagcaga

and
and

tccaccatgtcatcctcaaa
cttgccaccagtgccattat

3.3.6 ELISA assay
Blood samples were collected, centrifuged and plasma was stored at -80 °C. Samples were
assayed for human ASM concentration using ELISA kit (Cloud –Clone Corp., Houston, TX,
USA) according to the manufacturer’s protocol.
3.3.7 Statistical analysis
Data are presented as mean ± S.E.M. Results were analyzed for statistical significance by the
Student’s t-test (GraphPad Prism 7, GraphPad Software, San Diego, CA). Significance was
established at P < 0.05.

	
  
	
  

117

Table 2. Clinical characteristics of control and diabetic subjects involved in the circadian
study.
Table. 1 Clinical characteristics of control and diabetic subjects involved in the circadian study.

Subject
#

Gender

Age

Diabetes
duration

HbA1C

CVD

Retinopathy

Medications

Diabetic 1

Female

67

T2D
4 months

7.5

No

No

Glucophage

Diabetic 2

Female

65

T2D 20
years

8.8

No

No

Lantus

Diabetic 3

Male

76

T2D

7.4

No

No

Glucophage

Diabetic 4

Female

64

T2D

6.3

No

No

Actos

Diabetic 5

Female

67

T2D

6.4

No

No

Glucophage

Diabetic 6

Female

68

T2D 9
years

6.8

No

No

Diabeta

Diabetic 7

Female

48

T2D years

6.0

No

No

Metformin

Diabetic 8

Female

59

T2D 10
years

7.7

No

No

Humulin

Control 1

Female

43

No

No

Control 2

Female

57

No

No

Control 3

Female

68

No

No

Control 4

Male

43

No

No

Inclusion criteria: Individuals between the ages of 21 and 65 years were eligible to participate.
Exclusion criteria: Subjects were excluded for the following reasons: a) evidence of ongoing acute or chronic infection (HIV, Hepatitis B or C, or tuberculosis); b
ongoing malignancy; c) cerebral vascular accident or cerebral vascular procedure; d) current pregnancy; e) history of organ transplantation; f) presence of a graft; g
uremic symptoms, such as an estimated glomerular filtration rate of less than 20 cc/min (by Modification of Diet in Renal Disease equation), or an albumin of less than 3.6
(to avoid malnutrition as a confounding variable); h) a history of smoking; and i) anemia.

	
  
	
  

118

Table 2. (cont’d)
Inclusion criteria: Individuals between the ages of 21 and 65 years were eligible to participate.
Exclusion criteria: Subjects were excluded for the following reasons: a) evidence of ongoing
acute or chronic infection (HIV, Hepatitis B or C, or tuberculosis); b) ongoing malignancy; c)
cerebral vascular accident or cerebral vascular procedure; d) current pregnancy; e) history of
organ transplantation; f) presence of a graft; g) uremic symptoms, such as an estimated
glomerular filtration rate of less than 20 cc/min (by Modification of Diet in Renal Disease
equation), or an albumin of less than 3.6 (to avoid malnutrition as a confounding variable); h) a
history of smoking; and i) anemia.

	
  
	
  

119

Table 3. Clinical characteristics of control and diabetic HRECs donors and CD34+ CACs
subjects involved in the study.
Table. 2 Clinical characteristics of control and diabetic HRECs donors and CD34+ CACs subjects involved in the study

HRECs
donor #

	
  
	
  

Gender

Age

Diabetes
duration

CVD

Retinopathy

Nephropathy

Neuropathy

Medications

Diabetic 1

Male

71

T2D 18 years

Yes

Moderate
NPDR

Yes

No

Insulin

Diabetic 2

Male

66

T2D 16 years

Yes

Sever
NPDR

Yes

No

Insulin
Humalog
and
Humalin-N

Diabetic 3

Female

50

T2D 15 years

No

Sever
NPDR

No

No

Insulin

Diabetic 4

Male

73

T2D 6 years

Yes

Mild
NPDR

Yes

Yes

Insulin

Diabetic 5

Male

71

T2D 20 years

Yes

Moderate
NPDR

Yes

No

Insulin

Diabetic 6

Male

62

T2D 22 years

Yes

Mild
NPDR

Yes

No

Insulin

Diabetic 7

Male

70

T2D 15 years

Yes

Moderate
NPDR

Yes

No

Insulin

Control 1

Female

58

No

No

No

No

No

Control 2

Male

56

No

Yes

No

No

No

Control 3

Male

71

No

No

No

No

No

Control 4

Female

52

No

No

No

No

No

120

Table 3. (cont’d)

CD34+ CACs
subject #

	
  
	
  

Gender

Age

Diabetes
duration

HbA1C

CVD

Retinopathy

Nephropathy

Neuropathy

Diabetic 1

Male

55

T2D 10 years

14

No

Sever
NPDR

No

No

Diabetic 2

Female

58

T2D

6.5

No

Moderate
NPDR

No

No

Diabetic 3

Female

59

T2D

7.1

No

Moderate
NPDR

No

No

Diabetic 4

Male

41

T2D

14

No

PDR

No

No

Control 1

Female

57

No

No

No

No

No

Control 2

Male

49

No

No

No

No

No

Control 3

Female

39

No

No

No

No

No

Control 4

Female

58

No

No

No

No

No

121

3.4 Results
3.4.1 ASM expression level is increased in HREC, CD34+ CACs and blood plasma of
diabetic donors
To determine whether human diabetic tissues exhibited the same increase in ASM as we
observed in animal models (15,33), we measured ASM expression level in human RECs, CD34+
CACs and plasma samples in both diabetic and control donors. ASM expression level was
significantly increased in all three tissue types in diabetic compared to control donors (Fig. 16A,
B and C).

A)

0.015
0.010
0.005
0.000

Control plasma

Diabetic HRECs

Diabetic CD34 +

Diabetic plasma

*

0.012
0.011
0.010
0.009
0.008
0.008
0.007
0.006
0.005
0.004

2.0

*

ASMase conc. (ng/ml)
(relative to control)

0.020

C)
Control CD34 +

ASMase/cyclophilin

ASMase/cyclophilin

0.025

B)
Control HRECs

0.00006
0.00004
0.00002
0.00000

*

1.5
1.0
0.5
0.0

Figure 16. Diabetes induced increase in ASM expression.
Total RNA was isolated and the transcript level of ASM was analyzed by qRT-PCR in (A)
HRECs, (B) CD34+ cells and (C) plasma from diabetic donors (n=4-7) compared with control
donors (n=3-4). Data are means ± SEM. *P < 0.05, significantly different as determined by
Student’s t–test. Abbreviations: ASM, acid sphingomyelinase; HRECs, human retinal endothelial
cells.

	
  
	
  

122

3.4.2 Diabetes induces decrease in HREC tube formation
As shown above and previously demonstrated, HRECs isolated from diabetic donors have high
ASM activity and expression level. To evaluate the effect of diabetes-induced increase in ASM
on HRECs function, we performed tube formation assay to measure the ability of retinal
endothelial cells to form blood-vessel-like tubular structure. Tube formation by HRECs isolated
from healthy control retinas was compared to cells isolated from retinas with signs of NPDR as
determined by post-mortem retinal imaging (Fig. 17A). A significant decrease in tube area was
observed in HRECs from retinas with signs of NPDR as compared to control HRECs (Fig. 17B).

	
  
	
  

123

A)

B)
Control HRECs
Diabetic HRECs

Tube Area µ m 2

500000

*

400000
300000
200000
100000
0

Figure 17. Diabetes impairs tube formation capacity of HRECs.
(A) Postmortem imaging of human retina. Control retina with well-organized blood vessels
(left), Diabetic retina with signs of NPDR; intraretinal hemorrhages and microaneurysms (right).
(B) An in vitro tube formation assay was performed in control (n=4) and diabetic (n=7) HRECs
using Matrigel Matrix 96-well plate. Representative images of tube-like structures are shown.
The cells were stained with Qtracker 525 (green), images were taken in 10x magnification and
total tube areas were calculated using MetaMorph software system. Quantification of tube area is
shown on far right. Data are means ± SEM. *P < 0.05, significantly different as determined by
Student’s t –test. Scale bar = 50 µm.

	
  
	
  

124

3.4.3 Diabetes induced increase in ASM is associated with CD34+ CACs dysfunction
To determine the role of ASM expression in diabetes-induced defect in CD34+ CACs function,
we seeded CD34+ CACs isolated from both control (low ASM) and diabetic (high ASM) subjects
with HRECs and examined whether the level of ASM expression in CACs affect their ability to
incorporate into the endothelial tubes formed by the HRECs.	
   Interestingly, CD34+ CACs seeded
alone did not form tube-like structures, but they did incorporate into tubes formed by HRECs
when co-cultured with retinal endothelial cells (Fig. 18C). Increased incorporation into tubes
formed by diabetic HRECs was observed for the control CD34+ CACs (low ASM) compared to
diabetic CACs (high ASM; Fig. 18B). As expected, control HRECs exhibited robust tube
formation. Incorporation of CACs into control HREC tubes was not affected by the level of
ASM in CACs (Fig. 18A). These data demonstrate that high ASM expression levels in CD34+
CACs correlate with impaired incorporation ability, whereas CACs expressing lower levels of
ASM display enhanced in vitro incorporation.

	
  
	
  

125

A)
Control HRECs

Control HRECs + Control CD34+

Control HRECs + Control CD34+

Control HRECs + Diabetic CD34+

% of CD34 +incorporation

Control HRECs + Diabetic CD34+
10
8
6
4
2
0

B)
Diabetic HRECs + Control CD34+

Diabetic HRECs + Control CD34+
+
Diabetic HRECs + Diabetic CD34

Diabetic HRECs + Diabetic CD34+

% of CD34 +incorporation

Diabetic HRECs

C)

CD34+ alone

ns

25

****

20
15
10
5
0

CD34+ co-cultured with HRECs
CD34+

HRECs

Overlay

Cd

Figure 18. Reduced incorporation of diabetic CD34+ CACs into diabetic HRECs tubes.
Tube formation by HRECs (Qtracker 525, green) isolated from control (A) or diabetic (B)
donors either without CACs (left panel), or co-incubated with control (middle panel) or diabetic
(right panel) CACs (Qtracker 655, red) is shown. Quantification of % of CD34+ CACs
incorporation into HRECs tubes is shown on far right. Data are means ± SEM (n= 4-7). *** P <
0.0001, significantly different from control as determined by Student’s t –test; not significant at
P > 0.05. Scale bar = 50 µm.

	
  
	
  

126

Figure 18 (cont’d)
(C) CD34+ CACs alone were not able to form tube-like structures (left panel), but incorporated
into HREC tubes, forming tube-like structures, when co-cultured with HRECs (right three
panels). Abbreviations: CACs, circulating angiogenic cells.

	
  
	
  

127

3.4.4 Diabetes induced increase in ASM is associated with loss of circadian release of CD34+
CACs
We have previously demonstrated that normal diurnal pattern of CACs release from the bone
marrow is critical for efficient repair of retinal vasculature in rodents (7). Increase in ASM
activity in CACs in diabetic animal models was associated with decreased membrane fluidity
and impaired migration leading to increased CAC retention and loss of circadian release from the
bone marrow (34). We next determined the effect of diabetes on circadian release of CD34+
CACs in diabetic patients. Peripheral blood of type 2 diabetic individuals was collected every 2 h
for 24 h and analyzed for the number of CD34+ CACs by flow cytometry and compared with
control subjects. The dash line is the model fitted curve for individual subjects and bold curve is
the fitted curve for population (Fig. 19 A and B). In agreement with previous studies, healthy
individuals had a peak of circulating CD34+ cells in the middle of the night, representing the rest
phase for humans (Fig. 19A), however this peak of CD34+ release was lost in T2D subjects (Fig.
19B)

	
  
	
  

128

A)

B)

Figure 19. Loss of circadian release of diabetic CD34+ CACs.
Peripheral blood was collected every 2 h for a total of 24 h from (A) control (n=4) and (B)
diabetic (n=8) subjects and was analyzed for the number of CD34+ CACs by flow cytomtery. (A)
In control subjects, there is a clear peak of circulating CD34+ CACs that occurred in the middle
of the night. (B) Rhythmic CD34+ CACs release pattern was blunted in Type 2 diabetic patients.
The dash line is the model fitted curve for individual patients and bold curve is the fitted curve
for population.

	
  
	
  

129

3.5 Discussion
Diabetic retinopathy is a sight threatening complication of diabetes with limited treatment
strategies. Understanding the key molecular factors involved in the disease is important for
developing therapeutic targets to prevent progression into ocular neovascularization and
blindness. ASM is shown to be a key element in inflammatory signaling through ceramidemediated signal transduction (41,42). Diabetes–induced increase in ASM activity has been
shown to modulate inflammatory response in mature retinal endothelial cells (43), however,
there is no direct experimental evidence showing ASM effect on endothelial function. Here we
demonstrated that HRECs isolated from type 2 diabetic subjects with signs of NPDR had altered
retinal endothelial cell function with impaired capacity to form tube-like structures when
compared with control HRECs. The deficiency in tube formation was associated with increased
expression of ASM in diabetic HRECs. This is consistent with previous studies showing ASM–
mediated endothelial cell apoptosis in various tissues including retina, lungs and gastrointestinal
tract (30,31,33,44), and ASM antiangiogenic effect in tumor treatment (45). Endothelium is the
major source of ASM production in the body with endothelial cells synthesizing 20 times as
much ASM as any other cell type (32); thus, it is not surprising that ASM plays a major role in
endothelial cells function. In agreement with previous studies (32), endothelial cells had very
high ASM expression level, which was further increased in diabetes. Moreover, increased ASM
level in plasma further reflects the increased production and secretion of ASM by activated
endothelial cells.
Retinal vascular repair and revascularization is aided by CACs (4,13,16,46–49). We
examined the effect of diabetes-induced increase in ASM on the ability of CD34+ CACs to
incorporate and thus help in repair of defective

	
  
	
  

vascular-like tube structures formed by

130

diabetic HRECs. We demonstrated that diabetic CD34+ CACs, with high ASM, showed minimal
incorporation into the defective tubes formed by diabetic HRECs. Non-diabetic CD34+ CACs
with low ASM showed robust incorporation. These results are in line with other studies showing
that diabetic CACs are defective in proliferation, migration, adhesion, differentiation, and
participation in vascular regeneration process (13,14,16,46,50,51). Interestingly, no significant
difference was observed between non-diabetic and diabetic CD34+ CACs incorporation into
control HRECs. This is consistent with our previous studies demonstrating lack of incorporation
of control CACs into healthy vasculature that does not require repair in non-diabetic (13). We
have previously demonstrated that high	
  level	
  of	
  ASM	
  inversely	
  correlates	
  with	
  migration	
  and	
  
reparative	
   capacity	
   of	
   CACs,	
   with	
   the	
   inhibition	
   of	
   diabetes-­‐induced	
   ASM	
   resulting	
   in	
  
improved	
  migration	
  and	
  retinal	
  vascular	
  regeneration	
  in	
  diabetic	
  mouse	
  model	
  (15,34).
In this study, we compared effect of ASM on functional capacity of healthy and diabetic CACs
in human on control and diabetic HRECs. We revealed that the defective incorporation of
diabetic CD34+ CACs was associated with high ASM. As previously shown, high ASM activity
leads to accumulation of membrane ceramides. Accumulation of ceramide results in decreased
membrane fluidity, cell rigidity and defective migration which can explain defective
incorporation of diabetic CACs (15,34,52) and likely supports their defective release from the
bone marrow into the systemic circulation.
Although several lines of evidence demonstrate that ASM expression is an important factor for
progenitor cell release from the marrow, migration, proliferation and homing to the injured
vasculature; other factors beyond ASM activation are also known to be involved in diabetesinduced CAC dysfunction. These include bone marrow neuropathy and low level of
neurotransmitters production; increase in TGF

	
  
	
  

beta leading to decreased NO bioavailability

131

and diminished expression of CXCR-4 (7,53). These factors affect chemoattraction to SDF1 and
lead to increase in stem cell quiescence in the bone marrow niche. In health ASM works in
concert with other factors to maintain optimal bone marrow stem cell production and release,
however activation of ASM along with other factors disrupts this balance in diabetes
(7,34,54,55).
The physiological release of bone marrow progenitor cells including CD34+ CACs into
the circulation is not constant, but follows circadian oscillations. These oscillations are regulated
by sympathetic signaling to bone marrow stromal cells, which results in CXCL12 (SDF-1) down
regulation and progenitor egress from the bone marrow; all of these processes occur in circadian
pattern under control of clock genes (56). Diabetic bone marrow neuropathy with disruption of
circadian rhythm may contribute to endothelial progenitor cell dysfunction. Wang et al. have
demonstrated that mutation in circadian gene Per2 leads to reduced endothelial cell progenitor
mobilization and revascularization (57). We investigated whether circadian release of CD34+
CACs is altered in diabetic subjects. Peripheral blood was collected every 2 h for 24 h from both
control and diabetic subjects. In agreement with previous studies (58–60), normal rhythmic
oscillations were observed in control subjects with a clear peak of circulating CD34+ CACs cells
in the middle of the night (resting phase), optimal time of regeneration in humans. Similar to our
observations in diabetic rat model (7). This study revealed that circadian fluctuation of CD34+
CACs was disrupted with rhythmicity lost in diabetic individuals.

Importantly, we have

previously demonstrated that, similar to endothelial cells, the increase in ASM activity in CACs
in diabetic animal models was associated with increased ceramide levels, decreased membrane
fluidity and impaired migration. This impairment in migration, combined with diminished
sympathetic signaling to BM may lead to

	
  
	
  

increased CAC retention in the bone marrow

132

and loss of circadian release of the progenitor cells as demonstrated in this study.
This study underscores the deleterious effect of high ASM levels on the vascular repair
potential of both mature retinal endothelial cells and circulating angiogenic cells in diabetes.
Correcting this defect could treat vasodegeneration, enhancing vessel repair and thus preventing
progression into PDR stage.

	
  
	
  

133

REFERENCES

	
  
	
  

134

REFERENCES

1. Hammes HP, Feng Y, Pfister F, Brownlee M. Diabetic retinopathy: Targeting vasoregression.
Diabetes. 2011;60(1):9–16.
2. Lorenzi M, Gerhardinger C. Early cellular and molecular changes induced by diabetes in the
retina. Diabetologia. 2001;44(7):791–804.
3. Wong K. Defining diabetic retinopathy severity. Diabet Retin Evidence-Based Manag.
2010;10:105–20.
4. Grant MB, May WS, Caballero S, Brown G a J, Guthrie SM, Mames RN, et al. Adult
hematopoietic stem cells provide functional hemangioblast activity during retinal
neovascularization. Nat Med. 2002;8(6):607–12.
5. Kocher AA, Schuster MD, Szabolcs MJ, Takuma S, Burkhoff D, et al. Neovascular ization of
ischemic myocardium by human bone - marrow – derived angioblasts p revents
cardiomyocyte apoptosis , reduces remodeling and improves cardiac function. NAt
Med.200;7(4):430-6.
6. Kalka C, Masuda H, Takahashi T, Kalka-Moll WM, Silver M, Kearney M, et al.
Transplantation of ex vivo expanded endothelial progenitor cells for therapeutic
neovascularization. Proc Natl Acad Sci U S A. 2000;97(7):3422–7.
7. Busik J V, Tikhonenko M, Bhatwadekar A, Opreanu M, Yakubova N, Caballero S, et al.
Diabetic retinopathy is associated with bone marrow neuropathy and a depressed peripheral
clock. J Exp Med. 2009;206(13):2897–906.
8. Kusuyama T, Omura T, Nishiya D, Enomoto S, Matsumoto R, Takeuchi K, et al. Effects of
treatment for diabetes mellitus on circulating vascular progenitor cells. J Pharmacol Sci.
2006;102:96–102.
9. Fadini GP, Pucci L, Vanacore R, Baesso I, Penno G, Balbarini A, et al. Glucose tolerance is
negatively associated with circulating progenitor cell levels. Diabetologia. 2007;50(10):2156–
63.
10. Hazra S, Jarajapu YPR, Stepps V, Caballero S, Thinschmidt JS, Sautina L, et al. Long-term
type 1 diabetes influences haematopoietic stem cells by reducing vascular repair potential and
increasing inflammatory monocyte generation in a murine model. Diabetologia.
2013;56(3):644–53.
11. Loomans CJ, de Koning EJ, Staal FJ, Rookmaaker MB, Verseyden C, de Boer HC, et al.
Endothelial progenitor cell dysfunction: a novel concept in the pathogenesis of vascular
complications of type 1 diabetes. Diabetes.
2004;53(1):195–9.
	
  
	
  

135

12. Loomans C, van Haperen R, Duijs JM, Verseyden C, de Crom R, et al. Diffrentaion of bone
marrow-derived endothelial progenitor cells is shifted into a proinflammatory phenotype by
hyperglycemia. Mol Med. 2009;15(5–6):152-9.
13. Caballero S, Sengupta N, Afzal A, Chang KH, Li Calzi S, Guberski DL, et al. Ischemic
vascular damage can be repaired by healthy, but not diabetic, endothelial progenitor cells.
Diabetes. 2007;56(4):960–7.
14. Segal MS, Shah R, Afzal A, Perrault CM, Chang K, et al. Nitric oxide cytoskeletal-induced
alterations reverse the endothelail progenitor cell migratory defect associated with diabtes.
Diabetes.2006;55(1):102-9.
15. Tikhonenko M, Lydic TA, Opreanu M, Li Calzi S, Bozack S, McSorley KM, et al. N-3
Polyunsaturated Fatty Acids Prevent Diabetic Retinopathy by Inhibition of Retinal Vascular
Damage and Enhanced Endothelial Progenitor Cell Reparative Function. PLoS One.
2013;8(1):1–10.
16. Tamarat R, Silvestre J-S, Le Ricousse-Roussanne S, Barateau V, Lecomte-Raclet L, Clergue
M, et al. Impairment in ischemia-induced neovascularization in diabetes: bone marrow
mononuclear cell dysfunction and therapeutic potential of placenta growth factor treatment.
Am J Pathol. 2004;164(2):457–66.
17. Kern TS. Contributions of inflammatory processes to the development of the early stages of
diabetic retinopathy. Exp Diabetes Res. 2007;2007:95103.
18. Kowluru RA. Role of interleukin-1β in the pathogenesis of diabetic retinopathy. Br J
Ophthalmol. 2004;88(10):1343–7.
19. Kowluru RA, Odenbach S. Role of interleukin-1β in the development of retinopathy in rats:
Effect of antioxidants. Investig Ophthalmol Vis Sci. 2004;45(11):4161–6.
20. Joussen AM, Poulaki V, Mitsiades N, Kirchhof B , Koizumi K, et al. Nonsteroidal antiinflammatory drugs prevent early diabetic retinopathy via TNF-alph suppression.FASEB J.
2002;16(13):438-40.
21. Joussen AM, Doehmen S, Le ML, Koizumi K, Radetzky S, Krohne TU, et al. TNF-alpha
mediated apoptosis plays an important role in the development of early diabetic retinopathy
and long-term histopathological alterations. Mol Vis. 2009;15(August 2007):1418–28.
22. Demircan N, Safran BG, Soylu M, Ozcan a a, Sizmaz S. Determination of vitreous
interleukin-1 (IL-1) and tumour necrosis factor (TNF) levels in proliferative diabetic
retinopathy. Eye (Lond). 2006;20(12):1366–9.
23. Joussen AM, Poulaki V, Qin W, Kirchhof B, Mitsiades N, Wiegand SJ, et al. Retinal vascular
endothelial growth factor induces intercellular adhesion molecule-1 and endothelial nitric
oxide synthase expression and initiates early
diabetic retinal leukocyte adhesion in vivo.
	
  
	
  

136

Am J Pathol. 2002;160(2):501–9.
24. Adamiec-Mroczek J, Oficjalska-Młyńczak J. Assessment of selected adhesion molecule and
proinflammatory cytokine levels in the vitreous body of patients with type 2 diabetes--role of
the inflammatory-immune process in the pathogenesis of proliferative diabetic retinopathy.
Graefes Arch Clin Exp Ophthalmol. 2008;246(12):1665–70.
25. Miyamoto K, Khosrof S, Bursell SE, Rohan R, Murata T, Clermont AC, et al. Prevention of
leukostasis and vascular leakage in streptozotocin-induced diabetic retinopathy via
intercellular adhesion molecule-1 inhibition. Proc Natl Acad Sci U S A. 1999;96(19):10836–
41.
26. Semeraro F, Cancarini A, Dell’Omo R, Rezzola S, Romano MR, Costagliola C. Diabetic
retinopathy: Vascular and inflammatory disease. J Diabetes Res. 2015;2015:1-16.
27. Liu P, Anderson RG. Compartmentalized production of ceramide at the cell surface.J Biol
Chem.1995;270(45):27179-85.
28. Mathias S, Younes A, Kan C, Orlow I, Joseph C, Kolesnick RN, et al. Activation of the
Sphingomyelin Signaling Pathway in Intact EL4 Cells and in a Cell-Free System by IL-1β.
Science. 1993;259(5094):519–22.
29. Schutze S, Potthoff K, Machleidt T. TNF Activates Phospholipase Sphingomyelin NF-KB by
Phosphatidylcholine-Specific C-Induced “ Acidic ” Breakdown. Cell.1992;71(5):765–76.
30. Santana P, Peña LA, Haimovitz-Friedman A, Martin S, Green D, McLoughlin M, et al. Acid
sphingomyelinase-deficient human lymphoblasts and mice are defective in radiation-induced
apoptosis. Cell. 1996;86(2):189–99.
31. Haimovitz-Friedman A, Cordon-Cardo C, Bayoumy S, Garzotto M, et al. Lipopolysaccharide
induces disseminated endothelial apoptosis requiring ceramide generation. J Exp Med.
1997;186(11):1831-41.
32. Marathe S, Schissel SL, Yellin MJ, Beatini N, Mintzer R, Williams KJ, et al. Human Vascular
Endothelial Cells Are a Rich and Regulatable Source of Secretory Sphingomyelinase. J Biol
Chem. 1998;273(7):4081–8.
33. Opreanu M, Tikhonenko M, Bozack S, Lydic TA, Reid GE, McSorley KM, et al. The
unconventional role of acid sphingomyelinase in regulation of retinal microangiopathy in
diabetic human and animal models. Diabetes. 2011;60(9):2370–8.
34. Chakravarthy H, Navitskaya S, O'Reilly S, Gallimore J, Mize H, et al. Role of Acid
Sphingomyelinase in Shifting the Balance Between Proinflammatory and Reparative Bone
Marrow Cells in Diabetic Retinopathy. Stem cells. 2016;34(4):972-83.
35. Bhatwadekar AD, Yan Y, Qi X, Thinschmidt
	
  
	
  

137

JS, Neu MB, Li Calzi S, et al. Per2 mutation

recapitulates the vascular phenotype of diabetes in the retina and bone marrow. Diabetes.
2013;62(1):273–82.
36. Bhatwadekar AD, Yan Y, Stepps V, Hazra S, Korah M, Bartelmez S, et al. MiR-92a corrects
CD34+ cell dysfunction in diabetes by modulating core circadian genes involved in progenitor
differentiation. Diabetes. 2015;64(12):4226–37.
37. Roy S, Sala R, Cagliero E, Lorenzi M. Overexpression of fibronectin induced by diabetes or
high glucose: phenomenon with a memory. Proc Natl Acad Sci U S A. 1990;87(1):404–8.
38. Grant MB, Caballero S, Tarnuzzer RW, Bass KE, Ljubimov a V, Spoerri PE, et al. Matrix
metalloproteinase expression in human retinal microvascular cells. Diabetes.
1998;47(8):1311–7.
39. Kowluru RA, Santos JM, Mishra M. Epigenetic modifications and diabetic retinopathy.
Biomed Res Int. 2013;2013:635284.
40. Busik J V., Mohr S, Grant MB. Hyperglycemia-Induced reactive oxygen species toxicity to
endothelial cells is dependent on paracrine mediators. Diabetes. 2008;57(7):1952–65.
41. Gulbins E, Grassmé H. Ceramide and cell death receptor clustering. Biochim Biophys Acta.
2002;1585(2–3):139–45.
42. Marchesini N, Hannun YA. Acid and neutral sphingomyelinases: roles and mechanisms of
regulation. Biochem Cell Biol. 2004;82(1):27–44.
43. Opreanu M, Lydic T a, Reid GE, McSorley KM, Esselman WJ, Busik J V. Inhibition of
cytokine signaling in human retinal endothelial cells through downregulation of
sphingomyelinases by docosahexaenoic acid. Invest Ophthalmol Vis Sci. 2010;51(6):3253–
63.
44. Paris F, Fuks Z, Kang A, Capodieci P, Juan G, Haimovitz-friedman A, et al. Endothelial
apoptosis as the the Primary Lesion Initiating Intestinal Radiation Damage in Mice.
Science.2001;293(5528):293–7.
45. Garcia-barros AM, Paris F, Cordon-cardo C, Lyden D. Tumor Response to Radiotherapy
Regulated by Endothelial Cell Apoptosis.Science .2003;300(5622):1155–9.
46. Schatteman GC, Hanlon HD, Jiao C, Dodds SG, Christy BA. Blood-derived angioblasts
accelerate blood-flow restoration in diabetic mice. J Clin Invest. 2000;106(4):571–8.
47. Asahara T, Masuda H, Takahashi T, Kalka C, Pastore C, Silver M, et al. Bone marrow origin
of endothelial progenitor cells responsible for postnatal vasculogenesis in physiological and
pathological neovascularization. Circ Res. 1999;85(3):221–8.
48. Asahara T, Murohara T, Sullivan A, Silver
	
  
	
  

M, Van R, Asahara T, et al. Isolation of

138

Putative Progenitor Endothelial Cells for Angiogenesis Science.1997;275 (5302):964–7.
49. Takahashi T, Kalka C, Masuda H, Chen D, Silver M, et al. Ischemia - and cytok ine - induced
mobilization of bone marrow- derived endothelial progenitor cells fo r neovascularization.Nat
Med. 1999;5(4):434-8.
50. Tepper OM, Galiano RD, Capla JM, Kalka C, Gagne PJ, Jacobowitz GR, et al. Human
endothelial progenitor cells from type II diabetics exhibit impaired proliferation, adhesion,
and incorporation into vascular structures. Circulation. 2002;106(22):2781–6.
51. Chen Y, Lin S, Lin F, Wu T, Tsao C. High Glucose Impairs Early and Late Endothelial
Progenitor Cells by Modifying Nitric Oxide – Related but Not Oxidative Stress – Mediated
Mechanisms. Diabetes. 2007;56(June):1559–68.
52. Catapano ER, Arriaga LR, Espinosa G, Monroy F, Langevin D, López-Montero I. Solid
character of membrane ceramides: A surface rheology study of their mixtures with
sphingomyelin. Biophys J. 2011;101(11):2721–30.
53. Bhatwadekar AD, Guerin EP, Jarajapu YPR, Caballero S, Sheridan C, Kent D, et al. Transient
inhibition of transforming growth factor-β1 in human diabetic CD34+ cells enhances vascular
reparative functions. Diabetes. 2010;59(8):2010–9.
54. Chakravarthy H, Beli E, Navitskaya S, O’Reilly S, Wang Q, Kady N, et al. Imbalances in
mobilization and activation of pro-inflammatory and vascular reparative bone marrow-derived
cells in diabetic retinopathy. PLoS One. 2016;11(1):1–21.
55. Wang Q, Navitskaya S, Chakravarthy H, Huang C, Kady N, Lydic TA, et al. Dual AntiInflammatory and Anti-Angiogenic Action of miR-15a in Diabetic Retinopathy.
EBioMedicine. 2016;11:138–50.
56. Méndez-Ferrer S, Lucas D, Battista M, Frenette PS. Haematopoietic stem cell release is
regulated by circadian oscillations. Nature. 2008;452(7186):442–7.
57. Wang CY, Wen MS, Wang HW, Hsieh IC, Li Y, et al. Increased vascular senescence and
impaired endothelial progenitor cell function mediated by mutation of circadian gene Per2.
Circulation 2008; 118(21):2166-73.
58. Tsinkalovsky O, Smaaland R, Rosenlund B, Sothern RB, Hirt A, Steine S, et al. Circadian
variations in clock gene expression of human bone marrow CD34+ cells. J Biol Rhythms.
2007;22(2):140–50.
59. Mendez-Ferrer S, Andrew C, Merad M, Frenett PS. Circadian rhytms influence hematopoietic
stem cells.Curr Opin Hematol. 2009;16(4):235–42.
60. Abrahamsen JF, Smaaland R, Sothern RB, Laerum OD. Variation in cell yield and
proliferative activity of positive selected
human CD34+ bone marrow cells along the
	
  
	
  

139

circadian time scale. Eur J Haematol. 1998;60(1):7–15.

	
  
	
  

140

Chapter 4. Summary and future perspectives
4.1 Summary

•

The studies presented in chapter 2 are conceptually innovative, showing that not only
does dyslipidemia participate in diabetic retinopathy but also it has a role in blood retinal
barrier (BRB). We explore a novel concept that ELOVL4 plays an important role in the
blood retinal barrier (BRB) maintenance via the production and incorporation into the
tight junctions of specific species of ceramides previously found in skin permeability
barrier; omega-linked acyl-VLC ceramides. We proposed that diabetes induced downregulation of retinal ELOVL4 results in reduced production of very long chain ceramides.
This in turn contributes to increased retinal vascular permeability.

Using Bovine retinal endothelial cells (BRECs) as a model of inner blood retinal barrier,
overexpression of hELOVL4 using adenoviral vector system significantly decreased
basal permeability and prevented VEGF and IL-1β-induced increase in permeability.
Importantly, hELOVL4 overexpression prevented VEGF-induced decrease in occludin
expression and border staining of tight junction proteins; ZO-1 and claudin-5.
Interestingly, knock down of ELOVL4 was associated with increased paracellular
permeability. Ceramides were found to co-localize with tight junction complexes.
Lipidomic analysis of tight junction isolates revealed the presence of VLC ceramides.

In order to determine the effect of ELOVL4 on diabetes-induced blood retinal barrier
dysfunction in vivo, we used streptozotocin mouse model of diabetic retinopathy treated

	
  
	
  

141

with intravitreal injections of hELOVL4-AAV2. We demonstrated that AAV2 mediated
delivery of hELOVL4 prevents diabetes-induced increase in retinal vascular
permeability. These findings are consistent with the results obtained in BRECs,
supporting the curative potential of hELOVL4 vector.

•

Continuing the studies in Chapter 2 where we investigated the role of diabetes-induced
dyslipidemia in retinal vascular permeability, Chapter 3 describes the effect of
dyslipidemia on the function of both cell types involved in revascularization; mature
retinal endothelial cells and circulating angiogenic cells (CACs).

Using primary cultures of human retinal endothelial cells (HRECs) isolated from control
and diabetic donor tissue and human CD34+ CACs from control and diabetic patients. We
demonstrated that diabetes induced ASM activation, a key enzyme of sphingolipid
metabolism, in retinal endothelial cells and CD34+ CACs. Diabetes induced ASM
upregulation has deteriorative effect on both retinal endothelial cell and CD34+ function.
Where diabetic HRECs with high ASM showed defective ability to form blood-vessellike tubular structure. Moreover, diabetic CD34+ CACs with high ASM showed defective
incorporation into endothelial tubes.	
  

•

Collectively, these data indicate that dysregualtion of retinal sphingolipid metabolism
paly role in the development of diabetic retinopathy. Normalizing the ratio between probarrier long chain ceramides and pro-inflammatory and pro-apoptotic short chain
ceramides

	
  
	
  

through

the

control

of

ELOVL4 and ASM may represent a novel

142

therapeutic strategy for the prevention of diabetic visual impairment.
4.2 Future perspectives
1. We demonstrated the presence of omega-linked acyl VLC ceramides in native tight junction.
However, it is important to identify the changes in the abundance of these species of VLC
ceramides in cells transduced with or without recombinant ELOVL4 and in ELOVL4 knock
down cells.
2. Several studies have shown the protective anti-apoptotic effect of VLC ceramides. In this
study we have postulated that ELOVL4 produced VLC ceramides may act as bridges between
tight junction proteins stabilizing these proteins against degradation. However, it would be useful
to identify if there are different mechanisms by which ELOVL4 can reduce paracellular
permeability through neutralizing cytokine or growth factor induced increase in permeability.
3. The exact confirmation of VLC ceramides in the tight junction structure is not known. It
would be important to visualize the tight junction structure in its native environment to show the
exact configuration of lipid in interaction with tight junction protein. This can be achieved using
the cryo-electron microscopy technique.
4. Our lab has previously demonstrated that increase in ASM activity in diabetic CACs in animal
model was associated with accumulation of membrane short chain ceramides, thereby decreasing
membrane fluidity and impairs migration of CACs. Hence it is important to confirm the effect of
high ASM on membrane ceramide levels and membrane rigidity in human CD34+ CACs isolated
from diabetic individuals.

The potential use of gene therapy using AAV serotype 2 as a vector for the treatment of
different retinal diseases is being evaluated in

	
  
	
  

several clinical trials. In this context, a gene-

143

based therapy involving intravitreal injection of AAV2 expressing hELOVL4 in diabetic patients
during their preclinical stage could have significant utility in the treatment of diabetic
retinopathy. We examined the beneficial effect of ELOVL4 in vitro, using the bovine retinal
endothelial cells as model of the inner blood retinal barrier, and in vivo through STZ induced
diabetic mice as a model of type I diabetes.
Before progressing to clinical trials in humans we need to perform additional, confirmatory
preclinical studies addressing basic questions on AAV2-mediated hELOVL4 expression; dosing,
safety, toxicity, tissue distribution after intravitreal administration.

As we examined its

protective effect on a type I diabetic model, it would be important to test its beneficial effect on
several animal models of diabetic retinopathy; type II diabetic mice to confirm its effect in both
types of diabetes. Then shifting to large animal model of Diabetic Retinopathy (DR), such as
diabetic pigs. The pig model has the advantage that it posses retinal vascular system very similar
to that of humans. Most importantly, the pig model develops signs of DR within 18 week after
induction of diabetes compared to the dog model and rhesus monkey model, which develop
retinopathy after years of diabetes induction.
After these additional studies, including safety, tolerability and therapeutic dosing, a pilot study
could be carried in small number of human subjects to ensure safety and tolerability before
progressing into phase I clinical trial. The available evidence that rAAV mediated-gene therapy
holds great promise in treatment of different retinal diseases with great safety and long lasting
efficacy, suggests that delivery of hELOVL4 could serve as a novel gene therapeutic target for
DR patients.

	
  
	
  

144

Overall, our studies indicate that correcting sphingolipid metabolism in diabetic retina and CACs
could treat vasodegeneration, enhance vessel repair and thus prevent progression into PDR stage.

	
  

	
  
	
  

145