Content uploaded by Barry Hudson
Author content
All content in this area was uploaded by Barry Hudson
Content may be subject to copyright.
The RAGE Axis in Early Diabetic Retinopathy
Gaetano R. Barile,
1
Sophia I. Pachydaki,
1
Samir R. Tari,
1
Song E. Lee,
1
Christine M. Donmoyer,
1
Wanchao Ma,
1
Ling Ling Rong,
2
Loredana G. Buciarelli,
2
Thoralf Wendt,
2
Heidi Ho¨rig,
2
Barry I. Hudson,
2
Wu Qu,
2
Alan D. Weinberg,
3
Shi Fang Yan,
2
and Ann Marie Schmidt
2
PURPOSE. The receptor for advanced glycation end products
(AGEs) has been implicated in the pathogenesis of diabetic
complications. This study was conducted to characterize the
role of the RAGE axis in a murine model of nonproliferative
diabetic retinopathy (NPDR).
M
ETHODS. The retinas of hyperglycemic, hyperlipidemic
(HGHL, apolipoprotein E
⫺/⫺
db/db) mice were examined for
the development of early retinal vascular lesions of NPDR and
compared to littermates at 6 months of age. Neural function
was assessed with electroretinography. Immunohistochemis-
try, real-time RT-PCR, autofluorescence, and ELISA studies
were used to localize and quantify the AGE/RAGE axis. Soluble
RAGE, a competitor of cellular RAGE for its ligands, was ad-
ministered to assess the impact of RAGE blockade.
R
ESULTS. Early inner retinal neuronal dysfunction, manifested
by prolonged latencies of the oscillatory potentials and b-wave,
was detected in hyperglycemic mice. HGHL mice exhibited
accelerated development of acellular capillaries and pericyte
ghosts compared with littermate control animals. AGEs were
localized primarily to the vitreous cavity and internal limiting
membrane (ILM) of the retina, where they were intimately
associated with the footplates of RAGE-expressing Mu¨ller cells.
AGE accumulation measured by ELISA was increased within
the retinal extracellular matrix of hyperglycemic mice. AGE
fluorescence and upregulation of RAGE transcripts was highest
in the retinas of HGHL mice, and attenuation of the RAGE axis
with soluble RAGE ameliorated neuronal dysfunction and re-
duced the development of capillary lesions in these mice.
C
ONCLUSIONS. In early diabetic retinopathy, the RAGE axis,
comprising the cellular receptor and its AGE ligands, is ampli-
fied within the retina and is accentuated along the vitreoretinal
interface. Antagonism of the RAGE axis in NPDR reduces neu-
rovascular perturbations, providing an important therapeutic
target for intervention. (Invest Ophthalmol Vis Sci. 2005;46:
2916–2924) DOI:10.1167/iovs.04-1409
D
iabetic retinopathy, the leading cause of irreversible blind-
ness in the working population in the Western world,
encompasses both vascular and neural dysfunction.
1
Diabetes
mellitus leads to alterations in the perfusion and permeability
of the retinal vasculature, resulting in retinal ischemia and/or
edema, with loss of reading vision when these events occur in
the central macular region.
2
Diabetic retinopathy is also a
degenerative disease of the neural retina, associated with alter-
ations in neuronal function before the onset of clinical vascular
disease.
3
In advanced, proliferative diabetic retinopathy, an
angiogenic, VEGF-mediated response with retinal neovascular-
ization ensues, placing the eye at further risk of severe visual
loss due to the development of vitreous hemorrhage or trac-
tion retinal detachment.
4
Although many cases of diabetic
retinopathy may be amenable to treatment with laser photo-
coagulation or vitrectomy, such efforts may not prevent irre-
versible vascular or neuronal damage, thereby underscoring
the need for early intervention.
The duration and severity of hyperglycemia is the single
most important factor linked to the development of diabetic
retinopathy. The degree of hyperglycemia is the major alter-
able risk factor for both the development and progression of
diabetic retinopathy, in both types 1 and 2 diabetes, as seen in
the Diabetes Control and Complications Trial (DCCT)
5
and in
the UK Prospective Diabetes Study (UKPDS),
6
respectively.
Additional established risk factors for the acceleration of dia-
betic retinopathy include hypertension and hyperlipidemia,
with several clinical studies demonstrating benefit in the treat-
ment of diabetic retinopathy with intensive blood pressure
control and lipid-lowering therapy.
7–13
One metabolic conse-
quence of chronic hyperglycemia is the accelerated formation
of advanced glycation end products (AGEs), the accumulation
of which in diabetic tissues is enhanced not only by elevated
glucose but also by oxidant stress and inflammatory stimuli.
14
In the setting of diabetic retinopathy, AGEs, especially N
-
(carboxymethyl)lysine (CML) adducts, have been detected
within retinal vasculature and neurosensory tissue of diabetic
eyes.
15
Multiple consequences of AGE accumulation in the
retina have been demonstrated, including upregulation of
VEGF, upregulation of NF-
B, and increased leukocyte adhe-
sion in retinal microvascular endothelial cells.
16–18
In diabetic
patients, AGEs also accumulate within the vitreous cavity and
may result in characteristic structural alterations sometimes
referred to as diabetic vitreopathy.
19,20
Support for a role for
AGEs as a contributing factor in the pathogenesis of diabetic
retinopathy has been drawn from studies in animals with in-
hibitors of AGE formation.
21,22
In a 5-year study in diabetic
dogs, administration of aminoguanidine prevented retinopa-
thy. Similar beneficial effects in the retinal vasculature of dia-
betic rats have been observed with other inhibitors of AGE
formation, including pyridoxamine and benfotiamine.
23,24
AGEs exert cell-mediated effects via RAGE, a multiligand
signal-transduction receptor of the immunoglobulin superfam-
ily.
25
Coinciding with pathologic changes in tissues, RAGE
From the Departments of
1
Ophthalmology and
2
Surgery, College
of Physicians and Surgeons, and the
3
Department of Biostatistics,
Mailman School of Public Health, Columbia University, New York,
New York.
Supported by the Juvenile Diabetes Research Foundation, the
Columbia Irving Center for Clinical Research, the Eye Surgery Fund,
the Surgical Research Fund, and Research to Prevent Blindness. GRB is
a recipient of the Florence and Herbert Irving Clinical Research Scholar
Award and the Glaubinger Scholar Award. AMS is a recipient of the
Burroughs Wellcome Fund Clinical Scientist Award in Translational
Research.
Submitted for publication December 2, 2004; revised March 10,
2005; accepted March 29, 2005.
Disclosure: G.R. Barile, None; S.I. Pachydaki, None; S.R. Tari,
None; S.E. Lee, None; C.M. Donmoyer, None; W. Ma, None; L.L.
Rong, None; L.G. Buciarelli, None; T. Wendt, None; H. Ho¨rig,
None; B.I. Hudson, None; W. Qu, None; A.D. Weinberg, None; S.F.
Yan, None; A.M. Schmidt, None
The publication costs of this article were defrayed in part by page
charge payment. This article must therefore be marked “advertise-
ment” in accordance with 18 U.S.C. §1734 solely to indicate this fact.
Corresponding author: Gaetano R. Barile, Harkness Eye Institute,
635 West 165th Street, Box 94, New York, New York 10032;
grb17@columbia.edu.
Investigative Ophthalmology & Visual Science, August 2005, Vol. 46, No. 8
2916
Copyright © Association for Research in Vision and Ophthalmology
expression increases dramatically, with AGE ligands further
upregulating receptor expression to magnify local cellular re-
sponses.
26
RAGE also binds the proinflammatory mediators,
the S100/calgranulins, and amphoterin,
27,28
and is an endothe
-
lial cell adhesion receptor capable of promoting leukocyte
recruitment through interaction with the integrin Mac-1.
29
Consequences of ligand-RAGE interaction include increased
expression of vascular cell adhesion molecule (VCAM)-1, vas-
cular hyperpermeability, enhanced thrombogenicity, induc-
tion of oxidant stress and abnormal expression of eNOS, all
pathogenetic mechanisms that potentially contribute to the
ischemic and vasopermeability events of diabetic retinopa-
thy.
30,31
Based on these considerations, we examined the RAGE axis
in a newly characterized murine model of nonproliferative
diabetic retinopathy. We first bred hyperlipidemic apoE
⫺/⫺
mice into the hyperglycemic db/db background, observing
that hyperlipidemia accelerates structural vascular changes in
diabetic retinas that exhibit neuronal dysfunction. We localized
and quantified the RAGE axis—specifically, AGE ligands and
their cellular receptor RAGE, in the eyes of these mice. The
findings provide new insights into the role of the RAGE axis in
the pathogenesis of early diabetic retinopathy.
METHODS
Generation of the Mouse Colony
To generate the apoE
⫺/⫺
db/db mice, apoE
⫺/⫺
mice were first back
-
crossed six generations into mice heterozygous for the diabetes spon-
taneous mutation (Lepr
db
). As the homozygous db/db mouse is sterile,
we ultimately bred apoE
⫺/⫺
db/m offspring to generate apoE
⫺/⫺
db/db mice. Initially, male mice heterozygous for the diabetes sponta-
neous mutation (Lepr
db
) in the leptin receptor gene on chromosome
4 (BKS.Cg-m
⫹/⫹
Lepr
db
, former name C57BLK/J-m
⫹/⫹
Lepr
db
, Type
JAX GEMM TM Strain; Spontaneous Mutation Congenic, stock no.
000642; Jackson Laboratory, Bar Harbor, ME) were crossed with fe-
male mice homozygous for the Apoe
tm1Unc
mutation in chromosome 7
(B6.129P2-Apoe
tm1Unc
, former name C57BL/6J-Apoe
tm1Unc
, Type JAX
GEMM TM Strain; Targeted Mutation Congenic, stock no. 002052;
Jackson Laboratory) at approximately 8 weeks of age. All mice were
fed normal rodent chow (5053; PMI Nutrition International, Inc., St.
Louis, MO) and exposed to a 12-hour light-dark cycle. All offspring
were heterozygous for the apoE mutation. The genotype of their
offspring was identified by PCR with primers from Invitrogen Corp.
(Carlsbad, CA). The heterozygous mice from different parents were
again crossed at 8 weeks of age. Mice homozygous for the Apoe
tm1Unc
mutation and heterozygous for the Lepr
db
mutation (apoE
⫺/⫺
db/m)
were used as breeders and were crossed with one another to breed the
double-knock-out apoE
⫺/⫺
db/db mice. Control mice were littermates
obtained from the same colony: apoE
⫹/⫹
db/m mice (homozygous for
the wild type allele Apoe
tm1Unc
and heterozygous for the db mutation)
which are normoglycemic, nonobese littermates; apoE
⫹/⫹
db/db mice
(homozygous for the wild-type allele Apoe
tm1Unc
and homozygous for
the Lepr
db
mutation) which are hyperglycemic, normolipidemic litter
-
mates. Glucose measurements were performed during the course of
generation of the colony with a glucometer (Freestyle; Therasense,
Alameda, CA). Cholesterol measurements were then performed (Infin-
ity Cholesterol Liquid Stable Reagent kit; Thermo Electron Corp.,
Waltham, MA). The generation of the colony and all experiments were
performed in agreement with the ARVO Statement for the Use of
Animals in Ophthalmic and Vision Research and were approved by the
Institutional Animal Care and Use Committee at Columbia University.
Elastase Retinal Digest
An elastase digest with histopathologic vascular analysis was per-
formed on 35 mice at age 6 months, including analysis of the following
phenotypes: apoE
⫹/⫹
db/m (n ⫽ 7; normoglycemic, normolipidemic
[NGNL]); apoE
⫺/⫺
db/m (n ⫽ 8; normoglycemic, hyperlipidemic
[NGHL]); apoE
⫹/⫹
db/db (n ⫽ 7; hyperglycemic, normolipidemic
[HGNL]); and apoE
⫺/⫺
db/db (n ⫽ 13; hyperglycemic, hyperlipidemic
[HGHL]). At the time of death, the eyes were enucleated and placed in
10% formalin for 2 days. After fixation, the retina was gently dissected
from the neurosensory retina under microscopic observation. The
neurosensory retina was placed in distilled water overnight to remove
fixative. The elastase digestion method described by Laver et al.
32
was
then performed. After the vascular specimen was mounted on a slide,
periodic acid-Schiff and hematoxylin staining of the vascular network
and nuclei was performed. The specimens were then analyzed by
microscope with digital capture (Axioskop 2 Plus; Carl Zeiss MicroIm-
aging Inc., Thornwood, NY) for the presence of acellular capillaries
and pericyte ghosts. Acellular capillaries were at least one-third thick-
ness of normal capillary width, and intercapillary bridges were ex-
cluded from analysis.
33
The examiner was masked to the nature of the
specimen during the assessment of pathology. As vascular lesions may
be distributed nonuniformly, the entire retina was scanned during this
process, and images were pasted into a single image (Photoshop, ver.
7.0; Adobe Systems Inc., San Jose, CA) to obtain an image of wholem-
ounted retina for area calculations. The virtual area of each prepared
retina was measured (OphthaVision Imaging System, ver. 3.25; MRP
Group Inc., Lawrence, MA). The number of acellular capillaries and
pericyte ghosts for each digest was divided by the area scanned. The
data obtained were analyzed with frequency and descriptive statistics,
as described below.
Electrophysiology
Electroretinograms (ERGs) were performed on the following age-
matched, 6-month-old littermates: NGNL wild-type (apoE
⫹/⫹
db/m; n
⫽ 18), NGHL (apo E
⫺/⫺
db/m; n ⫽ 11), HGNL (apoE
⫹/⫹
db/db; n ⫽
8), and HGHL (apoE
⫺/⫺
db/db; n ⫽ 14) mice. The mice were dark
adapted overnight before each experiment, and the ensuing proce-
dures were performed under dim red light in a darkroom. The mice
were anesthetized with a mixture of 50 mg/kg ketamine and 5 mg/kg
xylazine administered intraperitoneally. The right eye pupil was dilated
with drops of 2.5% phenylephrine hydrochloride and 0.5% tropicam-
ide. The electroretinogram (ERG) responses were amplified and aver-
aged by a computerized data-acquisition system (PowerLab; ADInstru-
ments, Colorado Springs, CO). Once anesthetized, the mouse was
placed on a heated block, and body temperature was maintained near
37°C. The mouse was placed in a centered position at the edge of a
Ganzfeld dome. A rectal thermometer was placed in the mouse and
checked throughout the recording. A ground electrode was inserted in
the right leg, and the reference electrode was inserted in the forehead.
The data collected and analyzed included all the above and tempera-
ture of the animal during the experiment, a- and b-wave latency and
amplitude, oscillatory potential (OP)1, OP2, and OP3 implicit times
and amplitudes, as previously described.
34,35
The data obtained were
analyzed with frequency and descriptive statistics, as described later.
Immunochemical Staining
Eyes of 6-month-old mice were fixed overnight in 4% phosphate-
buffered paraformaldehyde and embedded in paraffin. The 4-mm par-
affin sections were deparaffinized and heated in citrate buffer with a
microwave for 15 minutes. After pretreatment with PBS containing 5%
normal goat serum (Jackson ImmunoResearch Laboratories Inc., West
Grove, PA), 0.5% BSA, and 0.1% Triton X-100 for 30 minutes at room
temperature (RT), sections were incubated with anti-mouse RAGE
antibody
36
(1:100), anti-AGE antibody
36
(1:100), anti-vimentin anti
-
body (1:200, Santa Cruz Biotechnology Inc., Santa Cruz, CA), anti-glial
fibrillary acidic protein (GFAP) antibody (1:100, Chemicon Interna-
tional, Inc., Temecula, CA), or anti-CD31 antibody (1:200, BD-Pharmin-
gen, San Diego, CA) for 1 hour at RT and then overnight at 4°C. After
they were rinsed with PBS, the sections were incubated for 1 hour at
RT with secondary antibody conjugated to Alexa Fluor 488 (Molecular
Probes Inc., Eugene, OR) or Alexa Fluor 546 (Molecular Probes Inc.).
All antibodies were diluted in PBS containing 0.5% goat serum, 0.5%
BSA, and 0.1% Triton X-100. Rabbit or chicken serum was used instead
IOVS, August 2005, Vol. 46, No. 8 RAGE in Early Diabetic Retinopathy 2917
of primary antibody for a negative control. The retina was examined
with a microscope (Eclipse E800; Nikon Instruments Inc., Meville, NY)
equipped with a confocal laser scanning system (Radiance2000; Bio-
Rad Laboratories, Hercules, CA). Images were captured and processed
(LaserSharp 2000 software; Bio-Rad Laboratories).
Autofluorescence and ELISA of Retinal AGEs
Five mice from each group were euthanatized. Whole retinas were
homogenized in 0.1 mL of PBS with 0.1% Triton X-100 at 0°C. Samples
were centrifuged at 20,000g for 5 minutes at 4°C. Protein concentra-
tion was determined with BSA used as a standard. The protein level in
the supernatant was adjusted to 1.6 mg/mL and used for a cellular
protein autofluorescence assay. The pellet, mostly extracellular matrix
(ECM), was washed with 20 mM phosphate buffer (pH 7.0), with 10
mM EDTA, and digested with 20
L of 25 U/mL papain (P5306,
Sigma-Aldrich, St. Louis, MO) in 20 mM phosphate buffer (pH 7.0) 10
mM EDTA, and 20 mM cysteine at 37°C. After 24 hours, another 20
L
of papain solution was added, and the incubation was continued for 24
hours. The supernatant was used for the measurement of ECM
autofluorescence and ELISA of AGEs after appropriate dilution. Fluo-
rescence intensities were measured on a multiwell plate reader (Cyto-
Fluor 4000; Applied Biosystems [ABI], Foster City, CA) using 360 ⫾
40/460 ⫾ 40-nm excitation/emission wavelengths. These excitation/
emission wavelengths allow for detection of well-defined AGEs.
37,38
Fluorescence was expressed in fluorescence intensity per 0.1 mg
cellular protein or its equivalent retinal size for ECM. For immuno-
chemical measurement of AGEs in ECM, a noncompetitive ELISA was
used. The wells (96-well Nunc-Immuno Plate; Nalge Nunc Interna-
tional, Rochester, NY) were coated with BSA control, AGE-BSA stan-
dard,
36
and biological samples in 0.1 mL of 50 mM carbonate buffer
(pH 9.6) at 4°C overnight. The wells were then washed with PBS
containing 0.05% Tween-20 (washing buffer) and blocked at room
temperature with 0.3 mL of 1% BSA and 5% rabbit serum in PBS
(blocking buffer) for 1 hour. After they were washed, the wells were
incubated with anti-AGE antibody
36
in blocking buffer for 3 hours at
room temperature, followed by washing and secondary antibody (rab-
bit anti-chicken IgY-HRP; Biomeda Corp, Foster City, CA) for 1 hour at
room temperature. The wells were then washed again and developed
with 0.1 mL of peroxidase substrates (o-phenylenediamine tablets;
Sigma-Aldrich) in the dark at room temperature. The absorbance at 490
nm was measured after adding 0.05 mL of blocking solution (2 M
H
2
SO
4
) at 10 minutes.
Quantitative Real-Time PCR
At least five mice of each group were euthanatized. Retinas were
isolated and stored in pairs at ⫺80°C in preservative (RNAlater; Am-
bion, Inc., Austin, TX). Total RNA was then prepared (RNeasy Minikit;
Qiagen, Inc., Valencia, CA). After quantification at OD
260
, total RNA
was analyzed (RNA Nano LabChips; 2100 Bioanalyzer; Agilent Tech-
nologies, Palo Alto, CA), to assess RNA quality. Only samples showing
minimal degradation were used. cDNA was synthesized (TaqMan Re-
verse Transcription Reagents Kit; ABI) according to the manufacturer’s
instructions. Primers and probes for

-actin and RAGE were designed
on computer (Primer Express; ABI). To confirm specific amplification
of the target mRNA, an aliquot of the PCR product was analyzed by gel
electrophoresis. The sequences of the primers and probe were as
follows: for

-actin, 5⬘-ACG GCC AGG TCA TCA CTA TTG-3⬘ (forward),
5⬘-TGG ATG CCA CAG GAT TCC AT-3⬘ (reverse), and 5⬘-6FAM-ACG
TCT ACC AGC GAA GCT ACT GCC GTC-TAMRA-3⬘ (probe); and for
RAGE, 5⬘-GGA CCC TTA GCT GGC ACT TAG A-3⬘ (forward), 5⬘-GAG
TCC CGT CTC AGG GTG TCT-3⬘ (reverse), and 5⬘-6FAM-ATT CCC GAT
GGC AAA GAA ACA CTC GTG-TAMRA-3⬘ (probe) (ABI). Real-time PCR
was conducted on a sequence-detection system (Prism 7900HTl ABI),
and results were analyzed by the 2
⫺⌬⌬C
T
method.
39
Experiments were
repeated three times, and statistical analysis was performed as de-
scribed later.
Administration of Soluble RAGE
Soluble (s)RAGE, the extracellular two-thirds of the receptor, binds
AGEs and interferes with their ability to bind and activate cellular
RAGE. Preparation, characterization, and purification of sRAGE was
performed with a baculovirus expression system using Sf9 cells (BD-
Clontech, Palo Alto, CA; Invitrogen Corp.) as previously described.
36
Purified murine sRAGE (a single-band of ⬃40 kDa, by Coomassie-
stained SDS-PAGE) was dialyzed against PBS; made free of detectable
endotoxin, based on the Limulus amebocyte assay (E-Toxate; Sigma)
after passage onto gel columns (Detoxi-Gel; Pierce Chemical Co.,
Rockford, IL); and sterile-filtered (0.2
m). We administered daily
doses of 100
g sRAGE based on previous dose–response studies.
27
FIGURE 1. Retinal elastase digest results among diabetic, hyperlipid-
emic, and littermate control mice at age 6 months. The development of
acellular capillaries (A, arrow; B) was accelerated in the retinas of
HGHL mice, with significantly more acellular capillaries present per
unit area than in NGNL or NGHL and HGNL mice. Pericyte ghosts (C,
arrow; D) were also increased in the retinas of HGHL mice compared
with NGNL littermates at age 6 months. Capillary outpouching (E,
arrow), suggesting early microaneurysm formation, was observed in
the retinal vasculature of HGHL mice. An intercapillary bridge, a
normal feature of retinal digests not included in analysis, was also
visible (arrowhead). Results are expressed as the mean ⫾ SEM. *P ⬍
0.05; **P ⬍ 0.01. Scale bar, 50
m.
TABLE 1. Glucose and Cholesterol Level at Euthanatizing*
NGNL NGHL HGNL HGHL
Glucose (mg/dL) 121.3 ⫾ 28.5 (15) 113.8 ⫾ 24.2 (18) 452.6 ⫾ 109.4 (13) 455.5 ⫾ 68.4 (10)
Cholesterol (mg/dL) 62.5 ⫾ 14.4 (5) 471.2 ⫾ 72.4 (15) 201.3 ⫾ 30.4 (5) 955.6 ⫾ 149.1 (15)
Data are expressed as the mean ⫾ SD (n).
* Mice were euthanatized at six months.
2918 Barile et al. IOVS, August 2005, Vol. 46, No. 8
Statistical Analysis
To analyze the vascular, neuronal, and experimental data among the
four groups, we used a two-factor analysis of variance (ANOVA) model.
The two factors considered were glucose (normal/high) and lipid
(normal/high). Interactions were tested for all analyses, but none was
found. A one-way ANOVA was also used to compare the four groups in
analyzing the AGE ELISA and autofluorescence data and the RAGE
q-PCR data. For the experiment involving treatment with sRAGE, a
one-way ANOVA was used to compare the three groups: NGNL, HGHL,
and sRAGE. If a difference was found among the groups (P ⬍ 0.05), a
post hoc analysis using the Duncan test was performed. All data were
analyzed on computer (SAS system software; SAS Institute Inc., Cary,
NC).
RESULTS
Effect of Hyperlipidemia on the Development of
Vascular Lesions of Early Diabetic Retinopathy in
Hyperglycemic Mice
The serum levels of glucose and cholesterol for each of the four
groups are presented in Table 1. We first examined the impact
of the introduction of hyperlipidemia into the hyperglycemic
db/db background on vascular properties in the retina. At age
6 months, the retinas of HGHL (apoE
⫺/⫺
db/db) mice dis
-
played the most significant capillary lesions of NPDR (Fig. 1).
Whereas the eyes of HGNL mice exhibited some development
of acellular capillaries within the retina, only the eyes of HGHL
mice had a significantly higher number of acellular capillaries
than all other groups (Fig. 1B). The development of pericyte
ghosts was detectable in both the HGNL and NGHL pheno-
types, but only in the HGHL mice was there a significant
difference from the NGNL control animals (Fig. 1D). Only in
HGHL mice was there evidence of capillary outpouching con-
sistent with early microaneurysm formation (Fig. 1E).
Electrophysiologic Neural Dysfunction of the
Inner Retina of Hyperglycemic Mice
Electrophysiologic testing at age 6 months revealed that hyper-
glycemia resulted in early inner retinal dysfunction of the
retina detected by prolongation in the latencies of the b-wave
and the OPs (Table 2). Specifically, there were significant
hyperglycemia-induced delays in the implicit time of the b-
wave and OP1, OP2, and OP3 (see Table 4). The ERG ampli-
tudes were not significantly affected in this study, with hyper-
glycemic mice demonstrating a statistically significant decline
only in the amplitude of OP1 (Tables 3, 4). Hyperlipidemia
alone did not induce statistically significant differences in any
of the parameters recorded and studied (Table 4).
The RAGE Axis at the Vitreoretinal Interface
RAGE expression was predominantly localized to glial cells of
the inner retina. Most of the RAGE-expressing cells within the
neural retina were consistent with the distribution of Mu¨ller
cells—particularly their internal footplates. In merged images,
RAGE-positive cells of the inner retina colocalized with vimen-
tin expression, confirming Mu¨ller cell expression (Fig. 2).
GFAP expression in astrocytes of the inner retina revealed no
evidence of colocalization with adjacent RAGE expression of
Mu¨ller cell processes and footplates (Figs. 3A–C). Expression of
RAGE was also detected adjacent to the microvasculature,
suggesting intimate neurovascular localization for RAGE in the
circulation of the inner retina (Figs. 3D, 3E). AGEs were prom-
inently detected within the vitreous cavity of the eye and
particularly along the vitreoretinal interface including the in-
ternal limiting membrane (Figs. 4B, 4F). AGEs were consis-
tently detected within the lens capsule and Bruch’s membrane
and occasionally within the basement membrane of the micro-
vasculature (not shown). In AGE and RAGE merged images,
AGE was localized to vitreous fibrils and the internal limiting
membrane, where there was close apposition to the footplates
of RAGE-expressing Mu¨ller cells (Fig. 4).
RAGE and Its AGE Ligands in NPDR
We next quantified the RAGE axis in this murine model of
NPDR. As AGEs can accumulate within cellular protein as well
as within the proteins of ECM, we assayed the autofluores-
cence of AGEs independently. As shown in Table 5, there was
not a significant difference among groups with regard to AGE
autofluorescence in the cellular protein. In contrast, AGE
autofluorescence increased in ECM in the setting of hypergly-
cemia, but only the retinas of HGHL mice had a significant
difference in fluorescence when compared with NGNL mice.
To quantify AGEs further in the ECM, we performed a non-
competitive ELISA. This study revealed significantly increased
AGE formation in the retinal ECM of hyperglycemic mice, both
HGNL and HGHL (Fig. 5A). As RAGE expression may be am-
plified in the setting of its ligands,
40
RAGE mRNA expression
from whole retina was then examined by quantitative real-time
PCR for each group. RAGE mRNA expression was increased in
TABLE 2. ERG Latencies in Mice at Age 6 Months
NGNL (n ⴝ 18) NGHL (n ⴝ 10) HGNL (n ⴝ 8) HGHL (n ⴝ 14)
b-Wave 32.0 ⫾ 2.0 32.4 ⫾ 4.0 35.3 ⫾ 3.5 34.5 ⫾ 2.9
OP1 23.4 ⫾ 1.4 23.0 ⫾ 2.1 25.4 ⫾ 1.9 24.6 ⫾ 1.7
OP2 32.0 ⫾ 2.0 31.7 ⫾ 3.2 34.8 ⫾ 2.5 33.8 ⫾ 2.2
OP3 42.6 ⫾ 2.9 42.9 ⫾ 5.4 45.4 ⫾ 3.2 44.9 ⫾ 3.1
⌺OPs 98.0 ⫾ 6.2 97.5 ⫾ 10.6 105.6 ⫾ 7.3 103.3 ⫾ 6.8
Data are expressed as the mean latency ⫾ SD (ms).
TABLE 3. ERG Amplitudes in Mice at Age 6 Months
NGNL (n ⴝ 18) NGHL (n ⴝ 10) HGNL (n ⴝ 8) HGHL (n ⴝ 14)
b-Wave 568.4 ⫾ 163.2 517.0 ⫾ 141.0 462.5 ⫾ 138.9 489.3 ⫾ 186.5
OP1 234.6 ⫾ 62.9 210.5 ⫾ 84.4 173.5 ⫾ 52.6 175.6 ⫾ 67.2
OP2 244.5 ⫾ 89.8 206.8 ⫾ 96.6 219.4 ⫾ 54.7 202.9 ⫾ 68.2
OP3 96.2 ⫾ 50.6 80.0 ⫾ 39.8 109.9 ⫾ 32.6 96.0 ⫾ 50.0
⌺OPs 575.3 ⫾ 191.8 497.3 ⫾ 212.1 502.8 ⫾ 109.7 474.5 ⫾ 168.2
Data are expressed as the mean amplitude ⫾ SD (mV)
IOVS, August 2005, Vol. 46, No. 8 RAGE in Early Diabetic Retinopathy 2919
the retinas of hyperglycemic mice (glucose effect for two-
factor ANOVA: P ⬍ 0.01); a significant increase was observed
in HGHL mice compared with each group of normoglycemic
mice (Fig. 5B). These studies demonstrate that the RAGE axis
comprising the cellular receptor and its AGE ligands is ampli-
fied in the diabetic retina, particularly in eyes with significant
capillary lesions of NPDR (HGHL mice).
Effect of Antagonism of RAGE on the Vascular
Lesions of Diabetic Retinopathy and Neuronal
Dysfunction at 6 Months of Age
Based on the upregulation of AGEs and RAGE in the HGHL
group, we tested the potential contribution of RAGE to the
pathogenesis of vascular and neuronal perturbation. Murine
sRAGE was administered to 10 HGHL mice from age 8 weeks to
age 6 months. The number of acellular capillaries per 10 mm
2
in the retinal digest of treated mice was significantly less than
those observed in nontreated mice (Fig. 6A). In addition, there
were significantly fewer pericyte ghosts in the retinas of
treated mice than in those of nontreated mice (Fig. 6B). Elec-
trophysiologic studies demonstrated that prophylactic treat-
ment with sRAGE reduced retinal neuronal dysfunction, with a
statistically significant (P ⬍ 0.05) reduction in the hyperglyce-
mia-induced latency delays observed in OP2, OP3, and 兺OPs
(summation of OPs) at 6 months of age (Table 6). Treatment
with sRAGE had no significant effect on the amplitudes of the
b-wave and OPs (data not shown).
DISCUSSION
The pathogenesis of diabetic retinopathy remains complex,
but prolonged hyperglycemia is essential in the development
of anatomic retinal vascular lesions in human diabetic retinop-
athy and most animal models of diabetic retinopathy.
41
In this
context, we investigated the db/db mouse, a well-character-
ized murine model of hereditary, insulin-resistant diabetes first
detected in the progeny of the C57BLKS/J strain at the Jackson
Laboratory and later characterized as being deficient in leptin
receptor signaling.
42
While the db/db mouse develops neurop
-
athy and nephropathy, the anatomic retinal vascular findings,
apart from basement membrane thickening, are less dramatic.
Our previous anatomic studies revealed acellular capillaries
and pericyte ghosts at age 8 months in db/db mice, but these
anatomic findings were variable and inconsistently present
(Barile GR, et al. IOVS 2000;41:ARVO Abstract 2156). Hyper-
lipidemia is associated with the severity of diabetic retinopa-
thy,
8–10
and successful treatment of hyperlipidemia in diabetic
patients may retard the progression of retinopathy or improve
it.
11–13
For these reasons, we investigated the influence of
hyperlipidemia on the retinal findings of the db/db mouse
model of diabetes mellitus, ultimately crossing it with mice
carrying a mutation in the apoE gene that leaves them devoid
of functioning apoE protein. We observed that the classic
anatomic retinal lesions of nonproliferative diabetic retinopa-
thy developed at the highest rate in HGHL mice, when com-
pared with the other groups, consistent with the burgeoning
notion that hyperlipidemia accelerates the retinal vascular dis-
ease of diabetes mellitus. These results further support increas-
ing evidence that dyslipidemia in diabetes mellitus indepen-
dently contributes to the pathogenesis and severity of diabetic
retinopathy, possibly through amplification of inflammatory
mechanisms.
43,44
Whereas diabetic retinopathy is classically a microvascular
disease of the retinal capillaries, diabetes may impair retinal
neuronal function before the onset of visible vascular lesions.
Numerous psychophysical and electrophysiological studies
demonstrate early retinal neuronal dysfunction in diabetes mel-
litus, before the onset of the classic microvascular lesions of
diabetic retinopathy.
45,46
In particular, Bresnick and Palta
47
and Bresnick
48
have emphasized that alterations in the OPs of
the electroretinogram better predict the development of high-
risk proliferative retinopathy than do clinical fundus photo-
graphs. Pathologic quantification of neural loss by Barber et
al.
49
showed apoptosis of retinal neurons and retinal atrophy,
with loss of inner retinal thickness and cell bodies, in both
diabetic rats and humans. Several investigators have noted
other retinal neuronal alterations in early diabetes, including
GFAP activation and glutamate transporter dysfunction in Mu¨l-
ler cells.
50,51
In our study, chronic hyperglycemia caused sig
-
nificant implicit time delays of OPs at 6 months that are com-
parable to those in previous studies of diabetes,
52
whereas
hyperlipidemia did not influence these electrophysiologic pa-
rameters. In conjunction with the histopathologic vascular
changes observed, this study supports the concept of early
diabetic retinopathy as a neurovascular disease of the retina,
with physiologic disturbances of neuronal function accompa-
nying traditional microvascular capillary pathologic disease.
It was in these contexts that we examined the RAGE axis in
this newly characterized murine model of NPDR. Not surpris-
ingly, we observed prominent AGE localization within the
vitreous cavity. The increased AGE formation in the vitreous
cavity of diabetic eyes has been postulated to increase collagen
cross-linking and cause vitreous changes characteristic of dia-
betic eyes, well-recognized phenomena sometimes referred to
as diabetic vitreoschisis or vitreopathy.
19,20
An additional find
-
ing of our study was prominent AGE accumulation along the
vitreoretinal interface, specifically posterior vitreous cortex
and the internal limiting membrane. Similar to the vitreous
cavity, the accumulation of AGEs at the vitreoretinal interface
may result in structural alterations that promote mechanical
traction in this region. Vitrectomy procedures are sometimes
performed to remove tractional effects that promote diabetic
macular edema. The localization of AGEs along the vitreoreti-
nal interface is consistent with the concept of a structurally
altered posterior hyaloid and internal limiting membrane capa-
ble of promoting subclinical vitreomacular disease in early
diabetic retinopathy. AGEs may also exert nontractional, recep-
tor-mediated effects through the RAGE axis. In this regard, an
intriguing finding of our study is the localization of RAGE
primarily to the Mu¨ller cells that extend from the internal
limiting membrane to the external limiting membrane of the
retina. The anatomically close apposition of an AGE-laden in-
ternal limiting membrane with the RAGE-expressing footplates
suggests that a possible physiologic benefit of diabetic vitrec-
tomy is the removal of AGE ligands from the posterior vitreous
TABLE 4. Two-Factor ANOVA Analysis of ERG Data from Tables 2
and 3
Glucose Effect Lipid Effect Interaction
b-Wave
Latency 0.004 0.805 0.516
Amplitude 0.174 0.799 0.422
OP1
Latency 0.001 0.216 0.649
Amplitude 0.021 0.588 0.516
OP2
Latency 0.001 0.376 0.675
Amplitude 0.550 0.225 0.661
OP3
Latency 0.031 0.933 0.744
Amplitude 0.283 0.271 0.934
⌺OPs
Latency 0.004 0.545 0.685
Amplitude 0.376 0.324 0.643
Data are probabilities.
2920 Barile et al. IOVS, August 2005, Vol. 46, No. 8
FIGURE 2. RAGE expression in the
retina of NGNL and HGHL mice.
RAGE immunofluorescence (A, D,
green) colocalizes with vimentin (B,
E, red), a marker of Mu¨ller cells (ar-
rows) in both NLNG and HGHL mice
(C, F). The extension of Mu¨ller cells
from the internal to the external lim-
iting membranes of the neurosensory
was highlighted with RAGE’s expres-
sion (A, D). ILM, internal limiting
membrane; IPL, inner plexiform lay-
er; INL, inner nuclear layer; ONL,
outer nuclear layer; ELM, external
limiting membrane. Scale bar, 50
m.
FIGURE 3. RAGE, GFAP, and CD31
immunohistochemistry of the retina
of HGHL mice. RAGE expression was
prominent in Mu¨ller cell processes,
particularly their internal footplates
(A, D; green, arrowheads) and was
not observed in adjacent astrocytes
(B, C; red, arrows). The intimate va-
soglial relationship of the RAGE-ex-
pressing Mu¨ller cell (D, green) with
the vascular endothelium of a retinal
capillary (E, red) is observed in (F).
ILM, internal limiting membrane; IPL,
inner plexiform layer; INL, inner nu-
clear layer. Scale bar, 25
m.
FIGURE 4. RAGE (green) and AGE
(red) immunohistochemistry of the
vitreoretinal interface in NGNL mice
(A–C) and HGHL mice (E–G). AGEs
are detected within the vitreous cav-
ity, posterior vitreous cortex, and in-
ternal limiting membrane of the ret-
ina (B, F, red). The internal
footplates of RAGE-expressing Mu¨ller
cells (A, E, green) are immediately
adjacent to AGEs in the internal lim-
iting membrane (C, G). Controls (D,
H). Vit, vitreous cavity; ILM, internal
limiting membrane; GCL, ganglion
cell layer; IPL, inner plexiform layer;
INL, inner nuclear layer. Scale bar, 25
m.
IOVS, August 2005, Vol. 46, No. 8 RAGE in Early Diabetic Retinopathy 2921
cortex and internal limiting membrane, downregulating the
proinflammatory RAGE axis in adjacent Mu¨ller cells.
The localization of RAGE to Mu¨ller cells raises exciting
possibilities for novel roles of these cells in the pathogenesis of
diabetic retinopathy. The specific RAGE-dependent mecha-
nisms by which AGEs may alter Mu¨ller cell structure and
function are the subject of future study. These cells are well
known to display a varied repertoire of structural and physio-
logic properties in the retina. The contact of vasoglial neuronal
tissue and especially Mu¨ller cells with underlying capillaries in
the retina suggests a potential pathophysiologic relationship in
diabetic retinopathy, once suggested by Ashton in his Bowman
lecture and supported by several recent studies.
53
In the set
-
ting of diabetes, alteration of the glutamate transporter, in part,
it is speculated, by oxidation; increased expression of GFAP
suggestive of reactive gliosis; and striking upregulation of VEGF
all have been detected in Mu¨ller cells.
54
Indeed, in vitro anal
-
yses suggests that incubation of cultured Mu¨ller cells with
AGEs upregulate expression of VEGF.
55
Mu¨ller cell ischemia
induces phosphorylation of extracellular signal-regulated ki-
nase (ERK) MAPKs in these cells,
56
again suggesting that a wide
array of changes in gene expression may ensue in these cells
when perturbed. The possible RAGE-dependence of these phe-
nomena remains to be determined, but the intimate relation-
ship of RAGE-expressing Mu¨ller cells with underlying vascular
endothelium suggests a potential role for Mu¨ller cell RAGE in
neurovascular dysfunction.
In addition to the specific localization of RAGE and its AGE
ligands in our study, we observed that AGEs accumulate in the
neurosensory retina with associated amplification of cellular
RAGE in the setting of hyperglycemia and early diabetic reti-
nopathy. The diversity by which AGEs may form on the amino
groups of proteins, lipids, and DNA is reflected in the variety of
locations that these products may accumulate during hypergly-
cemia, including the serum, ECM, and intracellular cyto-
plasm.
19
In this regard, it is noteworthy that we did not detect
significantly different AGE levels by fluorescence studies
within cellular proteins among the hyperlipidemic and hyper-
glycemic phenotypes. Instead, the retinas with the most severe
capillary disease had the highest levels of AGEs detected within
the ECM, both by fluorescence and ELISA studies. Hyperglyce-
FIGURE 5. Retinal AGE ELISA and RAGE mRNA transcripts. Retinal
AGEs accumulated in the retinas of hyperglycemic mice: Both HGNL
mice and HGHL mice had significantly increased AGEs compared with
NGNL littermates (A). RAGE mRNA expression in the retina was
increased in the setting of hyperglycemia and AGE accumulation.
RAGE transcripts were highest in the retinas of HGHL mice, with a
nearly twofold elevation compared with basal levels in NGNL litter-
mates as well as a significant increase compared with NGHL mice (B).
Results are expressed as the mean ⫾ SEM. *P ⬍ 0.05; **P ⬍ 0.01.
FIGURE 6. Effect of RAGE antagonism on vascular changes in HGHL
mice. sRAGE-treated mice had significantly less acellular capillaries (A)
and pericyte ghosts (B) in the retina compared with untreated HGHL
mice. Treatment of these mice also reduced the latency delays ob-
served in the OPs, with a significant reduction in the implicit times of
OP2, OP3 and ⌺OPs (the summation of OPs). *P ⬍ 0.05.
TABLE 5. Retinal AGE Fluorescence Intensities
NGNL NGHL HGNL HGHL
Cellular protein 1357 ⫾ 149 1666 ⫾ 182 1122 ⫾ 194 1181 ⫾ 161
ECM 2801 ⫾ 673 2342 ⫾ 531 3713 ⫾ 1229 5259 ⫾ 715*
Data are expressed as the mean autofluorescence ⫾ SE.
* Significant difference (P ⬍ 0.05) compared to the NGNL group.
2922 Barile et al. IOVS, August 2005, Vol. 46, No. 8
mia was the most important contributor to the development of
these AGEs, as HGNL mice also exhibited increased AGE accu-
mulation in the ECM in these studies (though this increase was
shown to be significant in this group only by ELISA). Consistent
with a role for RAGE ligands such as AGEs in the development
of retinopathy, we detected significant upregulation of RAGE
transcripts in the retinas of HGHL mice that had the highest
AGE accumulation and retinal disease. The amplification of
RAGE in the setting of its ligands is consistent with the known
biology of RAGE in other organ systems, and this property mag-
nifies the effect of the RAGE axis in local cellular responses.
26,40
Importantly, in this study, antagonism of the RAGE axis
ameliorated both neuronal dysfunction and vascular disease.
The electrophysiologic benefit we observed suggests that
RAGE contributes to neuronal dysfunction in the diabetic ret-
ina. The mechanisms of OP generation in the normal retina, the
associated alterations observed in these neuronal responses in
diabetic eyes, and the extent to which altered Mu¨ller cell
glutamate metabolism, signaling, and gene expression contrib-
ute to perturbation of these signals remains to be determined.
Antagonism of RAGE also reduced the progression of vascular
lesions of diabetic retinopathy in HGHL mice. This vascular
effect may relate to an a priori neuronal benefit to RAGE-
expressing Mu¨ller cells, but the ample data on AGE toxicity and
perturbation in retinal vascular endothelial cells also suggests
that antagonism of circulating serum AGEs with sRAGE may
reduce these perturbations and the resultant anatomic disease.
The precise neurovascular mechanisms altered with ligand
interaction with RAGE in the retina remain the subject of
current and future investigations, but the amelioration of neu-
rovascular features of diabetic retinopathy observed in this
study identifies the RAGE axis as an important therapeutic
target in the prevention and treatment of diabetic complica-
tions in the retina.
References
1. Frank RN. Diabetic retinopathy. N Engl J Med. 2004;350:48 –58.
2. Gardner TW, Antonetti DA, Barber AJ, LaNoue KF, Levison SW.
Diabetic retinopathy: more than meets the eye. Surv Ophthalmol.
2002;47(suppl 2):S253–S262.
3. Lieth E, Gardner TW, Barber AJ, Antonetti DA. Retinal
neurodegeneration: early pathology in diabetes. Clin Exp Ophthal-
mol. 2000;28:3–8.
4. Aiello LP, Avery RL, Arrigg PG, et al. Vascular endothelial growth
factor in ocular fluid of patients with diabetic retinopathy and
other retinal disorders. N Engl J Med. 1994;331:1480 –1487.
5. 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;329:977–986.
6. The 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;352:837– 853.
7. The UK Prospective Diabetes Study (UKPDS) Group. Tight blood
pressure control and risk of macrovascular and microvascular
complications in type 2 diabetes: UKPDS 38. BMJ. 1998;317:703–
713.
8. Kissebah AH, Kohner EM, Lewis B, Siddiq YK, Lowy C, Fraser TR.
Plasma-lipids and glucose/insulin relationship in non-insulin-requir-
ing diabetics with and without retinopathy. Lancet. 1975;1:1104 –
1108.
9. Chew EY, Klein ML, Ferris FL III, et al. Association of elevated
serum lipid levels with retinal hard exudate in diabetic retinopa-
thy. Early Treatment Diabetic Retinopathy Study (ETDRS) Report
22. Arch Ophthalmol. 1996;114:1079 –1084.
10. Davis MD, Fisher MR, Gangnon RE, et al. Risk factors for high-risk
proliferative diabetic retinopathy and severe visual loss: Early
Treatment Diabetic Retinopathy Study Report #18. Invest Oph-
thalmol Vis Sci. 1998;39:233–252.
11. Sen K, Misra A, Kumar A, Pandey RM. Simvastatin retards progres-
sion of retinopathy in diabetic patients with hypercholesterolemia.
Diabetes Res Clin Pract. 2002;56:1–11.
12. Gordon B, Chang S, Kavanagh M, et al. The effects of lipid lower-
ing on diabetic retinopathy. Am J Ophthalmol. 1991;112:385–391.
13. Fried LF, Forrest KY, Ellis D, Chang Y, Silvers N, Orchard TJ. Lipid
modulation in insulin-dependent diabetes mellitus: effect on mi-
crovascular outcomes. J Diabetes Complications. 2001;15:113–
119.
14. Schleicher ED, Wagner E, Nerlich AG. Increased accumulation of
the glycoxidation product N(epsilon)-(carboxymethyl)lysine in hu-
man tissues in diabetes and aging. J Clin Invest. 1997;99:457– 468.
15. Hammes HP, Alt A, Niwa T, et al. Differential accumulation of
advanced glycation end products in the course of diabetic retinop-
athy. Diabetologia. 1999;42:728 –736.
16. Lu M, Kuroki M, Amano S, et al. Advanced glycation end products
increase retinal vascular endothelial growth factor expression.
J Clin Invest. 1998;101:1219 –1224.
17. Tolentino MJ, Miller JW, Gragoudas ES, et al. Intravitreous injec-
tions of vascular endothelial growth factor produce retinal isch-
emia and microangiopathy in an adult primate. Ophthalmology.
1996;103:1820–1828.
18. Moore TC, Moore JE, Kaji Y, et al. The role of advanced glycation
end products in retinal microvascular leukostasis. Invest Ophthal-
mol Vis Sci. 2003;44:4457– 4464.
19. Stitt AW. The role of advanced glycation in the pathogenesis of
diabetic retinopathy. Exp Mol Pathol. 2003;75:95–108.
20. Sebag J. Abnormalities of human vitreous structure in diabetes.
Graefes Arch Clin Exp Ophthalmol. 1993;231:257–260.
21. Kern TS, Engerman RL. Pharmacological inhibition of diabetic
retinopathy: aminoguanidine and aspirin. Diabetes. 2001;50:
1636–1642.
22. Gardiner TA, Anderson HR, Stitt AW. Inhibition of advanced gly-
cation end-products protects against retinal capillary basement
membrane expansion during long-term diabetes. J Pathol. 2003;
201:328–333.
23. Stitt A, Gardiner TA, Alderson NL, et al. The AGE inhibitor pyri-
doxamine inhibits development of retinopathy in experimental
diabetes. Diabetes. 2002;51:2826 –2832.
24. Hammes HP, Du X, Edelstein D, et al. Benfotiamine blocks three
major pathways of hyperglycemic damage and prevents experi-
mental diabetic retinopathy. Nat Med. 2003;9:294–299.
25. Schmidt AM, Vianna M, Gerlach M, et al. Isolation and character-
ization of two binding proteins for advanced glycosylation end
products from bovine lung which are present on the endothelial
cell surface. J Biol Chem. 1992;267:14987–14997.
26. Yan SF, Ramasamy R, Naka Y, Schmidt AM. Glycation, inflamma-
tion, and RAGE: a scaffold for the macrovascular complications of
diabetes and beyond. Circ Res. 2003;93:1159–1169.
27. Hofmann MA, Drury S, Fu C, et al. RAGE mediates a novel proin-
flammatory axis: A central cell surface receptor for s100/calgranu-
lin polypeptides. Cell. 1999;97:889 –901.
28. Hori O, Brett J, Slattery T, et al. The receptor for advanced glyca-
tion end products (RAGE) is a cellular binding site for amphoterin:
mediation of neurite outgrowth and co-expression of RAGE and
amphoterin in the developing nervous system. J Biol Chem. 1995;
270:25752–25761.
29. Chavakis T, Bierhaus A, Al-Fakhri N, et al. The pattern recognition
receptor (RAGE) is a counterreceptor for leukocyte integrins: a
novel pathway for inflammatory cell recruitment. J Exp Med.
2003;198:1507–1515.
T
ABLE 6. sRAGE Effect on ERG Latencies at Age 6 Months
NGNL (n ⴝ 18) HGHL (n ⴝ 14) sRAGE (n ⴝ 10)
b-Wave 32.0 ⫾ 2.0 34.5 ⫾ 2.9 33.3 ⫾ 2.5
OP1 23.4 ⫾ 1.4 24.6 ⫾ 1.7 23.6 ⫾ 1.4
OP2 32.0 ⫾ 2.0 33.8 ⫾ 2.2 32.2 ⫾ 2.0*
OP3 42.6 ⫾ 2.9 44.9 ⫾ 3.1 42.0 ⫾ 3.0*
⌺OPs 98.0 ⫾ 6.2 103.3 ⫾ 6.8 97.7 ⫾ 6.1*
Data are expressed as the mean latency ⫾ SD (ms).
* Significant (P ⬍ 0.05) compared to HGHL group.
IOVS, August 2005, Vol. 46, No. 8 RAGE in Early Diabetic Retinopathy 2923
30. Schmidt AM, Hori O, Chen JX, et al. Advanced glycation end
products interacting with their endothelial receptor induce ex-
pression of vascular cell adhesion molecule-1 (VCAM-1) in cul-
tured human endothelial cells and in mice: a potential mechanism
for the accelerated vasculopathy of diabetes. J Clin Invest. 1995;
96:1395–1403.
31. Wautier JL, Zoukourian C, Chappey O, et al. Receptor-mediated
endothelial cell dysfunction in diabetic vasculopathy: soluble re-
ceptor for advanced glycation end products blocks hyperperme-
ability in diabetic rats. J Clin Invest. 1996;97:238 –243.
32. Laver NM, Robison WG Jr, Pfeffer BA. Novel procedures for iso-
lating intact retinal vascular beds from diabetic humans and animal
models. Invest Ophthalmol Vis Sci. 1993;34:2097–2104.
33. Kuwabara T, Cogan DG. Studies of retinal vascular patterns. I.
Normal architecture. Arch Ophthalmol. 1960;64:904 –911.
34. Algvere P. Clinical studies on the oscillatory potentials of the
human electroretinogram with special reference to the scotopic
b-wave. Acta Ophthalmol (Copenh). 1968;46:993–1024.
35. Segawa M, Hirata Y, Fujimori S, Okada K. The development of
electroretinogram abnormalities and the possible role of polyol
pathway activity in diabetic hyperglycemia and galactosemia. Me-
tabolism. 1988;37:454–460.
36. Park L, Raman KG, Lee KJ, et al. Suppression of accelerated
diabetic atherosclerosis by the soluble receptor for advanced gly-
cation endproducts. Nat Med. 1998;4:1025–1031.
37. Yin D. Biochemical basis of lipofuscin, ceroid, and age pigment-
like fluorophores. Free Radic Biol Med. 1996;21:871–888.
38. Meerwaldt R, Graaff R, Oomen PH, et al. Simple non-invasive
assessment of advanced glycation endproduct accumulation. Dia-
betologia. 2004;47:1324–1330.
39. Livak KJ, Schmittgen TD. Analysis of relative gene expression data
using real-time quantitative PCR and the 2(-Delta Ddelta C(T))
method. Methods. 2001;25:402– 408.
40. Li J, Schmidt AM. Characterization and functional analysis of the
promoter of RAGE, the receptor for advanced glycation end prod-
ucts. J Biol Chem. 1997;272:16498 –16506.
41. Engerman RL, Kern TS. Retinopathy in animal models of diabetes.
Diabetes Metab Rev. 1995;11:109 –120.
42. Chua SC Jr, Chung WK, Wu-Peng XS, et al. Phenotypes of mouse
diabetes and rat fatty due to mutations in the OB (leptin) receptor.
Science. 1996;271:994–996.
43. Lyons TJ, Jenkins AJ, Zheng D, et al. Diabetic retinopathy and
serum lipoprotein subclasses in the DCCT/EDIC cohort. Invest
Ophthalmol Vis Sci. 2004;45:910 –918.
44. Chen W, Jump DB, Grant MB, Esselman WJ, Busik JV. Dyslipide-
mia, but not hyperglycemia, induces inflammatory adhesion mol-
ecules in human retinal vascular endothelial cells. Invest Ophthal-
mol Vis Sci. 2003;44:5016 –5022.
45. Tzekov R, Arden GB. The electroretinogram in diabetic retinopa-
thy. Surv Ophthalmol. 1999;44:53– 60.
46. Han Y, Bearse MA Jr, Schneck ME, Barez S, Jacobsen CH, Adams AJ.
Multifocal electroretinogram delays predict sites of subsequent
diabetic retinopathy. Invest Ophthalmol Vis Sci. 2004;45:948 –
954.
47. Bresnick GH, Palta M. Predicting progression to severe prolifera-
tive diabetic retinopathy. Arch Ophthalmol. 1987;105:810– 814.
48. Bresnick GH. Diabetic retinopathy viewed as a neurosensory dis-
order. Arch Ophthalmol. 1986;104:989 –990.
49. Barber AJ, Lieth E, Khin SA, Antonetti DA, Buchanan AG, Gardner
TW. Neural apoptosis in the retina during experimental and hu-
man diabetes. Early onset and effect of insulin. J Clin Invest.
1998;102:783–791.
50. Zeng XX, Ng YK, Ling EA. Neuronal and microglial response in the
retina of streptozotocin-induced diabetic rats. Vis Neurosci. 2000;
17:463–471.
51. Mizutani M, Gerhardinger C, Lorenzi M. Mu¨ller cell changes in
human diabetic retinopathy. Diabetes. 1998;47:445– 449.
52. Shirao Y, Kawasaki K. Electrical responses from diabetic retina.
Prog Retin Eye Res. 1998;17:59 –76.
53. Ashton N. Bowman lecture: the blood-retinal barrier and vaso-glial
relationships in retinal disease. Trans Ophthalmol Soc UK. 1965;
85:199–230.
54. Amin RH, Frank RN, Kennedy A, Eliott D, Puklin JE, Abrams GW.
Vascular endothelial growth factor is present in glial cells of the
retina and optic nerve of human subjects with nonproliferative
diabetic retinopathy. Invest Ophthalmol Vis Sci. 1997;38:36– 47.
55. Hirata C, Nakano K, Nakamura N, et al. Advanced glycation end
products induce expression of vascular endothelial growth factor
by retinal Mu¨ller cells. Biochem Biophys Res Commun. 1997;236:
712–715.
56. Roth S, Shaikh AR, Hennelly MM, Li Q, Bindokas V, Graham CE.
Mitogen-activated protein kinases and retinal ischemia. Invest
Ophthalmol Vis Sci. 2003;44:5383–5395.
2924 Barile et al. IOVS, August 2005, Vol. 46, No. 8