Intravitreal Triamcinolone Acetonide Inhibits Breakdown of the Blood-Retinal Barrier through

20 Pages·2007·0.62 MB·English

Diabetes Publish Ahead of Print, published online January 3, 2008
Intravitreal Triamcinolone Acetonide Inhibits Breakdown of the Blood-
Retinal Barrier through Differential Regulation of VEGF-A and its Receptors
in Early Diabetic Rat Retinas

Xinyuan Zhang1,2 MD, Shisan Bao3 MD PhD , Donna Lai4 PhD,
Robert W Rapkins1, Mark C Gillies1 MD, PhD

1Retinal Therapeutic Research Group, Save Sight Institute,
Department of Clinical Ophthalmology, University of Sydney, Level 2, South Block,
Sydney Eye Hospital, Macquarie St, Sydney NSW 2000, Australia.
2Beijing Institute of Ophthalmology, Ophthalmic Center, Tongren Hospital,
Beijing University of Medical Science, 100730, PR China
3Department of Pathology, D06, University of Sydney, 2006
4Molecular Biology Facility, Bosch Institute, F13, University of Sydney,2006

Running Title: TA inhibits BRB Breakdown

Corresponding author:
Dr Xinyuan Zhang
Save Sight Institute, Department of Clinical Ophthalmology, Level 2, South Block,
Sydney Eye Hospital, Macquarie St, Sydney NSW 2000, Australia.
[email protected]

Received for publication 16 July 2007 and accepted in revised form 19 December 2007

Copyright American Diabetes Association, Inc., 2008
TA inhibits BRB Breakdown
ABSTRACT

Objective: To elucidate the mechanism of the unique beneficial effect of intravitreal steroid
therapy on diabetic macular edema, we investigated the effect of locally administered
triamcinolone acetonide (TA) on the expression of vascular endothelial growth factor (VEGF-A)
and its receptors in streptozotocin (STZ) - induced diabetic rat retinas. We then correlated the
expression of these proteins with breakdown of the blood-retinal barrier (BRB).

Research Design and Methods: 32 eyes of 16 diabetic and non-diabetic rats were divided into
four groups. TA was injected into the vitreous of the right eye and saline was injected into the left
eye (control) 3.5 weeks after induction of diabetes. Retinas were harvested 48h following
treatment. mRNA and protein expression of VEGF-A, VEGF-A receptor 1 (FLT-1) and VEGF-A
receptor 2 (FLK-1) were determined by real time RT-PCR and immunohistochemistry. BRB
permeability was quantitated by measuring extravasated endogenous albumin and retinal
thickness.

Results: Diabetes induced retinal thickness and albumin extravasation were significantly reduced
in TA-treated diabetic retinas to a level that was similar to sham-treated non-diabetic eyes. A
close correlation between albumin leakage and increased expression of both Vegf-a and Flk-1
was noted in the diabetic retinas. TA down-regulated the expression of Vegf-a and Flk-1, but up-
regulated the expression of Flt-1. TA did not alter the expression of these genes in non-diabetic
retinas.

Conclusions: Intravitreal injection of TA stabilizes the BRB in association with regulation of Vegf-
a, Flk-1 and Flt-1 expression in early diabetic rat retinas.

KEYWORDS. VEGF-A/VEGF-A Receptors, Triamcinolone Acetonide, Diabetic Retinopathy,
Blood-Retinal Barrier

ABBREVIATIONS. AMD: age related macular degeneration, AEC: 3-amino-9-
ethylcarbazole chromogen, BRB: blood-retinal barrier, DAB: diaminobenzidine, DR: diabetic
retinopathy, FLK-1: fetal liver kinase-1; vascular endothelial growth factor receptor-2, FLT-1:
fms-like tyrosine kinase-1; vascular endothelial growth factor receptor-1, I.A.U: image arbitrary
units, IVTA: intravitreal injection of triamcinolone acetonide, PFA: paraformaldehyde, qPCR:
quantitative polymerase chain reaction, REST: relative expression software tool, RT: reverse
transcription, STZ: streptozotocin, TA: triamcinolone acetonide, TBS: Tris buffered saline, UDG:
Uracil-DNA Glycosylase, VEGF-A: vascular endothelial growth factor A,

2
TA inhibits BRB Breakdown
A
s the prevalence of diabetes mellitus cascade that results in increased
progressively rises throughout the microvascular permeability and angiogenesis
world, improving the treatment of (11). Inhibition of the VEGF-A over-
macular edema, a common cause of loss of expression is thus a potential treatment for
vision from diabetic retinopathy (DR), has DR and AMD (12; 13).
become a major goal of ophthalmic research Triamcinolone acetonide (TA) is an
(1). Laser treatment has been proven to intermediate acting corticosteroid suspension.
reduce the risk of further loss of vision in Intravitreal injections of TA (IVTA) have
patients with clinically significant macular shown remarkable efficacy for the treatment
edema, but is less effective in restoring vision of advanced diabetic retinopathy, in particular
once it already has deteriorated. Moreover, diabetic macular edema (14; 15). IVTA has
laser treatment is an intrinsically destructive been shown to inhibit VEGF-induced
procedure that itself can lead to loss of vision breakdown of the blood-retinal barrier in
through progressive enlargement of laser rabbit, and leukostasis in rat models of DR
scars with time (2). (16; 17). In a two year, double-masked,
Breakdown of the blood-retinal barrier placebo-controlled, randomized clinical trial,
(BRB) is an early event in the pathogenesis of eyes with recalcitrant diabetic macular edema
DR (3). When local compensatory involving the fovea that were treated with
mechanisms are overcome, frank vasogenic IVTA had twice the chance of improving
macular edema develops. The anatomical vision and half the risk of further loss.
correlate responsible for the BRB is the tight However, a high incidence of cataract and
junction-associated proteins between retinal elevated intraocular pressure was found in
vascular endothelial cells (inner BRB) and IVTA-treated eyes (18). If the unique
retinal pigment epithelial cells (outer BRB). beneficial effect of intravitreal steroid therapy
It is strongly suggested that the inner BRB on macular edema is to be distinguished from
(rather than the outer BRB) is the primary site its high risk of local adverse events (glaucoma
of the vascular leak that results in diabetic and cataract), the precise mechanism of action
macular edema both in human and animal of TA needs to be elucidated. In this study, we
studies (4; 5). In the retinas of rodents, the determined the effect of IVTA on the
inner BRB has been shown to become expression of VEGF-A and its receptors in a
permeable to high molecular weight rodent model of early DR.
molecules in as little as two weeks after
induction of diabetes (6). RESEARCH DESIGN AND METHODS
Vascular endothelial growth factor-A Animals. Male Sprague-Dawley (SD) rats
(VEGF-A)/vascular permeability factor is (aged 8-10 weeks, ARC Laboratories, Perth,
believed to mediate increased vascular Australia), weighing 240-250g, were used in
permeability and co-localize with this study. The animals were cared for in
extravasated albumin in major retinal diseases accordance with The Association for Research
such as DR (7; 8) and age-related macular in Vision and Ophthalmology statement for
degeneration (AMD) (9). VEGF-A activates the use of Animals in Ophthalmic and Vision
two tyrosine-kinase receptors, fms-like Research. All animals were housed with
tyrosine kinase-1 (FLT-1, vascular endothelial environmental enrichment with free access to
growth factor receptor-1) and fetal liver food and water. Experimental procedures
kinase-1 (FLK-1, vascular endothelial growth were conducted under aseptic conditions and
factor receptor-2;) (10), which were approved by the Animal Ethics
autophosphorylate to initiate a signaling Committee of the University of Sydney.
3
TA inhibits BRB Breakdown
Induction of Diabetes. Diabetes was induced Victoria, Australia) in a volume of 5.6µl
with streptozotocin (STZ) (Sigma) in SD rats (10mg/ml) was injected into the vitreous of
according to the published protocol with the right eye with a 32-gauge microinjector
slight modifications (19; 20). Rats were given (Hamilton Co., Reno, NV) under a dissecting
a single dose of 65mg/kg STZ i.p. in 100mM microscope (17). The contra-lateral eye,
citrate buffer (pH 4.5); weight and age- which received an equal volume of saline,
matched non-diabetic control rats received served as a sham-treated control. A 30-gauge
vehicle only. Eight animals were used per needle was used to make a punch incision
group for each set of experiments. After STZ 1mm posterior to the temporal limbus, and a
injection, body weights were recorded daily, microinjector needle then inserted through the
and diabetic status was confirmed by tail vein incision, approximately 1.5mm deep, angled
blood glucose estimation using an automated toward the optic nerve.
Accu-Chek glucometer (Roche Diagnostics, Preparation of retinas for
Indianapolis, IN, USA) after 48h. Rats with histopathology/immunohistochemistry and
non-fasting blood glucose levels greater than real time RT-PCR. To detect VEGF-A, FLT-
13.8mmol/ml were deemed diabetic; those 1, or FLK-1 antigen, the chest cavity was
with lower values were excluded from this opened following anesthesia as described
study. Maintenance of a diabetic state was above. A 20-gauge perfusion cannula was
confirmed by weekly tail vein blood glucose introduced into the aorta via the left ventricle.
measurements. Each rat was perfused with 60ml PBS at 37°C
Treatment. Intravitreal treatment with TA or to clear the retinal vessels then perfused with
saline was performed 3.5 weeks after 40ml 2% paraformaldehyde (PFA) in 0.1M
induction of diabetes. Rats were divided into phosphate buffer (PBS), pH7.4, after
4 experimental treatment groups: achieving drainage from the right atrium.
1. Diabetic eyes with IVTA treatment (IVTA- Eyes were collected and fixed in 2% PFA
treated diabetic retinas, n=8, right eyes) overnight at 4°C. The fixed eyes were
2. Diabetic eyes with sham treatment (sham- processed routinely and embedded in paraffin
treated diabetic retinas, n=8, left eyes) wax. To quantitate BRB breakdown
3. Non-diabetic eyes with IVTA treatment (immunochemical analysis of albumin
(IVTA-treated non-diabetic retinas, n=8, right leakage), rats were euthanized without
eyes) perfusion.
4. Non-diabetic eyes with sham treatment For real-time RT-PCR studies, each retina
(sham-treated non-diabetic retinas, n=8, left was dissected and frozen immediately in
eyes) liquid nitrogen after perfusion with 100ml
Intravitreal Injection of TA. STZ-induced PBS. Retinas were stored at -80°C until
diabetic or non-diabetic animals were extraction of RNA.
anesthetized by isoflurane inhalation and a 4:1 RNA Isolation and Characterization. Total
mixture of ketamine hydrochloride (80mg/kg) RNA was isolated from individual retinas
and xylazine hydrochloride (4mg/kg) i.m. using TRI reagent (Sigma) and purified with a
Pupils were dilated with one drop each of RNeasy Mini kit (Qiagen, Victoria, Australia)
0.5% tropicamide and 2.5% phenylephrine according to the manufacturer’s instructions.
hydrochloride. For additional topical RNA concentrations, A /A and A /A
260 280 260 230
anesthesia, 0.4% Benoxinate-HCl of eye drop ratios were measured using a NanoDrop
was applied followed by 0.5% of levofloxacin spectrophotometer. RNA integrity was
ophthalmic solution to prevent infection. A assessed with an Agilent 2100 Bioanalyzer
single dose of 56µg TA (Kenacort-A 40, using an Agilent RNA 6000 Nano LabChip®
4
TA inhibits BRB Breakdown
kit (Agilent Technologies, Palo Alto, Morphological analysis. Retinal thickness
California, USA). was evaluated by using Image-Pro Plus 4.5
Two step real-time RT- PCR. Reverse (IPP4.5, Media Cybernetics, Silver Spring,
transcription (RT) was carried out using a MD, USA) on 5(cid:31)m hematoxylin and eosin
SuperScript III First-Stand Synthesis System (H&E) stained sections. The eyes were cut
for RT-PCR kit (Invitrogen, Australia). RNA vertically through the center of the cornea and
(1(cid:31)g) was reverse-transcribed using 2.5µM optic nerve, and both halves were embedded
Oligo(dT) primer after DNase I treatment face down. Care was taken to keep this
according to the manufacturer’s instructions. cutting angle consistent for all eyes measured
A non-RT-control (minus RT) was performed for retinal thickness. Images of retinal
for each RNA sample to test genomic DNA sections were obtained, and the thickness of
contamination. The RT reaction was both the retinas and sclera was measured
performed at 50°C for 50min after denaturing using IPP with a reference micrometer
at 65°C for 5min. The RT reactions were (Olympus, Australia) under the same
terminated by incubation at 85°C for 5min. conditions. Twenty lines were manually
Real time PCR was performed using a drawn by an observer masked to the status of
Platinum SYBR Green qPCR SuperMix-UDG the rat from which the eye was obtained, from
kit (Invitrogen, Australia) on a Rotor-Gene the inner limiting membrane to the tips of the
3000 real time thermal cycler according to the photoreceptor outer segments for each of 20
manufacturer’s instruction. The same real sequential fields. The mean scleral width was
time PCR cycling condition was used for all also calculated and found to be similar
target genes: 50°C for 2min (UDG reaction), amongst the 4 groups, indicating that
95°C for 2 min, 45 cycles of 95°C, 15 sec; differences in retinal thickness were not due
60°C 20 sec; 72°C 20 sec. Melting curve to systematic sampling error (not shown).
analysis and agarose gel electrophoresis were Immunohistochemical detection of retinal
performed to confirm the purity of the qPCR VEGF-A, FLK-1, FLT-1 and albumin.
products. All samples were performed in Standard indirect immunohistochemical
duplicate and replicate data were collected for methods were used to detect leakage of
further comparative analysis. The rat gene albumin and the expression of VEGF-A,
specific primer pairs used for PCR are listed FLK-1 and FlT-1 (21). The paraffin-
in Table 1. embedded retinal sections (5(cid:31)m）were de-
Relative quantification and statistical waxed and dehydrated. The sections were
analysis. The BestKeeper program then quenched with H O blocker (Dako,
2 2
(http://www.gene-quantification.info/) was Australia) for 10 min at room temperature,
utilized to determine the better reference gene and washed three times in Tris-buffered saline
between two widely used housekeeping genes (TBS) pH 7.6.
(Gapdh & β-actin) for our study. For detection of albumin leakage in the
The relative quantification analysis was retinas, the sections were incubated with
performed using the Relative Expression sheep anti-human albumin polyclonal
Software Tool (REST-MCS). The statistical antibody (Abcam, Cambridge, UK) at a 1:500
significance of variations in gene expression dilution for 1 hr at room temperature. The
was analyzed by the Pair Wise Fixed sections were further incubated with rabbit
Reallocation Randomization Test provided by anti-sheep HRP-Streptavidin immunoglobulin
the REST software. P values of <0.05 were (Abcam, Cambridge, UK) for 30min at room
considered significant. Data are shown as temperature, after three washes with TBS.
mean ± SEM. The sections were developed with 3-amino-9-
5
TA inhibits BRB Breakdown
ethylcarbazole chromogen (AEC, Dako, sections prior to automate counting. The
Australia) for 3 min, and counterstained with intravascular space was excluded by the
hematoxylin. analysis. The ratio of the area of positive
For detection of VEGF-A, FLK-1 and signal vs. the total retinal area in each field
FLT-1, the sections were treated with antigen was determined with IPP 4.5, and expressed
retrieval solution (Dako, Australia) at 90°C as image arbitrary units (I.A.U).
for 10 min, subsequently cooled to room All images analyzed with IPP4.5 were
temperature, then quenched with H O evaluated by an experienced masked
2 2
blocker. Sections were incubated with rabbit histopathologist.
anti-human polyclonal primary antibody Statistics for immunostaining. All data are
[VEGF-A (1:100); FLK-1 (1:200) or FLT-1 presented as means ± SEM. One-way
(1:50) antibody] (Abcam, Cambridge, UK) ANOVA was performed for statistical
for 1 h at room temperature. The sections analysis (GraphPad Prism version 4.0, San
were further incubated with goat anti-rabbit Diego, CA, USA). Differences between
HRP-Streptavidin immunoglobulin (Abcam, group means with P values of less than 0.05
Cambridge, UK) for 30 min at room were regarded as being statistically different.
temperature. Slides were developed with
diaminobenzidine (DAB, Dako, Australia) for RESULTS
5 min, and counterstained with hematoxylin. Retinal gene transcription of Vegf-a, Flk-1
An isotype-matched negative control was and Flt-1.
used for all the antibodies (data not shown). Total RNA yield and quality. The average
Quantification and statistical analysis of yields of total RNA extracted from retinas in
immunohistochemical staining. The different groups were comparable. The ratios
peroxidase activity was visualized using of A /A and A /A for all RNA
260 280 260 230
diaminobenzidine (DAB). The numbers of samples were above 2 (2.10±0.005 and
VEGF-A, FLT-1 and FLK-1 positive stained 2.03±0.002 respectively), indicating limited
cells were counted per high power field protein and organic contaminants. RNA
(400X) by using IPP4.5 as described integrity number (RIN), obtained
previously (22; 23). Briefly, the threshold of automatically by the Agilent 2100 Expert
positive signal was defined for each antibody software, was similar between the diabetic
for all the sections following different and non-diabetic groups (8.7±0.2 vs.
treatments. Color signal above the threshold 8.4±0.01, p = 0.88).
for each antibody defined was deemed to be A single PCR fragment was obtained for
positive; whereas any signal below the each gene-specific primer pair with the
threshold was regarded as negative. Analyses expected length as determined by 1.5%
consisted of 20 fields per retina. The entire agarose gel electrophoresis. This was
retina was measured with reference to a confirmed with a unique peak in the melting
micrometer (Olympus, Australia) under the curve analysis for each real time PCR reaction
same condition sequentially. The number of (data not shown).
positive signals vs. total cell numbers within β-actin was selected as a reference gene
each field of the retina was determined with for this study based on the results obtained
IPP 4.5, and was expressed as positive cell from BestKeeper software (Table 2-3).
numbers/mm, as described previously (24). Relative changes in Vegf-a, Flk-1, Flt-1
Extravasated albumin was visualized with mRNA expression. Diabetes and TA treatment
AEC Chromogen. The threshold of positive differentially regulated the transcription of
staining of albumin was defined for all Vegf-a and its receptors. Transcription of Vegf-
6
TA inhibits BRB Breakdown
a and Flk-1 was significantly increased by positive cells/mm, P<0.0001) (Figure 2B). In
2.5-fold and 1.8-fold (P =0.001, P contrast to the elevated levels of VEGF and
Vegf-a Flk-
=0.031) respectively, but Flt-1 was down- FLK-1, FLT-1 expression was reduced by
1
regulated 4.3-fold (P =0.005) in sham- 37.5% in diabetic retinas (16.94 ±0.76
Flt-1
treated diabetic retinas compared to sham- positive cells/mm vs. 27.09±0.56 positive
treated non-diabetic retinas (Figure 1A). Vegf- cells/mm, P<0.0001) (Figure 2C). The
a and Flk-1 transcription was significantly number of FLT-1 positive cells in IVTA-
down-regulated 2.9-fold and 2.0-fold treated diabetic retinas increased by 1.2-fold
respectively (P =0.030, P =0.028), compared to the sham-treated diabetic retinas
Vegf-a Flk-1
however Flt-1 was significantly up-regulated (20.87±0.73 positive cells/mm vs. 16.94±0.76
1.3-fold (P =0.002) with IVTA treatment positive cells/mm, P=0.0002) (Figure 2C), but
Flt-1
(Figure 1B). There was no statistically was still significantly less than both TA-
significant difference in the transcription of treated or Sham-treated non-diabetic retinas
Vegf-a and Flk-1 or Flt-1 between sham- (20.87±0.73 positive cells/mm vs. 26.53±0.67
treated and IVTA-treated non-diabetic groups positive cells/mm; 20.87±0.73 positive
(P =0.519, P =0.679, P =0.909) cells/mm vs. 27.09±0.56 positive cells/mm,
Vegf-a Flk-1 Flt-1
(Figure 1C). P<0.0001 for both). No significant difference
Retinal protein expression of VEGF-A, was detected in the expression of VEGF-A
FLK-1 and FLT-1. The expression of VEGF- and its receptors (FLK-1 and FLT-1) between
A and its two receptors (FLK-1 and FLT-1) IVTA-treated and sham-treated non-diabetic
was found predominantly in the ganglion cell retinas (P =0.2641, P =0.4792, P
VEGF-A FLK-1 FLT-
layer, inner nuclear layer and outer plexiform =0.5222) (Figure 2A-2C).
1
layer of all four treatment groups. In non- Retinal thickness analysis. There was no
diabetic retinas, immunoreactivity against significant difference in retinal thickness
VEGF and FLK-1 was weak in these layers in between IVTA and sham-treated control eyes
contrast to in the diabetic retinas. The (147.28±1.23(cid:31)m vs. 147.60±1.40(cid:31)m,
expression of VEGF-A was increased by 1.8- P=0.8627) (Figure 3A and 3B). There was a
fold in the sham-treated diabetic retinas 10% reduction in the thickness of the IVTA-
(37.44±0.74 positive cells/mm vs. 20.40 treated diabetic retinas compared with the
±0.44 positive cells/mm, P<0.0001) compared sham-treated diabetic retinas
with sham-treated non-diabetic retinas. The (148.20±1.06(cid:31)m vs. 163.25±1.78(cid:31)m,
number of VEGF-A positive cells was P<0.001) (Figure 3C, D and E). Sham-treated
reduced by 32.1% in the IVTA-treated diabetic retinas were significantly thicker than
diabetic retinas in contrast to the sham-treated sham-treated non-diabetic retinas
diabetic retinas (25.41±0.62 positive cells/mm (163.25±1.78(cid:31)m vs. 147.60±1.40(cid:31)m,
vs. 37.44 ±0.74 positive cells/mm, P<0.001) P<0.001) (Figure 3 E).
(Figure 2A). Similarly, FLK-1 expression was The Effect of IVTA on albumin leakage.
increased by 1.6-fold in the sham-treated Extra-vascular albumin was scarcely detected
diabetic retinas compared with sham-treated in the inner nuclear layer, outer nuclear layer
non-diabetic retinas (53.17±1.08 positive and outer plexiform layer of non-diabetic
cells/mm vs. 32.95 ±0.47 positive cells/mm, retinas (Figure 4A and 4A1). In IVTA-treated
P<0.0001). The number of FLK-1 positive and sham-treated diabetic rats, albumin
cells was significantly decreased by 28.5% in leakage was observed mainly in the inner
the diabetic retinas with IVTA treatment nuclear, outer plexiform and outer nuclear
compared to sham-treated diabetic retinas layers (Figure 4B, B1, C and C1). There was
(37.99±0.71 positive cells/mm vs. 53.17±1.08 3.6-fold higher albumin expression in the
7
TA inhibits BRB Breakdown
sham-treated diabetic retinas than the non- increased albumin leakage, supports the
diabetic controls (17.20±1.78 I.A.U vs. hypothesis that FLK-1 is another important
4.84±1.58 I.A.U, P<0.0001) (Figure 4E). permeability factor and regulator of BRB
However, albumin expression in the extra- dysfunction in diabetes. These data are
vascular space was significantly reduced by consistent with a report that FLK-1 plays a
63.5% in IVTA-treated diabetic retinas, critical role in breakdown of the brain barrier
compared to sham-treated diabetic retinas (26).
(6.27±0.46 I.A.U vs. 17.20±1.78 I.A.U, The expression of another VEGF receptor,
P<0.0001) (Figure 4E).There was no FLT-1, was unexpectedly decreased in
difference between the IVTA-treated non- diabetic retinas, and increased following
diabetic retinas vs. sham-treated non-diabetic IVTA treatment. This may be a result of
retinas (4.81±1.58 I.A.U vs. 4.84±1.48 I.A.U, competition between FLK-1 and FLT-1. FLT-
P=0.9590) (Figure 4E). 1 on endothelial cells may function as a
negative regulator of FLK-1, acting as a so-
DISCUSSION called “decoy” receptor which regulates the
In this study we have demonstrated that activity of VEGFs on vascular endothelium
IVTA inhibits the increased expression of by sequestering VEGF-A and thus rendering it
Vegf-a and Flk-1 in early diabetic rat retinas. less available to bind to FLK-1 (27). It is
By contrast, Flt-1 expression was decreased speculated that triggering of FLT-1 induces
in early diabetic retinas but was up-regulated weak or undetectable ligand-dependent
following IVTA treatment. A strong tyrosine phosphorylation in porcine aortic
correlation of albumin leakage and retinal endothelial cells, whereas FLK-1 binding
thickness with the expression of Vegf-a and its induces a strong response (28). Furthermore,
receptors was observed. These findings VEGF-A mutants that bind selectively to
indicate that one likely mechanism of action FLK-1 are able to induce mitogenesis and
of IVTA on diabetic macular edema is chemotaxis in endothelial cells, and vascular
through the differential regulation of VEGF-A permeability and angiogenesis in vivo,
and its receptors. whereas VEGF-A mutants which selectively
VEGF-A is a major regulator of bind to FLT-1 do not have the activities stated
angiogenesis and vascular permeability in above (29; 30). In contrast to our finding,
development and injury. The involvement of increased expression of Flt-1 in the 6 month
one of its receptors, FLK-1 in angiogenesis STZ induced diabetic rat retinas was reported
has been widely demonstrated (11), but few by Hammes et al (31). In their study,
studies have addressed its role as a mediator increased Flt-1 mRNA levels detected by in
of the BRB permeability. Ojima et al situ hybridization were found in the ganglion
demonstrated that VEGF-induced FLK-1 cell and both nuclear layers of diabetic
phosphorylation and its downstream signaling samples only. This apparent discrepancy
cascade correlated with BRB breakdown and maybe explained by the different durations of
was inhibited by EphA2/ephrinA1 both in diabetes studied.
vitro and in vivo (25). In our study, the direct Diabetic macular edema is consistently
relationship between the expression of FLK-1 reduced after IVTA in human clinical trials, as
and the vascular permeability/breakdown of is sensitively demonstrated by Optical
the BRB was further investigated in the Coherence Tomography (32). In this study,
diabetic retina. The over-expression of the reduction of VEGF-A levels by IVTA
VEGF-A and FLK-1 that was found in the treatment that we observed is consistent with
diabetic retina in this study in association with a reduction in levels of extravasated albumin
8
TA inhibits BRB Breakdown
from the retinal vessels. This provides strong they are currently in advanced clinical
evidence that anti-VEGF therapy is an development for diabetic retinopathy, and
effective strategy for diabetic BRB more specific VEGF inhibitors are in the
dysfunction. Indeed, a recent phase II study pipeline. Presently one of the disadvantages
suggested that the VEGF-A inhibitor, of these drugs is that hey need to be injected
pegaptanib sodium, may also be efficacious in every 4-6 weeks, compared with every 6
the treatment of diabetic macular edema(12). months for triamcinolone (18), however it is
Two hypotheses have been proposed to likely that longer acting preparations will
explain the clinically established beneficial appear in due course. While VEGF inhibitors
effect of corticosteroids on diabetic macular do not cause corticosteroid-associated side
edema and the restoration of the BRB. Firstly, effects, they may confer an increased risk of
corticosteroids may inhibit VEGF-A signaling systematic vascular adverse events (39).
by blocking cytokine production, inducing There is also a suspicion that VEGF inhibitors
leukocyte apoptosis (33) and, secondly, that may eventually cause retinal atrophy by
corticosteroids can directly induce the barrier blocking the cytokine’s recently discovered
properties of inter-endothelial tight junction neuroprotective action(40). On the other hand,
complexes (34). A recent study suggests that an increase in systemic adverse events has not
corticosteroids may have relatively rapid been detected in patients receiving intravitreal
effects, such as a decrease in paracellular flux steroid therapy, and there is some evidence to
and an increase in transepithelial resistance, suggest that steroids may have a protective
which correlates with an increase in tight effect on photoreceptors (41) which could
junction immunoreactivity at the cell border. potentially offset any neurotoxic effect it
A concomitant decrease in occludin might have through inhibiting VEGF-A. How
phosphorylation, potentially countering the the known adverse events of steroids will
effects of VEGF-A has also been noted (35). balance against the as yet largely unknown
Steroid treatment also appears to increase the safety profile of specific VEGF inhibitors is a
expression of at least some tight junction- subject that warrants careful clinical research
associated proteins (36). in future, as does the issue as to whether a
By providing a deeper understanding of long-acting steroid combined with a specific
the mechanisms of action of intravitreal VEGF inhibitor in low doses might deliver
steroids on retinal disease, the results of this the safest and most efficacious clinical
study may be used to refine their clinical use. response.
In a placebo-controlled, randomized clinical The integrity of the blood-retinal barrier
trial, IVTA treatment doubled the chance of can be assessed by various techniques, each
vision improvement and halved the risk of of which has advantages and limitations (42).
further loss in eyes with advanced diabetic Extravasated albumin has been demonstrated
macular edema. However there was a to be precise indicator of the permeability of
significant risk of local adverse events, with the blood-retinal and blood-brain barriers (21;
55% of eyes requiring surgery to remove 43). IPP is a sensitive image analysis system
cataract over 2 years, and 44% requiring eye that has been used to quantify subtle changes
drops to control elevated intraocular pressure in a number of studies (24; 44). In this study,
(18). Anti-VEGF antibodies are another immunolocalization of extravasated albumin
potential treatment for diabetic retinopathy. was identified at the light microscopic level
Ranibizumab (37; 38) and pegaptanib sodium and quantified using IPP, providing a useful
(13) have already been approved in the U.S. technique with which to study the effect of
for treatment of subretinal neovascularisation; IVTA on DR.
9
TA inhibits BRB Breakdown
Rodent models do not mimic all aspects of therapy of low dose anti-VEGF drugs and
human diabetes, but such models are essential steroids may offer advantages over anti-
for the development of new therapeutic VEGF drugs alone for the treatment of DR
approaches. While the short duration of that is resistant to conventional management.
diabetic injury in this model (3.5 weeks)
imposes some limitations on the interpretation ACKNOWLEDGEMENTS
of the data, similar findings relating to the Preliminary data were presented at the
function of VEGF-A and its receptors on the Australian Ophthalmic and Vision Science
BRB breakdown have also been observed in Meeting, Dec 2005. Dr. Xinyuan Zhang is
humans with diabetes, providing justification supported by a “University of Sydney – China
for our studies in rodents (45). Scholarship Council” scholarship. The study
This study adduces data that support the was also supported by the Lowy Medical
hypothesis that intravitreal corticosteroid Research Institute. The authors appreciate the
treatment improves diabetic retinal edema by technical assistant from the pathological
regulating the pathological expression of immuno-histo lab of the University of
VEGF-A and its receptors. We also show that Sydney.
IVTA does not alter physiological VEGF-A,
FLK-1 and FLT-1 levels in the normal retina.
This raises the possibility that combination

10