Abstract
Purpose.:
We investigated the efficacy of intravitreal injections of anti-VEGF agents in a rat model of anterior ischemic optic neuropathy.
Methods.:
We applied laser-induced photoactivation on the optic nerve head after intravenously administered rose bengal (RB). The rats immediately received an intravitreal injection of either ranibizumab or PBS. The density of retinal ganglion cells (RGCs) was calculated using retrograde FluoroGold labeling. Visual function was assessed by flash visual–evoked potentials (FVEP). We investigated TUNEL assays of the retinal sections and ED1 staining of the optic nerve.
Results.:
After treatment, the RGC densities in the anti-VEGF–treated rats were not statistically significant different from those of the PBS-treated rats (57.0% vs. 40.0% in the central retinas; 39.8% vs. 33.6% in midperipheral retinas, both P > 0.05). Measurements of FVEP showed no statistically significant differences in preserved latency or amplitude of the P1 wave between anti-VEGF and PBS groups (latency 131 ± 15 ms versus 142 ± 14 ms, P = 0.157; amplitude 34 ± 12 μv versus 41 ± 13 μv, P = 0.423). Assays of TUNEL showed that there was no statistical difference in the number of apoptotic cells in the RGC layers between anti-VEGF and PBS groups (7.0 ± 0.8 cells/high-power field [HPF] versus 7.8 ± 1.3 cells/HPF; P = 0.275). In the optic nerves, we did not observe statistically significant differences in ED1-positive cells/HPF between anti-VEGF and PBS groups (P = 0.675).
Conclusions.:
Intravitreal injections of anti-VEGF did not have a protective effect in the rat model of anterior ischemic optic neuropathy.
Nonarteritic anterior ischemic optic neuropathy (NAION) is the most common type of ischemic optic neuropathy in the elderly. Currently, there is no accepted treatment to rescue vision in patients with NAION. Although a previous study showed that systemic steroid treatment was effective in patients with NAION in whom the initial visual acuity was below 20/70,
1 the complications of systemic steroid use are a concern in clinical practice.
2 Intravitreal steroid injection in patients with NAION has been reported in some small case series with no control group, and local steroid treatment has been shown to improve visual acuity in some cases.
3–6
Cellular inflammation was shown to play a major early role following infarct in a primate NAION model.
7 In human patients with NAION, venous insufficiency causing initial disc edema associated with the creation of a compartment syndrome has been reported to be a possible pathogenesis.
8 In the retina, VEGF is essential for angiogenesis, and it also promotes vascular permeability in hypoxic conditions or vascular inflammation.
9 Agents of anti-VEGF such as bevacizumab (Avastin; Genentech, Inc., San Francisco, CA, USA) and ranibizumab (Lucentis; Novartis, Inc., Basel, Switzerland) have been reported to inhibit VEGF signaling and thereby potentially decrease optic disc edema.
10 Agents of anti-VEGF have also been shown to have an anti-inflammatory effect.
11 Therefore, it is reasonable to hypothesize that anti-VEGF agents may have therapeutic effects in NAION. However, the few studies having reported anti-VEGF treatment in patients with NAION show conflicting results. Improvements in visual acuity have been reported in some cases,
12,13 whereas a randomly controlled trial of intravitreal injections of triamcinolone combined with anti-VEGF agents in patients with NAION (15 injected eyes, 17 controls) reported no significant difference between the two groups.
14 Another nonrandomized controlled clinical trial in patients with acute NAION showed no difference between intravitreal bevacizumab (1.25 mg) and natural history in functional and anatomic improvements.
15 To date, no benefits of anti-VEGF treatment have been reported in patients with NAION.
In optic neuropathy, there is a concern that inhibition of VEGF function may cause neurodegeneration of retinal neurons. Thaler et al.
16 reported that the retinal ganglion cell (RGC) count in healthy rat eyes was essentially unchanged from those of control animals after the administration of both low and high concentrations of bevacizumab, ranibizumab, or pegaptanib (Macugen; Eyetech Pharmaceuticals, Inc., New York, NY, USA) for 2 months. In addition to inflammation role of VEGF, VEGF-A indicated to be a critical neuroprotectant with evidence of significant loss of RGCs after chronic anti-VEGF treatment in an ischemia/reperfusion model.
17 Vascular endothelial growth factor was reported to protect the neurodegeneration of RGCs via activating ERK-1/2 and Akt pathways in a model of optic nerve (ON) axotomy.
18 However, anti-VEGF agents were reported to contribute to RGC preservation in a mouse model of ON crush.
19 In an animal laser-induced choroidal neovascularization, no significant reduction was found in retinal function, the retinal level of mRNA for ganglion cell–specific genes, and RGC axon count after blockage of VEGF for 7 months in transgenic mice.
20
To the best of our knowledge, anti-VEGF agents have never been evaluated in a rat model of anterior ischemic optic neuropathy (rAION). Inasmuch as the results of in vivo animal experiments cannot be directly transferred to clinical NAION, physiologic and histologic changes in cellular inflammation in rAION are similar to those in early NAION.
7,21 In this study, we aimed to investigate whether intravitreal anti-VEGF therapy has anti-inflammatory and antiapoptotic effects, and whether it has the ability to rescue RGCs after the induction of rAION.
The rats received one intravitreal injection of ranibizumab (2 μL per injection, 10 mg/1 mL, 20 rats; Lucentis) or PBS (as the controls; 2 μL per injection, 20 rats) in the right eye 1 day after the induction of rAION. The final intraocular concentration used in the rats was no less than that used in routine clinical practice.
16 We used 33-G needles (Hamilton7747-01 with a Gastight syringe, IA2-1701RN 10-μL SYR; Hamilton Co., Hamilton, KS, USA) to perform the intravitreal injections. Intraocular pressure (IOP) was measured using a Tono-Pen (Reichert Technologies, Depew, NY, USA) 1 day after the intravitreal injections. The 20 control rats received a sham operation (laser treatment without intravenous rose bengal [RB]; Sigma-Aldrich Corp., St. Louis, MO, USA] injections). The rats were euthanized 4 weeks post infarct by CO
2 insufflation. Density of RGC was measured by retrograde labeling with FluoroGold (Fluorochrome, LLC, Denver, CO, USA), and visual function was assessed by flash visual-evoked potentials (FVEP) 4 weeks post infarct. Assays of TUNEL of the RGC layer and immunohistochemistry (IHC) of ED1 (markers of macrophage) expression in the ON were also conducted.
We used a sample size of six rats in every test per group that would have 77% probability of achieving a significant result, given the 1.13 effect size of the observed anti-VEGF versus PBS groups in RGC study (Cohen's d with α = 0.05 in G*Power Software; Heinrich-Heine-Universität Düsseldorf, Dusseldorf, Germany). After performing a Mann-Whitney U test, an experimental group size of approximately n = 16 has been required to have power of 80% in RGC study. In order to follow-up the principles of the three R's (replacement, reduction, and refinement) for humane animal research, we finally determined the n = 6 in every test per group.
Retinal protein samples containing 30 μg protein were separated on 12% sodium dodecyl sulphate-polyacrylamide gels and transferred to polyvinylidene difluoride membranes (Millipore Corp., Bedford, MA, USA). The membranes were incubated in Tris-buffered saline and Tween 20 ([TBST], 0.02-M Tris base, pH 7.6, 0.8% NaCl, 0.1% Tween 20) supplemented with 5% dry skimmed milk for 60 minutes to block nonspecific binding. After rinsing with TBST buffer, the samples were incubated with primary antibodies to VEGF (1:200, Santa Cruz Biotechnology, Inc., Dallas, TX, USA) and GAPDH (glyceraldehyde 3-phosphate dehydrogenase; 1:1000; Sigma-Aldrich Corp.) at 4°C overnight. The membranes were washed twice with TBST buffer followed by incubation with Biotin horseradish peroxidase–conjugated appropriate rabbit anti-mouse IgG secondary antibodies (1:10,000; Jackson ImmunoResearch Laboratories) at room temperature for 1 hour. The blots were then washed with TBST. The specific immune complexes were detected by ECL plus Western blotting reagents (GE; RPN2232; n = 3 in each group).
The results of the current study demonstrate that intravitreal injections of anti-VEGF did not have an anti-inflammatory effect on ONs or a rescuing effect on RGCs after the induction of rAION.
It has been reported that some patients develop NAION within 4 to 60 days after intravitreal injections of anti-VEGF agents, and a transient rise in intraocular pressure after the injection has been postulated to be a possible mechanism.
28 In our model, we used a 33-G needle with a 10-μL syringe (Hamilton Co.) to prevent trauma or IOP fluctuations from the injections. The average IOP the day after the injection of the anti-VEGF agent was 9.8 mm Hg, which was not significantly different to the IOP of the sham group (9.3 mm Hg) in our study. However, we did not note secondary IOP elevations after the intravitreal injections.
In an ischemia/reperfusion brain model, reduced tissue oxygen tension was reported to trigger VEGF expression and increase protein and mRNA levels of VEGF and its receptors (Flt-1, Flk-1/KDR), resulting in brain edema.
29 Antagonism of VEGF has been shown to reduce ischemia/reperfusion-related brain edema and injury, implicating VEGF in the pathogenesis of stroke and related disorders.
30,31 However, in a rat model of focal cerebral embolic ischemia, the systemic administration of VEGF after a longer period (after 48 hours) was found to markedly enhance angiogenesis in ischemic brains and reduce neurologic deficits during stroke recovery.
32 Vascular endothelial growth factor is known to have protective and antiapoptotic effects on brain neurons in hypoxic conditions through lessening caspase-3 activation and antiexcitotoxic effects, and it has been implicated to be an endogenous neuroprotective factor.
33–36
In our rAION model, immediate anti-VEGF intravitreal injections did not significantly rescue RGCs after ischemic optic neuropathy insults compared with the results of an ON crush model.
18 The dose and type of anti-VEGF agent were different between the two studies (the present study: ranibizumab 20 μg/2 μL; Rappoport's study: bevacizumab 75 μg/3 μL). In addition, the time course of RGC death has been reported to be longer in an rAION model (2–3 weeks) than in a crush model (7 days), and this delay in RGC death in rAION suggests that a potential treatment window does exist for anti-VEGF therapy.
37,38 In retina ischemia/reperfusion models, anti-VEGF treatment has been proven to inhibit vascular permeability without effecting apoptosis or cell degeneration.
39 In a model of ON axotomy, VEGF was found to protect neurodegeneration of RGCs via activating ERK-1/2 and Akt pathways.
14 To date, the therapeutic role of anti-VEGF agents in an ischemic ON model is uncertain. In the Western blot analysis of retinas after rAION, no significant changes of VEGF level were noted in our results, which is similar to a study that reported minimal changes in VEGF mRNA levels in the retinas of mice at 1, 3, and 21 days following rAION induction.
40 These observations imply that an rAION model may not involve angiogenesis and that VEGF plays a minor role in rAION.
We previously reported that recombinant human granulocyte colony-stimulating factor has dual actions of antiapoptosis for RGC survival and anti-inflammation in ONs in an rAION model.
24 Granulocyte colony-stimulating factor has been shown to stimulate neurogenesis in adult rat brains via the actions of VEGF and signal transducer and activator of transcription activation.
41 Vascular endothelial growth factor-A165a (VEGF-A165a) has been shown to modulate inflammatory pathways, resulting in upregulation of intercellular adhesion molecule 1 (ICAM-1) in retinal vascular endothelial cells.
42 Vascular endothelial growth factor-A165b has also been shown to inhibit tumor necrosis factor-α–mediated upregulation of ICAM-1 expression and increase monocyte–retinal pigment epithelium adhesion, suggesting an anti-inflammatory property of VEGF-A165b in the eye.
43 Extrinsic macrophages (blood-borne macrophages, which invade the damaged tissue) have been reported to be identifiable 3 days post induction in ONs by the presence of ED1 (CD68)-positive macrophages/microglia, indicating significant blood–brain barrier disruption.
25 In the current study, inflammatory activity in the ONs did not show statistically significant decrease after anti-VEGF treatment as evidenced by persistent infiltration of ED1+ cells, suggesting that intravitreal injections of anti-VEGF agents have no anti-inflammatory effect on ONs in rAION. The histologic studies of the ONs also demonstrated that inflammatory cell infiltration and vacuolization of axons were not changed by anti-VEGF treatment. To ensure that the dose of anti-VEGF agent in our model was sufficient, it was no less than used in clinical practice.
16 Visual functional evaluation by FVEP did not show statistically significant improvements in the amplitude and latency of the P1 wavelet after intravitreal injection of anti-VEGF agents in our rAION model.
In conclusion, we demonstrated that early intravitreal injections of an anti-VEGF agent did not have statistically significant protective effects on RGCs and ONs in our rAION model, as evidenced by RGC morphometry, TUNEL apoptotic assay, inflammation of the ONs, and functional assessment of FVEP.
The authors thank Malcolm Higgins for his assistance in editing and Su-Zen Chen for her illustration preparation and data assistance.
Supported by the Tzu Chi General Hospital (TCRD102-22).
Disclosure: T.-L. Huang, None; C.-H. Chang, None; S.-W. Chang, None; K.-H. Lin, None; R.-K. Tsai, None