Background

Although trabeculectomy is an effective method of reducing the intraocular pressure (IOP) in primary open-angle glaucoma (POAG), current methods lack the high level of predictability that we expect from other types of ocular surgery, such as phacoemulsification. All types of glaucoma filtration surgery that function by draining aqueous to the sub-Tenon space, including trabeculectomy, some types of nonpenetrating surgery, and aqueous shunts, depend on external resistance created by wound healing. The predictability in the outcome of these procedures is hampered by variability in the wound-healing response. Hypotony results when healing is inadequate, whereas a return to high IOP occurs when healing is excessive.

With trabeculectomy, healing under the scleral flap may play a role in some cases, but overall, Tenon fibroblasts are regarded as the main effector cells in the initiation and mediation of most wound healing and fibrotic scar formation. After trabeculectomy and other types of glaucoma filtration surgery, the development of conjunctival and episcleral fibrosis occurs as a result of progressive fibroblast migration, proliferation, collagen deposition, and angiogenesis at the site of filtration. Histologic studies have shown that the maximum proliferation of subconjunctival fibroblasts occurs in the third to fifth postoperative day. Pharmacologic enhancement of trabeculectomy using 5-fluorouracil (5-FU) and mitomycin C (MMC) has improved rates of success significantly. However, the nonspecific mechanism of action of these agents may result in widespread cell death and thin-walled avascular blebs that are susceptible to leakage, infection, and dysaesthesia. With lower levels of exposure, the frequency of bleb-related complications are less, but filtration failure is more common. This variability in outcome effectively presents a tradeoff in which safer procedures are less effective and more effective procedures are less safe. A degree of predictability is required for filtration surgery that is similar to that observed with phacoemulsification. Modulators of wound healing that result in more predictable outcomes may assist the achievement of this goal.

In 1994, Wong and associates reported that angiogenesis inhibitors demonstrated marked inhibitory effects on both human Tenon fibroblast proliferation and migration in vitro. This study was performed before the availability of the anti-VEGF monoclonal antibodies, but raised the possibility that the latter may have a role in the control of wound healing after filtering surgery. Since then, there has been ample evidence that VEGF has a pivotal role in the wound-healing response. This review examines the evidence for the use of bevacizumab and other anti-VEGF monoclonal antibodies in the control of wound healing after glaucoma filtration surgery. Not only does VEGF seem to drive fibrosis via angiogenesis, but also there is evidence of a more direct effect on fibroblastic activity. VEGF has been shown to be a mediator in the signal transduction cascade leading to fibroblast migration and proliferation. As a powerful inducer of angiogenesis, VEGF also promotes early migration of inflammatory cells and fibroblasts. Delayed wound healing is a known complication of the systemic use of bevacizumab, and at the time of writing, delayed wound healing is listed as a warning on the manufacturer’s prescribing information for bevacizumab (Avastin; Genentech, San Francisco, California, USA).

Anti–Vascular Endothelial Growth Factor Monoclonal Antibodies

Bevacizumab is a recombinant humanized monoclonal immunoglobulin (Ig) G1 antibody that binds to and inhibits the biologic activity of human VEGF in vitro and in vivo assay systems. Bevacizumab contains human framework regions and the complementarity-determining regions of a murine antibody that binds to VEGF. Bevacizumab has a molecular weight of 149 kD and is produced in the Chinese-hamster–ovary mammalian-cell expression system (Avastin full prescribing information; available at: http://www.gene.com/gene/products/information/pdf/avastin-prescribing.pdf ; accessed December 28, 2010). It is approved by the United States Food and Drug Administration for the treatment of metastatic colorectal and breast cancer. However, like other antiscarring agents such as 5-FU and MMC, bevacizumab is not approved for use in the eye. There are few reports of use of other anti-VEGF monoclonal antibodies, such as ranibizumab (Lucentis; Genentech), and none using pegaptanib (Macugen; OSI Pharmaceuticals, Long Island, New York, USA) in glaucoma surgery, so the main focus of this article is bevacizumab.

Anti–Vascular Endothelial Growth Factor Monoclonal Antibodies

Bevacizumab is a recombinant humanized monoclonal immunoglobulin (Ig) G1 antibody that binds to and inhibits the biologic activity of human VEGF in vitro and in vivo assay systems. Bevacizumab contains human framework regions and the complementarity-determining regions of a murine antibody that binds to VEGF. Bevacizumab has a molecular weight of 149 kD and is produced in the Chinese-hamster–ovary mammalian-cell expression system (Avastin full prescribing information; available at: http://www.gene.com/gene/products/information/pdf/avastin-prescribing.pdf ; accessed December 28, 2010). It is approved by the United States Food and Drug Administration for the treatment of metastatic colorectal and breast cancer. However, like other antiscarring agents such as 5-FU and MMC, bevacizumab is not approved for use in the eye. There are few reports of use of other anti-VEGF monoclonal antibodies, such as ranibizumab (Lucentis; Genentech), and none using pegaptanib (Macugen; OSI Pharmaceuticals, Long Island, New York, USA) in glaucoma surgery, so the main focus of this article is bevacizumab.

Aqueous and Plasma Vascular Endothelial Growth Factor and Vascular Endothelial Growth Factor Receptors

Aqueous VEGF levels have been reported to be elevated consistently in eyes of POAG patients undergoing either cataract surgery or trabeculectomy when compared with eyes without comorbidity undergoing cataract surgery. For example, in 1998, Tripathi and associates reported aqueous mean ± standard error of the mean VEGF concentrations of 0.257 ± 0.043 ng/mL for 20 eyes undergoing cataract surgery alone, 0.726 ± 0.204 ng/mL for 28 POAG eyes, and 29.276 ± 7.350 ng/mL for 12 eyes with neovascular glaucoma. The mean level in POAG eyes was significantly higher than that in those with only age-related cataract ( P < .05). Both Hu and associates and Li and associates reported similar findings to those of Tripathi and associates, but they differed in that VEGF was detected in all of the aqueous humor samples in both studies, in contrast to 15 of 28 POAG samples and 4 of 20 cataract samples reported by Tripathi and associates. Both Hu and associates and Li and associates attribute their higher rate of VEGF detection in aqueous to the use of a double-antibody sandwich enzyme-linked immunosorbent assay technique with a greater sensitivity than that used by Tripathi and associates. Upregulated levels of VEGF in glaucomatous eyes in these studies may be related to the disease, but also may be related to other factors such as glaucoma medication or case selection, because these were not controlled studies. However, it is interesting that the observation of elevated VEGF in POAG was consistent in 3 studies.

Two high-affinity VEGF tyrosine kinase receptors have been identified: fms -like tyrosine kinase 1 and kinase domain receptor. These are found predominantly on vascular endothelial cells. However, Li and associates demonstrated that both fms -like tyrosine kinase 1 and kinase domain receptor are expressed on human and rabbit Tenon fibroblasts at both the mRNA and protein level using quantitative real-time reverse-transcriptase polymerase chain reaction, and it seems that there is a high degree of cross-reactivity between rabbit and human VEGF.

Animal Studies

Esson and associates, using a microarray technique, reported a large-scale gene expression analysis of bleb tissue biopsies from rats 2, 5, and 12 days after glaucoma filtration surgery. They found elevated VEGF gene expression as well as other cytokines in the postoperative bleb. Li and associates reported an increase in aqueous humor VEGF protein concentration after trabeculectomy in rabbits that did not returned to baseline after 30 days. This upregulation was suppressed partially for up to 6 days by a single injection of bevacizumab into the subconjunctival space (200 μL of 25 mg/mL) and anterior chamber (100 μL of 25 mg/mL) at the time of trabeculectomy. They did not observe a significant benefit in terms of IOP lowering up to 29 days after surgery, but they did report a larger bleb area using the Moorfields bleb classification in eyes treated with bevacizumab. Suppression of aqueous VEGF also was demonstrated by Miyake and associates, who reported that 1.25 mg bevacizumab injected into the vitreous of 3 cynomolgus macaque eyes not only resulted in a reduction in aqueous VEGF concentrations to an undetectable level from 1 to 28 days after injection, but also that the aqueous VEGF concentration did not return to preinjection levels until 42 days after injection.

Memarzadeh and associates randomized 42 rabbits to trabeculectomy using a much more aggressive regimen of 7 subconjunctival injections of 1.25 mg bevacizumab, 5 mg 5-FU, or 0.1 mL balanced salt solution, given during the first 14 postoperative days. They reported no differences in mean IOP throughout the study period, but bevacizumab did significantly improve the outcome of filtration surgery as defined by bleb survival ( P < .05) compared with the 5-FU and balanced salt solution groups. Bleb height and bleb area also were increased. All blebs in the 5-FU group had failed by day 10, although 1 control group bleb remained functional at day 10. In the bevacizumab groups, all but 2 blebs had failed by postoperative day 16. The mean number of survival days was 7.4, 6.9, and 16.0 days for the control, 5-FU, and bevacizumab groups, respectively. Eyes treated with bevacizumab also had less scarring on histologic examination at postoperative days 10 and 20 ( P = .009).

How and associates reported 100% bleb survival in the rabbit model at 28 days when subconjunctival bevacizumab (2.5 mg) was combined with 5-FU (5 mg) and found 100% bleb survival at day 28 in this group. In contrast, 50% of blebs survived to day 28 in the bevacizumab only group, 25% in the 5-FU–only group, and 0% in the control group (0.1 mL phosphate-buffered saline). Conjunctival vascularity was reduced significantly in the bevacizumab-only and combined 5-FU groups. Specimens also were obtained from the rabbits and were stained for collagen fibers. Interestingly, control and 5-FU–treated eyes showed densely packed mature collagen I, compared with the 2 groups treated with bevacizumab, which demonstrated a more loose arrangement of collagen fibers in the subconjunctival space. There was also a significant reduction in the transcript levels of collagen I and fibronectin mRNA in the bevacizumab-treated groups.

A limitation of animal studies of glaucoma filtration surgery in general is the difficulty in obtaining demonstrable IOP differences as primary outcomes, requiring the use of bleb morphologic features instead.

In Vitro Studies

Two in vitro studies confirm the direct effect of bevacizumab on human Tenon fibroblast proliferation. Li and associates reported that the addition of VEGF induced a significant increase in human ( P = .04) and rabbit ( P = .02) Tenon fibroblast proliferation in tissue culture. This was inhibited subsequently by administration of bevacizumab in a dose-dependent manner. O’Neill and associates demonstrated that in an in vitro model of wound healing, bevacizumab disrupts fibroblast proliferation, inhibits collagen gel contraction ability, and induces fibroblast cell death at concentrations of more than 7.5 mg/mL in serum-free conditions.