Antifibrotic Activity of Bevacizumab on Human Tenon's Fibroblasts In Vitro

Abstract

Purpose.: To evaluate the effect of the anti–VEGF-A monoclonal antibody bevacizumab on primary human Tenon's capsule fibroblasts (HTFs) in an in vitro model of wound healing.

Methods.: Fibroblasts were cultured in RPMI media, and bevacizumab was administered at a concentration ranging from 0.25 to 12.5 mg/mL. Fibroblast viability and cell death were assessed using the MTT colorimetric assay, lactate dehydrogenase assay, BrdU assay, and live/dead assay. Fibroblast contractility was assessed in floating collagen gels. Morphologic changes were assessed by transmission electron microscopy. Antifibrosis activities were compared with 5-fluorouracil.

Results.: Bevacizumab induced a significant dose-related reduction of HTF cell number at 12.5 mg/mL at 72 hours (P < 0.05). Under serum-free conditions, bevacizumab induced significant fibroblast cell death at concentrations greater than 7.5 mg/mL (P < 0.05). Bevacizumab caused a moderate inhibition of fibroblast gel contraction from baseline (P < 0.05). Scanning electron microscopy revealed marked vacuolization in bevacizumab-treated fibroblasts.

Conclusions.: Bevacizumab disrupted fibroblast proliferation, inhibited collagen gel contraction ability, and induced fibroblast cell death at concentrations greater than 7.5 mg/mL in serum-free conditions. These results demonstrated that bevacizumab inhibited a number of fibrosis activities in culture. These activities may underpin the antifibrosis effect proposed in vivo.

Excess scarring at the site of a filtering bleb is the most common cause of failure after glaucoma filtration surgery. ^{1 –3} Tenon's fibroblasts are the main effector cells in the initiation and mediation of wound healing and fibrotic scar formation after trabeculectomy. ³ Success rates of filtration surgery have significantly improved with the use of adjunctive antifibrotic agents such as 5-fluorouracil (5-FU) and mitomycin C (MMC). ^{4 –6} These agents, however, can induce significant cell death in treated tissues ⁷ that may contribute to antifibrosis activity but simultaneously predispose to potentially sight-threatening complications, including hypotony, wound leak, blebitis, and endophthalmitis. ⁸ Despite these treatments, a number of patients continue to scar excessively, and improved methods for titrating and controlling the wound healing response are sought. 

Angiogenesis, the formation of new capillaries from preexisting blood vessels, occurs in both health and disease. It is implicated in tumorigenesis and metastasis, ⁹ rheumatoid arthritis, ¹⁰ and blinding ocular conditions including proliferative diabetic retinopathy ¹¹ and choroidal neovascularization in age-related macular degeneration (ARMD). ¹² Angiogenesis is also a critical component of wound healing because it allows early migration of inflammatory cells and fibroblasts into the wound. The family of vascular endothelial growth factors (VEGF) has been identified as the primary regulators of angiogenesis ^{13 –15} and VEGF-A as the primary regulator driving angiogenesis. As a result multiple antiangiogenic agents, targeting VEGF-A or its receptor VEGFR2, have been developed to specifically target and treat VEGF-A–driven ocular pathology. ¹⁴  

Anti-VEGF monoclonal antibodies have been developed to treat solid tumors ^{16 –18} and now also form part of the clinical management of ocular neovascular diseases. ¹⁹ Bevacizumab (Avastin; Genetech, Inc., South San Francisco, CA) is a full-length humanized monoclonal antibody directed against all isoforms of VEGF-A (VEGF-121, -145, -165, -183, -189, and -209) and is approved by the US Food and Drug Administration for the treatment of metastatic colorectal and metastatic breast cancer. ^17,18 It binds and neutralizes all human VEGF-A isoforms and bioactive proteolytic fragments. ²⁰ Wound healing complications, including early delayed wound closure or dehiscence in anastomosis (after colorectal surgery), have been reported after the administration of intravenous bevacizumab. ^21,22 This dehiscence may occur several months after the original surgical anastomoses, suggesting bevacizumab may induce long-term inhibition of wound healing. 

Intravitreal use of bevacizumab has a good safety profile in humans, ²³ and there are several reports of its use in proliferative diabetic retinopathy, ARMD, inflammatory ocular neovascularization, and neovascular glaucoma. ^{24 –30} There are also isolated case reports of its use in glaucoma filtration surgery, particularly in neovascular glaucoma. ^31,32 However, to date, the efficacy of bevacizumab in filtration surgery, specifically the effect of bevacizumab on human Tenon's fibroblasts (HTFs) in culture, is largely unknown. 

In this study, we describe the effects of the anti–VEGF-A monoclonal antibody bevacizumab on HTF in an in vitro model of wound healing. 

Methods

Human Tenon's Fibroblast Explant Culture

HTFs were propagated from explanted subconjunctival Tenon's capsule isolated during glaucoma filtration surgery, as described previously. ³³ The tenets of the Declaration of Helsinki were observed, institutional human ethics committee approval was granted, and written informed consent was obtained from all patients. Explanted tissue was attached to the bottom of a six-well plate (Greiner Bio-One, Jena, Germany) with a sterile coverslip and overlaid with RPMI (Sigma-Aldrich, St. Louis, MO). All culture media were supplemented with l-glutamine 2 mM, penicillin 100,000 U/L, and streptomycin 10 mL/L (all Sigma-Aldrich). For propagation of fibroblasts, the media were also supplemented with fetal calf serum (FCS; 10% of final volume; JRH Biosciences, Lenexa, KS). HTFs were routinely cultured in RPMI media, as described. Once the monolayers reached confluence, fibroblasts were passaged and subcultured in 175-cm² tissue culture flasks (Greiner Bio-One). Cells were incubated at 37°C/5% CO₂ in a humidified incubator. 

Fibroblast Proliferation and Viability Studies

The MTT assay is a convenient and accurate way of determining mammalian cell viability. This colorimetric assay quantifies the number of metabolically active cells based on the cleavage of the yellow tetrazolium salt MTT (3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyltetrazolium bromide) to purple formazan crystals. The assay was performed according to the manufacturer's instructions (Cell Proliferation Kit 1; Roche Diagnostics, Basel, Switzerland). Fibroblasts were plated at 5000 cells/well in 96-well plates and incubated, on attachment, in 50 μL media with bevacizumab (Pharmatel Fresenius Kabi, Bad Homborg, Germany) diluted to concentrations of 0.25, 2.5, 5, 7.5, 10, and 12.5 mg/mL for 24, 48, and 72 hours. Experiments were performed in both 10% FCS media and serum-free media conditions. Viability experiments were repeated comparing bevacizumab with two isotype control antibodies: a humanized monoclonal IgG1κ chimeric antibody in which variable regions are deimmunized and directed toward fibrin degradation products (proprietary of and kindly supplied by Agen Ltd., Melbourne, Australia) and a nonhumanized IgG isotype. 

Cell death was quantified using the lactate dehydrogenase (LDH) assay, based on the measurement of LDH activity released from the cytosol of dying cells into the supernatant. Fibroblasts were plated at 5000 cells/well in 96-well plates and were treated with increasing concentrations of bevacizumab to 12.5 mg/mL, as described. Experiments were performed in serum-free media conditions because of intrinsic LDH activity in serum. LDH was measured only at the 72-hour time point. At 72 hours, 80 μL supernatant (media) was transferred to a new 96-well plate, taking care not to disturb the monolayer of fibroblasts. The assay was performed according to the manufacturer's instructions (Cytotoxicity Detection Kit; Roche Diagnostics), and 1% Triton X-100–treated cells served as positive controls for 100% cell death. 

Fibroblast proliferation was measured with a 5-bromo-2-deoxyuridine (BrdU) assay that quantitates BrdU uptake into newly synthesized DNA of replicating cells. HTFs (1 × 10⁴) were plated into separate wells of a 24-well plate and incubated with BrdU after bevacizumab treatment for 24 and 48 hours. Proliferating cells were detected by BrdU labeling directed by the BrdU detection kit manufacturer's protocol (Kit I; Roche Diagnostics). Five hundred to 1000 cells were randomly counted in each well. The average of three wells per sample was counted. The BrdU labeling index was calculated as the percentage of BrdU-positive cells. Representative images were captured using an epifluorescent inverted phase-contrast microscope (TE2000S; Nikon, Tokyo, Japan) at 100× magnification. 

Viability/cytotoxicity assay (Live/Dead; Invitrogen Molecular Probes, Carlsbad, CA) was used to determine the ratio of live and dead fibroblasts after bevacizumab treatment. Dye concentrations for calcein AM and ethidium homodimer-1 (EthD-1) were optimized according to the kit's protocol before experimentation. HTFs were incubated in serum-containing and serum-free RPMI. HTFs were treated with bevacizumab (10 mg/mL) in accordance with the previous protocol. Control wells were composed of 1% digitonin (for 100% cell death) and 50% PBS mixture with culture media; 48-well plates were then incubated at 24, 48, and 72 hours. At each time point, the corresponding wells were washed with 1 mL PBS to remove any esterase activity present in serum-supplemented RPMI because serum esterase can hydrolyze calcein AM and cause increased extracellular fluorescence. The fibroblasts were then incubated with the dyes calcein AM and EthD-1 at room temperature for 30 minutes. Cells were washed with PBS and viewed under a fluorescence microscope. Five random fields from each well were photographed and counted. 

Floating Collagen Contraction Studies

To assess the influence of bevacizumab on fibroblast contraction, we measured the contraction of fibroblast-seeded collagen gels. First, fibroblasts were resuspended at a density of 5 × 10⁵ cells/mL in either 10% FCS (JBH Biosciences, Lenexa, KS) or concentrated serum-free medium. Each gel was made from 125 μL of 4× concentrated medium and 200 μL dialyzed collagen 0.1% acetic acid and mixed gently. Sodium hydroxide (0.1 M) was then added to the gel to return the solution to physiological pH and to precipitate the collagen. Thirty microliters of fibroblasts (at 5 × 10⁵ cells/mL) were then seeded into the neutralized gel and resuspended briefly. This was added to a well of a 48-well plate and incubated at 37°C for 15 minutes to set. Three hundred microliters of bevacizumab (PFK 25 mg/mL) diluted in RPMI (10% FCS, L-glutamate, penicillin/streptomycin) to concentrations of 2.5, 7.5, and 12.5 mg/mL was added to the wells containing solidified collagen gels. The gels were then gently detached from the wells, and culture medium was added. The free-floating fibroblast-populated collagen gels were incubated for 7 days. Images of collagen gels were taken with a digital camera (CyberShot DSC-S700; Sony, Tokyo, Japan) on days 0, 1, 2, 3, and 7. The images were then measured and assessed using ImageJ software (developed by Wayne Rasband, National Institutes of Health, Bethesda, MD; available at http://rsb.info.nih.gov/ij/index.html). 

Electron Microscopy

Cells treated with bevacizumab (10 mg/mL), and controls in serum and in serum-free RPMI were prepared as follows. Suspended HTFs were fixed in 2.5% glutaraldehyde in 0.1 M cacodylate buffer and 2% osmium tetroxide in distilled water for 30 minutes, respectively, and spun at 3000 rpm for 4 minutes to obtain a single-cell pellet. The pellet was then washed 2 × 10 minutes with distilled water and spun down to reach a concentrated pellet. Then 1% agar was poured over the pellet and was allowed to solidify at 4°C, after which it was sliced into pieces of 1 mm³ and processed in 3% uranyl acetate, followed by dehydration with acetone for routine electron microscopy. Each piece of pellet was embedded in an individual mold, and the resin was cured in a 60° oven for 24 hours. Specimen sections 50- to 90-μm thick were cut using a microtome (Ultracut; Reichert-Jung, Wetzlar, Germany) and a diamond knife (Diatome, Biel, Switzerland), collected on 200-mesh copper grids, and poststained in 5% uranyl acetate in ethanol (10 minutes) and Reynold's lead citrate (5 minutes). Specimens were viewed with a transmission electron microscope (CM10; Philips, Eindhoven, The Netherlands) operating at 60,000 V. Negatives were converted to digital images (Photoshop 7; Adobe, Mountain View, CA). 

Statistical Analysis

Where practical, each laboratory experiment was performed with at least three replicates per treatment group. Data were assumed to follow the normal distribution. Statistically significant differences between two data points were analyzed using the paired t-test. Statistically significant differences among three or more data points were analyzed using the repeated-measures one-way ANOVA, followed by the Bonferroni post test in some cases. Data were deemed statistically significant at P < 0.05. All analysis was performed using Microsoft software packages (Excel; Microsoft, Redmond, WA). 

Results

Proliferation Inhibition of HTFs by Bevacizumab

The effect of bevacizumab on fibroblast proliferation was quantified with MTT assay to determine the number of viable cells. A statistically significant decrease in fibroblast number was observed at a concentration of bevacizumab of 12.5 mg/mL (P < 0.05) in cells with 10% FCS (Fig. 1A). In serum-free conditions, there was a significant reduction in fibroblast number at a concentration of bevacizumab of 5.0 mg/mL to 12.5 mg/mL (P < 0.05; Fig. 1B). Interestingly, there was a statistically significant increase in fibroblasts at a concentration of 2.5 mg/mL bevacizumab in serum-free conditions. (P < 0.05). Proliferation rates were also measured using BrdU incorporation. This demonstrated significant inhibition at 48 hours with bevacizumab (10 mg/mL) compared with controls (P < 0.05; Fig. 2). 

Figure 1.

 

Bevacizumab-treated HTFs (continuous 72-hour) MTT assay. (A) 10% FCS media conditions. Reduction in MTT absorbance at 12.5 mg/mL bevacizumab in cells in media with 10% FCS (P < 0.05). *P < 0.05 with respect to cells in media with 10% FCS with no bevacizumab treatment. (B) Serum-free conditions MTT assay. Reduction in MTT absorbance at concentrations of 5 mg/mL to 12.5 mg/mL in cells in media (P < 0.05). Significant increase in absorbance at a concentration of 2.5 mg/mL bevacizumab compared with control cells (P < 0.05; n = 3). *P < 0.05 with respect to cells in serum-free media with no bevacizumab treatment (control).

Figure 2.

 

Bevacizumab inhibition of HTF proliferation as measured with BrdU assay. HTFs were (A) untreated cells and (B) cells incubated in bevacizumab (10 mg/mL). Nonproliferating cells stained blue, and proliferating cells stained green. (C) Histogram of BrdU positivity in bevacizumab-treated cells. Values plotted are mean ± SD (n = 3). *P < 0.05 with respect to untreated cells.

Bevacizumab-Induced Fibroblast Cell Death

The LDH assay was then used to assess cumulative cell death 72 hours after bevacizumab treatment. In serum-free conditions, bevacizumab at 7.5 mg/mL or greater had a significant cytotoxic effect on HTFs (P < 0.001, paired t-test). Treatment effect appeared to plateau beyond 7.5 mg/mL (Fig. 3A). This was correlated with a reduction in cell numbers determined by MTT assay beyond 7.5 mg/mL bevacizumab (Fig. 3B). 

Figure 3.

 

Bevacizumab-treated HTFs (continuous 72 hours) in serum-free conditions, (A) LDH assay, and (B) MTT assay. (A) Cell death measured by LDH assay (cumulative cell death). *P < 0.05 with respect to cells with no bevacizumab treatment. (B) Cell viability measured by MTT assay (n = 3). *P < 0.05 with respect to cells with no bevacizumab treatment.

We next investigated the effect of serum on bevacizumab-induced cell death using a live/dead assay in bevacizumab (10 mg/mL), and digitonin-treated cells (positive controls) in serum-containing and serum-free conditions given that LDH assays cannot be performed in serum-containing conditions because of the intrinsic activity of LDH in serum. At 10 mg/mL, bevacizumab induced no significant cell death in 10% FCS to 72 hours (Fig. 4A). However, it induced almost 100% fibroblast cell death at 24 hours in serum-free conditions. (Figs. 4B, 4C). 

Figure 4.

 

Effect of bevacizumab on HTF cell death. Live/dead assay was used to score the number of live (green) to dead (red) cells at 24 and 72 hours after treatment. (A) HTFs were incubated in RPMI ± bevacizumab in cycling cells compared with digitonin control. (B) HTF in serum-free media ± bevacizumab in noncycling cells compared with digitonin control. (C) Histogram of cell death in bevacizumab-treated cells. Bevacizumab (10 mg/mL) induced significant fibroblast cell death at 72 hours in serum-free conditions (n = 3).

To confirm that the effect on fibroblast viability was specific to bevacizumab and not a nonspecific consequence of high antibody concentration, we compared proliferation assays of bevacizumab with two isotype control antibodies. Experiments were performed in serum-free conditions comparing the effect of 7.5 mg/mL bevacizumab (known to inhibit proliferation in serum-free conditions) with 7.5 mg/mL isotype control antibodies. The humanized control antibody had no significant effect on the number of viable cells. There was a statistically significant decrease in the number of viable fibroblasts with bevacizumab-treated cells only (P < 0.05; Fig. 5). 

Figure 5.

 

Bevacizumab-treated HTF and humanized isotype control antibody-treated HTF (continuous 72 hours) in serum-free conditions MTT assay (n = 3). *P < 0.05 with respect to cells in serum-free media with no treatment.

Bevacizumab Effect on HTF-Mediated Collagen Gel Contraction

Bevacizumab at 12.5 mg/mL had a moderate inhibitory effect on collagen gel contraction compared with untreated cells (P < 0.001, two-way ANOVA; Fig. 6A). A dose response was seen between 2.5 mg/mL and 12.5 mg/mL bevacizumab treatment (Fig. 6B). Under serum-free conditions, 12.5 mg/mL bevacizumab increased the inhibition of collagen gel contraction (78% collagen gel area) compared with controls (27% collagen gel area; Fig. 6C). Inhibition of HTF contraction with bevacizumab was increased in serum-free conditions. 

Figure 6.

 

Floating collagen gel contraction studies. (A) Floating collagen gel contraction studies, 7 days (n = 3; P < 0.05, two-way ANOVA). (B) Floating collagen gel contraction studies seeded with HTF prepared in 100% serum and treated with bevacizumab concentrations of 2.5 mg/mL, 7.5 mg/mL, and 12.5 mg/mL (n = 3). (c) Floating collagen gel contraction studies, 7 days (HTFs resuspended in serum-free medium), bevacizumab 12.5 mg/mL (n = 3; P < 0.001, two-way ANOVA).

Morphologic Changes in HTF after Bevacizumab Treatment: Phase-Contrast Microscopy and Transmission Electron Microscopy

Analysis of cell morphology under phase-contrast (Fig. 7) and the live/dead assay (Figs. 4A–C) showed that under serum-free conditions, 10 mg/mL bevacizumab induced high levels of cell death. To assess the mode of cell death, scanning electron microscopy examining the nuclear morphology of HTF after bevacizumab (10 mg/mL) treatment was performed (Fig. 8). In bevacizumab-treated fibroblasts, there was the characteristic morphologic appearance seen in serum-free cells. Transmission electron microscopy images revealed significant vacuolization of internal HTF cell cytoplasm with complete lysis of HTF cell membranes and cell debris. 

Figure 7.

 

Effect of serum and bevacizumab on HTF morphology (n = 3).

Figure 8.

 

Scanning electron microscopy images of HTFs. (A) HTFs incubated in RPMI with 10% serum showing normal cellular structure and normal nucleus and cytoplasm. (B) HTFs incubated in serum-free RPMI. (C) HTFs incubated in bevacizumab 10 mg/mL with 10% serum showing significant vacuolization of cytoplasm (arrows). (D) HTFs incubated in serum-free bevacizumab 10 mg/mL showing only cellular debris.

Discussion

In the present study, we demonstrated that the anti–VEGF-A monoclonal antibody bevacizumab induced potent antifibrotic activity through significant reduction in fibroblast viability, inhibition of HTF proliferation, induction of low levels of cell death, and inhibition of cell-mediated collagen gel contraction in vitro. The concentrations of bevacizumab required to induce cell death were higher in serum-containing conditions than in serum-free conditions because VEGF in serum binds bevacizumab and inactivates it. 

Excessive postoperative scarring is the most common cause of failed glaucoma filtration surgery. Intraoperative MMC and 5-FU are clinically used to inhibit fibrosis and improve surgical outcomes. ^{4 –6} However, despite their use, a significant failure rate persists, with an associated increase in other postoperative complications. VEGF is a key molecule in the wound healing response. ³⁴ Humanized anti–VEGF-A monoclonal antibodies (bevacizumab and ranibizumab) are used widely in ophthalmology to inhibit angiogenesis in ARMD and to enable a good safety profile with ocular administration. ^{25 –28}  

Angiogenesis plays a vital role in wound repair, facilitating wound closure by enabling inflammatory cells and fibroblasts to migrate to the wound and by providing the vascular scaffold for granulation tissue formation. ³⁴ Vascular remodeling occurs because of the carefully balanced interplay of proangiogenic and antiangiogenic factors. Both angiogenic agonists and antagonists have been identified at various stages of wound repair. It is now widely established that VEGF-A is responsible for normal vasculogenesis, hemangiogenesis, and lymphangiogenesis. ^{13,35 –37} Despite this, however, relatively little attention has been given to the reported antifibrotic effects of VEGF-A. 

In addition to the expected elevated levels of VEGF-A in ocular fluids, diabetic retinopathy, and other retinal vascular disorders, ¹¹ raised VEGF-A levels in the aqueous humor of patients with nonneovascular glaucoma have also been reported. ^38,39 This increase in VEGF in the aqueous humor of glaucoma patients may contribute to postoperative inflammation and fibrosis. VEGF-A receptors have also recently been shown to be expressed in HTF. ³⁹ Thus, targeting the VEGF-A molecule would appear to be a plausible method of reducing the post-operative scarring response after glaucoma filtration surgery. 

VEGF inhibition has also been shown to attenuate fibrosis in a murine model of allergic airway disease, ⁴⁰ to induce a profibrogenic gene expression profile in glomerular endothelial cell lines, ⁴¹ and to reduce fibrosis in cutaneous wounds of adults. ⁴² Recent animal studies using the rabbit model of trabeculectomy, a model with a known vigorous wound healing response, have shown subconjunctival bevacizumab significantly improved glaucoma filtration surgery success and bleb survival ^39,43 because of the combined inhibition of angiogenesis during the initial phase of healing and the reduced fibrosis at later stages of wound repair. Li et al. ³⁹ found a single dose of 0.75 mg bevacizumab (0.3 mL of 25 mg/mL) given immediately after surgery significantly reduced the density of blood vessels during the early stages of wound healing and reduced collagen deposition in the later stages. Memerzadeh et al., ⁴³ using seven injections of 1.25 mg bevacizumab (0.05 mL of 25 mg/mL) during the first 14 days after surgery found it reduced collagen and elastic fiber deposition, reduced fibroblast differentiation into myofibroblasts, and resulted in a loss of fibroblast mitotic activity. Recent clinical reports have explored the use of bevacizumab after glaucoma filtering surgery, indicating that the agent can be administered safely at the time of surgery or in the postoperative period. ^{32,44 –46} In the largest of these, Grewal et al., ⁴⁴ using 1.25 mg bevacizumab (0.05 mL of 25 mg/mL) immediately after trabeculectomy, found improved bleb survival at 6 months. This study, however, was conducted in a small sample size, and follow-up was conducted over a short-period, nonrandomized, noncontrolled small case series of 12 patients only. 

Our findings show that bevacizumab, through the inhibition of VEGF-A, reduces wound healing and scar formation at the level of the HTF through the combined inhibition of fibroblast proliferation and induced fibroblast cell death in addition to its effect on angiogenesis. This, in conjunction with previous animal studies ^39,43 and small case reports in humans, supports the notion that pharmacologic neutralization of VEGF-A with the administration of bevacizumab offers a potentially safe and effective adjunctive therapy to prevent the failure of trabeculectomy. It is enticing to consider that late bleb failure may be treated with anti-VEGF therapy in future studies to address optimization and dosage scheduling, safety profile, and long-term sequelae needed. 

The findings from this study provide strong evidence that VEGF-A is a key mediator for the development of conjunctival vascularization and subconjunctival fibrosis. Furthermore, in parallel studies, we have also found that the combined delivery of bevacizumab and 5-FU offers a synergistic elevated antifibrotic effect when compared with their independent usage both in vitro and in vivo (JGC, unpublished data, 2008). Thus, bevacizumab potentially works synergistically with 5-FU to deliver a more profound effect on fibrosis. It is proposed that the antiangiogenic and antifibrotic effects of bevacizumab could provide a magnified inhibitory effect on the wound healing response. 

In conclusion, to date anti–VEGF-A treatment has targeted pathologic angiogenesis for both systemic and ocular neovascular disorders. We have shown that bevacizumab, through the inhibition of VEGF-A, exerts a potent antifibrotic effect through the inhibition of fibroblast proliferation, induction of fibroblast cell death, and inhibition of cell-mediated collagen gel contraction. These findings support the notion that adjunctive treatment with the VEGF-A inhibitor bevacizumab has the potential to improve surgical outcomes after glaucoma filtration surgery with greater safety and efficacy. 

Footnotes

 Disclosure: E.C. O'Neill, None; Q. Qin, None; N.J. Van Bergen, None; P.P. Connell, None; S. Vasudevan, None; M.A. Coote, None; I.A. Trounce, None; T.T.L. Wong, None; J.G. Crowston, None