Complementary effects of bevacizumab and MMCin the improvement of surgical outcome afterglaucoma filtration surgery

Tine Van Bergen,1 Evelien Vandewalle,2 Lieve Moons1,3 and Ingeborg Stalmans1,2

1Laboratory of Ophthalmology, KU Leuven, Leuven, Belgium2Department of Ophthalmology, University Hospitals Leuven, Leuven, Belgium3Unit of Animal Physiology and Neurobiology, Biology Department, KU Leuven, Leuven, Belgium

ABSTRACT.

Purpose: To determine the optimum administration route of bevacizumab after

glaucoma filtering surgery (GFS) and to investigate whether a reduced dose of

mitomycin-C (MMC) in combination with bevacizumab could improve surgical

outcome with a reduced incidence of side-effects.

Methods: Plasma levels of bevacizumab were determined via ELISA after

intracameral (IC), subconjunctival (SC) and intravitreal (IV) injections in mice,

subjected to a mouse model of GFS. Application of MMC was compared to

bevacizumab (SC, 25 lg) and to the combined use of both adjuvants. Surgical

sponges soaked in MMC 0.02% or 0.01% were exposed to the sclera for 1 or

2 min. Treatment outcome was studied by bleb investigation.

Results: The three administration routes of bevacizumab equally improved

surgical outcome. The VEGF antibody was detected at relatively high levels in

plasma shortly after IV injection, whereas it was minimally absorbed after IC

and SC injections. Both bevacizumab (SC) and MMC 0.02% (2 min) similarly

increased bleb area. As compared to MMC, the combined injection with

bevacizumab induced an additional effect on surgical outcome. Exposure of

MMC 0.02% for 1 or 2 min together with bevacizumab equally improved

surgical outcome, but 2 min application induced corneal toxicity. The combined

use of bevacizumab and 1-min MMC 0.01% also improved surgical outcome

compared to monotherapy, although to a lesser extent than the combination with

1-min MMC 0.02%.

Conclusions: Adjunctive bevacizumab not only enhances the beneficial effect of

MMC on surgical outcome, but also allows reducing the administration time of

MMC 0.02%, thereby eliminating its toxic effects on the cornea.

Key words: bevacizumab – glaucoma filtration surgery – mitomycin-C (MMC) – vascular endo-

thelial growth factor (VEGF) – wound healing

Acta Ophthalmol. 2015: 93: 667–678ª 2015 Acta Ophthalmologica Scandinavica Foundation. Published by John Wiley & Sons Ltd

doi: 10.1111/aos.12766

Introduction

Different antimitotics, such as 5-fluo-rouracil (5-FU) and mitomycin-C(MMC), still serve as the gold-standard

treatment to reduce scar formationafter glaucoma filtration surgery(GFS). These agents indeed signifi-cantly improve the surgical success rate

by enhancing bleb survival. However,due to their non-specific action andmechanism, these antimitotics carry arisk of vision-threatening complica-tions (Lama & Fechtner 2003). Withlower levels of exposure, bleb-relatedcomplications are less frequent, butfiltration failure is more common. Thisvariability in outcome effectively pre-sents a trade-off in which safer proce-dures are less effective and moreeffective procedures are less safe(Mathew & Barton 2011). Newapproaches may allow modulation ofwound healing resulting in more pre-dictable surgical outcomes. Findingdrugs with pleiotropic working mecha-nisms would be a major benefit,because this might offer improvedtherapeutic opportunities. We indeedshowed that inhibition of vascularendothelial growth factor (VEGF) –particularly the non-selective inhibitors– could improve the surgical outcomeby reducing angiogenesis and fibrosisin a rabbit model of glaucoma surgery(Li et al. 2009; Van Bergen et al. 2011)and in a prospective randomized clin-ical trial (Vandewalle et al. 2014).These insights may open new perspec-tives for glaucoma surgery, althoughsome questions remain unanswered.However, a few small comparativestudies have been performed (Nilforu-shan et al. 2012; Sengupta et al. 2012;Anand & Bong 2015), it remainsunclear whether antifibrotics, such asbevacizumab, can replace the use ofantimitotics in clinical practice orshould rather be used as an adjunctive

667

Acta Ophthalmologica 2015

to antimitotics. It has been suggestedthat administration of MMC can leadto an upregulation of factors that arepro-inflammatory, pro-angiogenic and/or pro-fibrotic. Interestingly, Wong andcoworkers (oral communication ofunpublished results) demonstrated thatMMC application results in an upregu-lation of VEGF in aqueous humour(AH) of patients after glaucomasurgery. Therefore, associating MMCwith drugs that block these upregulatedpro-fibrotic growth factors may offerpromising complementary efficacy andmay enable to lower the dose of MMCand thus its related side-effects.

This study aimed at further explor-ing the hypothesis that a combinedapplication of antimitotic and antiscar-ring agents may have complementaryeffects. Therefore, we first comparedthe effect of MMC versus the mono-clonal anti-VEGF antibody, bev-acizumab, in a mouse model of GFS.Secondly, we investigated the effect ofcombined administration of MMC andbevacizumab on the surgical outcome.Importantly, we also checked whetherthe dose and/or administration time ofMMC could be lowered when used incombination with VEGF inhibition, toreduce its side-effects while maintainingits efficacy.

Materials and Methods

All experimental animal procedureswere performed in accordance withthe standards in the Association forResearch in Vision and OphthalmologyStatement for the Use of Animals inOphthalmic and Vision Research andthe EC Directive 86/609/EEC for ani-mal experiments. All experiments werealso approved by the InstitutionalAnimal Care and Research AdvisoryCommittee of KU Leuven.

Mouse model of glaucoma filtration

surgery

Mouse glaucoma surgery model

In this study, C57BL/6J mice (8–10 weeks old; Charles River Laborato-ries, Lyon, France) were used. Generalanaesthesia was induced via 10 times-diluted (60 mg/kg final dose) sodiumpentobarbital (Nembutal, 60 mg/ml;CEVA, Sante Animale, Brussels, Bel-gium) injected intraperitoneally. Asdescribed in the literature, filteringsurgery was performed on both eyes

resulting in a filtration bleb (Seet et al.2010, 2011; Van Bergen et al. 2013a).Shortly, the conjunctiva was first sur-gically dissected to expose the underly-ing sclera, and a small filtrationsubconjunctival space was created byrunning the surgical scissors under-neath the dissected conjunctiva. Next,a 30-gauge needle was used to make anincision through the sclera into theanterior chamber of the eye to allowAH to escape into the subconjunctivalspace. Finally, the conjunctiva wasclosed at the limbus by suturing overthe newly created fistula (using a 10-0nylon suture; Alcon). A topical combi-nation preparation containing steroidsand antibiotics (Tobradex; SA Alcon-Couvreur, Puurs, Belgium) was admin-istered at the end of surgery.

Clinical investigation of blebs

Every 2 days, bleb investigation wasperformed by taking bleb pictures fromanaesthetized mice. Herefore, a digitalCanon Eos 5D Mark II camera with a100 mm lens was used at a magnifica-tion of 49. Using commercial software(KS300; Zeiss, Brussels, Belgium), blebarea and survival were determined onthese images. The appearance of ascarred and flat bleb at two consecutivemeasurements was defined as bleb fail-ure (Van Bergen et al. 2013a). Alsocorneal opacity was graded on thesepictures at day 1, 3, 5, 7, 13, 21 and 29after surgery, according to a previouslypublished 0–4 scoring (Yoeruek et al.2008; Lee et al. 2009), ranging from 0as a clear/transparent cornea to 4 ascomplete corneal opacity.

Mouse treatment regimens

In all experiments, an analytic sciencesyringe (SGE Analytic Science, Austin,TX, USA) and glass capillaries with adiameter of 50–70 lm at the end wereused to perform the injections of bev-acizumab and NaCl (1 ll). All theinjections were controlled by theUMP3I Microsyringe Injector andMicro4 Controller (all from WorldPrecision Instruments Inc., Hertford-shire, UK). Of note, the use of NaCl ascontrol for bevacizumab in variouscancer and even eye models has alreadybeen reported. Furthermore, there wereno differences between animals treatedwith NaCl or an irrelevant monoclonalIgG antibody, which suggest that bothcan be used as a negative control (Lee

et al. 2009) (Folkesson et al. 1995;Santoro et al. 2009; Huang et al.2011). The surgical sponges soakedwith either MMC or NaCl were placedon the exposed sclera for 1–2 min, andafter removing the sponge, the ocularsurface and subconjunctival space wereextensively rinsed with 2 ml of NaCl.

Optimal route of administration of bev-

acizumab

In the first experiment (n = 15 mice pergroup), mice were divided into threedifferent groups to investigate the mostoptimal administration route of bev-acizumab. Immediately after surgery,the VEGF antibody (1 ll, 25 lg;Genentech, Inc., San Francisco, CA,USA) was injected intracamerally (IC)in one eye of the first group, subcon-junctivally (SC) in the second group andthe third group received an intravitreal(IV) injection of bevacizumab. In everygroup, NaCl 0.9% (B. Braun Medical,Bethlehem, PA,USA) was administeredin the contralateral eye, which served asnegative control. In a second experi-ment, a group of three mice per condi-tion was used to investigate systemicabsorption of bevacizumab. One eye ofeach mouse was operated and treatedwith bevacizumab (IC, SC or IV injec-tions) and the other, non-operated eyewas used to collect blood at 30 min, 1 h,4 h, 8 h, 1 day, 3 days, 7 days and14 days after injection. Blood was col-lected in heparin (5000 IU/ml; LeoPharma, Ballerup, Denmark) filledtubes (1/50 dilution in NaCl), centri-fuged for 5 min at 7000 revolutions permin (rpm) and stored at �20°C untilanalysis. An enzyme immunoassay forthe quantitative determination of bev-acizumab (ABIN1540255; Matriks Bio-tek Laboratories, Ankara, Turkey) witha detection limit of 30 ng/ml was usedto investigate bevacizumab levels in theplasma samples.

Effect of bevacizumab on surgical outcome

in comparison with MMC

In the third experiment (n = 20 miceper group), the effect of MMC (KyowaHakko Kirin Co., Ltd., Princeton, NJ,USA) on surgical outcome was com-pared to the effect of bevacizumab. Oneeye was treated with a SC injection ofthe VEGF inhibitor (1 ll; 25 lg), com-bined with a surgical sponge soaked inNaCl. A surgical sponge soaked withMMC 0.02% (2 min) and a SC injec-tion of NaCl was administered to the

668

Acta Ophthalmologica 2015

other eye. A control group received acombination of an injection and surgi-cal sponge of NaCl in both eyes.

Effect of bevacizumab on surgical outcome

in combination with MMC

To investigate the complementaryeffect of bevacizumab (25 lg) and theantimitotic agent (2-min administra-tion of MMC 0.02%), mice weredivided into three different groups inthe fourth experiment (n = 10 mice pergroup). To perform the experimentwith a masked observer, all eyes weretreated with a combined sponge andinjection application. Moreover, allseparate conditions were compared toeach other. To check whether theconcentration or administration timeof the antimitotic could be lowered, afifth experiment was set-up (n = 10mice per group). Similar groups as inthe fourth experiment were used, but a

surgical sponge soaked in MMC 0.02%and MMC 0.01% was exposed to thesclera for 1 and 2 min, respectively.

An overview of the different treat-ment groups is provided in Table 1.

(Immuno)histological and microscopical

analyses

Tissue preparation

Mice were killed by cervical dislocationon postoperative days 8, 14, 30 and 34after surgery (n = 5–10 per time-point).After enucleating both eyes, whole eyeswere fixed in 1% paraformaldehyde(Merck,Darmstadt,Germany) overnightand rinsed three times for 5 min inphosphate-buffered saline (PBS). Serialparaffin sections were cut (7 lm) in fiveseries of five glass slides. To localize thebleb (area of analysis), haematoxylin andeosin (H&E) staining was performed onthe first slide from each series.

Staining for inflammatory and endothelial

cells and for collagen deposition

Inflammation in the bleb was evaluatedusing an immunostaining for theinflammatory cell marker CD45 andblood vessel density was analysed usinga staining for CD31, as previouslydescribed by our laboratory (Van Ber-gen et al. 2013a). Rat anti-mouseCD45 (1/100; 553076; Pharmingen;Erembodegem, Belgium) or rat anti-mouse CD31 antibody (1/500; 557355;Pharmingen), respectively, was used toincubate the eye sections overnight.The next day, the slides were incubatedwith rabbit anti-rat biotin labelled-antibody (1/300; Dako, Heverlee, Bel-gium). Antibody binding was visual-ized by fluorescent staining using theTSATM Cyanin 3 system (RenaissanceTSATM Indirect; NEL704A; Waltham,MA, USA), and slides were mountedwith Prolong Gold with 40,6-diamidi-no-2-phenylindole (DAPI; MolecularProbes, Eugene, OR, USA). SiriusRed staining was performed to analysedeposition of collagen in the bleb.Slides were deparaffinised, washed andplaced in Sirius Red solution (DirectRed 80 and 1.3% picric acid solution;both Sigma Aldrich, Diegem, Belgium)for 60 min. Sections were then placedin 0.01N HCl (Prolabo, Leuven, Bel-gium) for 2 min, dehydrated andmounted with DPX mounting medium(Prosan, Merelbeke, Belgium).

Microscopic analysis

As described above, the bleb waslocalized on the first serial slide, basedon the H&E staining. The consecutiveslides were used to perform the differ-ent (immuno)histological stainings,and analysis of the different woundhealing processes was only performedin the bleb. For every staining, oneslide per eye (five sections per slide) wasanalysed and different eyes were aver-aged. A Leica Microsystems micro-scope (Wetzlar, Germany), equippedwith a digital camera (Axiocam MrC5;Carl Zeiss, Oberkochen, Germany),was used to obtain images of thedifferent stainings, based on previousreports (Li et al. 2009; Van Bergenet al. 2011, 2013a,b). Shortly, pictureswere taken at a magnification of 209, aresolution of 2584 9 1936 pixels, andmorphometric analyses were per-formed using commercial software(KS300; Zeiss). The density of leuco-cytes and blood vessels was determined

Table 1. Overview of different treatment groups.

Mouse experiments

1. Optimal route of administration (D14) One eye Contralateral eye

Efficacy Injection Injection

First group (n = 15) IC bevacizumab IC NaCl

Second group (n = 15) SC bevacizumab SC NaCl

Third group (n = 15) IV bevacizumab IV NaCl

Safety Injection No treatment

First group (n = 3) IC bevacizumab Blood collection

Second group (n = 3) SC bevacizumab Blood collection

Third group (n = 3) IV bevacizumab Blood collection

One eye Contralateral eye

Sponge SC injection Sponge SC injection

2. Bevacizumab versus MMC 0.02% – 2 min (D30)

First group (n = 20) NaCl Bevacizumab MMC NaCl

Second group (n = 20) NaCl NaCl NaCl NaCl

3. Bevacizumab with MMC 0.02% – 2 min (D30)

First group (n = 20) MMC NaCl NaCl NaCl

Second group (n = 20) MMC bevacizumab MMC NaCl

Third group (n = 20) NaCl NaCl MMC Bevacizumab

4. Lower concentration and administration time of MMC (D34)

First group (n = 10 for MMC 0.02% –0.01% for 1 and 2 min)

MMC NaCl NaCl NaCl

Second group (n = 10 for MMC 0.02% –0.01% for 1 and 2 min)

MMC bevacizumab MMC NaCl

Third group (n = 10 for MMC 0.02% –0.01% for 1 and 2 min)

NaCl NaCl MMC Bevacizumab

Rabbit experiment

1. VEGF expression after MMC

One eye Contralateral eye

Sponge Sponge

First group (n = 10) MMC 0.02% – 2 min NaCl

IC, intracameral; SC, subconjunctival; IV, intravitreal.

669

Acta Ophthalmologica 2015

by calculating, respectively, the CD45-positive and the CD31-positive area asa proportion of the bleb area. Thepercentage of the area of Sirius Red-stained mature collagen fibres to thetotal wound area was measured todetermine collagen deposition in thebleb. Polarized light was used to dis-tinguish mature (appearing bright yel-low and orange) from immature(green) collagen fibres.

Determination of VEGF levels in a rabbit

model of glaucoma filtration surgery

Rabbitmodel of glaucomafiltration surgery

New Zealand rabbits (n = 10; 12–14 weeks; animal facility of KU Leu-ven, Belgium) were anaesthetized viaintramuscular injection of 50 mg/mlketamine (Ketalar; Pfizer, Ann Arbor,MI, USA) and 2% sedative (Rompun;Bayer Health Care, Pittsburgh, PA,USA), as previously described (Li et al.2009; Van Bergen et al. 2011, 2013b).Surgery was performed on both eyes,using a technique similar to trabecu-lectomy in humans (Vandewalle et al.2014). In brief, a partial thickness, 9/0corneal traction suture (Alcon) wasplaced superiorly and the eye waspulled down. A fornix-based conjunc-tival flap was performed in the upperpart of the eye, and the sclera wasexposed. One eye was treated withMMC, whereas the other eye servedas a control and was treated with NaCl.A surgical sponge soaked with eitherMMC 0.02% or NaCl 0.9% was placedon the exposed sclera for 2 min, andafter removing the sponge, the oculartissue was abundantly washed withNaCl (Table 1). A limbus-based half-scleral-thickness flap was dissecteduntil clear cornea was reached. AKhaw titanium punch was used toremove a standard size part of thetrabecular meshwork (TM), to allowdrainage of the AH, resulting in afiltering bleb. A peripheral iridectomywas performed, and both flaps wereclosed with two 10-0 nylon sutures(Alcon). Tobradex ointment (SA Al-con-Couvreur) was applied at the endof the procedure.

Detection of VEGF protein levels in aque-

ous humour samples

Samples (200 ll) of AH were collectedusing a BD microfine insulin syringe(30G; Becton Dickinson, Erembode-gem, Belgium) from both eyes on the

day before and at different time-points(days 1, 3 and 7) after GFS. Allsamples were stored immediately at�20°C until analysis and VEGF pro-tein levels in the AH samples weredetermined using a double-antibody‘sandwich’ ELISA (R&D Systems,Minneapolis, MN, USA), with a detec-tion limit of 15.6 pg/ml.

Statistical evaluation

The Student’s t-test was used to analyseall morphometric data for independentsamples. Data at individual time-pointswere studied usingmixedmodel analysisfor repeated measures. Kaplan–Meiersurvival analysis using the log-rank testwas performed for bleb failure (all usingGraphPad Prism 5.3, La Jolla, CA,USA). p < 0.05 was considered statisti-cally significant. Data are represented asmean � standard error of the mean(SEM), unless otherwise stated.

Results

Optimal route of administration of

bevacizumab

The most appropriate administrationroute of the VEGF antibody afterglaucoma surgery is still uncertain(Mathew & Barton 2011). Therefore,surgical outcome after a single IC, SCand IV injections of the VEGF inhib-itor was compared in a mouse modelof GFS. The three routes of adminis-tration of the VEGF antibody wereable to significantly improve bleb area(n = 15; p < 0.001; mixed model analy-sis for repeated measures; Fig. 1A)and bleb survival (n = 15; p < 0.05;log-rank test; Fig. S1A) compared totheir respective controls (NaCl) until14 days after surgery. A direct com-parison between the three groupsshowed no significant difference inbleb area (n = 15; p = 0.95) and sur-vival (n = 15; p = 1.00), suggestingthat administration of the VEGFinhibitor via the various routes equallyimproves surgical outcome. Moreover,IC, SC or IV bevacizumab injectionssimilarly reduced the process of angio-genesis at day 14 (n = 10; p > 0.05between different conditions; Student’st-test; Fig. S1B). Importantly, bev-acizumab was detected at relativelyhigh levels in plasma shortly after IVinjection (6.35 lg/ml after 30 min),whereas a minimal bevacizumab

absorption was detected only fromday 4 after SC (0.69 lg/ml) or IC(2.46 lg/ml) administration. In allgroups, the plasma levels of bev-acizumab remained detectable and sta-ble until day 14 (4.07; 0.71 and0.58 lg/ml, respectively) (Fig. 1B).

Thus, SC and IC injections of bev-acizumab offer an effective way toimprove surgical outcome after GFS,with less systemic absorption comparedto IV injection. Based on these resultsand on practical considerations relatedto injection in mice, SC injection wasselected as the administration route forthe following experiments.

Equal improvement of surgical outcome by

bevacizumab and MMC

To elucidate whether bevacizumab is asefficient as MMC to reduce scar forma-tion after glaucoma surgery, the effectof the VEGF inhibitor on the surgicaloutcome was compared to that of theantimitotic agent in a mouse trabecu-lectomy model. Bevacizumab (25 lg)or MMC 0.02% (2-min sponge) appli-cation was found to be equally effectivein improving surgical outcome until30 days after surgery. Indeed, theVEGF antibody was able to signifi-cantly improve bleb area over 30 dayswith 73 � 6% compared to controleyes (n = 20; p < 0.001), whereasMMC induced a 75 � 5% improve-ment (n = 20; p = 0.003). There was nosignificant difference between bothgroups (n = 20; p = 0.12; Fig. 2A).Also bleb survival was prolonged afterboth treatments, as shown in the Kap-lan–Meier survival curve, with 43% ofthe blebs surviving until postoperativeday 30 after bevacizumab or MMCadministration (n = 20; p = 0.87),whereas all blebs had failed by thattime-point in the control group (n = 20;p < 0.001; Fig. 2A). Moreover, eyestreated with bevacizumab and MMCshowed an equal reduction in bloodvessel area on day 14 (with 5 � 0.08%;p = 0.06) and fibrosis on day 30 (with11 � 0.20%; p = 0.10), as compared tothe control eyes (n = 5–10; p < 0.001;Fig. 2B). No differences in inflamma-tory responses were seen in any of thethree groups (data not shown). Impor-tantly, grading corneal opacity indi-cated that 2-min administration ofMMC 0.02% significantly increasedcorneal toxicity as compared to NaCl(p = 0.04), whereas this was not the

670

Acta Ophthalmologica 2015

case after treatment with bevacizumab(p = 0.20; Table 2; Fig. 2C).

Upregulation of VEGF in aqueous humour

after MMC administration

As it has been suggested that treatmentof MMC can lead to an upregulation offactors that are pro-inflammatory, pro-angiogenic and/or profibrotic, we inves-tigated VEGF levels in AH after MMCadministration. Unfortunately, themouse trabeculectomy model could notbe used, because the volume of murine

AH that can be collected (�4 ll per eye)is too small to perform an ELISA.Therefore, the rabbit model of GFSwas used, because 200 ll of AH per eyecan be collected in this animal. Surgerywas performed on both eyes and one eyewas treated with MMC 0.02% (2 min),whereas the other eye received NaCl.Samples (200 ll) of AH were collectedand ELISA showed that after surgery,aqueous VEGF levels were almost dou-bled in MMC-treated eyes on day 1, ascompared to the control eye treatedwithNaCl (n = 10; p = 0.01; Fig. S2). Aque-

ous VEGF levels at day 3 and 7 werecomparable to baseline.

Thus, bevacizumab was equallyeffective in improving surgical out-come, but with a better safety profilecompared to MMC 0.02%. Moreover,VEGF was upregulated in the AH ofMMC-treated eyes.

Complementary effects of combination of

MMC and bevacizumab

As it has been suggested that associatingMMCwith drugs that block upregulated

(A) (B)

Fig. 1. The most optimal administration route of bevacizumab. (A) Bleb area (n = 15; p < 0.05) was significantly improved after a single intracameral

(IC), subconjunctival (SC) and intravitreal (IV) injections of the VEGF inhibitor, compared to their respective controls. (B) Determination of plasma

levels of bevacizumab after IC, SC or IV injections revealed high values in the plasma from 30 min until 14 days after IV injection, whereas a minimal

absorption of the VEGF inhibitor was detected from day 4 until day 14 after IC and SC injections.

671

Acta Ophthalmologica 2015

profibrotic growth factors may offerpromising complementary efficacy, thecombinationof bevacizumabandMMC0.02% (2 min) was investigated in themouse model of GFS. Clinical investi-gation of the bleb showed that the bleb

area was increased after the combina-tion therapy with 82 � 4%, comparedto the eyes treated with the saline solu-tion (n = 20; p < 0.001; Fig. 3A), until30 days after surgery. All blebs treatedwith the combinationwith bevacizumab

survived until 30 days after surgery,whereas all blebs failed in the controlgroup, as shown in the Kaplan–Meiersurvival curve (n = 20; p < 0.001;Fig. 3A). This improvement in bleb areawas associatedwith a reduction in blood

Table 2. Corneal opacity: MMC 0.02% – 2 min (day 30).

Days after surgery 1 3 5 7 13 21 29

NaCl 0.05 � 0.04 0.03 � 0.03 0.03 � 0.03 0.03 � 0.03 0.00 � 0.00 0.00 � 0.00 0.00 � 0.00

Bevacizumab*,† 0.20 � 0.06 0.14 � 0.05 0.18 � 0.06 0.02 � 0.02 0.00 � 0.00 0.00 � 0.00 0.00 � 0.00

MMC‡,§ 0.66 � 0.10 0.67 � 0.10 0.94 � 0.12 0.63 � 0.10 0.57 � 0.15 0.74 � 0.23 0.47 � 0.24

MMC + Bevacizumab¶ 0.84 � 0.11 0.87 � 0.12 0.50 � 0.14 0.26 � 0.10 0.40 � 0.11 0.45 � 0.16 0.60 � 0.15

Scoring � SEM; n = 20 mice/compound.

* Overall p = 0.20 (bevacizumab versus NaCl).† Overall p = 0.003 (bevacizumab versus combination).‡ Overall p = 0.04 (MMC versus NaCl).§ Overall p = 0.96 (MMC versus combination).– Overall p = 0.0006 (combination versus NaCl).

(A)

(B)

(C)

Fig. 2. Bevacizumab and mitomycin-C (MMC) administrations were equally effective in improving surgical outcome in the mouse trabeculectomy

model. (A) The VEGF inhibitor (bev; SC; 25 lg) was able to improve bleb area (n = 20; p = 0.12) and bleb survival (n = 20; p = 1.00) in a similar

way, as compared to application of MMC 0.02% for 2 min. (B) Blood vessel density at day 14 (D14) (n = 5; p = 0.06) and fibrosis at day 30 (D30)

(n = 10; p = 0.10) were similarly reduced after both treatments. (C) Representative macroscopic postoperative photographs of the blebs, showing

corneal toxicity associated with administration of the antimitotic agent.

672

Acta Ophthalmologica 2015

vessel density on day 14 with 7 � 1%and fibrosis on day 30 with 17 � 0.26%(n = 5–10; p < 0.001 compared toNaCl; Fig. 3B). Importantly, a directcomparison showed that the combina-tion therapy was able to induce a signif-icant additional effect in theimprovement of surgical outcome com-pared to MMC administration. Indeed,as compared to monotherapy of MMC,bleb area was significantly increasedwith 5 � 1% (n = 20; p < 0.001;Fig. 3A), with a further decrease inblood vessel density (with 3 � 0.42%on day 14; p = 0.002) and collagendeposition (with 5 � 0.21% on day 30;p = 0.001; Fig. 3B). Notably, cornealtoxicity was still present in the eyesthat received combined treatment, ascompared to saline-treated eyes

(p = 0.0006). No differences in toxicitywere seen between the combined treat-ment and MMC 0.02% monotherapy(p = 0.96; Table 2; Fig. 3C).

Thus, administration of bev-acizumab together with MMC 0.02%may have complementary effects, prob-ably or at least partly because of theobserved upregulation of VEGF afterMMC treatment. However, the use ofthe combination therapy was still asso-ciated with corneal toxicity.

Lowering dose and/or administration time

of MMC in combination with bevacizumab

As the combination therapy of bev-acizumab and 2-min administration ofMMC 0.02% was still associated withcorneal toxicity, we investigated

whether lowering the dose (0.01%)and/or administration time of MMC(1 min) could diminish these side-effects. Clinical investigation was per-formed until all blebs in the MMC-treated eyes were failed. We showedthat 1- or 2-min application of MMC0.02% was able to equally improvesurgical outcome in the mouse trabe-culectomy model until 34 days aftersurgery (n = 10; p = 0.65). Indeed, blebarea and bleb survival improved in asimilar way, compared to their respec-tive controls (n = 10; p < 0.001;Figs 4A-B and S3A-B). This was asso-ciated with an equal reduction in col-lagen deposition on postoperative day34, compared to the control eyes(n = 10; p < 0.001; Fig. 4A-B), withno significant differences between the

(A)

(B)

(C)

Fig. 3. Administration of MMC combined with bevacizumab showed complementary effects in the improvement of the surgical outcome in the

mouse trabeculectomy model. (A) As compared to monotherapy of MMC, combination therapy (combi) of the antimitotic agent (MMC 0.02% –2 min) together with the VEGF inhibitor (bev; SC; 25 lg) was able to induce an additional increase in bleb area (n = 20; p < 0.001). (B) This use of

the combined therapy was also associated with a further decrease in blood vessel density on day 14 (D14) (n = 5; p = 0.002) and collagen deposition

on day 30 (D30) (n = 10; p = 0.001), versus MMC treatment. (C) Representative macroscopic postoperative photographs of the blebs showing that

corneal toxicity was also present with the use of the combined therapy of bevacizumab and MMC 0.02% for 2 min.

673

Acta Ophthalmologica 2015

two conditions (n = 10; p = 0.90).Importantly, similar effects were seenwhen these conditions of MMC werecombined with bevacizumab. Adminis-tration of bevacizumab and MMC0.02% for 1 or 2 min equally improvedbleb area and survival, compared totheir respective controls (n = 10;p < 0.001; Figs 4A-B and S3A-B), withno significant differences between them(n = 10; p = 0.51). Analysis of the Sir-ius Red staining on day 34 showed anequal decrease in fibrosis between thetwo conditions, compared to the neg-ative controls (n = 10; p < 0.001;Fig. 4A-B), but they were not signifi-cantly different (n = 10; p = 0.84).Importantly, eyes treated with MMC0.02% for 2 min, with and withoutbevacizumab, clearly induced cornealtoxicity (p = 0.03 and p = 0.02, respec-tively, compared to NaCl; Table 3).This was not noticed after 1-minadministration (p < 0.05; Table 4),suggesting that this condition is safercompared to the longer exposure timeof 2 min (Fig. 4D).

As it has been described in theliterature that the dose of MMC mightbe more important than the adminis-tration time (Megevand et al. 1995;Schnyder et al. 1995; Thornton et al.2008), application of MMC 0.01% for1 min was chosen to be compared toMMC 0.02% (1 or 2 min), rather than2-min exposure of MMC 0.01%.Administration of MMC 0.01% for1 min did not induce corneal toxicity(Table S1) and was able to significantlyimprove bleb area and bleb survival,compared to NaCl (n = 10; p < 0.001;Figs 4C and S3C). However, Kaplan–Meier survival curves showed that allblebs after MMC therapy failed at day30 (Figs 4C and S3C). Thus, the clin-ical results showed that this lower dosewas less effective as compared to the 1-and 2-min application of MMC 0.02%(n = 10; p < 0.05 for all conditions).Importantly, the combined therapy ofbevacizumab and 1-min exposure ofMMC 0.01% significantly improvedsurgical outcome versus monotherapy(p < 0.001 versus 1-min MMC 0.01%),although to a lesser extent than thecombination with MMC 0.02% for1 min (n = 10; p < 0.001). Indeed,60% of the blebs treated with thecombination therapy of bevacizumabwith 1-min 0.01% MMC failed at day34 (Figs 4C and S3C), whereas allblebs survived after the combined

treatment with the higher concentra-tion of the antimitotic agent. More-over, the 1-min administration ofMMC 0.01% together with bev-acizumab was not able to induce anadditional decrease in fibrosis com-pared to monotherapy of MMC(n = 6; p = 0.35; Fig. 4C), suggestingthat halving the concentration andadministration time of MMC with theVEGF inhibitor is not as effective asthe higher dose/administration time.

Thus, halving the administrationtime of MMC 0.02% to 1 min incombination with bevacizumab seemedto be equally effective in improvingsurgical outcome compared to thecombination therapy with 2-min0.02% MMC, without inducing cor-neal toxicity. The combination ofMMC with the VEGF inhibitor offersthe possibility to reduce the adminis-tration time of MMC 0.02%, therebyreducing corneal toxicity while main-taining the beneficial effects on surgicaloutcome. A further reduction in dosedid not offer superior antifibroticeffects over monotherapy, as comparedto the combination therapy with MMC0.02%.

Discussion

Since the early 1980s, antimitotics suchas 5-FU and MMC have been increas-ingly used as an adjunctive to trabecu-lectomy and have become the goldstandard to reduce scar formation.However, the non-specific mechanismof action of these agents can result insevere side-effects (Lama & Fechtner2003). Therefore, there is still a needfor alternative strategies to preventfiltration failure. We already showedthat pharmacological enhancement oftrabeculectomy using VEGF inhibitorswas able to significantly improve ratesof surgical success. A single injection ofbevacizumab at the time of trabeculec-tomy improved the surgical outcomeby reducing postoperative angiogenesisduring the initial phase and collagendeposition at later stages in a rabbitmodel of trabeculectomy (Li et al.2009). These results were recently con-firmed in a prospective, randomized,double-masked, placebo-controlledtrial performed by our group in theUniversity Hospitals of Leuven. Weshowed that a single intracameralinjection of bevacizumab at the end oftrabeculectomy significantly reduces

the need for additional interventionsduring the first year of follow-up aftertrabeculectomy (Vandewalle et al.2014). Although the results from thisclinical trial are highly relevant andvery promising, some questions remainunanswered. It is unclear whether theintracameral administration route isthe optimal one. Furthermore, thestudy was not designed to answer thequestion whether bevacizumab couldreplace MMC in clinical practice orshould rather be used as an adjunctiveto MMC.

As the most appropriate route ofadministration of bevacizumab afterGFS is still unclear, different adminis-tration routes were first compared. Ourdata showed that the three routes ofVEGF antibody injection were able toequally improve surgical outcome. Bev-acizumab was detected at relativelyhigh levels in plasma shortly after IVinjection, whereas the other routes ofadministration only induced a minimalbevacizumab absorption from day 4.The levels after SC and IC injectionscan be considered as negligible, becausethey are comparable to the minimaleffective inhibitory concentration ofbevacizumab. It is indeed describedthat bevacizumab loses its VEGF-inhibiting properties at a concentrationof 1 lg/ml (Klettner & Roider 2008).These data suggest that the SC and ICadministration routes induce less sys-temic absorption of the VEGF inhibi-tor as compared to IV injection.Uptake of bevacizumab in the bloodafter IV injection has already beendescribed in the literature (Heiduschkaet al. 2007; Miyake et al. 2010; Wuet al. 2010; Chuang et al. 2011; Satoet al. 2012) and our results are in linewith those of Kim et al. 2009; whoreported that plasma levels of theVEGF antibody after IV administra-tion can represent up to 30% of theinjected dose in rodents . Heiduschkaet al. described that the systemicabsorption of bevacizumab after IVinjection could be explained by pene-tration of the VEGF antibody from thevitreous into the inner retina, photore-ceptors and retinal pigment epithelium,presumably regulated by active trans-port of the M€uller cells. Upon arrivalin the choroidal blood vessels, theVEGF inhibitor is substantially trans-ferred into the systemic blood circula-tion. Moreover, they also showed thatbevacizumab was able to cross the

674

Acta Ophthalmologica 2015

(A)

(B)

(C)

(D)

Fig. 4. The effect of lowering the dose and/or exposure time of MMC on the surgical outcome. (A-B) One- and two-min exposure of MMC 0.02%

both improved bleb area, compared to their respective controls (n = 10; p < 0.001). Both conditions induced an equal reduction in collagen deposition

of 10% on postoperative day 34, compared to the control eyes (n = 10; p < 0.001). Administration of bevacizumab and MMC 0.02% for 1 or 2 min

also equally improved surgical outcome by reducing collagen deposition with 15%, versus their respective controls (n = 10; p < 0.001). (C)

Administration of MMC 0.01% for 1 min, combined with bevacizumab, induced an additional effect in bleb area as compared to MMCmonotherapy

(n = 10; p = 0.002). However, these results were not as efficient as the combined therapy (combi) with 0.02% MMC. Using 1-min 0.01% MMC and

bevacizumab was not able to induce an additional decrease in fibrosis compared to monotherapy of MMC (n = 6; p = 0.35). (D) Representative

macroscopic postoperative photographs of the blebs showing that none of the eyes treated with MMC for 1 min in combination with bevacizumab

showed corneal toxicity, as compared to the longer exposure time of 2 min.

675

Acta Ophthalmologica 2015

blood–retinal barrier immediately afterthe injection with subsequent deliveryinto the blood circulation (Heiduschkaet al. 2007). We also believe that thesystemic uptake can be enhanced byreflux of bevacizumab into the subcho-roidal space due to intravitreal injec-tion, which will increase absorption ofthe VEGF antibody into the choroidalblood vessels and blood circulation.Based on our results and on the feasi-bility of intra-ocular injection in mice,SC injection was selected over IC andIV administration route in this study.

To answer the open questionwhether anti-VEGF should be seen asa replacement for or as a complementto MMC, we first compared the effectof MMC versus the monoclonal anti-VEGF antibody, bevacizumab, on sur-gical outcome in a mouse model ofGFS (Seet et al. 2011). The effect onsurgical outcome of a SC injection ofbevacizumab (25 lg) was compared toa surgical sponge soaked in MMC0.02%, exposed to the sclera for2 min (in analogy to current clinicalpractice). The administration of anti-VEGF antibody or MMC was equallyeffective in improving surgical outcomeand both similarly reduced angiogene-sis and fibrosis in the filtering bleb.Comparable to what has been reportedin other studies, we did not observe aneffect on inflammation. Indeed, it hasalready been shown that both bev-acizumab and MMC do not affect the

bleb infiltration of inflammatory cells(Minguini et al. 2000; Chou et al. 2007;Wilgus et al. 2008; Li et al. 2009). Wealso showed that bleb survival wassimilarly prolonged after both treat-ments with a 100% survival of blebs inthe two groups on day 30.

Importantly, from clinical practice,we know that the use of MMC can beassociated with various complicationsrelated to cell toxicity, amongst whichcorneal cells (Sihota et al. 1998). In thisstudy, corneal opacity was graded as ameasure of corneal toxicity, using apreviously described scoring system(Yoeruek et al. 2008; Lee et al. 2009).Toxicity was indeed noticed in micetreated with MMC, whereas this wasnot the case after bevacizumab admin-istration. So, bevacizumab was clearlyassociated with a better safety profilecompared to MMC. These results arein line with those reported by Senguptaet al., who showed that, in a random-ized controlled clinical trial with 38patients, a SC injection of bevacizumab(1.25 mg) was equally effective inreducing IOP compared to MMC0.03% administration. No toxic effectsof bevacizumab were seen in the cor-neal epithelium or endothelium and theanti-VEGF antibody had a relativelybetter safety profile, as compared toMMC at the 6-month follow-up. Thesafety and tolerability of the VEGFinhibitor was shown by the absence oflocal toxicity or of intra-ocular inflam-

mation and by the lack of any systemicadverse effects. Moreover, they alsoshowed that bevacizumab soaked in asponge was effective in improving sur-gical outcome; however, it appeared tohave no advantages over MMC of SCinjection of bevacizumab (Senguptaet al. 2012). Also the group of Nilforu-shan et al. showed that, in a prospec-tive randomized, comparative study of34 patients, no significant differenceswere detected between SC bevacizumab(2.5 mg) and MMC (0.02% for 3 min)in terms of bleb morphologic features.This study also indicated that the use ofSC bevacizumab in association withtrabeculectomy was safe and effective(Nilforushan et al. 2012).

Although these results demonstratethat bevacizumab andMMC are able toimprove surgical outcome, it is stillnecessary to broaden the therapeuticapproach, because treatment with asingle agent may lead to drug resistance.Indeed, because of the complexity ofthe wound healing process (Lama &Fechtner 2003; Georgoulas et al. 2008),it is possible that targeting one moleculemay be insufficient to reach the level thatis necessary for predictable glaucomasurgery due to upregulation of othergrowth factors. As it has been suggestedthat treatment of MMC can lead to anupregulation of factors that are pro-inflammatory, pro-angiogenic and/orprofibrotic, VEGF levels in AH of oper-ated rabbits after MMC administration

Table 3. Corneal opacity: MMC 0.02% – 2 min (day 34).

Days after surgery 1 3 5 7 13 21 29

NaCl 0.10 � 0.07 0.20 � 0.08 0.00 � 0.00 0.00 � 0.00 0.00 � 0.00 0.00 � 0.00 0.00 � 0.00

MMC*,‡ 0.65 � 0.17 0.90 � 0.19 0.90 � 0.19 0.85 � 0.15 0.65 � 0.13 0.25 � 0.08 0.40 � 0.10

MMC + Bevacizumab† 0.75 � 0.20 0.70 � 0.19 0.70 � 0.19 0.70 � 0.13 0.85 � 0.08 0.75 � 0.08 0.70 � 0.20

Scoring � SEM; n = 10 mice/compound.

* Overall p = 0.02 (MMC versus NaCl).† Overall p = 0.07 (MMC versus combination.‡ Overall p = 0.03 (combination versus NaCl).

Table 4. Corneal opacity: MMC 0.02% – 1 min (day 34).

Days after surgery 1 3 5 7 13 21 29

NaCl 0.10 � 0.07 0.10 � 0.07 0.00 � 0.00 0.00 � 0.00 0.00 � 0.00 0.00 � 0.00 0.00 � 0.00

MMC*,† 0.20 � 0.11 0.20 � 0.11 0.10 � 0.07 0.10 � 0.07 0.05 � 0.05 0.00 � 0.00 0.05 � 0.05

MMC + Bevacizumab‡ 0.15 � 0.08 0.10 � 0.07 0.10 � 0.07 0.05 � 0.05 0.05 � 0.05 0.05 � 0.05 0.05 � 0.05

Scoring � SEM; n = 10 mice/compound.

* Overall p = 0.96 (MMC versus NaCl).† Overall p = 0.88 (MMC versus combination).‡ Overall p = 0.93 (combination versus NaCl).

676

Acta Ophthalmologica 2015

were investigated. Of note, the rabbitmodel was used, because the volume ofmurine AH is too small to perform anELISA. We showed that MMCenhanced the postoperative VEGFupregulation on day 1 compared to thecontrol eyes. These results are in keepingwith previous publications on the upreg-ulation of VEGF after administration ofantimitotic agents in different cancermodels (Riedel et al. 2004; Verma et al.2011) and more specifically in glaucomapatients (unpublished data from thegroup of Tina Wong, oral communica-tion). These results suggest that associ-ating MMC with VEGF inhibitors, thatcan block the upregulated VEGF, mayoffer complementary efficacy. Indeed,recent evidence supports the combineduse of MMC and antifibrotic agents. Aprevious small study by Kahook com-bining MMC and anti-VEGF therapyshowed beneficial effects in improvingsurgical outcome after glaucoma surgery(Kahook 2010). Therefore, we furtherexplored the combination of MMC andbevacizumab on the surgical outcome ina mouse model of GFS described by thegroup of Tina Wong (Seet et al. 2011).The combination therapy of bev-acizumab and MMC was able to inducea significant additional improvement ofsurgical outcome compared to MMCmonotherapy, associated with a furtherreduction in blood vessel density and incollagen deposition. These results arecomparable to previously publisheddata, which also demonstrated that thecombination therapy of bevacizumaband 5-FU offers superior antifibroticeffect over monotherapy in a rabbitmodel of GFS (How et al. 2010) and ina non-randomized, retrospective, com-parative study with 61 patients (Freiberget al. 2013).

Importantly, the use of the combi-nation therapy of bevacizumab and thehighest dose of MMC (0.02% – 2 min)was associated with corneal toxicity,similar to MMC (0.02% – 2 min)monotherapy. Therefore, we investi-gated the effect on surgical outcomeafter reducing the dose and/or admin-istration time of MMC. Halving theexposure time of MMC 0.02% to1 min in combination with bev-acizumab seemed to be equally effectiveas the combination therapy with 2-minapplication of MMC 0.02%, withoutinducing any corneal toxicity. Thus,bevacizumab together with MMCmakes is possible to halve the exposure

time of MMC, without inducing cor-neal side-effects. Halving the dose ofMMC to 0.01%, however, reduced theefficacy of the adjuvant. The literatureindeed suggests that a short intra-operative application of MMC is aseffective as a long exposure (Megevandet al. 1995; Schnyder et al. 1995),whereas a low dose of MMC is lesseffective as a higher concentration ofMMC (Thornton et al. 2008). More-over, the combination therapy of 1 min0.01% MMC with the VEGF inhibitordid not induce an additional antifib-rotic effect compared to monotherapy,suggesting that this condition is not themost optimal to be used in combina-tion with bevacizumab.

In this study, treatment outcome wasmainly measured by morphologicalinvestigation of the filtering bleb. How-ever, it is known that increased fibrosismay result from accumulation of extra-cellular matrix (ECM) structural pro-teins, such as matrix metalloproteinases(MMP), fibronectin and laminin.Therefore, as recently described, adetailed evaluation of ECM degrada-tion by immunohistochemical stainingsmight indicate bleb failure and could behelpful to distinguish functioning fromnon-functioning blebs (Valimaki & Uu-sitalo 2014, 2015). One should alsoremain cautious with extrapolating offindings from animal research to thecare of human disease, as it is knownthat animal models never form precisereplicas of the human pathology. Thehuman connective tissue is, for exam-ple, less homogenous as compared toanimals and, very often, has beenexposed to antiglaucomatous dropsfor quite some time before surgery.The chronic use of topical antiglauco-ma agents in glaucoma patients isknown to result in long-term morpho-logical effects on the conjunctiva andTenon’s capsule (Sherwood et al. 1989;Broadway et al. 1993, 1994). Indeed,conjunctival inflammation associatedwith an increased number of inflamma-tory cells and fibroblasts has been notedafter long-term use of glaucoma drops,which may increase the risk of blebfailure (Broadway et al. 1994). As such,these factors might possibly influencethe results in human filtration surgeries.

Conclusions

In conclusion, this study indicates thatbevacizumab can enhance the benefi-

cial effects of MMC on scarring afterglaucoma surgery, and allows to reducethe administration time of MMC0.02%, thereby improving its safetyprofile while maintaining the beneficialeffects on surgical outcome.

References

Anand N & Bong C (2015): Deep sclerectomy

with bevacizumab and mitomycin C: a

comparative study. J Glaucoma 24: 25–31.Broadway D, Grierson I & Hitchings R (1993):

Adverse effects of topical antiglaucomatous

medications on the conjunctiva. Br J Oph-

thalmol 77: 590–596.Broadway DC, Grierson I, O’Brien C &

Hitchings RA (1994): Adverse effects of

topical antiglaucoma medication. I. The

conjunctival cell profile. Arch Ophthalmol

112: 1437–1445.Chou SF, Chang SW & Chuang JL (2007):

Mitomycin C upregulates IL-8 and MCP-1

chemokine expression via mitogen-activated

protein kinases in corneal fibroblasts. Invest

Ophthalmol Vis Sci 48: 2009–2016.Chuang LH, Wu WC, Yeung L, Wang NK,

Hwang YS, Chen KJ, Kuo JZ & Lai CC

(2011): Serum concentration of bevacizumab

after intravitreal injection in experimental

branch retinal vein occlusion. Ophthalmic

Res 45: 31–35.Folkesson HG, Matthay MA, Hebert CA &

Broaddus VC (1995): Acid aspiration-

induced lung injury in rabbits is mediated

by interleukin-8-dependent mechanisms. J

Clin Investig 96: 107–116.Freiberg FJ, Matlach J, Grehn F, Karl S &

Klink T (2013): Postoperative subconjuncti-

val bevacizumab injection as an adjunct to

5-fluorouracil in the management of scarring

after trabeculectomy. Clin Ophthalmol 7:

1211–1217.Georgoulas S, Dahlmann-Noor A, Brocchini S

& Khaw PT (2008): Modulation of wound

healing during and after glaucoma surgery.

Prog Brain Res 173: 237–254.Heiduschka P, Fietz H, Hofmeister S et al.

(2007): Penetration of bevacizumab through

the retina after intravitreal injection in the

monkey. Invest Ophthalmol Vis Sci 48:

2814–2823.How A, Chua JL, Charlton A, Su R, Lim M,

Kumar RS, Crowston JG & Wong TT

(2010): Combined treatment with bev-

acizumab and 5-fluorouracil attenuates the

postoperative scarring response after exper-

imental glaucoma filtration surgery. Invest

Ophthalmol Vis Sci 51: 928–932.Huang H, Shen J & Vinores SA (2011):

Blockade of VEGFR1 and 2 suppresses

pathological angiogenesis and vascular leak-

age in the eye. PLoS ONE 6: e21411.

Kahook MY (2010): Bleb morphology and

vascularity after trabeculectomy with intra-

vitreal ranibizumab: a pilot study. Am J

Ophthalmol 150: 399–403.e391.

677

Acta Ophthalmologica 2015

Kim H, Robinson SB & Csaky KG (2009):

FcRn receptor-mediated pharmacokinetics

of therapeutic IgG in the eye. Mol Vis 15:

2803–2812.Klettner A & Roider J (2008): Comparison of

bevacizumab, ranibizumab, and pegaptanib

in vitro: efficiency and possible additional

pathways. Invest Ophthalmol Vis Sci 49:

4523–4527.Lama PJ & Fechtner RD (2003): Antifibrotics

and wound healing in glaucoma surgery.

Surv Ophthalmol 48: 314–346.Lee SH, Leem HS, Jeong SM & Lee K (2009):

Bevacizumab accelerates corneal wound

healing by inhibiting TGF-beta2 expression

in alkali-burned mouse cornea. BMB Rep

42: 800–805.Li Z, Van Bergen T, Van de Veire S et al.

(2009): Inhibition of vascular endothelial

growth factor reduces scar formation after

glaucoma filtration surgery. Invest Ophthal-

mol Vis Sci 50: 5217–5225.Mathew R & Barton K (2011): Anti–vascularendothelial growth factor therapy in glau-

coma filtration surgery. Am J Ophthalmol

152: 10–15 e12.

Megevand GS, Salmon JF, Scholtz RP &

Murray AD (1995): The effect of reducing

the exposure time of mitomycin C in glau-

coma filtering surgery. Ophthalmology 102:

84–90.Minguini N, Monteiro de Carvalho KM,

Akaishi PM & De Luca IM (2000): Histo-

logic effect of mitomycin C on strabismus

surgery in the rabbit. Invest Ophthalmol Vis

Sci 41: 3399–3401.Miyake T, Sawada O, Kakinoki M, Sawada T,

Kawamura H, Ogasawara K & Ohji M

(2010): Pharmacokinetics of bevacizumab

and its effect on vascular endothelial growth

factor after intravitreal injection of bev-

acizumab in macaque eyes. Invest Ophthal-

mol Vis Sci 51: 1606–1608.Nilforushan N, Yadgari M, Kish SK & Nassiri

N (2012): Subconjunctival bevacizumab ver-

sus mitomycin C adjunctive to trabeculecto-

my. Am J Ophthalmol 153: 352–357 e351.

Riedel F, Gotte K, Goessler U, Sadick H &

Hormann K (2004): Targeting chemother-

apy-induced VEGF up-regulation by VEGF

antisense oligonucleotides in HNSCC cell

lines. Anticancer Res 24: 2179–2183.Santoro L, Boutaleb S, Garambois V et al.

(2009): Noninternalizing monoclonal anti-

bodies are suitable candidates for 125I

radioimmunotherapy of small-volume peri-

toneal carcinomatosis. J Nucl Med 50:

2033–2041.Sato T, Wada K, Arahori H, Kuno N,

Imoto K, Iwahashi-Shima C & Kusaka S

(2012): Serum concentrations of bev-

acizumab (avastin) and vascular endothe-

lial growth factor in infants with

retinopathy of prematurity. Am J Oph-

thalmol 153: 327–333 e321.

Schnyder CC, Bernasconi O, Mermoud A &

Faggioni R (1995): Comparative study of

administration time of mitomycin C in

trabeculectomy: 2.5 or 5 minutes?. Klin

Monbl Augenheilkd 206: 307–311.Seet LF, Lee WS, Su R, Finger SN, Crowston

JG & Wong TT (2011): Validation of the

glaucoma filtration surgical mouse model

for antifibrotic drug evaluation. Mol Med

17: 557–567.Seet LF, Su R, Barathi VA et al. (2010):

SPARC deficiency results in improved sur-

gical survival in a novel mouse model of

glaucoma filtration surgery. PLoS ONE 5:

e9415.

Sengupta S, Venkatesh R & Ravindran RD

(2012): Safety and efficacy of using off-label

bevacizumab versus mitomycin C to prevent

bleb failure in a single-site phacotrabeculec-

tomy by a randomized controlled clinical

trial. J Glaucoma 21: 450–459.Sherwood MB, Grierson I, Millar L & Hitch-

ings RA (1989): Long-term morphologic

effects of antiglaucoma drugs on the con-

junctiva and Tenon’s capsule in glaucoma-

tous patients. Ophthalmology 96: 327–335.Sihota R, Sharma T & Agarwal HC (1998):

Intraoperative mitomycin C and the corneal

endothelium. Acta Ophthalmol Scand 76:

80–82.Thornton I, Xu M & Krueger RR (2008):

Comparison of standard (0.02%) and low

dose (0.002%) mitomycin C in the preven-

tion of corneal haze following surface

ablation for myopia. J Refract Surg 24:

S68–S76.Valimaki J & Uusitalo H (2014): Immunohis-

tochemical analysis of extracellular matrix

bleb capsules of functioning and non-func-

tioning glaucoma drainage implants. Acta

Ophthalmol 92: 524–528.Valimaki J & Uusitalo H (2015): Matrix

metalloproteinases (MMP-1, MMP-2,

MMP-3 and MMP-9, and TIMP-1, TIMP-

2 and TIMP-3) and markers for vasculari-

zation in functioning and non-functioning

bleb capsules of glaucoma drainage

implants. Acta Ophthalmol 93: 450–456.Van Bergen T, Jonckx B, Hollanders K et al.

(2013a): Inhibition of placental growth fac-

tor improves surgical outcome of glaucoma

surgery. J Cell Mol Med 17: 1632–1643.Van Bergen T, Marshall D, Van de Veire S,

Vandewalle E, Moons L, Herman J, Smith

V & Stalmans I (2013b): The Role of LOX

and LOXL2 in Scar Formation After Glau-

coma Surgery. Invest Ophthalmol Vis Sci

54: 5788–5796.Van Bergen T, Vandewalle E, Van de Veire S,

Dewerchin M, Stassen JM, Moons L &

Stalmans I (2011): The role of different

VEGF isoforms in scar formation after

glaucoma filtration surgery. Exp Eye Res

93: 689–699.Vandewalle E, Abegao Pinto L, Van Bergen T

et al. (2014): Intracameral bevacizumab as

an adjunct to trabeculectomy: a 1-year

prospective, randomised study. Br J Oph-

thalmol 98: 73–78.Verma A, Degrado J, Hittelman AB, Wheeler

MA, Kaimakliotis HZ & Weiss RM (2011):

Effect of mitomycin C on concentrations of

vascular endothelial growth factor and its

receptors in bladder cancer cells and in

bladders of rats intravesically instilled with

mitomycin C. BJU Int 107: 1154–1161.Wilgus TA, Ferreira AM, Oberyszyn TM,

Bergdall VK & Dipietro LA (2008):

Regulation of scar formation by vascular

endothelial growth factor. Lab Invest 88:

579–590.Wu WC, Lai CC, Chen KJ, Chen TL, Wang

NK, Hwang YS, Yeung L & Li LM (2010):

Long-term tolerability and serum concen-

tration of bevacizumab (avastin) when

injected in newborn rabbit eyes. Invest

Ophthalmol Vis Sci 51: 3701–3708.Yoeruek E, Ziemssen F, Henke-Fahle S, Tatar

O, Tura A, Grisanti S, Bartz-Schmidt KU &

Szurman P (2008): Safety, penetration and

efficacy of topically applied bevacizumab:

evaluation of eyedrops in corneal neovascu-

larization after chemical burn. Acta Oph-

thalmol 86: 322–328.

Received on December 12th, 2014.

Accepted on April 11th, 2015.

Correspondence:

Ingeborg Stalmans

University Hospitals Leuven

Department of Ophthalmology

Kapucijnenvoer 33

B-3000, Leuven, Belgium

Tel: +32-16-33-23-72

Fax: +32-16-33-23-67

Email: Ingeborg.Stalmans@uzleuven.be

The authors wish to thank Sofie Beckers,

Martine Leijssen and Ann Verbeek for their

technical support. This work was supported

by IWT (Agentschap voor Innovatie door

Wetenschap en Technologie Vlaanderen),

FWO (Fonds voor Wetenschappelijk Ond-

erzoek Vlaanderen) and FRO (Fund for

Research in Ophthalmology).

Supporting Information

Additional Supporting Informationmay be found in the online version ofthis article:

Table S1. Corneal opacity: MMC0.01% – 1 min (day 34).Figure S1. Bleb survival and bloodvessel density after intracameral (IC),subconjunctival (SC), intravitreal (IV)injection of bevacizumab.Figure S2. Aqueous VEGF levels afterMMC administration.Figure S3. Bleb survival after loweringdose and/or exposure time of MMC.

678

Acta Ophthalmologica 2015