Evaluation of Pirfenidone as a New Postoperative Antiscarring Agent in Experimental Glaucoma Surgery

Abstract

Purpose.: To investigate whether topical administration of pirfenidone eye drops could be used to prevent postoperative scarring in a rabbit model of experimental glaucoma filtration surgery.

Methods.: In a randomized, controlled, masked-observer study, 40 rabbits underwent trabeculectomy in the right eyes and randomly received postoperative administration of 0.1% or 0.5% pirfenidone, perioperative mitomycin C (0.25 mg/mL), or no treatment. Bleb characteristics and functions were evaluated over a period of 4 weeks. The animals were killed on days 7, 14, and 28. Histopathology and immunohistochemistry were performed to determine the amount of scarring and fibrosis. Ocular toxicity was assessed by the Draize test, histopathology, and electron microscope.

Results.: The four treatment groups were similar with respect to intraocular pressure and anterior chamber depth. Pirfenidone 0.5% significantly prolonged bleb survival, and the blebs were larger and higher than those in the control group (P < 0.05); the 0.1% pirfenidone concentration was less effective. Furthermore, the histology and immunohistology results showed that the 0.5% pirfenidone and mitomycin C groups had less scarring at days 7 to 28 than did the controls. Toxicity assessments showed that pirfenidone did not damage the rabbit eyes.

Conclusions.: Postoperative use of 0.5% pirfenidone eye drops was associated with improved trabeculectomy bleb survival in a rabbit model. Pirfenidone eye drops may be a safe and effective antiscarring agent in glaucoma filtration surgery.

Postoperative subconjunctival wound healing is the major cause of late bleb failure after glaucoma filtration surgery. ^1,2 The use of local antimetabolite agents—5-fluorouracil (5-FU) and mitomycin C (MMC)—have been shown to prevent postoperative scarring, but the use of these agents are limited by the potential of blinding complications such as bleb leakage, hypotony, and infective endophthalmitis. ^{3 –5} Moreover, some patients, especially those with scarring risk factors (e.g., youth, aphakia, active anterior segment neovascularization, and inflammation), ⁶ may still develop late bleb failure, despite the use of high concentrations of 5-FU or MMC. These difficulties underscore a need to develop effective, safe, and well-tolerated agents that are able to modulate wound healing after glaucoma filtering surgery (e.g., trabeculectomy). 

Pirfenidone (5-methyl-1-phenyl-2-[1H]-pyridone; PFD) is a novel agent that has shown antifibrotic potential in animal models and clinical trials. ^{7 –9} PFD could downregulate a series of cytokines, including transforming growth factor (TGF)-β, ¹⁰ connective tissue growth factor (CTGF), ⁷ platelet-derived growth factors (PDGF), ¹¹ and tissue necrosis factor (TNF)-α. ¹² All these cytokines are important players in the wound-healing process. PFD could also scavenge reactive oxygen species (ROS) and thus inhibit tissue repair. ¹³ PFD's antifibrotic effects and its safety have been established in organs such as lung, ^14,15 liver, ⁸ and kidney. ⁷ In human retinal pigment epithelial (RPE) cells, PFD halts TGF-β1-induced fibronectin synthesis. ¹⁶ In addition, PFD could inhibit the interleukin-1-induced increase in tissue inhibitor of metalloproteinase(TIMP)-1 and has antifibrotic effects on orbital fibroblasts in patients with thyroid-associated ophthalmopathy (TAO). ¹⁷  

Our group showed that PFD can prevent human Tenon's fibroblast proliferation, migration, and contraction in vitro. ¹⁸ We also found that the effect of PFD on human Tenon's fibroblasts may be achieved through inhibition of mRNA and protein expression of TGF-β isoforms. ¹⁸ In the present study, we investigated whether PFD could suppress conjunctival scarring in an experimental model of glaucoma surgery in vivo. 

Materials and Methods

Animals

All animal procedures and methods used for securing the animal tissue complied with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research and our institutional guidelines. 

Forty New Zealand White rabbits aged between 12 and 14 weeks and weighing 2.0 to 2.5 kg were used. Experimental animals were housed in individual cages with a 12/12-hour light/dark cycle and fed ad libitum. All rabbits were acclimatized for 1 week before the experiments started. 

Surgical Procedure

Trabeculectomy was performed based on a standard study protocol: The general anesthesia induced intramuscular injections of ketamine hydrochloride (50 mg/kg; Gutian Pharmaceutical Company, Fujian, China) plus chlorpromazine hydrochloride (25 mg/kg; Gutian Pharmaceutical Company); local anesthesia was induced with proxymetacaine hydrochloride drops (Alcaine 0.5%; Alcon USA, Fort Worth, TX). A limbus-based conjunctival flap was produced 5 mm from the limbus; a limbus-based rectangular scleral flap was outlined with a triangular steel blade. A partial-thickness scleral flap was dissected carefully, starting 2 mm behind the limbus and continuing until the blade was just visible in the anterior cornea stroma. Thus, a 2-mm² square opening was made into the anterior chamber. A peripheral iridectomy was then performed. Tenon's capsular was closed with interrupted sutures, and a conjunctival incision separately closed with a 10-0 nylon suture (Alcon, USA) to give a water-tight closure. All glaucoma filtration surgeries were performed by a single surgeon with experience in using the rabbit model. 

All rabbits underwent filtration surgery to the right eye only and were randomly allocated to one of four treatment groups: Group A (n = 10) received 0.1% PFD eye drops six times per day; group B (n = 10) received 0.5% PFD eye drops six times per day; group C (n = 10) received an antimetabolism drug administered with a small (2 × 4-mm) lint impregnated with 0.25 mg/mL mitomycin C (MMC) on the site of the incision for 3 minutes; and group D (n = 10), the control group, received no treatment. 

Preparation and Administration of PFD

PFD (P2116; Sigma-Aldrich, St. Louis, MO) was dissolved in sterile water (AquaPro; Microgen, West Caldwell, NJ) at 1 (0.1%) and 5 (0.5%) mg/mL. Thiomersal was used as an eye drop preservative. One drop was administered in rabbit eye six times daily. 

Clinical Evaluations on Surgical Effects

Before and after the surgery (days 1, 3, 7, 10, 14, and 28), intraocular pressure (IOP) was measured in both eyes with an applanation tonometer (Tono-Pen Avia; Reichert, Inc., Depew, NY) under topical anesthesia at 7 to 8 AM. The mean reading of three measurements was record. The observer was not aware of the treatment. Slit lamp examination was performed to identify anterior chamber depth, which was recorded as deep (+2), shallow (+1), or flat (0). Filtration bleb survival was taken as the primary efficacy end point. Bleb failure was defined as the appearance of a flat, vascularized, and scarred bleb in the presence of a deep anterior chamber. According to the bleb size measurement method described by Cordeiro et al., ¹⁹ we calculated the bleb area with caliper measurements of width and depth and we graded the bleb height as defined above. 

Histopathologic Evaluations

At 1, 2, and 4 weeks, animals (3, 3, and 4 rabbits, respectively, at each time point) were killed by a lethal intravenous injection of excess phenobarbitone, and the tissues were processed for histopathology. Both eyes were enucleated. The upper lid was left intact, attached to the globe to preserve the architecture of the superior fornix and conjunctival tissues around the drainage site. All the eyes were fixed in formalin acetic acid (FAA) alcohol solution for 24 hours, stored in 70% alcohol, and fixed in paraffin wax. Sequential 5-μm sections of the operative wound site were prepared and stained with hematoxylin and eosin (H&E) to obtain a general impression of total cellularity; with Masson trichrome, to determine collagen deposition with α-smooth muscle actin (α-SMA) immunohistochemistry, to identify myofibroblast cells, and with proliferating cell nuclear antigen (PCNA) immunohistochemistry, to reveal recent cell division. For immunohistologic studies, the number of positively stained proliferating cells per five high-powered fields was counted. 

Toxicity Assessment

The Draize eye test ¹⁵ was used to evaluate the irritation and ocular surface toxicity of topical PFD solutions. Solutions (50 μL) of 0.1%, 0.5%, and 1% PFD were administered into the rabbit eyes, and we used the slit lamp imaging system to evaluate conjunctiva congestion, cornea opacity, and iris congestion. 

H&E-stained sections at high magnification were used to evaluate the cell and tissue structure of corneal epithelium, conjunctival epithelium, ciliary body, and retina. 

We also compared the electron micrograph characteristics of eyes treated with postoperative PFD and untreated eyes at day 28. In each group, three rabbits were used for electron microscopic evaluations. The reason for choosing this time point is that the toxic effects may be at maximum in the end, compared with other time points. The corneas were excised and postfixed in 2.5% glutaraldehyde for at least 2 hours; rinsed in PBS three times for 10 minutes; dehydrated in an ethanol gradient; critical-point dried; and gold coated before observation in a scanning electron microscope (model XL-30E SEM; Philips, Eindhoven, The Netherlands). The endothelium was counted, and the morphology was noted. 

The ciliary body and retina-choroid tissues were dissected into 1-mm³ sections and postfixed in phosphate buffered 2.5% glutaraldehyde overnight. Then, tissue fragments were postfixed in 1% osmium tetroxide and embedded in Epon epoxy resin. The sections were stained with uranyl acetate and lead citrate. Semithin sections were evaluated by light microscopy; ultrathin sections were stained with uranyl acetate and lead citrate. Sections were examined with transmission electron microscope (model CM10; Philips). The ultrastructure of ciliary body and retina-choroid tissues were examined in detail. 

Statistical Analysis

Statistical analysis was performed to determine the differences in bleb survival times, IOP, and cell count among the four groups (SPSS 13.0 software; SPSS Inc., Chicago, IL). ANOVA test was performed, and significance was assumed if P < 0.05. With a sample size of three rabbits per group, we have a >90% statistical power (95% confidence interval; α = 0.05) to detect a 5.3-day (SD 2.2) difference in bleb survival time and a 2.3-mm Hg (SD 1.2) difference in IOP level between the groups. 

Results

Slit Lamp Observations

The fate of the experimental and control eyes was evaluated clinically on the designated days. No case of endophthalmitis was observed. Corneal edema around the wound was detected as a transient finding in a few animals. Both the control and treated groups showed no sign of cataract during the period. The anterior chamber was flat in most of the animals 1 day after surgery. Over the next 7 days, the anterior chamber gradually deepened. No significant difference was found between the four groups in the time taken for the anterior chamber to deepen (P = 0.782). 

Bleb

The mean bleb survival times (mean ± SD) were 14.5 ± 2.1, 20.3 ± 4.6, 19.0 ± 2.0, and 12.8 ± 2.2 days in the A, B, C, and D groups, respectively. Survival curves of blebs are shown in Figure 1. A Kaplan-Meier analysis showed a significant difference in the survival distributions among the four comparison groups (log rank = 19.633; P < 0.01). The 0.5% PFD group and the MMC group significantly improved bleb survival compared with the 0.1% PFD group and the untreated control group (P < 0.05). There were no significant differences in bleb area and bleb height in the four groups within 7 days after the surgery, but both the 0.5% PFD and the MMC group showed a significantly larger bleb area than the 0.1% PFD and the control group (P < 0.05; Fig. 1). Figure 2 shows the appearances of the filtration blebs on day 14. Treatment with 0.5% PFD (group B) was associated with typically well-formed, elevated, diffuse, fleshy looking blebs compared with the flat, scarred blebs in the A and D groups. 

Figure 1.

 

Filtering blebs after glaucoma filtration surgery. Filtering blebs in 0.1% PFD (n = 4) and the control (n = 4) group almost failed within 2 weeks. Treatment with 0.5% PFD (n = 4) markedly increased the bleb survival period. In the MMC group (n = 4), the bleb survived nearly 3 weeks. A Kaplan-Meier analysis showed a significant difference in the survival distributions among the four groups (log rank = 19.633; P < 0.01).

Figure 2.

 

Filtering blebs on day 14 after the surgery. Arrows: surgical area and bleb margin. In the 0.5% PFD group (B), the bleb remained diffusely elevated, with little dilation of blood vessels in the local conjunctiva. MMC treatment (C) resulted in a characteristically avascular and thin bleb. Both 0.1% PFD treatment (A) and no treatment (D) resulted in bleb failure, indicated by a flat, scarred, and vascularized bleb.

Intraocular Pressure

IOP was measured between 7 and 8 AM on the designated days (before the surgery and on days 1, 3, 7, 10, 14, 21, and 28 after the surgery). For every group, IOP decreased in the early stage after surgery and slowly increased with time. The mean IOP (SD) levels among groups A, B, C, and D were 13.0 (1.4), 12.7 (2.5), 12.5 (1.3), and 13.5 (2.2), respectively, before the surgery. These readings were 8.6 (1.6), 8.9 (1.3), 9.8 (2.2), and 8.3 (2.6), respectively, on day 1; 11.9 (1.8), 11.3 (1.0), 11.1 (1.7), and 11.0 (1.1) on day 7; and 13.8 (0.5), 12.0 (0.3), 12.5 (0.7), and 12.5 (0.7) on day 28 after surgery. There were no significant differences between all the groups over the entire study period (P > 0.05 for all the postoperative time points). 

Histopathologic Features

Histologic analysis of the specimens was performed at the center of the sclerotomy site, as indicated by the location of iridectomy. In groups A and D, histologic profiles revealed numerous activated fibroblasts, moderate inflammatory cell infiltration, and much more frequent vascular patterns in the subconjunctival area at day 7. At day 14, the increased level of collagen deposition (fibrosis) and dense subconjunctival tissues became evident. At day 28, dense fibrocellular scar tissue and closed sclerotomy was observed in control group D and 0.1% PFD group. Groups B and C showed only a mild fibrotic response and depositions of collagen in the subconjunctival spaces. Inflammatory cell infiltration and fibroblast proliferation were obviously lower compared with those in groups A and D. The subepithelial connective tissue was loosely arranged and contained histologically clear subconjunctival filtration space. Even when the scleral wound healed by day 28, the residues of the fistula areas were still present in groups B and C (Figs. 3B, 3C). Masson trichrome staining showed much looser, hypocellular connective tissue and rare blue collagen deposition in the MMC group. In group B, similar to group C, the subconjunctival area contained loose tissues and a few cells. Collagen deposition was much lower than in groups A and C, which contained more collagen deposition and hypercellular fibrotic tissues (Figs. 3E–H). 

Figure 3.

 

Histologic characteristics of the filtration site on day 28. Tissues shown in (A–D) were stained with H&E. In the 0.1% PFD treatment (A) and control groups (D), subconjunctival scarring was noted with fibrocellular scar tissue and closed sclerotomy. By contrast, in the 0.5% PFD group (B), the subepithelial connective tissues were loosely arranged, and residue was seen in the fistula area. In the MMC group (C), there were a few inflammatory cells and thin subconjunctival tissue, and a fistula tract could also be seen. C, conjunctiva; SS, sclerostomy site; and CB, ciliary body. (E–H) Tissues stained with Masson trichrome. The quantity of collagen deposition (fibrosis) was associated with staining density and extent (blue). The control (H) and 0.1% PFD (E) group stained more densely than the (F) 0.5% PFD- and (G) MMC-treated groups.

PCNA immunohistochemistry staining results are shown in Figure 4 and Table 1. Cell division was higher in groups A and D than in B and C. Groups B and D demonstrated a reduction in the number of PCNA-expressing cells compared with the control group and the 0.1% PFD-treated eyes (Fig. 4). At days 14 and 28, positive cell counts in the B and C groups were significantly lower than in the A and D groups (P < 0.05), but there was no significant difference between the B and C groups (P > 0.05; Table 1). 

Figure 4.

 

Immunohistochemical staining of filtration blebs on day 28. There was a reduction in the number of PCNA-expressing cells in 0.5% PFD (B) and MMC (C) treatment groups compared with the 0.1% PFD (A) and control (D) groups. Arrows: PCNA expression in the filtration areas with higher magnification.

Table 1.

 

Cells Expressing PCNA among the Four Treatment Groups

Group	Day 7 after Surgery (n = 3)	Day 14 after Surgery (n = 3)	Day 28 after Surgery (n = 4)
A	50.2 ± 25.4	55.5 ± 4.0	67.1 ± 7.6
B	31.6 ± 14.1	34.6 ± 10.1* †	23.3 ± 10.1* †
C	24.6 ± 4.7	13.6 ± 1.5* †	10.6 ± 2.1* †
D	57.8 ± 15.1	50.2 ± 5.7	61.1 ± 13.6

 Data are expressed as mean ± standard deviation. n = number of animals. The groups are as described in the text.

*  PCNA expressing cells among these groups are significantly different from group A (P < 0.05).

†  PCNA expressing cells among these groups are significantly different from group D (P < 0.05).

α-SMA is the hallmark of myofibroblast generation and the development of fibrotic tissue. In all groups, immunohistochemical analysis showed that α-SMA-expressing cells in the subconjunctival space were still positive on day 28, even though they were fewer than on days 7 and 14. Immunohistochemical staining of the filtration blebs showed a reduced number of α-SMA-expressing cells in groups B and C than in groups A and D (Fig. 5). 

Figure 5.

 

Immunohistochemical staining of filtration blebs on day 28. Fibroblast differentiation to myofibroblast phenotype is characterized by the expression and assembly of α-SMA into stress fibers. The 0.5% PFD (B) and MMC (C) treatment groups showed reduced expression of α-SMA compared with the 0.1% PFD (A) and control (D) groups. (A, D, red arrows) Areas with increased α-SMA expression, contrasting with the rare expression in the same representative areas in (B) and (C). Black arrows: normal α-SMA expression by the ciliary muscle cells and conjunctival vascular smooth muscle cell.

Toxicity

After topical administration, no significant changes (e.g., conjunctival redness, chemosis, or discharge) were observed in groups A and B by slit lamp examination. The cornea was clear and integrated. The iris was normal without congestion and swelling. The Draize test score was 0 at all times. Histologic analysis at the light microscopic level showed that there was no obvious damage to the corneal epithelium, conjunctival epithelium, ciliary body, and retina (Fig. 6). Scanning electron microscopy showed the corneal endothelial cells in the four groups (Fig. 7). Morphologically, there was no significant damage to the corneal endothelial cells, and there was no significant difference in cell count among the four groups (F = 2.27; P = 0.134). Under the transmission electron microscope, there were no obvious differences in the ultrastructure of the ciliary body among the four groups (Fig. 8). The mitochondria appeared normal, and there were no apparent differences among the four treatment groups. The RPE microvilli appeared normal and circumvoluted the photoreceptor. The retinal layers were clear, and there were no signs of swelling, edema, degeneration, and necrosis. No pathologic change in the retina was found in any of the groups. 

Figure 6.

 

Histologic analysis of rabbit eyes 28 days after the administration of PFD eye drops. H-E stain. (A1–D1) Clear corneal architecture with tight cell–cell junction without necrosis or exfoliation of cells. (A2–D2) Integrated conjunctival epithelium and clear cell structure. (A3–D3) Ciliary processes. (C2) Relatively thin conjunctival epithelium in the MMC-treated group. There were sparse vacuoles in the pigmented and nonpigmented epithelia among the four treatment groups. The ciliary epithelium appeared to have a normal thickness and was regularly arranged. There was no sign of bleeding or necrosis. (A4–D4) Entire retinal layers were visible and appeared normal. There was no structural retinal damage in all treated eyes. Magnification, ×400.

Figure 7.

 

Scanning electron microscopic images of corneal endothelium. The 0.1% PFD (A) and control (D) groups showed integrated hexagonal cells of uniform size with interdigitations of the cell borders, and numerous microvilli appeared as multiple protrusions on the cell surface. The 0.5% PFD treatment group (B) showed a few irregular size cells and unclear cell borders. (C) The MMC treatment group showed a small number of edematous endothelial cells and a reduced number of microvilli. Magnifications, ×3000.

Figure 8.

 

Transmission electron microscopy of ciliary epithelium. (A–D) The mitochondria appeared normal, with no apparent differences among the four treatment groups.

Discussion

Postoperative scarring leading to late failure of glaucoma filtration surgery remains a major barrier to long-term IOP control and arrest of optic neurologic damage. ^{20 –22} We have shown that PFD has an inhibitory effect on proliferation, migration, and collagen contraction of human Tenon's fibroblasts in vitro. ¹⁸ The present study extended our finding by showing that PFD, as a postoperative topical eye drop, could exert its antifibrotic effect on a rabbit model and thus improve the outcome of glaucoma filtration surgery. Although we showed that 0.5% PFD had a similar antifibrotic effect with MMC, its topical administration may be safer, more flexible, and better tolerated. On the other hand, MMC and 5-FU are more likely to be associated with blinding complications, such as leaky blebs and endophthalmitis. ^{3 –5} Taken together, our study showed that PFD may serve as a promising antifibrotic agent in glaucoma filtration surgery. 

TGF-β is a pivotal mediator of the wound-healing process; inhibiting TGF-β may be an effective means of preventing postoperative scarring after trabeculectomy. ^2,3 Although the antifibrotic mechanism of PFD in the eye remains unclear, our previous in vitro results showed that PFD could downregulate expression of TGF-β mRNA and protein in human Tenon's fibroblasts. ¹⁴ PFD has been demonstrated to play anti-inflammatory and antifibrotic roles in a range of cells (e.g., hepatic stellate cells, renal fibroblasts, myometrial cells, and leiomyoma cells) in vitro and in many organs (e.g., kidney, ²⁰ lung, ²¹ and liver ²² ) in vivo. Apart from in vitro and in vivo experiments, PFD has been applied in several clinical trials among patients with fibrotic disorders, ²³ including hypertrophic cardiomyopathy, pulmonary fibrosis, neurofibromatosis, uterine leiomyoma, idiopathic scleroderma, sclerosing peritonitis, and radiation-induced fibrosis. 

There were no significant differences in IOP among the study groups. This is not surprising, as the preoperative IOP was not elevated, and thus IOP may not be a reliable indicator of treatment effect. ²² The tonometer (Tono-Pen; Reichert) may be inaccurate in measuring the rabbit's IOP level, especially when the variance is rather small at low-teens IOP. ²⁴ The tonometer may have also underestimated the IOP level in some cases. The wound-healing response of the surgical site in the rabbit is usually much stronger than in humans, ^25,26 limiting our ability to observe a functional bleb in a reasonable period. ²⁷ Given that histologic analysis may provide a more precise assessment of the wound-healing response, ²³ we used bleb survival and histologic analysis as primary outcomes. In our histologic findings, treatment of 0.5% PFD or MMC resulted in mild fibrotic responses and depositions of collagen in the subconjunctival spaces, as well as reduction in tissue cellularity, recent cell division (PCNA), and collagen deposition. The antifibrotic effect of 0.5% PFD or MMC persisted during the whole postsurgery period. The antiscarring effect of 0.1% PFD was not as strong as 0.5% PFD, since the effects of PFD may be dose dependent. 

Based on the Trypan blue viability test, our previous work showed that PFD exhibited its antiscarring effect on human Tenon's fibroblasts at nontoxic concentrations. ¹⁸ In the present study, 0.1% and 0.5% PFD showed no apparent toxicity in cornea, ciliary body, and retina. Thus, 0.5% PFD may be safe for use in the eye. The toxic effects of MMC have been reported previously, but were not apparent in this study; it may be that the MMC concentration was lower (0.25 mg/mL) and the duration of administration was short. 

In summary, PFD reduced the wound-healing response in surgical sites in a rabbit model. PFD eye drops may provide a safer and more convenient antifibrotic treatment than MMC. A long-term administration of PFD eye drops may further illustrate its antiscarring role after glaucoma filtration surgery. 

Footnotes

 Supported by Guangdong Science and Technology Fund Grant 2009B030801193 and Medical Scientific Research Foundation of Guangdong Province, China Grant B2009082.

Footnotes

 Disclosure: H. Zhong, None; G. Sun, None; X. Lin, None; K. Wu, None; M. Yu, None

The authors thank Geoffrey Arden (City University London) for paper revisions.