Corneal Neovascularization and Ocular Irritancy Responses in Dogs Following Topical Ocular Administration of an EP4-Prostaglandin E 2 Agonist

Abstract

Prostaglandin receptor agonists have intraocular pressure–lowering effects in humans and are of interest in the treatment of glaucoma. The prostanoid receptor agonist PF-04475270 is a potent and selective agonist of the prostaglandin E₂ receptor EP4. This paper characterizes the toxicity associated with topical ocular administration of PF-04475270 in beagles. Dogs were given PF-04475270 topically to the eye on a consecutive daily dosing schedule for one or four weeks followed by a one-or four-week reversal period, respectively. Clinical observations, ophthalmic, and laboratory parameters were recorded. Necropsies were conducted at the end of the dosing and recovery phases, and histologic examinations performed. Corneal neovascularization that was considered adverse was observed at doses of ≥1.0 μg/eye and was not reversed by the end of the recovery phase. Dogs dosed with ≥0.25 μg/eye developed a dose-related conjunctival hyperemia that persisted throughout the reversal period. Corneal neovascular cells stained positive with EP4 and the endothelial biomarker Factor VIII-vWF. Other histopathology findings observed at doses of ≥1.0 μg included single-cell necrosis and neutrophils in the cornea, inflammatory cell infiltrates in the iris/ciliary body, and iridal endothelial cell hypertrophy. A resolving acute to subacute inflammation in the iris/ciliary body was observed after the four-week recovery period.

Introduction

The various prostaglandins (PGs) of the D, E, F, and I types are cyclooxygenase (COX) metabolites of the unsaturated fatty acid arachidonic acid (Sugimoto and Narumiya 2007). Synthesized immediately in response to a variety of stimuli by a variety of cells, PGs act locally and mediate diverse actions. Receptors mediating prostanoid actions are designated DP, EP, FP, and IP according to the natural prostanoid type they preferentially bind: PGD₂, PGE₂, PGF_2α, and PGI₂, respectively (Breyer et al. 2001; Narumiya, Sugimoto, and Ushikubi 1999). Four receptors for PGE₂ have been characterized based on pharmacological analysis and are designated EP1, EP2, EP3, and EP4 (Sugimoto and Narumiya 2007). Topical administration of prostaglandins PGE₂ or PGF_2α into the eyes of animals was reported to lower intraocular pressure (IOP), and since then, PGs have become a therapeutic target of interest for glaucoma (Bito 2001; Bito et al. 1983; Stern and Bito 1982). Many glaucoma therapies are now directed toward the hypotensive properties of prostaglandin analogs (Woodward et al. 2004) and prostaglandin receptor agonists (Hellberg et al. 2002; Nilsson et al. 2006; Prasanna et al. 2009; Stjernschantz et al. 2000; Woodward et al. 2009). The primary mechanisms by which prostamide mimetics and FP, EP1, and EP2 receptor agonists are reported to lower IOP is by increasing uveoscleral outflow (Hellberg et al. 2002; Nilsson et al. 2006; Stjernschantz et al. 2000; Toris et al. 1997). Recently, an EP4 receptor agonist was reported to induce ocular hypotension in monkeys by a mechanism that does not appear to involve uveoscleral outflow (Woodward et al. 2009).

PF-04475270 (Pfizer, Inc.) is a topically administered, selective EP4 receptor agonist (Prasanna et al. 2009) and a prodrug of the active drug moiety CP-734432. CP-734432 stimulated cAMP production in vitro, underwent rapid hydrolysis by corneal homogenates, demonstrated good permeability in a human corneal epithelial model, and exhibited highest exposure levels in the cornea followed by mid-level exposures in the aqueous humor and lower levels of exposure in the iris and ciliary body in the dog (Prasanna et al. 2009). When given once daily at ≥0.25 μg/eye in the normotensive beagles, PF-04475270 lowered IOP by 30% to 45% at twenty-four hours postdose relative to the vehicle control and was associated with a minimal to mild conjunctival hyperemia (Prasanna et al. 2009). Because PF-04475270 lowered IOP as early as two hours postdose and maximally at twenty-four hours (Prasanna et al. 2009), preclinical safety studies were pursued to evaluate the safety profile of this novel EP4 receptor agonist.

We report here the nonclinical safety profile of PF-04475270 as characterized in beagle dogs. The objective of the work included the characterization of systemic exposure and target organ toxicity following ocular topical instillation of PF-04475270 for one week and up to four weeks in beagles and verification of the reversibility of any treatment-related effects.

Materials and Methods

Two topical ocular instillation toxicity studies in dogs are described: one-and four-week toxicity studies followed by one-and four-week recovery periods, respectively.

Animals and Husbandry

Eight-month-old male beagle dogs were used for the one-week study. Older (11.4–12.3 months) male and female beagles weighing 7.1–10.2 kg and 6.5–10.1 kg, respectively, at the start of treatment, were used for the four-week study. All dogs for both studies were supplied by Marshall BioResources, North Rose, NY.

In the one-week study, standard procedure and conditions were applied for animal care, feeding, and maintenance of the rooms, caging, and environment in a facility accredited by the Association for Assessment and Accreditation of Laboratory Animal Care International. All procedures involving laboratory animals were reviewed and approved by the Institutional Animal Care and Use Committee associated with the facility in which animals were housed. The four-week study was conducted in compliance with the Food and Drug administration Good Laboratory Practice regulations (Code of Federal Regulation, Title 21, Part 58).

Drug

PF-04475270 was supplied for all studies by Pfizer Global Research and Development, La Jolla Laboratories (San Diego, CA, USA). All dose levels are represented as the free base equivalent mg/kg dose. PF-04475270 was formulated in 0.3%Tween 80, 0.02% benzalkonium chloride, 0.1% EDTA, 0.86% NaCl, and 10 mM citrate buffer at pH 5.5.

Study Design

In the one-week study, two male beagles per group were given vehicle or PF-04475270 in both eyes at 0, 0.5, 2.5, and 5.0 μg/eye and were euthanized at scheduled times at the end of the dosing phase on day 8, and at the end of the recovery phase on day 15. Toxicity was determined by assessing changes in clinical signs using a direct ophthalmoscope three times a day. Ocular irritation was scored in accordance with the Organization for Economic Cooperation and Development 1987 guidelines. Ophthalmic examinations were performed once daily in eyes dilated with phenylephrine (AK-Dilate) and cyclopentolate (AK-Pentolate, Akorn Inc, Somerset, NJ, USA) using a slit lamp and indirect ophthalmoscope. Eyes were fixed in 6% glutaraldehyde, processed to slides, and stained with hematoxylin and eosin or immunohistochemically for histologic evaluation. Blood samples for the assessment of levels of systemic exposure to PF-04475270 were collected from the central auricular artery on days 1 and 7 of the study at five, fifteen, forty-five, and ninety minutes post-dose. Hematology and clinical chemistry parameters were analyzed from blood samples procured on day 8 and at recovery on day 15. Necropsy was performed at the end of the dosing and recovery phases; eyes and eyelids were retained for histopathologic examinations.

In the four-week study, three to five male and female beagles were grouped as controls or PF-04475270–treated. The control dogs were given vehicle in the left eye, whereas the right eye remained naïve. The PF-04475270–treated dogs were given the compound at 0.25, 1.0, or 2.5 μg/dose in the right eye and given the vehicle in the left eye. The dose volume of either vehicle or PF-04475270 was 30 μL/eye/dose and doses were given once daily. Three animals/sex/group were euthanized after at least four weeks of treatment. Animals designated for recovery (two animals/sex in the vehicle and 2.5 μg/eye/dose groups) had a four-week dosing-free period following the last day of dose administration.

The animals were assessed twice daily (morning and afternoon) for mortality, abnormalities, and signs of pain or distress. Body weights were recorded during the pretest period and at least weekly thereafter. Electrocardiography was recorded once during the predose phase and during weeks 2 and 4 of the dosing phase. Post-dose ocular observations for evidence of ocular squinting were done immediately following dosing for the first week and then once weekly thereafter during the dosing phase at approximately two hours post-dose. Ocular irritation scoring (modified Draize system) was done once during the predose phase; on days 1, 2, and 29 of the dosing phase; daily for the first week of the recovery phase; and weekly thereafter. Ophthalmoscopic examinations were performed once during the predose phase; on days 1, 3, and 8 of the dosing phase; and approximately weekly thereafter during the dosing and recovery phases. The pupils of both eyes of each dog were dilated with a mydriatic agent, and the eyes were examined with an indirect ophthalmoscope, as well as grossly examined and graded using a modified MacDonald-Shadduck scoring system. Fluorescein staining examinations were performed during the predose phase; on days 2, 4, and 9 of the dosing phase; and approximately weekly thereafter during the dosing and recovery phases. Intraocular pressure (IOP) and central corneal thickness (pachymetry) measurements were done during the predose phase; on day 1 of the dosing phase; and approximately weekly throughout the study. Electroretinography was done once during the predose phase, during week 4 of the dosing phase, and in the week before euthanasia during the recovery phase.

Laboratory examinations (hematology, clinical chemistry, and urinalysis) were performed once during the predose phase and just prior to scheduled euthanasia at the end of the dosing and recovery phases. Blood samples for the assessment of levels of systemic exposure to PF-04475270 were collected from the central auricular artery on day 1 and in week 4 of the study at five, fifteen, and forty-five minutes, and two hours post-dose. Necropsy was conducted at the end of the treatment and recovery periods; selected organs were weighed, and histologic examinations were performed. Eyes were fixed in Davidson’s fixative, processed to slides, and stained with hematoxylin and eosin or immunohistochemically for histologic evaluation.

Immunohistochemistry

Eyes were sectioned at 3 μm, placed on charged slides, air-dried, and loaded onto the automated Bond system (Leica Microsystems). All solutions used for automated immunohistochemistry were from Leica Microsystems unless otherwise specified. Slides underwent deparaffinization, heat-induced epitope retrieval (H2 solution, twenty minutes), and staining online using a heat protocol. Slides were incubated ten minutes with Background Buster (Innovex Biosciences, Inc). Next, slides were incubated with rabbit EP4 polyclonal antibody (1:200; Cayman) or rabbit antihuman von Willebrand factor (1:2000; Sigma) for thirty minutes. The antibodies were labeled using an alkaline phosphatase polymer red detection kit (Vision Biosystems). Slides were counterstained with hematoxylin for two minutes, removed from the autostainer, and coverslipped with Micromount mounting medium (Surgipath).

Results

Clinical Signs of Toxicity

One-Week Study with a One-Week Recovery Period:

Treatment-related ocular irritancy responses were observed at ≥0.5 μg/eye/dose. At the 0.5-and 2.5-μg dose levels, a dose-related ocular irritancy response was observed immediately after dosing with some noticeable recovery before the next dose. A mild to moderate ocular irritancy response included congested scleral vessels, squinting reactions, and a tendency toward increased conjunctival hyperemia or “red eye” with repeat dosing. A minimal to mild reddening and protrusion of the nictitating membrane was observed as early as day 1 post-dose in dogs given 5.0 μg/eye and by day 3 in dogs given ≥2.5 μg/eye and persisted through final dosing. A minimal ocular discharge was evident by day 3 and lasted to the end of dosing on day 7 in dogs given ≥0.5 μg/eye/dose. Conjunctival hyperemia persisted through the recovery period in one dog given 2.5 μg/eye.

Four-Week Study with a Four-Week Recovery Period:

Ocular clinical signs following topical instillation of ≥1.0 μg PF-04472570 to the right eye included cloudy ocular discharge, closed eye, red conjunctiva, and squinting for five to ten seconds and up to one minute. The incidence of the majority of these clinical signs increased with dose.

Treatment-related ocular irritancy responses noted by the Draize scoring method were consistent with the McDonald-Shadduck scoring scheme and are summarized in Table 1. In general, all doses tested in both male and female dogs resulted in varying degrees of ocular irritation, with no significant sex differences. In the majority of dogs given ≥1.0 μg/eye, incomplete pupillary dilation or mild miosis occurred following topical ocular instillation of a mydriatic agent, although anisocoria prior to dilation was not evident. With the exception of one dog given 1.0 μg/eye, all dogs given ≥1.0 μg/eye PF-04472570 had swelling or congestion in the fold of the iris. Conjunctival hyperemia, conjunctival chemosis (swelling), and conjunctival discharge were noted by Draize scoring. Corneal neovascularization, conjunctival hyperemia, conjunctival chemosis, conjunctival discharge, aqueous flare, loss of corneal transparency, corneal opacity, iridal hemorrhage, blepharospasm, and subtle diffuse corneal edema were noted by McDonald-Shadduck scoring. Blepharospasm was observed in dogs given ≥0.25 μg/eye and was more noticeable in dogs given 2.5 μg/eye. Conjunctival hyperemia in dogs given ≥0.25 μg/eye did not diminish with repeated dosing, but increased in a dose-dependent manner and persisted through the recovery period.

Grading of the conjunctival hyperemia following treatment with PF-04475270 ranged from 1 to 3 and corresponded with a severity of minimal = 1, mild = 2, or moderate = 3. A grade 2 conjunctival hyperemia was observed in one of six dogs given 0.25 μg/eye, five of six dogs given 1.0 μg/eye, and ten of ten dogs given 2.5 μg/eye. On day 8 in the study, a grade 3 conjunctival hyperemia was observed in one dog given 2.5 μg/eye. Conjunctival chemosis ranged from grade 1 to 2. A grade 1 conjunctival chemosis was observed in four of four dogs given 1.0 μg/eye and all dogs given 2.5 μg/eye. A grade 2 conjunctival chemosis was observed in three of ten dogs given 2.5 μg/eye over the course of the study. Varying degrees of ocular discharge were identified in dogs given 2.5 μg/eye. Conjunctival chemosis and ocular discharge was not observed during the recovery phase of the study, but conjunctival hyperemia was slow to improve in dogs given 2.5 μg/eye. A grade 1 conjunctival hyperemia was observed in four of four dogs on day 14, three of four dogs on day 21, and one of four dogs on day 28 at the 2.5 μg/eye dose level during the recovery phase.

Slit-lamp biomicroscopy of the cornea revealed circumlimbal corneal neovascularization, corneal edema, and aqueous flare in dogs given PF-04475270. Two dogs given 1.0 μg/eye and eight dogs given 2.5 μg/eye PF-04475270 had a short (less than 2 mm in length) circumferential ciliary flush (neovascularization) into the peripheral cornea. During the dosing phase, corneal neovascularization was observed on day 9 in eight of ten dogs and on day 29 in nine of ten dogs given 2.5 μg/eye. Corneal neovascularization of similar clinical severity was observed in two of six dogs given 1.0 μg/eye on day 29. During the recovery phase, ghost vessels with no active blood flow were observed by day 6. Trace amounts of protein in the anterior chamber (aqueous flare) and mild corneal edema were observed during the dosing phase in dogs given ≥1.0 μg/eye and resolved during the recovery phase, indicating the reversibility of these treatment effects.

Central corneal thickness measurements (corneal pachymetry) prior to dosing were within normal limits for young beagle dogs in a laboratory setting. Normal central corneal thickness measurements ranged between 543 to 617 μm. Corneal thickness values were increased over the vehicle control group values for dogs given ≥1.0 μg/eye at various time points during the study. For dogs given 1.0 μg/eye/dose, corneal thickness was significantly increased in males on days 8, 15, and 29 and on days 8 and 22 of the dosing phase for females. For animals given 2.5 μg/eye/dose, corneal thickness was significantly increased on days 8, 15, 22, and 29 of the dosing phase for males and females. The central corneal thickness measurements in PF-04475270–treated eyes ranged from 623 to 717 μm. The central corneal thickness values were not significantly different in dogs given 0.25 μg/eye of PF-04475270 or vehicle before or after dosing. No difference in central corneal thickness between PF-04475270–treated, vehicle-treated, or naïve control eyes was noted during the recovery phase of the study, indicating reversibility of the PF-04475270–related effect.

Intraocular Pressure:

Treatment with PF-04475270 in beagles significantly decreased the intraocular pressure (IOP) by 5 to 6 mmHg when given 0.25 μg/eye or by 7 mmHg when given ≥1.0 μg/eye. Although PF-04475270 significantly decreased the IOP, there was no apparent dose-related response. A zero to 27% decrease in IOP was observed at six hours post-dose, and a maximal decrease in IOP ranging between 30% to 50% occurred in all dogs at twenty-four hours post-dose. During the recovery phase, the IOP returned to baseline levels twenty-four hours after the last dose.

Electroretinography (ERG), Electrocardiography (ECG), Macroscopic and Organ Weight, Clinical Chemistry, and Hematology Findings:

No abnormal findings attributable to the administration of PF-0447520 were observed.

Histopathologic and Immunohistochemistry Evaluations

The incidence of treatment-related histopathologic findings for the one-week and four-week studies is summarized in Table 2. All of these findings were reversible or tending toward reversibility, except for corneal neovascularization, at the end of the treatment-free recovery period. The changes considered to be the primary preclinical effects of PF-04475270 were located in the cornea, iris, conjunctiva, and nictitating membranes. Treatment-related histopathologic changes are described below and illustrated in Figures 1, 2, and 3.

Corneal neovascularization was observed after the first week of treatment at doses of ≥2.5 μg/eye (Table 2). This change occurred in the deep and superficial stroma of the peripheral cornea and was characterized by the presence of congested vessels lined by plump endothelial cells (Figure 1). Corneal neovascularization was not reversible after a one-week recovery period in a dog given 2.5 μg/eye (Table 2). Immunohistochemistry evaluation for EP4 expression in the eyes revealed cytoplasmic staining of neovascular endothelial cells (Figure 1) as well as resident blood vessels in the conjunctiva, retina and epithelium, endothelium, and stroma of the cornea (data not shown). Other treatment-related changes in the vicinity of the limbus included minimal multifocal neutrophilic infitrates in the corneal stroma and epithelium, single-cell necrosis in the corneal epithelium (Figure 2), and multifocal lymphoplasmacytic infiltrates in the iris and ciliary body (Figure 2).

The presence of neovascular endothelium in the peripheral cornea at lower doses after a four-week treatment period was confirmed using the endothelial cell biomarker Factor-VIII-von Willebrand (vWF). The cytoplasm of the corneal neovascular endothelial cells stained positively for Factor VIII-vWF (Figure 3). Another finding in the iris that was not observed in the one-week study was endothelial cell hypertrophy (Figure 3). Immunohistochemistry evaluation using Factor VIII-vWF confirmed an increased number of Factor-VIII-vWF–positive cells within the iris after four weeks of treatment (Figure 3). At the end of the four-week treatment-free period, the number of hypertrophied Factor-VIII-vWF–positive endothelial cells was similar to the control eye, indicating the reversibility of the finding. Other changes observed in the four-week study included minimal to mild neutrophilic infiltrates in the corneal epithelium and stroma, corneal epithelial degeneration, lymphoplasmacytic infiltrates in the iris (Figure 3), and ciliary body, acute to subacute inflammation in the palpebral and bulbar conjunctiva, and nictitating membrane. Although lymphoplasmacytic infiltrates in the iris and conjunctival inflammation were observed at the end of the recovery period, there was distinct evidence of ongoing recovery.

Toxicokinetics

In the one-week study, systemic exposure of the active moiety CP-734432 of the prodrug PF-04475270 was limited to ≤0.8 ng/mL for the 2.50 and 5.00 μg/eye dose levels at five and fifteen minutes post-dose and below the limit of detection at forty-five minutes post-dose. There was no evidence of systemic accumulation after seven consecutive days of dosing. In the four-week study, systemic exposure was low and below the limit of quantitation (<0.500 ng/mL) in most samples at most collection times.

Discussion

In dogs, PF-04475270 induced corneal neovascularization at doses of ≥1.0 μg/eye and conjunctival hyperemia or red eye at doses of ≥0.25 μg/eye, and these adverse findings were not reversible during the one-week or four-week recovery period. The safety margin for corneal neovascularization was approximately 5× and calculated by dividing the no observed adverse effect level (NOAEL) at 0.25 μg by the human predicted efficacious therapeutic dose of 0.05 μg. Thus, the safety margin for corneal neovascularization was considered insufficient for the targeted therapeutic indication. Because the red eye was not reversible during the one-week or four-week reversal period and other ocular irritancies were also observed during the four-week dosing and reversal phases, a NOAEL was not identified in either study. Irritation responses related to PF-04475270 were dose dependent and included eye squinting or blepharospasm, conjunctival hyperemia, nictitating membrane hyperemia, miosis, and occasional conjunctival chemosis. With the exception of conjunctival hyperemia, PF-04475270–related irritation responses were reversible or tending toward reversibility. The irritation responses that correlated with PF-04475270–induced histologic findings included peripheral corneal neovascularization, infiltrates of lymphocytes and plasma cells in the iris, and acute to subacute inflammation in the bulbar and palpebral conjunctiva and nictitating membrane. Other irritation responses including corneal edema, corneal opacity, aqueous flare, and ocular discharge were reversible and not associated with any histopathologic changes.

Topical instillation of PF-04475270 to the eyes of dogs induced varying degrees of ocular irritancy at doses ≥0.25 μg/eye. Red eye did not diminish with repeated dosing, increased in a dose-dependent manner, and persisted throughout the recovery period. Conjunctival hyperemia is a common side effect in patients with ocular hypertension given FP receptor agonists and prostaglandin analogues (Honrubia et al. 2009). Vasodilation and resultant increased blood supply to the conjunctival vessels may involve pharmacologic and/or inflammatory mechanisms that are separable. Investigation of the mechanism of conjunctival hyperemia related to treatment with FP receptor agonists and prostaglandin analogues in rabbits, dogs, and nonhuman primates has suggested a common signaling pathway that involves intracellular calcium and endothelial-derived, nitric oxide–mediated vasodilatation and was not associated with inflammation (Chen et al. 2005). The conjunctival hyperemia observed with PF-04475270 treatment in dogs was associated with minimal to mild acute to subacute inflammation in the bulbar conjunctiva and nictitating membranes, and the inflammatory response persisted during the treatment-free recovery period. In contrast, a dose-dependent conjunctival hyperemia observed in the eyes of Dutch-Belted rabbits treated topically with PF-04475270 for one month was not associated with conjunctival inflammation and resolved by day 7 during the thirty-day recovery phase (Pfizer Inc., data on file). These observations suggest that there may be a species-dependent sensitivity difference in the nature of the ocular response to PF-04475270. In addition to species differences, vasorelaxant properties and conjunctival inflammation are potential mechanistic explanations for conjunctival hyperemia.

Blepharospasm often occurred in dogs given doses ≥0.25 μg/eye and correlated with a mild to moderate bulbar and palpebral inflammation. Other ocular clinical findings including corneal edema, increased corneal thickness, and aqueous flare had resolved during the treatment-free recovery phase, with the exception of iridal hemorrhage and corneal neovascularization. The reversibility of the increased corneal thickness and corneal edema may have been a result of a transient corneal endothelial cell dysfunction, because the highest exposure level of the active drug moiety was found in the cornea (Prasanna et al. 2009).

Circum-limbal corneal neovascularization observed by slit-lamp examination of dogs given ≥1.0 μg/eye was confirmed histopathologically and suggested that topical instillation of PF-04475270 promoted angiogenesis in vivo. The mechanism by which corneal neovascularization occurred may involve both direct and indirect pharmacological effects. Others have suggested that the EP4 receptor plays a direct role in endothelial cell function. In vitro paradigms have suggested PGE₂ or the selective EP4 receptor agonist induced endothelial migration, tubulogenesis, ERK activation, and cAMP production in endothelial cells (Rao et al. 2007). The EP4 receptor agonist–induced endothelial cell migration was inhibited by an ERK inhibitor, thereby defining a functional link between PGE₂-induced migration and EP4-mediated signaling (Rao et al. 2007).

In vivo studies have suggested a role for the EP4 receptor in de novo vascularization. Neovascularization was observed in mouse corneas using an EP2 or EP4 agonist implants in a corneal micropocket assay (Kuwano et al. 2004) and in a subcutaneous sponge assay using EP4 agonist or PGE₂ implants (Rao et al. 2007). Because the technical procedures in these in vivo models incite an inflammatory response, a direct versus indirect role for the EP4 receptor in neovascularization could not be ascertained. Indirect inflammatory angiogenic effects could also be mediated through the release of angiogenic factors such as VEGF, IL-1β, bFGF, and/or matrix metalloproteases (Inoue et al. 2002; Pavlovic et al. 2006).

Topical administration of PF-04475270 in dogs induced single-cell necrosis of corneal epithelium, which in turn may have released neutrophil chemotactic factors into the tear film and recruitment of neutrophils via the tear film. Release of angiogenic factors from transmigrating neutrophils and/or injured corneal epithelial cells also likely induced the angiogenic process. The corneal edema and aqueous flare noted clinically could have occurred by transmigration of neutrophils from the conjunctiva to the corneal epithelium and into the corneal stroma, from where they could have further migrated into the anterior chamber. Alternatively, the neutrophils could have migrated from the limbal vessels directly to the anterior chamber and cornea.

Other factors such as release of matrix metalloproteases, bFGF, and the presence of an EP4 agonist PF-04475270 in the corneal stroma could have induced migration of endothelial cells into the corneal stroma from the nearby limbal vessels. The presence of hyperemic blood vessels and acute to subacute inflammation in the conjunctiva could have also delivered neutrophils and angiogenic factors from activated endothelial cells to the cornea via the tear film or blood stream.

Immunohistochemistry staining of eyes treated topically with PF-04475270 demonstrated cytoplasmic staining for EP4 and the endothelial biomarker Factor VIII-vWF in resident endothelial cells and in corneal neovascular endothelial cells. Positive Factor VIII-vWF immunostaining confirmed the presence of increased numbers of hypertrophied endothelial cells in the iris, suggesting that the active moiety of PF-04475270 reached the iris and exerted an effect either directly or indirectly on endothelial cell activation. Therefore, the presence of neovascular endothelial cells in the deep and superficial peripheral corneal stroma, clinical signs of aqueous flare, and hypertrophied endothelial cells in the iris suggested that PF-04475270 was likely converted to the active moiety in the corneal stroma and redistributed to the aqueous humor and/or iris and mediated a treatment effect.

The miosis in some dogs given ≥1.0 μg/eye of PF-04475270 was observed only following treatment with a mydriatic agent. This observation was consistent with the demonstrated selective and potent miotic effects of topically administered type F, and not type E, PGs in dog and cat eyes at doses sufficient to reduce IOP (Miranda and Bito 1989). The resistance to pupillary dilation could have been related to treatment effects in the anterior uvea of PF-04475270–treated eyes. One possible mechanism may involve a process that increases the outflow of aqueous humor. Increased outflow of aqueous humor may be mediated via direct stimulation of portions of the ciliary muscle that insert near the iridocorneal angle or via contraction of the ciliary muscle, which results in increased drainage via the trabecular meshwork or in larger trabecular spaces, respectively (Maggs, Miller, and Ofri 2008). Resistance to mydriatic sympathomimetics can also be associated with anterior uveitis (Maggs, Miller, and Ofri 2008). The subacute inflammatory process noted in the iris and ciliary body of dogs given PF-04775270 may likely explain this clinical observation. The finding of plasma cell and lymphocyte infiltrates in the iris and ciliary body also correlated with the incomplete pupillary dilatation observed during the recovery phase in a dog given 2.5 μg/eye.

The mechanism by which PF-04475270 lowered IOP is currently unknown, and the IOP decrease in dogs observed in this study was similar in magnitude to that previously reported using similar doses of PF-04475270 (Prasanna et al. 2009). Because PF-04475270 treatment lowered IOP in normotensive beagles, the selective prostanoid EP4-receptor agonist became a target of interest for glaucoma therapy. The predominant mechanism of action in lowering IOP for DP, EP2, and FP prostanoids involved increased uveoscleral outflow (Woodward et al. 2009). A report on another prototypical prostanoid EP4 agonist (3,7-diathiaPDGE1) suggested that the reduction in IOP likely involved trabecular outflow and not uveoscleral outflow or alteration of aqueous humor production (Woodward et al. 2009). In addition, the decrease in IOP induced by 3,7-diathiaPDE1 could not be explained by a marked decrease in episcleral venous pressure and episcleral vascular dilatation, because treatment was not associated with pronounced conjunctival hyperemia. Thus, the presence or absence of a pronounced conjunctival hyperemia associated with treatment with different EP4 agonists may be explained by differences in drug-to-tissue distribution. The contributory factors to the mechanism of IOP reduction by PF-04475270 may be multifactorial and may involve episcleral venous dilatation associated with decreased venous pressure, as well as with increased trabecular outflow.

In summary, this work described ocular changes in preclinical dog studies for an EP4-prostaglandin agonist and confirmed the lowering of IOP by such agents. This expected pharmacologic effect, as well as ocular irritancy responses, occurred in a dose-related manner, but the exact mechanisms by which the various changes were induced remains unknown. The effects observed with PF-04475270 may be owing to its selective EP4 agonist activity and/or an indirect induction of inflammatory angiogenic factors. The induction of corneal neovascularization in dogs after PF-04475270 treatment was demonstrated by slit-lamp microscopy, histopathology, and immunopositive staining with the EP4 and the inducible endothelial biomarker Factor VIII-vWF. Neovascularization can be induced by a variety of mechanisms that converge into a common pathway of corneal wound repair.