Topical Prostaglandin F2α Treatment Reduces Collagen Types I, III, and IV in the Monkey Uveoscleral Outflow Pathway FREE

TOPICAL prostaglandin F_2α treatment reduces intraocular pressure (IOP) in monkey eyes by increasing uveoscleral outflow.^{1 - 3} Likewise, IOP reduction in human eyes following topical treatment with latanoprost is associated with increased uveoscleral outflow.^{4 - 6}

Although the biologic basis for these changes is poorly understood, tracer studies have shown that aqueous humor flow within the uveoscleral pathway is through collagen-containing spaces between the ciliary muscle fiber bundles.^{7 - 8} Moreover, histologic examination of ciliary muscle in monkey eyes that had received repeated topical application of prostaglandin F_2α has revealed enlarged interstitial spaces among monkey ciliary muscle fiber bundles.^{9 - 10} It has been postulated that this may explain the washout of extracellular matrix from these spaces.^{9 - 10}

Subsequently, it was observed that introducing prostaglandins into the medium of human ciliary smooth muscle cultures reduced collagen types I and III—constituents of the stromal extracellular matrix that fill the spaces between the ciliary muscle fiber bundles—adjacent to the cells.^{11 - 14} Furthermore, we observed that the medium of similarly treated ciliary muscle cell cultures contained increased amounts of collagen-degrading matrix metalloproteinases (MMPs).^{15 - 16} Based on these observations, we hypothesized that reduced hydraulic resistance induced by MMP-mediated degradation of collagen types I and III in ciliary muscle may contribute to increased aqueous flow through the interstitial spaces of the ciliary muscle following topical prostaglandin treatment.¹⁷

To assess the validity of this hypothesis in situ, this study measured the effect of repeated topical prostaglandin F_2α isopropyl ester treatment on collagen types I, III, and IV within the monkey uveoscleral outflow pathway.

Prostaglandin F_2α isopropyl ester, which converts to prostaglandin F_2α as it passes through the cornea, was chosen for this study because it reliably reduces IOP in monkeys.^{1 - 2 ,9 ,18} Young adult cynomolgus monkeys (Macaca fascicularis) were studied according to established protocols at the University of Wisconsin, Madison. Measurement of IOP¹⁹ (prone position, eyes 4 cm above the heart) and slitlamp biomicroscopy were performed bilaterally under intramuscular ketamine hydrochloride anesthesia, 10 mg/kg on 2 occasions during the week before initiation of treatment. The IOPs ranged from 12 to 19 mm Hg. Eyes were biomicroscopically quiet. After the second IOP evaluation, blood-aqueous barrier integrity was confirmed by measuring the appearance and disappearance of fluorescence in the anterior chamber following intravenous fluorescein administration. For this purpose, intravenous fluorescein, 10 mg/kg of 10% solution, was administered via the saphenous vein under intramuscular ketamine hydrochloride anesthesia, 10 mg/kg. After washing with 2 mL of lactated Ringer solution, corneal and anterior chamber fluorescence was measured (Fluorotron Master Scanning Ocular Fluorophotometer equipped with an anterior segment adapter; Coherent Medical, Palo Alto, Calif) at 20, 40, 60, 90, 120, 180, and 240 minutes after injection. To qualify for further study, anterior chamber fluorescence levels and course of appearance and decay in both eyes had to be similar and within the range of values obtained for control eyes in a previous study.³ Monkeys who met these conditions were presumed to have an intact blood-aqueous barrier.³

The following week, each qualifying monkey received 2 µg of prostaglandin F_2α isopropyl ester (in 5 µL physiological saline) twice daily (morning and afternoon, approximately 7 hours apart) in one eye and 5 µL of vehicle in the other eye for 5 days. On days 1 through 3 and on the mornings of days 4 and 5, drops were administered to conscious monkeys that were restrained to maintain their bodies in a vertical position. The head was tilted so that the eye was looking up, and the drop was administered to the central cornea. The eyelid was held open during drop administration and for the next 30 to 60 seconds. After IOP measurements were taken on day 4, the afternoon drop was administered to the already ketamine-anesthetized monkey that was in a supine position with the eye looking up; drops and eyelid holding were as previously described.³ On the fourth day of treatment, IOP was measured at 3, 312, and 4 hours after the morning treatment. On the fifth day of treatment, slitlamp biomicroscopy was performed and IOP was measured at similar times. Two monkeys with less than a 5-mm Hg IOP decrease in the treated eyes (compared with pretreatment ipsilateral baseline and simultaneous contralateral control) were excluded from further study. Eight eyes of 4 monkeys (code numbers, K91, K344, K453, and K454) who met the criteria were obtained. Anterior chamber cells or flare were not observed during treatment of these monkeys. Measurements of IOP before and after treatment were compared using the Student t test. Differences were considered significant at P<.05.

The animals were killed following the IOP reading on day 5. After administration of an intravenous pentobarbital sodium overdose,²⁰ blood was removed by whole-body perfusion through the left ventricle with approximately 500 mL of lactated Ringer solution and 5% dextrose solution. The eyes were enucleated and hemisected through an equatorial plane. The anterior segments were immediately fixed in methacarn (60% methanol, 30% chloroform, and 10% glacial acetic acid) for 3 hours. Increased sensitivity of immunohistochemical staining of various antibodies has been demonstrated for many antigens after methacarn fixation.^{21 - 22} Fixed anterior segments were transferred to 100% cold ethanol and shipped overnight to the University of California, San Diego. The tissue specimens were embedded in paraffin, and sections were collected from the midsagittal region of each eye on slides precoated with a silane-based adhesive (Vectabond; Vector Laboratories, Burlingame, Calif). For histopathologic analysis, 3 to 4 sections from each eye were stained with hematoxylin-eosin. All procedures were conducted in accordance with the Association for Research in Vision and Ophthalmology Statement on the Use of Animals in Ophthalmic and Vision Research.

Anterior segment tissues were immunostained by a standardized protocol. Each step of the protocol was optimized as previously described.²³ Briefly, concentrations and incubation times of solutions used to remove intrinsic melanin were optimized to eliminate melanin while preserving the structural integrity of the tissue. Also, the concentration of each solution containing antibodies or horseradish peroxidase–conjugated streptavidin was optimized to obtain submaximal (nonsaturating) staining intensity, as determined by imaging densitometry (described herein). Finally, the incubation time with diaminobenzidine was optimized for each primary antibody to obtain the strongest signal (staining intensity) that was still increasing linearly with time. The elimination of saturating binding or development parameters from the protocol support the position that the observed changes in staining intensity reflected differences in tissue content of the target antigen.

Sections from the treated and control eyes were stained at the same time. Five sections, 10-µm thick, from each eye were heated to 56°C for 20 minutes, washed in 3 xylene changes to remove paraffin, and rehydrated through graded ethanols. The sections were treated with antigen retrieval solution (AR-10; Biogenex Laboratories, San Ramon, Calif) at 95°C for 5 minutes. After cooling, the sections were exposed to 3% hydrogen peroxide for 10 minutes to suppress endogenous peroxidase activity. To remove intrinsic melanin, sections were treated successively with aqueous potassium permanganate, 2.5 g/L, for 10 minutes, and oxalic acid, 5 g/L, for 3 minutes.^{24 - 26} After rinsing, the sections were blocked for 30 minutes with 0.1% bovine serum albumin (Sigma-Aldrich Corp, St Louis, Mo) and incubated overnight at 4°C with affinity-purified polyclonal rabbit antihuman collagen type I (dilution 1:50, T61547R; Biodesign International, Kennebunk, Me), monoclonal antimouse antibody to human collagen type III (dilution 1:60, HWD1.1; Biogenex Laboratories), or polyclonal antirabbit antibody to human collagen type IV (dilution 1:10, CIV22; Biogenex). Specificity of these antibodies has been previously confirmed.^{27 - 29} After rinsing, the sections treated with antibodies to collagen types I and IV were exposed to biotinylated goat antirabbit immunoglobulin for 20 minutes. The sections treated with the antibody to collagen type III were exposed to biotinylated goat antimouse immunoglobulin for 20 minutes. The sections were then rinsed and exposed to horseradish peroxidase–conjugated streptavidin for 20 minutes. Consecutively, each section was rinsed and incubated with 3,3′-diaminobenzidine chromogen for 10 minutes (HRP-DAB Super Sensitive Immunodetection System; Biogenex Laboratories). To facilitate comparability, sections from the control vehicle-treated and prostaglandin-treated eye of each monkey were immunostained at the same time.^{30 - 31} To serve as controls for nonspecific staining, sections from each eye were simultaneously processed by the same protocol but without the primary antibody.

Immunohistochemical staining intensity was directly measured using a high-resolution imaging densitometer as previously described.²³ Measurements from multiple sections stained at the same time facilitated assessment of measurement precision and permitted statistical comparison of differences among control and experimental eyes. Immunostained sections were scanned by placing the slides directly on the platen of an imaging densitometer (model GS-700; Bio-Rad, Hercules, Calif). Resolution of the scans was set to 1200 dots per inch (50-µm-wide pixels), and the scanning mode was set to transillumination. Optical density (OD) measurements of the immunostained sections were all less than 1.00 OD U. Because the densitometer can accurately measure ODs greater than 3.0 units (Bio-Rad specifications), these measurements were well within the appropriate range for accurate determinations. The scanned digital data were displayed in a masked fashion and analyzed using an image analysis program (Molecular Analyst, version 2.1; Bio-Rad). The OD along 2 line segments positioned perpendicular to the long axis of the ciliary body and near the widest region of the ciliary muscle was measured in each section using 2-dimensional imaging densitometry. The positioning of these line segments over the anterior segment tissues avoided any remaining cluster of pigment granules. A similar line segment was positioned perpendicular to the long axes of the iris root and the sclera adjacent to the ciliary body.

Mean OD scores along each line segment were determined.²³ For each eye, 10 scores were obtained from 5 midsagittal sections. Background OD was measured along one line segment nearby but not overlying the tissue. This OD score was considered baseline and subtracted from the original OD scores. Ciliary and iridial pigmented epithelial cells, which showed a high peak on an image densitometry map, were excluded. The specific OD scores along each line segment over the ciliary muscle, iris, and adjacent sclera were calculated by dividing the OD area score (OD × millimeters, by densitometer) by the length of the line segment (in millimeters) for that score. Mean specific OD scores from the prostaglandin-treated eye of each monkey were compared with corresponding scores from the contralateral control eyes using the paired Student t test. The unpaired Student t test was used to compare the mean of all mean OD scores of the prostaglandin-treated and control eyes. In each case, P< .05 was considered significant.

Inflammatory cells were graded in hematoxylin-eosin–stained sections from all experimental eyes on a 6-point scale as follows: 0, occasional nonclustered inflammatory cells, including lymphocytes and macrophages; 1, occasional clusters of inflammatory cells (<5 cells per cluster); 2, small clusters of inflammatory cells (5-10 cells per cluster); 3, large clusters of inflammatory cells (>10 cells per cluster); 4, large clusters of inflammatory cells with minor tissue reorganization; and 5, large clusters of inflammatory cells with major tissue reorganization. Results are given as mean ± SD.

The IOP measurements are shown in Table 1. Before treatment, IOP ranged from 12.0 to 16.7 mm Hg, with a mean of 14.3 ± 1.1 mm Hg (baseline 1 plus baseline 2 measurements). The difference between measurements taken for baseline 1 and baseline 2 were also insignificant (14.1 ± 0.7 and 14.4 ± 1.5 mm Hg, respectively; P=.59). When measurements from baselines 1 and 2 were combined, the starting mean IOP in the treated eyes was 14.7±1.1 mm Hg and in the control eyes was 13.8 ± 0.6 mm Hg. This difference was insignificant (P=.21). Also, IOP measurements from day 4 were insignificantly different from those taken on day 5 in both the treated eyes and the control eyes (P=.60 and .16, respectively); this suggests that the treatment effect had stabilized by day 4. However, comparing the mean of IOP measurements collected for baselines 1 and 2 with those collected on days 4 and 5 revealed that topical treatment with prostaglandin F_2α isopropyl ester reduced IOP by 9.0 ± 2.2 mm Hg (P<.001). In contrast, there was no significant IOP change in the contralateral eyes exposed to vehicle control (P=.29).

Collagen type I immunohistochemistry showed a staining of collagen fibrils within the interstitial spaces separated by bundles of ciliary smooth muscle cells (Figure 1, A). Collagen type III immunoreactivity was also observed within these spaces (Figure 1, B). Staining of both collagens was also observed in the stromas of the ciliary processes, iris, sclera, and cornea. This was similar to previous reports in cynomolgus monkey and human eyes.^{12 - 14} At higher magnification, immunostained collagen type I fibril bundles were seen throughout the ciliary muscle interstitial matrix (Figure 2, A). Collagen type III also stained fibril bundles; however, the fibrils appeared shorter in length than collagen type I fibrils (Figure 2, D). The intensity of collagen type I and collagen type III immunoreactivity within the ciliary muscle of the treated eye was less than in contralateral vehicle-treated eyes (Figure 2, B and E). Negligible staining was observed in the sections that were processed according to the same immunostaining protocols, except that the primary antibody step was omitted (Figure 2, C and F).

Collagen type IV immunoreactivity was observed in the basement membranes of ciliary muscle, ciliary processes, iris, and blood vessels in adjacent sclera (Figure 1, C). High-magnification examination of ciliary muscle showed that staining was restricted to the basement membranes of smooth muscle cell bundles (Figure 2, G). The intensity of collagen type IV immunostaining within the ciliary muscle was visibly less in prostaglandin-treated eyes than in vehicle-treated eyes (Figure 2, H). Negligible staining was observed in sections that were processed according to the same immunostaining protocol for collagen type IV, except that the primary antibody step was omitted (Figure 2, I).

The structural organization of the ciliary muscle and adjacent tissues was readily recognized in images from the imaging densitometer (Figure 3). This facilitated the placement of analysis lines over the ciliary muscle and precluded inadvertent scoring of the ciliary epithelium (Figure 3, B).

Collagen type I immunoreactivity in the ciliary muscle was reduced in the prostaglandin-treated eyes of each of the 4 monkeys (Figure 4). In addition to ciliary muscle, OD measurements were also collected in the iris root and sclera adjacent to the ciliary muscle. In each case, 2 analysis lines were positioned to traverse the long axis of these structures in the sections. As shown in Figure 5, means of the collagen type I immunoreactivity measurements in all 3 tissues were less in prostaglandin-treated eyes than in vehicle-treated eyes: 52% ± 7% (P<.001), 43% ± 4% (P=.01), and 43% ± 32% (P=.04) in the ciliary muscle, iris root, and sclera, respectively. Similarly, collagen type III immunoreactivity was significantly less in all 4 prostaglandin-treated monkey eyes. As shown in Figure 6, reductions of collagen type III were 45% ± 7% (P<.001), 38% ± 30% (P=.008), and 45% ± 13% (P=.02) in the ciliary muscle, iris root, and sclera, respectively. All 4 prostaglandin-treated monkey eyes showed less collagen type IV immunoreactivity in the ciliary muscle and iris root. As shown in Figure 7, mean ODs in the ciliary muscle and iris root were reduced by 34% ± 11% (P=.002) and 45% ± 6% (P=.003), respectively. As mentioned herein, the scleral stroma was minimally immunoreactive for collagen type IV. Because baseline measurements were collected from image areas without tissue, the small ODs measured in collagen type IV–immunolabeled sections most likely reflect the inherent OD of unstained tissue. The difference between these measurements in sclera was insignificant (P=.40).

Occasional inflammatory cells, mostly lymphocytes, were observed in one or both eyes of 3 of the 4 monkeys, and were all scored +1 (Table 2). In monkey K344, one loose cluster containing about 18 lymphocytes was observed in the stroma between the ciliary muscle head and adjacent ciliary epithelium in the treated eye. There was no tissue reorganization present in this eye, hence the assignment of a +3 score. Comparison of mean collagen type I reduction in K344 (53%) with the other monkeys showed that it was less than K454 (65%), in which the inflammation score in the treated eye was +1 (Figure 4).

These results support the hypothesis that topical prostaglandin treatment reduces collagen types I, III, and IV in the interstitial extracellular matrix of tissues comprising the monkey uveoscleral outflow pathway, including the ciliary muscle and adjacent sclera. These results are consistent with previous in vitro experiments^{11 ,17} showing that the introduction of prostaglandin F_2α into ciliary smooth muscle cultures reduced collagen types I and III within the cell layer.

The characteristic staining patterns for each collagen type (I, III, and IV) in this study are similar to previous reports^{12 - 14} of the distribution of these collagens in anterior segment tissues of monkey and human eyes. This supports the specificity of antibodies used in this densitometric analysis and validity of quantitative findings. This study found a reduction of collagen type IV in the ciliary muscle with prostaglandin treatment, whereas a previous in vitro study¹¹ found increased collagen type IV. This may be the result of differences between the in vivo and in vitro systems. Collagen type IV deposition relies on the accumulation of a sufficient concentration of activated collagen type IV molecules.³² In vitro, prostaglandin induction of collagen type IV biosynthesis lead to an accumulation of this molecule in the medium during the 3-day treatment phase. In vivo, however, the extracellular spaces of the ciliary muscle are constantly washed by aqueous humor, which leaves the eye via the uveoscleral route. Such aqueous flow may have a diluting effect that prevents newly synthesized collagen type IV from attaining the critical concentration necessary for incorporation into the basement membranes.

The present results do not address the mechanism by which collagen types I, III, and IV are reduced in the ciliary muscle following topical prostaglandin treatment. Previously, it was demonstrated that introduction of prostaglandin F_2α or its analogs into a culture of human ciliary smooth muscle cells increased the secretion of MMPs,^{15 - 16} including MMP-1 (interstitial collagenase), MMP-2 (72 kd gelatinase), MMP-3 (stromelysin-1), and MMP-9 (92 kd gelatinase). These MMPs cleave specific peptide sequences found in collagens and other extracellular matrix molecules.^{33 - 34} In addition, immunohistochemical studies have recently confirmed the presence of MMP-1 in ciliary muscle in vivo.³⁵ Thus, results showing reduction of collagen types I, III, and IV in the ciliary muscle following topical prostaglandin treatment may reflect the induction of MMP production by ciliary muscle cells in situ.

Three of the 4 monkeys had minimal signs of inflammation. In the fourth monkey, clusters of inflammatory cells were observed, although there was no tissue rearrangement. As seen in Figure 4, the magnitude of collagen reduction observed in this monkey (K344) was less than in another monkey (K454) that had minimal signs of inflammation in the treated eye (Table 2). Moreover, each prostaglandin-treated monkey eye had a substantial reduction of collagen types I, III, and IV when compared with their vehicle-treated controls. This suggests that collagen reduction with IOP-lowering topical prostaglandin treatment is not dependent on the presence of an inflammatory response.

Recently, Ocklind³⁶ reported that topical treatment of cynomolgus monkeys with 3 or 10 mg of latanoprost daily for 10 days reduced collagen type IV immunoreactivity in histologic sections of the ciliary muscle. This result is consistent with the reduction of collagen type IV with prostaglandin F_2α isopropyl ester observed in the present study. Unlike the present study, however, changes in immunoreactivity for either collagen type I or collagen type III were not found by Ocklind. Several factors may account for this difference. First, the active form of latanoprost is a relatively specific agonist for the FP receptor, whereas prostaglandin F_2α activates both the FP and the EP2 receptors.³⁷ Hence, the difference might reflect the additional contribution of EP2 receptor activation. Supporting this possibility, human ciliary smooth muscle cells exposed to equal concentrations of prostaglandin F_2α or latanoprost-free acid are induced to release more plasmin-generating activity with prostaglandin F_2α than with latanoprost.³⁶ A second factor that may also contribute to the differing results is that a subjective semiquantitative scoring method was used to evaluate the intensity of immunostaining in the latanoprost study, whereas direct measurement of staining intensity by imaging densitometry was used in the present study. Both studies agree that the intensity of staining for collagen types I and III is much less than for collagen type IV. Hence, it is possible that subtle differences in staining for collagen types I and III may be more readily identified using the densitometric method. Direct evaluation of immunostained sections from latanoprost-treated eyes using the densitometric method would address this possibility.

In addition to ciliary muscle, the present study reports prostaglandin-mediated collagen reductions in both the sclera adjacent to the ciliary body and the iris root. We recently found that normal human eyes contain MMP-1 immunoreactivity in the ciliary body, adjacent sclera, and iris root stroma.²³ This observation may be of particular interest because it has been shown previously³⁸ that tracers introduced into the anterior chamber were also found within the extracellular matrix adjacent to blood vessels and nerves as they penetrate the sclera. Like these tracers, MMPs released by ciliary muscle cells into the uveoscleral outflow pathway might be carried from the muscle to the sclera. Because a significant portion of aqueous humor proteins appear to arise from the ciliary body capillary bed and flow forward through the iris root,^{39 - 40} it is possible that MMPs arising from ciliary muscle cells may exert their effect in part within the iris root. Alternatively, local biosynthesis of MMPs may contribute to prostaglandin-mediated reduction of collagens in the iris and sclera. Further studies will be needed to evaluate these possibilities.

Accepted for publication February 23, 1999.

This study was supported in part by grants EY05990 (Dr Weinreb) and EY02698 (Dr Kaufman) from the National Eye Institute, Bethesda, Md; Foundation for Eye Research, San Diego, Calif (Dr Gaton); and unrestricted grants from Research to Prevent Blindness Inc, New York, NY. Drs Weinreb and Kaufman are Senior Scientific Investigators of Research to Prevent Blindness.

Corresponding author: Robert N. Weinreb, MD, University of California San Diego Glaucoma Center, 9500 Gilman Dr, La Jolla, CA 92093-0946 (e-mail: weinreb@eyecenter.ucsd.edu).