Plasminogen Activator Inhibitor-1 mRNA Is Localized in the Ciliary Epithelium of the Rodent Eye

Abstract

purpose. To identify in the adult and developing rodent eye cells expressing the gene encoding plasminogen activator inhibitor-1 (PAI-1), an important component of the fibrinolytic system.

methods. PAI-1 mRNA was localized in cryostat thin eye sections via in situ hybridization analysis using specific ³⁵S-labeled riboprobes. PAI-1 activity was tested in the aqueous humor using one-phase reverse zymography.

results. In the adult eye, PAI-1 mRNA was detected exclusively in epithelial cells of the ciliary processes, primarily in the apexes. In addition, PAI-1 activity was detected in the aqueous humor. PAI-1 mRNA was first found in the ciliary epithelium in embryonic day 18.5, when the ciliary body has reached an advanced developmental stage. PAI-1 mRNA was also detected in the ganglion cell layer of the retina at postnatal days 1 to 4, when angiogenesis takes place.

conclusions. During development, PAI-1 is likely to be involved in retina vascularization, in agreement with other cases of angiogenesis. Results for the adult eye indicate that the ciliary epithelium is the source for PAI-1 activity found in the aqueous humor. The results suggest that PAI-1 plays a role in balancing fibrinolysis and proteolysis specifically in the anterior segment of the eye, implying that PAI-1 overproduction in the ciliary epithelium could shift the balance against proteolysis and thus may interfere with aqueous outflow.

Plasminogen activator inhibitor-1 (PAI-1) is a secreted ∼50 kDa glycoprotein belonging to the serpin superfamily of serine protease inhibitors. PAI-1 is the principal physiological inhibitor of tissue-type (tPA) and urokinase-type (uPA) plasminogen activators (PAs). These are serine proteases that specifically convert the inactive proenzyme plasminogen into plasmin, the ultimate blood clot–dissolving enzyme. ^{1 2} Plasmin is a nonspecific trypsin-like protease that can also directly degrade diverse extracellular components and can activate proenzymes of matrix-degrading metalloproteases, thereby mediating extracellular proteolysis and cell adhesion. Plasmin can also extracellularly activate transforming growth factor-β (TGF-β) from its large latent form and can release growth factors bound to extracellular matrices. Apart from fibrinolysis, the PA/plasmin system has been implicated in various normal and pathologic events, including wound healing, angiogenesis, inflammation, restenosis, tumor metastases, and the maintenance of patency of tubular structures, such as the urinary tubules and vas deferens. ^{3 4 5 6}  

In humans, PAI-1 deficiency correlates with abnormal bleeding tendency, whereas elevated plasma PAI-1 has been associated with thrombotic diseases, with the acute phase response, and with non–insulin-dependent diabetes. ^{1 7} Increased PAI-1 synthesis has also been demonstrated in atherosclerotic plaques, suggesting that PAI-1 may play a role in atherogenesis as well, perhaps by promoting matrix deposition. ^{6 8} PAI-1 has also been implicated in PA-mediated non-fibrinolytic events, including angiogenesis, ⁹ cell migration, ⁵ inhibition of neointima formation, ⁶ and mediating uPA and tPA internalization/degradation through the multiligand surface receptor LRP/α₂MR. ⁵  

PAI-1 is normally produced in several species in vivo in various tissues and cell types. ^{2 10 11} The PAI-1 gene can respond to multiple biological modulators in vivo in mice or in cultured cells, usually at the transcriptional level. Inducers of PAI-1 gene expression include glucocorticoids, TGF-β, tumor necrosis factor-α (TNF-α), interleukin-1, the tumor promoter phorbal-myristate acetate (PMA), bacterial endotoxins, ^{2 12} and kainic acid, an analogue of the neurotransmitter glutamate. ¹¹ Conversely, PAI-1 gene expression has been shown to be reduced by interferons ¹³ and cAMP-elevating agents. ² Insulin is thought to be a major physiological regulator of PAI-1 in plasma. ⁷ However, the regulation of PAI-1 concentration in plasma is complex and not well understood. 

In the eye, tPA and uPA activities have previously been detected in different structures and fluids in several species, including human trabecular cells and meshwork ^{14 15} and aqueous humor. ^{16 17 18 19} In addition, tPA administration into the eye accelerates clot lysis, ¹⁶ introducing plasmin accelerated aqueous outflow, ²⁰ and reduced fibrinolytic activity was reported in the aqueous humor of patients with chronic simple glaucoma ²¹ and in chronic uveitis in the cat. ²² Accordingly, PAs were suggested to be involved in intraocular clot lysis, extracellular matrix dissolution, or aqueous outflow. ^{14 15 16 17 18 19 20 21 22}  

Less attention has been given to PAI-1 in the eye. So far, antigenic PAI-1 has been detected in the human aqueous humor; however, PAI-1 activity could not be detected. ^{17 18} To evaluate whether PAI-1 could play a role in the eye, we localized in this study PAI-1 gene expression in adult and developing murine eyes through in situ hybridization experiments and also tested PAI-1 activity in the aqueous humor. Most of our PAI-1 analysis has been conducted in the FVB/N mouse strain, because we previously used this strain to generate transgenic mice overproducing uPA in the ocular lens, carrying uPA cDNA linked to the αA-crystallin promoter. ²³ However, FVB/N mice exhibit retinal degeneration, resulting from an rd nonsense mutation in the gene encoding the β subunit of the rod photoreceptor cGMP phosphodiesterase. ²⁴ Therefore, to exclude effects of the rd mutation on PAI-1 expression, we also localized PAI-1 mRNA in the adult eye of nonmutated rodents. 

Methods

Animals

FVB/N mice (originally obtained from Dr. H. Westphal, the NIH), CD-1 mice (originally purchased from Charles River Labs, Wilmington, MA), and Wistar-derived rats were propagated and maintained at the Weizmann Institute Animal House. Experiments were conducted in accordance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. 

In Situ Hybridization

Adult female mice (3-months-old) were killed by deep anesthesia with ether, and eyes were immediately enucleated and frozen over dry ice. Newborn mice (day 4) were similarly killed, and their heads were excised and immediately frozen. Frozen samples were kept at −70°C for up to 2 weeks until sectioned. Thin cryostat sections (10- to 12-μm) were tested through in situ hybridization using ³⁵S-labeled riboprobes at the antisense and sense orientations, specific for mouse PAI-1 (derived from the 3.2 kb murine PAI-1 cDNA), as we described previously in detail. ^{11 25} Sections were kept under photographic emulsion for 12 to 14 days and lightly counterstained with cresyl violet or hematoxylin and eosin. 

One-Phase Reverse Zymography

Mice (3 months old) were killed as noted above, and the aqueous humor was extracted from both eyes (7–10 μl per eye) with a 29-gauge needle attached to a 1 ml syringe. PAI-1 activity was tested in one-phase reverse zymography as previously described. ^{11 26} Briefly, the indicated samples were electrophoresed in a sodium dodecyl sulfate (SDS)–polyacrylamide gel containing human plasminogen (purified from human plasma) and casein. The gel was then washed with 2.5% Triton X-100 to remove SDS, incubated for 4 hours at 37°C in a buffer containing standard human uPA (0.5 IU/ml; WHO International Laboratory, Holly Hill, Hampstead, London), and then stained with Coomassie brilliant blue. A parallel control gel was conducted without plasminogen and casein. This gel was incubated without any enzyme to distinguish between dark bands generated from inhibition of casein degradation and those resulting from just staining of sample proteins. 

Results

PAI-1 mRNA Localization in the Adult Rodent Eye

To identify cells producing PAI-1 in the adult eye, we tested cryostat thin eye sections from adult mice through in situ hybridization experiments using PAI-1–specific ³⁵S-labeled riboprobes at the antisense and sense orientations. Specific PAI-1 hybridization signals were detected exclusively in the ciliary body (Fig. 1A ). These signals were confined to the ciliary epithelium, but specifically to cells in the apexes of the ciliary processes (Figs. 1D 1E 1F) . PAI-1 hybridization signals appeared to be associated with both layers of the ciliary epithelium. The signals were seen with the antisense but not with the sense riboprobe (Fig. 1G) , thus indicating that they represented PAI-1 mRNA. These results were obtained with FVB/N, an inbred mouse strain often used to generate transgenic mice. Previously we derived from this strain transgenic mice overproducing uPA in the ocular lens ²³ ; therefore, we have also extensively studied PAI-1 in the parental mouse strain. However, FVB/N mice carry the rd mutation that leads to photoreceptor degeneration that can be detected by postnatal day 8. ²⁴ We therefore tested the FVB/N eye at younger postnatal ages and also localized PAI-1 mRNA in the ciliary body (Figs. 1B 1C 1K) . In addition, we confirmed the specific ciliary body localization of PAI-1 mRNA also for the nonmutated eyes of the CD-1 mouse (Figs. 1H 1I) and the Wistar rat (results not shown). 

PAI-1 mRNA Localization in the Developing Murine Eye

We also followed PAI-1 mRNA in the FVB/N eye throughout development from embryonic day 10.5 up to postnatal day 4. Hybridization signals in the ciliary body were first detected in embryos of 18.5 gestational days (Fig. 1J) , when the ciliary body shows evidence of differentiation, ²⁷ and signals substantially increased at postnatal day 4 (Fig. 1K) . As described above, hybridization signals remained high in the adult ciliary body, thus suggesting no interference of the rd mutation. The only other ocular tissue in which PAI-1 mRNA was detected was the retina, where signals were confined postnatally to the ganglion cell layer (Figs. 1C 1L) , which at that stage contains mainly postmitotic ganglion cells. ²⁷ Here, hybridization signals were first seen at postnatal day 1 and increased at postnatal day 4. We have not yet defined the specific cell type producing PAI-1 in the ganglion cell layer. 

PAI-1 Activity in the Aqueous Humor

To test whether PAI-1 mRNA in the ciliary body is translated into a biochemically active protein, we analyzed the aqueous humor of the adult eye for protease inhibitor activity by one-phase reverse zymography. ²⁶ This assay can visualize in crude biological samples activities of PA inhibitors (i.e., PAI-1 and plasminogen activator inhibitor-2 [PAI-2], a second specific albeit less potent PA inhibitor ³ ) and plasmin inhibitors. Thus, aqueous samples were electrophoresed in a SDS–polyacrylamide gel containing plasminogen and casein. To detect PAI-1, we coelectrophoresed conditioned media collected from murine and human hepatoma cell lines (Hepa lc17 and HepG2, respectively) previously shown to produce PAI-1 activity via this assay. ²⁶ After electrophoretic separation, the denaturing SDS was washed out to restore inhibitor activity, and the gel was incubated in the presence of uPA, which converted plasminogen throughout the gel to plasmin, which in turn degraded the casein. Subsequently the gel was stained with Coomassie brilliant blue. In this assay, activities inhibiting uPA or plasmin are seen as darkly stained bands on the lightly stained background of semidegraded casein. The results (Fig. 2A ) show that the aqueous samples contained a dark band corresponding to an ∼50 kDa protein comigrating with the Hepa PAI-1 when tested separately (lanes 3 and 5), or when mixed with the Hepa sample (lanes 4 and 6). We also conducted a control gel without plasminogen and casein to test whether the dark inhibitory band could represent just staining of sample proteins. No band corresponding to an ∼50 kDa protein could be seen in this control gel (Fig. 2B) . Notably, in the aqueous humor we did not detect activity of PAI-2 or ofα ₂-antiplasmin, the major plasmin inhibitor in blood. These two inhibitors have previously been detected in other biological samples via the one-phase reverse zymography, where they were clearly distinguished from PAI-1 by their sizes. ²⁶  

Reverse zymography containing plasminogen can also visualize proteolytic activities of PAs as bands lighter than the background. ²⁶ In the murine aqueous, a major ∼48 kDa uPA band and a minor ∼70 kDa tPA band were also seen (Fig. 2A , lanes 3 and 5). These two bands comigrated with known murine tPA and uPA and were not seen in the absence of plasminogen (data not shown), thus indicating that they represented PAs. 

Based on these results, we concluded that the murine aqueous humor contains activities of PAI-1, uPA, and tPA, with PAI-1 being the major inhibitor capable of balancing the PA/plasmin-mediated proteolysis. Notably, we also conducted in the aqueous humor the one-phase reverse zymography under conditions to detect trypsin inhibitors (i.e., plasminogen was omitted, and the gel was incubated with trypsin ²⁶ ). The results revealed several trypsin inhibitory bands that have not been further investigated (results not shown). 

Discussion

This study has shown that the PAI-1 gene is expressed in the ciliary epithelium and that PAI-1 activity is found in the aqueous humor of the murine eye (see Fig. 3 for a schematic represention). These findings, together with previous reports on uPA and tPA in the aqueous humor, ^{16 17 18 19} link the PA/plasmin system specifically to the anterior segment of the adult eye. Our results suggest that PAI-1 is involved in balancing fibrinolysis and proteolysis in the aqueous humor and imply that overproduction of PAI-1 in the ciliary epithelium will suppress proteolysis. Normal PAI-1 levels may be required to maintain a homeostatic state in the anterior segment with respect to the fibrinolytic capacity and the extracellular matrix composition. Increased PAI-1 levels in the aqueous humor could potentially interfere with clot lysis and with PA-mediated extracellular proteolysis in the aqueous channels. Such proteolysis could also involve activation of metalloprotease proenzymes previously detected in the aqueous humor and reported to be released from cultured ciliary smooth muscle cells. ^{28 29} Decreased extracellular proteolysis may result in the accumulation of excessive extracellular matrix, which if occurring in the trabecular meshwork could obstruct aqueous drainage and thus contribute to increasing the intraocular pressure (IOP). 

It is as yet unknown what the natural modulators of PAI-1 synthesis in the ciliary epithelium are, and whether ocular PAI-1 synthesis is enhanced in cases with elevated plasma PAI-1, such as the acute-phase response and non–insulin-dependent diabetes. ⁷ Two potent PAI-1 inducers are of particular interest in the eye context. The first is TGF-β, which was detected in the stroma of the human ciliary processes and in the aqueous humor, ^{30 31} suggesting that TGF-β may physiologically induce PAI-1 synthesis in the ciliary epithelium. The interrelation of PAI-1 and TGF-β in the ciliary body could be of interest, because PAI-1 can, in turn, exert a negative feedback on TGF-β production through inhibition of the plasmin-mediated activation of the large latent form of TGF-β. ⁴ The second inducer of interest is the synthetic steroid dexamethasone, which enhanced PAI-1 transcription in hepatic and endothelial cells. ² This anti-inflammatory drug can lead to IOP elevation, particularly in patients with primary open-angle glaucoma, which is the most frequent glaucoma form associated with elevated IOP. ³² Our results thus raise the possibility that dexamethasone-enhanced PAI-1 synthesis in the ciliary body could be among the factors contributing to dexamethasone-elevated IOP. 

Our finding that in the ciliary epithelium PAI-1 mRNA is confined to cells in the process apexes indicates for the first time that these cells are somewhat distinct from the rest of the epithelium. The nature and mechanism of this difference are not known. Interestingly, the ciliary body has recently been demonstrated to express a large group of genes, including genes encoding protease inhibitors. ³³ Collectively, these published data and the results presented here indicate the need to carefully balance proteolytic activities in the aqueous humor and the anterior segment of the eye. Notably, we did not detect PAI-1 mRNA in retinal pigment epithelial cells, which have been found previously to secrete PAI-1 when derived from the human eye and grown in culture. ¹² This discrepancy could reflect differences between the species and/or the states in vivo and in culture. 

In the present study, we also found that among developing eye tissues PAI-1 mRNA could be detected only in the ganglion cell layer of the retina at postnatal days 1 to 4 (we did not test later postnatal ages). This PAI-1 expression coincides with retinal angiogenesis ³⁴ and is in accordance with PAI-1 involvement in neovascularization in other physiological cases. ⁹ It is as yet unknown whether PAI-1 is involved in pathologic cases of ocular neovascularization such as after hypoxia, reperfusion, or diabetic retinopathy. It is of interest, however, that retinal microvessels from diabetic patients contain increased levels of PAI-1 mRNA as determined by solution hybridization. ³⁵  

In conclusion, our results indicate that PAI-1 is produced in the ciliary epithelium and found in the aqueous humor, suggesting that PAI-1 plays a role in balancing proteolysis specifically in the anterior segment of the eye. Our data raise the possibility that PAI-1 overproduction in the ciliary epithelium may inhibit proteolysis and thus could causally contribute to outflow obstruction, implying that suppression of PAI-1 synthesis could be a novel therapeutic approach to reduce the IOP. 

 Supported by the Dorot Science Fellowship Foundation administered by the Israeli Academy of Sciences and Humanities, by the Israel Ministry of Health, and by the Leo and Julia Forchheimer Center for Molecular Genetics at the Weizmann Institute of Science.

 Submitted for publication January 22, 1999; revised June 10 and November 1, 1999; accepted November 30, 1999.

 Commercial relationships policy: N.

 

Figure 1.

 

Localization of PAI-1 mRNA hybridization signals in adult and developing murine eyes. (A) A coronal section through the eye (excluding the lens) of a 3-month-old FVB/N mouse was hybridized with PAI-1 ³⁵S-labeled riboprobe at the antisense orientation. The photograph was taken in dark field illumination. (B, C) A section through the eye of a 4-day-old FVB/N mouse was hybridized as in (A), and photographed in bright field (B) or dark field (C) illumination. (D, E) The ciliary body of a section as in (A), but from a different eye, was photographed in bright field (D) or dark field (E) illumination. Note the association of hybridization signals specifically with the process apexes. (F) A higher magnification photograph of a ciliary process from a section analyzed as in (A). (G) The ciliary body in a section close to (F), but hybridized with PAI-1 ³⁵S-labeled sense riboprobe. (H, I) Ciliary body processes of 3-month-old CD-1 mice analyzed as in (A). (J) FVB/N ciliary body at embryonic day 18.5. (K) A higher magnification view of the left side ciliary body seen in (B). (L) A higher magnification view of the left side retina seen in (B). The cell indicated by the rod is magnified in the inset. Note hybridization signals in the ciliary body of the adult and developing eyes. No PAI-1–specific hybridization signals were seen in the lens or retinal pigment epithelium at any age in sections noted above and similar sections. C, cornea; R, retina; CB, ciliary body; CP, ciliary body processes; L, lens; GCL, ganglion cell layer. Arrows point to the ciliary body; arrowhead points to the ganglion cell layer of the retina. Scale bars, (A) 0.6 mm; (B, C) 0.7 mm; (D, E) 0.1 mm; (F, I, K) 11 to 13 μm; (G) 140 μm; (H, J, L) 30 μm.

Figure 2.

 

PAI-1 and PA activities in the aqueous humor. Samples were tested in one-phase reverse zymography as described in the text and the Methods section. Gel (A) included plasminogen and casein, and gel (B) did not include any protein substrate. The samples applied to the gels were as follows: lanes 1 and 2, samples (10 μl) of conditioned medium collected without serum from Hepa or HepG2 cells, respectively; lanes 3 and 5, undiluted aqueous humor samples from two different FVB/N mice (aq-1 and aq-2, 10 μl each in gel[ A]; 5 μl aq-1 in gel [B], lane 3); lanes 4 and 6, mixtures of aq-1 or aq-2 (5 μl each) with Hepa (5 μl). MW, Molecular weight; m., murine; h., human.

Figure 3.

 

Sites of PAI-1 synthesis and activity in the adult eye. Schematic representation of the anterior segment of the eye and aqueous inflow and outflow (according to ^{Ref. 32} ). Arrows indicate the sites of PAI-1 synthesis and secretion.

The authors thank Rene Abramovitz for conducting the one-phase reverse zymography.