Plasminogen Activators and Inhibitors in the Corneas of Mice Infected with Pseudomonas aeruginosa

Abstract

purpose. To characterize the presence of plasminogen activators and their inhibitors in the corneas during the inflammatory response in naïve and immunized mice intracorneally infected with Pseudomonas aeruginosa.

methods. RT-PCR was used to detect gene expression for plasminogen activators and their inhibitors in naïve and immunized mice. Immunoblot analysis, zymography, and ELISA were used to demonstrate the syntheses of these proteins.

results. Naïve mice intracorneally infected with P. aeruginosa showed a temporally enhanced expression of tissue-type plasminogen activator (tPA), urokinase-type plasminogen activator (uPA), its receptor (uPAR), and plasminogen activator inhibitors 1 and 2 (PAI-1 and PAI-2), over a several-day holding period. Immunized mice demonstrated a lower and shorter expression of these factors over the same period. Expression of these factors at the mRNA and protein levels may have been due to enzymes and inhibitors present in inflammatory cells and in resident corneal cells.

conclusions. These results show a correlation between the overexpression of the PA system in infected naïve mice as part of the inflammatory response, with eventual ocular destruction. Immunized mice exhibit a more balanced and shorter expression of the PA system, which may contribute to the restoration of corneal clarity.

The pathogen Pseudomonas aeruginosa is an opportunistic bacterium that causes severe corneal infection in humans. Many of the clinical features of the infection can be reproduced in a variety of experimental animals. ^{1 2 3 4} Previous studies from our laboratory have demonstrated that corneal clarity is not restored in naïve (susceptible) C57BL/6J mice after infection with P. aeruginosa strain 19660 (American Type Culture Collection, Rockville, MD [ATCC]) and that corneal perforation, shrinkage, or both eventually occur. ⁵ However, either active or passive immunization results in restoration of corneal clarity within a few days to a few weeks, despite a similar intensity of the initial infection in both naïve and immunized mice. ^{6 7 8} Peak levels of both polymorphonuclear leukocytes (PMNs) and bacteria are detected at approximately 5 to 7 days after infection, along with arachidonic acid metabolites and cytokines, and then begin to decrease. ^{8 9 10 11 12} An increase in macrophages also occurs, but peaks at a later stage in the inflammatory process. ^{9 13} Viable bacteria are no longer detectable in naïve mice by quantitative plate counts at 9 to 12 days after infection, whereas the infection is cleared sooner in immunized mice. ⁸ Other studies from our laboratory indicate that the expression of both bacterial alkaline protease and host metalloproteinases temporally correlates with the inflammatory response and the degree of immunity. ^{14 15}  

Resident and infiltrating cells within the cornea during infection have the ability to synthesize many proteases and their inhibitors, which may play a role in various physiological and ocular processes. One such proteolytic system is the plasminogen activator (PA) cascade, consisting of the serine proteases urokinase-type plasminogen activator (uPA) and tissue-type plasminogen activator (tPA). Both enzymes are found in either the single- or double-chain configuration, depending on their degree of activation, and they can convert the zymogen plasminogen to the active form, plasmin. ^{16 17} Once activated, plasmin has the ability to cleave many extracellular matrix (ECM) proteins such as laminin, fibronectin, and nonfibrillar collagens, as well as to activate certain matrix metalloproteinases (MMPs), which in concert with plasmin, can lead to stromal destruction. ^{18 19 20} Concomitantly, enzymatic expressions of tPA and uPA are putatively controlled by several serine protease inhibitors (serpins), including plasminogen activator inhibitors (PAI)-1 and -2. At present, the PA cascade has been implicated in neoplasia, neuronal plasticity, and ovulation, among other processes. ^{21 22 23} Of special interest to this study is that the serpins also play a role in inflammation, angiogenesis, and wound repair. ^{21 23} Consequently, the purpose of the present study was to characterize the corneal expressions of uPA, the receptor of uPA (uPAR), tPA, and both PAIs during infection of naïve mice and mice immunized with P. aeruginosa. 

Methods

Bacteria

Stock cultures of P. aeruginosa strain 19660 (ATTC) were stored at 4°C on tryptose agar slants (Difco Laboratories, Detroit, MI) and were used for the inoculation of 50 to 75 ml broth medium containing 5% peptone (Difco Laboratories) and 0.25% trypticase soy broth (BBL Microbiology Systems, Cockeysville, MD). Strain 19660 is hemolytic and lecithinolytic and produces exotoxin A, alkaline protease, and elastase under appropriate culture conditions. ¹⁰ Cultures were grown on a rotary shaker at 37°C for 16 to 18 hours, centrifuged at 8000 rpm at 4°C for 10 minutes, washed with normal saline (Baxter Travenol Laboratories, Deerfield, IL) and diluted to a concentration of 2 × 10¹⁰ CFU per milliliter. A standard curve was developed to relate viable counts to optical density at 440 nm. 

Infection of Animals

All animals were treated in accordance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. Age-matched naïve and immunized female C57BL/6J mice (Jackson Laboratories, Bar Harbor, ME), each weighing 18 to 22 g, were infected at 14 weeks of age, using a previously described method. ¹³ Naïve C57BL/6J have previously been classified as susceptible because corneal clarity is not restored within 30 days after infection, whereas immunized mice are considered resistant, because in most, corneal clarity is restored within a few days to a few weeks. ^{6 12} Immunization began at 6 weeks of age by administering 0.1 ml 10⁶ to 10⁷ heat-killed P. aeruginosa 19660 intraperitoneally weekly for 4 weeks. Animals were then rested for 4 weeks before corneal infection. 

Corneal Sample Collection and Processing

At selected time points after infection, mice were killed and corneas were excised. Immediately after isolation, corneas were routinely rinsed in sterile saline and then processed for the purposes of the different assays. Individual samples consisted of 12 pooled corneas per period. Uninfected mice (day 0) were treated similarly. Samples were collected at five time points for each assay, a total of 5 samples (including that from day 0) from both naïve and immunized mice were used in these studies after extensive pre-titrations. 

Reverse Transcription–Polymerase Chain Reaction

The pooled corneal samples were homogenized with TRIzol (1 ml; Gibco, Grand Island, NY) in a mortar (Coors Porcelain, Golden, CO) and incubated at room temperature for 5 minutes. Two hundred microliters chloroform was added to the extract, and the mixture was vortexed vigorously. The extract was centrifuged at 13,000 rpm at 4°C for 15 minutes. The aqueous phase (containing total RNA) was transferred to a new centrifuge tube. Six hundred microliters isopropyl alcohol was added and mixed with the aqueous phase. The mixture was incubated at room temperature for 15 minutes and centrifuged at 13,000 rpm at 4°C for 10 minutes. The supernatant was removed, and the total RNA precipitate was washed once with 75% ethanol and saved (in 75% ethanol at −20°C) for RT-PCR analysis. 

The total RNA was dissolved in water treated with diethyl pyrocarbonate (DEPC), and the concentration was measured by a spectrophotometer at 260 nm. The total RNA samples were treated with DNase I (0.2 U/μl; Ambion, Austin, TX) to remove possible DNA contamination. All the reagents needed for RT-PCR were purchased from Perkin Elmer (Norwalk, CT). Five hundred nanograms extracted total RNA from each sample was used as the standard amount for each assay. RT-PCR was performed sequentially in the same 0.65-ml RNase-free tubes under optimized conditions, as previously described. ¹³ Thirty cycles were selected for amplification of all target genes. The following specific primers for mouse uPA, uPAR, tPA, and PAIs were designed and prepared based on the available information for these mouse genes (GenBank, National Center for Biotechnology, Bethesda, MD; available in the public domain at http://www.ncbi.nlm.nih.gov). The primers, both forward and reverse, from 5′ to 3′ were as follows: mouse uPA: GCC CAC AGA CCT GAT GCT AT and TAG AGC CTT CTG GCC ACA CT; mouse uPAR: AGG TGG TGA CAA GAG GCT GT and AGC TCT GGT CCA AAG AGG TG; mouse tPA: TAC AGA GCG ACC TGC AGA GA and AAT ACA GGG CCT GCT GAC AC; mouse PAI-1: GCT GTA GAC GAG CTG ACA CG and ACG TCA TAC TCG AGC CCA TC; and mouse PAI-2: CAC CAC AGG GGG ATT ATT TG and TGG GAT TTC ACC TTT GGT TT. Finally, the amplified specific genes were revealed by electrophoresis on 1% agarose gels. A housekeeping gene (18S rRNA) was also amplified and used as an internal control for the comparison of all time samples. 

Immunoblot Analysis

Corneal samples were homogenized as described by Brown et al. ²⁴ After homogenization in 200 μl of 50 mM Tris-HCl (pH 7.4) containing 10 mM CaCl₂ and 0.25% Triton X-100, the samples were centrifuged at 9000 rpm at 4°C for 30 minutes. The concentrations of the total protein were measured with the BCA protein assay. ²⁵ Equal amounts of individual samples (5 μg protein) were mixed with 5 μl 4× sample loading buffer (0.125 M Tris-HCl [pH 6.8], 4% SDS, 40% glycerol, and 0.02% bromphenol blue), containing 0.1% β-mercaptoethanol, and boiled for 5 minutes. The samples and a mixture of prestained molecular weight markers (Bio-Rad, Richmond, CA) were electrophoresed on 12% SDS gels and subsequently transferred to nitrocellulose membranes. 

A fusion protein was used under reducing conditions as a marker for PAI-2 (gift of David Ginsburg, University of Michigan, Ann Arbor). The membranes were blocked for 30 minutes in blocking reagent (TBS containing 0.5% Tween 20, 3% nonfat milk, and 2% bovine serum albumin; Blotto, Santa Cruz Biotechnology, Santa Cruz, CA) and then incubated with the following specific primary antibodies: a polyclonal antimurine PAI-2 antibody (2.2 μg/ml, a generous gift from Dr. Ginsburg) and a polyclonal anti-human uPAR antibody (3 μg/ml; American Diagnostic) on a rocker at room temperature for 2 hours. The antibody to human uPAR also recognizes mouse uPAR. Parallel immunoblots without primary antibody treatment were processed as negative controls. Afterward, the blots were incubated with secondary antibodies conjugated with horseradish peroxidase (0.5 μg/ml; Boehringer-Mannheim, Indianapolis, IN) at room temperature for 1 hour. Finally, the blots of PAI-2 and uPAR were developed by a chemiluminescence kit (Amersham Pharmacia Biotech, Parsippany, NJ) and were visualized as dark bands. 

Casein-Agarose Plasminogen Zymography

Zymography was performed to detect the plasminogen activators uPA and tPA in corneas from both naïve and immunized mice harvested at several time points after infection up to 14 days. Corneal extracts were prepared according to Brown et al. ²⁴ and run on a 7.5% nonreducing SDS-PAGE discontinuous gel system. The plasminogen used in the zymography experiments had been previously treated with a 10-mM excess of the irreversible serine protease inhibitor Phe-Phe-Arg-chloromethylketone, which was subsequently removed by extensive dialysis. This procedure ensured that all plasmin activity would be due solely to the plasminogen activators present in the acrylamide gel overlay. The tPA stimulator was added to enhance the tPA activity of the samples, which is in general much less effective than urokinase in activating plasminogen in this system. The gels were then placed in a humid chamber that consisted of a tightly sealed plastic container containing moistened paper towels. The overlay was allowed to develop at 37°C for 4 hours and then maintained at 4°C overnight. The next morning the gel (with obvious uPA bands) was examined and then allowed to continue development at 37°C to bring out slower developing tPA bands. The resultant gel clearance was photographed against a black background. 

Total PAI-1 ELISA

A total murine PAI-1 assay (Molecular Innovation, Inc., Royal Oak, MI) was developed to measure the inhibitor levels in the corneal extracts. The monoclonal capture antibody H34G6 directed against murine PAI-1 (gift of Brad Schwartz, University of Illinois, Chicago) produced in PAI-1 knockout mice ²⁶ was coated at a concentration of 10 μg/ml on microliter plates (Immulon 2; Dynex Technologies, Chantilly, VA) at 4°C overnight. A standard curve was first generated using highly purified recombinant murine PAI-1. Briefly, the microliter plates were washed with 300 μl TBS buffer three times to remove excess capture antibody. One hundred microliters of the murine PAI-1 ranging from 0 to 50 ng/ml was added to the wells and incubated for 20 minutes, while rapidly mixing on a rotary shaking table. The plates were washed with 300 μl TBS buffer three times. One hundred microliters of the secondary rabbit antimurine polyclonal antibody (ASMPAI-GF; Molecular Innovations) was added at a 1:5000 dilution and incubated for 20 minutes on a shaker. The plates were washed with 300 μl TBS buffer three times. One hundred microliters biotinylated goat anti-rabbit IgG (supplied with the murine PAI-1 assay kit: MPAIKT; Molecular Innovations, Inc.) was added to the wells and incubated for 20 minutes. The plates were washed with 300μ l TBS buffer three times. One hundred microliters avidin alkaline phosphatase (supplied with the murine PAI-1 assay kit) was added to the wells and incubated for 20 minutes. The plates were washed and developed using commercially available chromogenic p-nitrophenylphosphate (PNPP) substrate solution (Sigma, St. Louis, MO). Linear rates were measured using a microtiter plate reader (Bio-Rad) at 405 nm. According to the manufacturer, the assay measures active PAI-1, latent PAI-1, and the PAI-1/uPA complex equally well. Corneal samples were measured in the same manner as the PAI-1 standards and the initial rates were then compared with the standard curve to calculate the total PAI-1 concentrations in the samples. 

Results

RT-PCR was used in initial studies to determine the corneal gene expression in both naïve and immunized mice ranging from day 0 (uninfected) to 11 days after infection. This time interval was selected because it reflects the increase and reduction of the inflammatory response, as previously noted. ^{11 12} Gene expression for uPA, uPAR, tPA, PAI-1, and PAI-2 was determined, and the results are shown in Figures 1 and 2 . Faint bands for each gene expression were noted in uninfected control specimens and sometimes showed up stronger in extracts from immunized mice, possibly because of a nonspecific increase in basal levels of uninfected, immunized animals compared with naïve animals. However, the subsequent response to infection in immunized mice was subdued in comparison with the response in naïve animals. These results suggest that each gene was constitutively present before infection. There appeared to be no substantive difference between scratched and unscratched corneas in both naïve and immunized mice (data not shown). 

Gene expression for uPA in naïve mice indicated a very faint band at day 0, with a strong enhancement after corneal infection that remained at very high levels through day 11. In contrast, uPA signals in immunized mouse corneas were very faint and did not appear to differ from those in uninfected controls. The cofactor-receptor of uPA (uPAR), which uPA requires for many in vivo functions, ²¹ yielded parallel results for naïve mice, as seen with naïve uPA results. However, there appeared to be peak expression at 1 day after infection in immunized mice, which gradually declined thereafter but was still detectable at day 11. The gene expression of uPAR overall was more intense in naïve mice than that in extracts from immunized mice. 

tPA mRNA expression in naïve mouse corneas showed a time-dependent bell-shaped response. A faint band was detected at day 0; peak expression was observed approximately 7 days after infection and became fainter at day 11. However, the pattern in immunized mice appeared to be approximately the same for all time points. Thus, gene expression in immunized animals of both uPA and tPA was not significantly enhanced by corneal infection. 

RT-PCR indicated that mRNA expression for PAI-1 and -2 was enhanced as a result of corneal infection by P. aeruginosa (Fig. 2) . In both PAI-1 and -2 determinations, gene expressions were greater in naïve mice than in immunized mice. A faint band was detected in extracts from uninfected naïve mice. However, gene expression for PAI-1 in naïve mice appeared to be relatively the same from day 1 to day 11, whereas there appeared to be a peak on day 1 in immunized mouse corneas, which dissipated at days 7 and 11. Naïve mice exhibited a peak of mRNA expression for PAI-2 at day 4 after infection and persisted at slightly lower levels up to 11 days, whereas mRNA levels of immunized mice also peaked at day 1, but persisted at lower levels up to day 11. 

To determine the relative enzymatic expression of uPA and tPA in uninfected and infected corneas, we used plasminogen-dependent caseinolytic zymography. The results can be seen in Figure 3 and indicate a bell-shaped response for uPA in extracts from both naïve and immunized mice. Expression of uPA appeared to peak between 5 and 7 days in naïve mice and between 3 and 5 days after infection in immunized mice. Immunized mice exhibited lower enzymatic activity at the peak time point (day 3) compared with naïve mice (days 5–7). Enzymatic activity for tPA was also detected, but required prolonged incubation of the gel before it was detected (12–24 hours’ incubation), whereas uPA was detected within 4 hours or less by this procedure. Again, tPA activity appeared to be greater overall in naïve mice than in immunized mice. Peak enzymatic activity was seen at day 7 after infection in naïve mice, which paralleled the response of uPA. Peak enzymatic activity of tPA in immunized mice occurred at day 3 and mirrored the response of uPA activity in immunized mice. 

uPAR expression was detected by immunoblot analysis, and the results can be seen in Figure 4 . Immunoblot analysis of extracts prepared from uninfected naïve and immunized mice yielded one major band of approximately 87 kDa, which completely disappeared after infection of naïve mice and initially in immunized mice. A new major band of approximately 17 kDa was detected at day 3, with a molecular weight consistent with the enzymatic release of the D1 component of uPAR, which is composed of three homologous domains (D1, D2, and D3). ^{27 28 29} The 17-kDa band diminished in intensity at days 5 and 7 and was no longer apparent by day 12 in naïve mice, which may be due to its further proteolytic degradation. Extracts from immune mice exhibited the same band, but at days 3 and 5, and then was no longer apparent. 

uPAR degradation by uPA, plasmin, PMN elastase, and exogenous chymotrypsin resulting in several bands is well established. ^{27 28 29} An additional major band of approximately 70 kDa was detected in extracts from naïve mice at day 3 and less so at day 5 in both naïve and immunized mice. The molecular weight of the two bands (17 and 70 kDa) seen in infected mice at day 3 appear to add up to the 87 kDa of the band seen at day 0. A fourth band obtained from only immune mice was detected at approximately 33 kDa and may either represent an additional cleavage site by a different enzyme or may be an artifact due to immunization, because it was seen in uninfected immunized mice, but not in naïve mice. Finally, the 87-kDa band thought to be intact uPAR and normally seen in uninfected mice became apparent again in immunized mice at days 7 and 12. The intensity of the band at day 12 was greater than that seen at day 7, suggesting a strong correlation between the ability to restore corneal clarity and the resynthesis of uPAR. 

To determine the PAI-1 levels in extracts from both naïve and immunized mice, we used an ELISA, which can detect both free and bound PAI-1. The results can be seen in Figure 5 and indicate a bell-shaped curve with a peak at day 9 and near maximum expression up to 14 days after infection. The results from infected immunized mice exhibited little or no significant induction of PAI-1 expression with an insignificant peak at days 2 through 6 after infection, compared with the results in naïve mice. 

PAI-2 expression over a 12-day period was also detected by immunoblot analysis and can be seen in Figure 6 . A recombinant PAI-2 fusion protein, which migrates at 47 kDa, was used as a positive control. A second, but slightly lower molecular weight band was also present and may reflect the presence of a partially cleaved form of the fusion protein. Detectable protein bands were obtained at all time points for both naïve and immunized mice. Peak expression in naïve mice was between days 5 and 7, whereas it occurred between days 3 and 5 in immunized mice. A second, but lower weight band was detected in most of the extracts from both naïve and immunized mice and may represent the N-terminal fragment upstream of a putative cleavage site of the reactive center. The temporal related intensity of the lower bands tended to mirror their corresponding 47-kDa bands. 

It is well established that serpins are “suicide” inhibitors and require some proteolytic degradation by their target proteases to express inhibitory activity. Thus, the lower band appears to reflect a partially cleaved 47-kDa band as a result of the infection and subsequent enzymatic activity. ³⁰ One of the characteristics of the inhibitory activity of serpins is their remarkable degree of conformational flexibility and their rapid polymerization in the absence of citrate. ³¹ Thus, the upper band(s), above 47 kDa, are frequently seen with serpins and may be due to polymerization or possibly to an enzyme–substrate complex. 

Discussion

A number of investigators have demonstrated the ocular expression of serine proteases of the plasminogen system along with their serpins. ^{32 33 34 35 36 37 38} In the present study we used RT-PCR to demonstrate corneal mRNA expression of uPA, uPAR, tPA, PAI-1, and PAI-2 in uninfected and infected mice. Corneal extracts from naïve mice in which corneal clarity was not restored indicated a temporal dependent upregulation of gene expression for each target gene. However, extracts from immunized mice expressed overall lower or less induction of the same genes with a shorter effective period, indicating a direct correlation between the ability to restore corneal clarity and less proteolysis. Concomitantly, there was also lower induction of the two inhibitors PAI-1 and -2. 

There is general agreement that the primary role of uPA involves ECM degradation (fibronectin, vitronectin, fibrin, and collagens) and cell migration and proliferation, whereas tPA appears to be more responsible for thrombolysis and cellular differentiation. However, uPA and tPA can substitute for each other in the generation of plasmin. ²¹ uPA may also be implicated in angiogenesis, priming of PMN superoxide anion release, and PMN chemotaxis. ³⁹ Of particular relevance to these studies is the infiltration of both PMNs and macrophages into the infected cornea. ⁹ The latter cell and possibly resident cells is responsible for production of both tPA and uPA. ^{21 33} Thus, initiation of these proteolytic events contributes to the eventual corneal destruction seen in naïve mice. 

During corneal infection of naïve mice, a complex series of biochemical and pathologic events are initiated that ultimately result in the destruction of the ECM. Stromal ulceration begins superficially in association with epithelial defects and the release of the plasminogen activators on the stromal surface, resulting in plasmin formation and the upregulation of a cascade of other proteolytic activities. ¹⁹ Once formed, plasmin can degrade fibrin, fibronectin, vitronectin, and laminin, the component of the subepithelial basement membrane. Concomitantly, there is active PMN infiltration into the stroma, along with activation of fibroblasts, both of which contribute to stromal collagen and glycosaminoglycan destruction. ¹⁸ There is also release of pharmacologic agents, such as platelet activating factor, cytokines, and arachidonic acid metabolites. ^{11 12 40 41} A large part of the corneal stroma becomes edematous along with neovascularization, a gradual increase of macrophages containing inclusion bodies, a loss of keratocytes and heavy fibrin deposition, and eventual ulceration. ¹⁹ All these events are complicated by the temporal increase of bacterial numbers and their potential virulence factors. 

Plasmin substrates include the glycoproteins that cover and shield collagen from degradation by MMPs, as well as the third component of complement that results in chemotactic components for PMNs. ⁴² In addition, plasmin can activate prostromelysin 1 (MMP-3), which can then activate other pro-MMPs, such as pro-MMP-9. ^{43 44} Plasmin can also activate several growth factors including TGF-β, which plays a critical role in regulation of TGF-β–mediated processes. ^{45 46} Differential processing of plasminogen by MMP-2 and plasmin can lead to active fragments such as angiostatin, which has antiangiogenic properties. ^{47 48}  

The concomitant upregulation of the uPA receptor/cofactor uPAR is normally derived from both infiltrating cell types, PMNs, and macrophages, ^{17 21} cell types that have been previously demonstrated to be present in experimental murine corneal infections by P. aeruginosa. ^{9 11 12} uPAR is a cell-membrane–anchoring binding protein for uPA, accumulating plasminogen activity at cell surfaces. Because uPA activity is dependent on uPAR expression, it was expected and experimentally verified in these studies that both uPA and uPAR would be expressed simultaneously, as detected by RT-PCR. 

It should be pointed out that the molecular weight of solubilized and purified human uPAR is approximately 50 to 65 kDa. ^{27 28} However, its molecular weight is highly variable from cell type to cell type, depending on its degree of glycosylation and its binding avidity to vitronectin, integrins, and thrombospondin. The 87-kDa band seen in our immunoblotting of extracts from the uninfected mice may have reflected uPAR bound to a membrane receptor fragment that has been identified to be the glycosylphosphatidyl inosital (GPI) anchor. ²⁹ Two recently discovered unrelated proteins to uPAR have been also shown to colocalize to GPI and they consist of the membrane-type MMP-17 and -25 (MT4-MMP and MT6-MMP). ^{49 50}  

The detection of the putative D1 component of uPAR by immunoblotting yielded a 16- to 17-kDa fraction, which migrated similarly to the 16-kDa fragment when purified human uPAR was treated with exogenous chymotrypsin. ²⁷ Although chymotrypsin is not present in the mouse cornea, other proteases, such as plasmin, uPA, and PMN elastase, may have been responsible for the uPAR degradation. ²⁹ The D1 and D2D3 fragments have previously been shown to induce chemotaxis, cell adhesion, and signal transduction–related processes. ²⁹ To the best of our knowledge, this is the first in vivo example of uPAR degradation during an ocular infection. 

The upregulation of tPA, uPA, and uPAR in these studies exhibited good correlation with the overall inflammatory response that was described in earlier manuscripts. ^{9 11 12} In those temporal studies, our inflammatory parameters were based on PMN and macrophage response, bacterial numbers, and release of arachidonic acid metabolites and certain cytokines. In addition, we demonstrated the upregulation of MMP-9 (gelatinase B) and bacterial alkaline protease in corneal extracts from infected naïve mice, whereas there was much less induction of both enzymes in extracts from immunized mice. ^{14 15} Recent findings in our laboratory have demonstrated a similar response for MT1-MMP, MT2-MMP, and MT3-MMP. ¹³  

The regulation of proteolytic enzymes in tissues by endogenous inhibitors is a critical requirement for the maintenance of homeostasis. Therefore, it was of particular relevance that we demonstrated the expression of PAI-1 and -2 in corneal extracts from infected mice that are considered to be the main inhibitors of plasminogen activators. ²¹ Upregulation of both PAI-1 and -2 can be attributed to the stimulation of infiltrating cells and possibly resident cells by lipopolysaccharide (LPS) of the invading organisms ^{17 23 51} and/or the host-released TNF-α and IL-1 that has been previously demonstrated in these and other ocular studies. ^{11 41} Plasmin release may have also occurred indirectly by bacterial protease activation of the Hageman factor-prekallikrein-kallikrein pathway. ⁵²  

Of the two serpins described herein, only PAI-2 has been detected constitutively in normal human corneas. ³⁶ However, PAI-1 has been extensively demonstrated in the aqueous humor and is thought to accelerate aqueous outflow and clot lysis. Masos et al. ⁵³ recently demonstrated its presence in the aqueous humor and ciliary epithelium of the rodent eye, but not in the cornea. In our studies using pooled corneas from a different strain of mice and a differing experimental procedure, we were able to demonstrate corneal gene expression of PAI-1 using RT-PCR and PAI-1 protein synthesis by ELISA in uninfected mice. It is apparent that expression of both PAI-1 and -2 was upregulated and easier to detect once intracorneal infection and subsequent inflammation occurred. It should be pointed out that other protease inhibitors have been demonstrated in human corneas such as α₁-antichymotrypsin,α ₂-macroglobulin,α ₂-antiplasmin, andα ₁-proteinase inhibitor. ^{33 54} Because PAI-1 is the principal physiological inhibitor of both tPA and uPA, it is not unreasonable to expect detection of PAI-1 in the cornea along with the plasminogen activator/plasmin system. 

In conclusion, we have demonstrated the constitutive presence of tPA, uPA, uPAR, PAI-1, and PAI-2 in murine corneal extracts from uninfected mice. Each factor was upregulated in infected naïve mice and to a lesser extent in immunized mice, probably because of both inflammatory and resident cells. In addition, there are other corneal endogenous inhibitors and proteases along with microbial proteases that are also present, but were not studied in this experimental model. ^{13 14 33} Thus, the results described herein are a reflection of the overall dynamic interrelationship between host and bacteria and represent a wide variety of biological processes affecting ECM turnover. From these studies, it is apparent that extensive upregulation of tPA and uPA, leading to plasmin formation results in extensive ECM damage, both directly and indirectly. Whereas the shorter and less intense inflammatory response in immunized animals plays a role in the restoration of corneal clarity. 

 Supported by National Eye Institute Grant EY-11757 and P30 EY-04068 from the National Institutes of Health, Bethesda, Maryland.

 Submitted for publication July 17, 2000; revised November 3, 2000 and February 7, 2001; accepted March 5, 2001.

 Commercial relationships policy: N (RSB, MK, ZD); I (DED); E (DED).

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Figure 1.

 

mRNA expression of uPA, uPAR, and tPA in mouse corneas infected with P. aeruginosa. Naive and immunized mice were infected with P. aeruginosa, and total RNA samples were prepared on days 0 (uninfected), 1, 4, 7, and 11. Equal amounts of total RNA from individual samples (500 ng) were used for RT-PCR. After reverse transcription, specific primers for mouse uPA, uPAR, and tPA were applied to the reaction system to amplify respective target genes. A housekeeping gene (18S rRNA) was also amplified and used as an internal control for the comparison of all time samples. Amplified genes were stained by ethidium bromide as bright bands in the dark background. Lane M: molecular marker (100-bp ladder).

Figure 2.

 

mRNA expression of PAl-1 and -2 in mouse corneas infected with P. aeruginosa. Amplification was conducted as in Figure 1 , using specific primers for mouse PAI-1 and -2.

Figure 3.

 

Expression of uPA and tPA proteins in mouse corneas infected with P. aeruginosa. Naïve and immunized mice were infected with P. aeruginosa, and samples were prepared on days 0 (uninfected), 3, 5, 7, and 12. Equal amounts (5 μl) of extracts from individual samples were loaded for electrophoresis, and the gel was incubated in different buffers to renature and develop uPA and tPA. uPA and tPA were visualized as transparent bands corresponding to their molecular weights. The gel was photographed against a black background to better view the bands. rU and rT stand for recombinants of uPA and tPA (human), respectively. Slight variation in the molecular weight of recombinant uPA and tPA with that of murine uPA and tPA is probably due to species differences.

Figure 4.

 

Expression of uPAR protein in mouse corneas infected with P. aeruginosa. Naïve and immunized mice were infected with P. aeruginosa, and samples were prepared on days 0 (uninfected), 3, 5, 7, and 12. Equal amounts (5 μg) of total protein from individual samples were loaded for electrophoresis. The blots were then incubated with specific antibodies that recognize mouse uPAR. Several species of uPAR were visualized as dark bands corresponding to their molecular weights.

Figure 5.

 

Total PAl-1 in the mouse corneas infected with P. aeruginosa. Naïve and immunized mice were infected with P. aeruginosa, and samples were prepared on days 0 (uninfected), 2, 6, 9, and 14. Equal amounts (5 μl) of extracts from individual samples were used for PAl-1 detection by ELISA. The assay was run in duplicate.

Figure 6.

 

Expression of PAI-2 protein in mouse corneas infected with P. aeruginosa. Naïve and immunized mice were infected with P. aeruginosa, and samples were prepared on days 0 (uninfected), 3, 5, 7, and 12. Equal amounts (5 μg) of total protein from individual samples were loaded for electrophoresis. The blots were then incubated with specific antibodies recognizing mouse PAI-2. PAI-2 was visualized as a dark band corresponding to its molecular weight of approximately 47 kDa.

The authors thank Daniel A. Lawrence and Julianna M. Berk for technical assistance.