Involvement of the Erk-MAP Kinase Pathway in TNFα Regulation of Trabecular Matrix Metalloproteinases and TIMPs

Abstract

purpose. TNFα is a strong modulator expression of trabecular meshwork (TM) matrix metalloproteinase (MMP) and tissue inhibitor (TIMP). Laser trabeculoplasty appears to rely on this process to restore normal aqueous humor outflow facility. Thus, studies were conducted to determine whether the extracellular signal-regulated kinase (Erk)-mitogen-activated protein (MAP) kinase signal-transduction pathway is involved.

methods. Porcine TM cells were treated with TNFα, and changes in MMPs and TIMPs were evaluated by zymography and Western immunoblot assay. Phosphospecific antibodies to proteins from the Erk pathway were used to evaluate responses to treatment with TNFα. Inhibitors of Mek, the kinase that activates Erk, and of protein kinase C (PKC) isoforms were used to define pathway involvement.

results. Treatment with TNFα increased MMP-1, -3, and -9 and TIMP-1, whereas expression of MMP-2 was not affected and expression of TIMP-2 was decreased. Erk and Mek were rapidly phosphorylated after treatment with TNFα, and c-Raf-1 showed a significant bandshift. A specific inhibitor of Mek blocked the TNFα induction of the MMPs and TIMPs and the phosphorylation of Erk. An inhibitor of the PKC-μ isoform, which also blocks the effects of MMP-TIMP of TNFα, did not affect phosphorylation of Erk.

conclusions. The components of this MAP kinase pathway in the TM are dramatically affected by TNFα and inhibition of Erk’s phosphorylation blocks the changes in MMP and TIMP expression. PKC μ, which is also required in this transduction process, does not appear to be upstream from Erk in the signaling cascade. Manipulation of this and related TM signal-transduction pathways may provide targets for developing improved glaucoma treatments.

The matrix metalloproteinase (MMP) family and their tissue inhibitors, the TIMPs, are integrally involved in regulating the turnover of TM extracellular matrix. ^{1 2 3 4} These proteinases have been implicated in a variety of pathologic conditions, including arthritis, angiogenesis, and metastatic invasion, ^{5 6 7 8} and, recently, in the regulation of aqueous humor outflow ^{9 10} and the mode of action of laser trabeculoplasty, a common laser treatment for glaucoma. ^{11 12 13} At least 2.47 million persons in the United States and 66.8 million worldwide are affected by glaucoma, a common blinding disease. ^{14 15} Understanding the regulation of the MMPs and TIMPs within the TM may thus have significant medical implications. 

The expression of MMP and TIMP is intricately regulated at the transcriptional level. ^{16 17 18} The promoter regions of the various MMP and TIMP genes contain a variety of simple and complex enhancer elements, and their expression is modulated by numerous growth factors, cytokines, steroids, integrin ligation, and other extracellular information and conditions. ^{16 17 19 20 21 22 23} In addition, activation of the latent proenzyme forms and inhibition by the TIMPs add additional layers of regulation. ^{3 7 17 18 24 25}  

The phorbol mitogen, 12-tetradecanoylphorbol-13-acetate (TPA), and the cytokines, TNF and IL-1, are among the strongest inducers of TM expression of MMP and TIMP that we have identified. ²⁶ The TM response to laser trabeculoplasty, which includes strong and sustained increases in TM expression of MMP, is mediated by IL-1β and TNFα. ¹³ Thus, the details of the increases in TM expression of MMP induced by these cytokines are of considerable potential therapeutic interest. Although the signal-transduction pathways involved in modulating expression of MMP and TIMP have been investigated in some detail in several other tissues, ^{20 27 28 29 30} little is known about this transduction in the TM. We have previously shown that PKC μ, but apparently not other isoforms, is involved in a required step in this TNFα signal-transduction pathway. ³¹ Additional PKC isoforms appear to be necessary for TPA to affect the expression of MMP. 

Another pathway frequently involved in regulation of MMP by various growth factors and cytokines in other tissues is the mitogen-activated protein (MAP) kinase pathway. ^{32 33 34 35} Recently, an involvement of the extracellular signal-regulated kinase (Erk) in the rapid TM MMP-2 secretion observed in response to PDGF-treatment has been reported. ³⁶ There are three well-defined and apparently parallel MAP kinase pathways: the p44-p42 or Erk 1/2 pathway (Fig. 1) ; the c-Jun N-terminal kinase-stress-activated protein kinase (JNK/SAPK) pathway; and the p38 Map kinase pathway. Because TNFα often mediates effects through these pathways, we evaluated the possible involvement of the Erk-MAP kinase cascade in transducing the TNFα induction of TM cell MMPs. 

Materials and Methods

Human eyes were obtained within 48 hours after death from the Portland Lion’s Eye Bank (Portland, OR). The donor eyes were managed in compliance with the provisions of the Declaration of Helsinki for research involving human tissue. Porcine eyes were obtained within a few hours after death from Carlton Packing (Carlton, OR), human recombinant TNFα was from R&D Systems (Minneapolis, MN); 12-tetradecanoylphorbol-13-acetate (TPA), leupeptin, aprotinin, pepstatin, and horseradish peroxidase-conjugated secondary antibodies were from Sigma-Aldrich (St. Louis, MO); GF 109203X (bisindolylmaleimide I), Gö 6976, Ro 31-8220, and PD 98059 were from CalBiochem (San Diego, CA); PicoGreen was from Molecular Probes (Eugene, OR); the MMP and TIMP antibodies were from Triple Point Biologics (Portland, OR). Different Erk, Mek, c-Raf-1, and Ras antibodies were from Promega (Madison, WI), Cell Signaling Technology (Beverly, MA), BD Transduction Laboratories (San Diego, CA), and Santa Cruz Biotechnology (Santa Cruz, CA), respectively; phosphospecific Erk (T202/Y204) and Mek (S217/S221) antibodies were from Cell Signaling Technology; Dulbecco’s modified Eagle medium (DMEM), antibiotics and antimycotics were from GibcoBRL (Grand Island, NY); fetal bovine serum was from HyClone (Logan, UT); and chemiluminescent detection kits were from NEN Life Sciences (Boston, MA). 

Cell Culture, Treatments, and Extractions

Porcine and human TM cells were cultured as previously described. ^{4 37} Except as specifically indicated, all the data shown are from porcine TM cells. All the key observations were replicated in humans, showing no significant species differences. TM cells were used as confluent monolayers at passage 3 and were maintained serum free during and for 48 hours before treatments. Approximately 15 separate porcine TM cell lines, each derived from a separate pool of the dissected TMs from 20 to 40 eyes, were used for these studies. All experiments presented were repeated at least three times, and representative gels were selected for presentation. DNA analysis (PicoGreen; Molecular Probes) to estimate cell density in parallel flasks was conducted for some studies as directed by the manufacturer. Because the differences between flasks were always less than ±10%, this analysis was not conducted for all studies. Ponceau S staining of all Western blots before probing provided further verification of uniform gel loading. MMP and TIMP analyses were conducted on culture media collected 24, 48, or 72 hours after treatments. Media samples were immediately centrifuged for 5 minutes at 4000g and 4°C ammonium sulfate was added to the supernatant, bringing it to 70% saturation. After incubation overnight at 4°C, the precipitate was collected by centrifugation at 15,000g for 30 minutes, resuspended in one thirtieth of the original volume and stored in aliquots frozen at −20°C until use. Freeze-thaw of samples was avoided. Analysis of signal transduction proteins and phosphoproteins was conducted on extracts of cells at the times indicated, using methods detailed previously. ³¹ Briefly, the medium in each T-75 flask was replaced with 0.5 mL 4°C-modified RIPA buffer ^{38 39} (2 mM EDTA, 2 mM EGTA, 1% NP-40, 0.5% sodium deoxycholate, 0.1% sodium dodecyl sulfate [SDS], 100 mM NaF, 1 mM sodium orthovanadate, 1 mM phenylmethylsulfonyl fluoride [PMSF], 20 μg/mL leupeptin, 20 μg/mL aprotinin, 20 μg/mL pepstatin, and 50 mM Tris [pH 7.5]) and flasks were immediately placed on ice. Cells were then scraped from the flasks, and the extract was sonicated on ice for 30 seconds using a micro tip and 50% maximum setting with a 400-W, 20,000-Hz sonifier (model 450; Branson, Danbury, CT). The sonicate was then centrifuged at 12,000g for 5 minutes and the aliquoted supernatant stored at −20°C until use. ³¹  

Zymograms and Western Immunoblots

Western immunoblots, transferred electrophoretically from standard SDS-polyacrylamide gels to polyvinylidene difluoride (PVDF) membranes, were probed with the indicated primary antibodies and detection used the appropriate secondary antibodies with conjugated horseradish peroxidase and chemiluminescence, according to the manufacturer’s instructions. Gelatin was use as the substrate to detect gelatinase A and B (MMP-2 and -9, respectively), or β-casein was used as the substrate to detect stromelysin (MMP-3) in the zymograms (substrate, SDS-polyacrylamide gels). ^{4 40}  

Results

Effects of TPA and TNFα on MMPs and TIMPs

TM cells respond to treatment with TPA or TNFα by increasing the expression of MMP-9, -3, and -1 and TIMP-1 (Fig. 2) . ^{26 31} Neither agent affected expression of MMP-2, and TNFα, but not TPA, decreased levels of TIMP-2. All these effects were dose- and time-dependent with detectable increases occurring in culture media between 1 and 10 ng/mL for both agents and by 24 hours after treatment (not shown). Stronger responses were observed with 25 ng/mL, and in general the expression changes increased to at least 72 hours. At the higher dose, the responses reached a maximum earlier and some declined by 72 hours. 

Effects of TPA and TNFα on TM Phosphorylation of Erk1/2 and Mek

Exposure of TM cells to TPA or TNFα induced a rapid phosphorylation of Mek and of Erk-1 and -2 (Figs. 3A 3B) . Phosphorylation of Erk was maximal by 5 minutes for both agents and declined gradually over time, although it remained elevated several fold above control at 24 hours, Mek phosphorylation was elevated strongly by 5 minutes after treatment with either agent. At 15 minutes after treatment with TPA, phosphorylation of Mek was higher yet. By 30 minutes after TPA or 15 minutes after TNFα was added, phosphorylation of Mek declined significantly. The protein levels of Mek (Fig. 3B) and Erk (not shown) did not change appreciably with these treatments at any times examined. 

Supraphosphorylation of c-Raf-1 after Treatment with TPA or TNFα

At 15 minutes after treatment with TPA or TNFα, c-Raf-1 showed a significant bandshift and migrated at a slower rate on gels (Fig. 3C) . This bandshift and appearance of a doublet were still apparent at 2 hours. The protein levels of A- and B-Raf did not change over the 24 hours evaluated (not shown). Minimal amounts of A-Raf were present in these cells, but B-Raf was present at appreciable levels. Neither showed any indication of a similar bandshift (not shown). Translocation studies of c-Raf-1, to evaluate its partitioning between particulate, cytosolic, and detergent-extractable membrane fractions were uninformative (not shown). c-Raf-1 was distributed approximately equally between cytosolic and membrane fractions, with or without these treatments. The same bandshifts apparent at 15 minutes, are observed in both fractions (not shown). 

Effects of Inhibition of Mek and PKC

The specific Mek activity inhibitor PD 98059 ⁴¹ blocked the effects of both TPA and TNFα on MMPs and TIMP-1 (Fig. 4) . The inhibitor was added 1 hour before the other treatments were begun. The response was dose dependent and time dependent, with the 10 μM dose producing moderate reduction in the effects and the 50-μM dose producing dramatic effects. Similar effects were seen at other doses of TPA and TNFα and at 24 and 48 hours. At 50 μM, PD 98059 alone had no effect on expression of MMP or TIMP. Higher doses of PD (100 or 200 μM) had more dramatic and rapid effects on the TPA- and TNFα-induced MMP and TIMP responses, but these doses also had effects when added alone (not shown). 

Pretreatment with PD 98059 also significantly reduced the phosphorylation of Erk after treatment with TPA or TNFα (Fig. 5A) but did not reduce the phosphorylation of Mek (data not shown). This Mek inhibitor also blocked, or at least radically diminished, the c-Raf-1 bandshift that was triggered by both TPA and TNFα (Fig. 5B) . 

We have demonstrated ³¹ that the effects of TPA on TM expression of MMP and TIMP are blocked by pretreatment for 1 hour with the specific PKC inhibitors Bis I (GF 109203X), Gö 6976, and Ro 31-8220, whereas only Gö 6976 is able to block completely the effects of TNFα on expression of MMP and TIMP. To determine whether the affected PKC isoforms were acting upstream from Mek and Erk, we evaluated the effects of these PKC inhibitors on phosphorylation of Mek and Erk. Pretreatment with high levels (200 nM), of Bis I, Gö, or Ro (the latter not shown) did not have significant effects on the TPA- or TNFα-induced phosphorylation of either Mek or Erk when evaluated at 5, 15, or 60 minutes or at 24 hours (Fig. 6) . 

In parallel experiments, the effects of these PKC inhibitors on the supraphosphorylation bandshift that occurred with c-Raf-1 in response to treatment with TPA or TNFα was evaluated. None of these PKC inhibitors (Bis, Gö, or Ro), blocked the c-Raf-1 bandshift that occurred at 15 minutes in response to TPA or TNFα (data not shown). 

Discussion

Based on the phosphorylation patterns observed and the ability of a very specific Mek inhibitor ⁴¹ to block the TNFα induction of the TM MMP/TIMPs, it seems highly likely that the Erk-MAP kinase pathway is critical in transducing the TNFα signal to increase MMP production in the TM. The cascade from c-Raf-1 to Mek to Erk ^{42 43} is commonly involved in cytokine or growth factor regulation of the expression of MMP. ^{33 44} The sustained elevation of phosphorylation of Erk-1 and -2 is particularly indicative of this process. ⁴⁵  

The transient Mek phosphorylation and the delayed c-Raf-1 supraphosphorylation are in agreement with the recognized mechanism of this pathway. ^{42 43 46} Inactive c-Raf-1 is recruited to the plasma membrane, which allows it to autophosphorylate and become active. ⁴⁷ Activated c-Raf-1 then recruits Mek, phosphorylating it on S217 and S221, which activates it. ⁴² Mek, a dual-specificity kinase, then recruits Erk, phosphorylating it at T202 and Y204, which activates it. Erk then shuts down the upstream portion of the cascade by phosphorylating other inhibitory sites on Mek, c-Raf-1, and several other upstream components. ⁴⁸ The supraphosphorylation of c-Raf-1 accounts for the bandshift observed at 15 minutes (Fig. 2C) . Erk is then thought to escape from this complex and to phosphorylate several downstream components and targets of the pathway, including several transcriptional activator proteins. 

The fact that the specific Mek activity inhibitor PD 98059 blocks phosphorylation of Erk, c-Raf-1 supraphosphorylation, and the changes in levels of MMP-TIMP, provides strong support for the hypothesis that this pathway is involved in transducing this signal. A recent report of the signal-transduction pathway(s) involved in the IL-1 induction of TM MMP-3 supports our conclusion. ⁴⁹  

Our earlier observation ³¹ that PKC-μ is also required in this signal-transduction pathway in the TM suggests a possible upstream involvement of PKC in this cascade. PKC has been shown to act upstream from Ras, which can recruit c-Raf-1 to the plasma membrane for activation. PKC can also act directly on c-Raf-1, circumventing a Ras involvement. ^{45 50} However, the inability of several PKC inhibitors with overlapping but separate isoform specificities to block phosphorylation of Erk argues strongly that PKC-μ is either acting downstream from Erk or is acting on a separate parallel pathway. If parallel pathways exist, they probably converge farther downstream. Thus, both PKC-μ and Mek-Erk are necessary, but not sufficient, for this signal transduction process. 

Although other mediating events have not been ruled out, the clinical mechanism whereby laser trabeculoplasty restores normal outflow facility for several years in many cases can be explained by a simple working model. TM laser burns trigger increases in IL-1β and TNFα, which are secreted within 8 hours after treatment. ¹³ These cytokines induce juxtacanalicular TM expression of MMP-3 and -9 ^{11 12} with an associated reduction in TIMP-2. ^{26 31} Increased TM MMP levels cause turnover and remodeling of the juxtacanalicular extracellular matrix, which reduces the outflow resistance ⁹ and restores normal intraocular pressure. Although the initial cause of the glaucoma has not been corrected, the outflow pathway has been temporarily rejuvenated, and in many cases it takes years for outflow facility to become problematic again. 

Although direct application of these cytokines is probably not a viable future therapy, identifying drugs that mimic laser trabeculoplasty seems worthwhile. Small molecules that modulate the signal-transduction pathways involved could be valuable as a substitute for laser treatment. Thus, manipulation of the Erk-MAP kinase pathway or of PKC-μ may provide therapies that can substitute for laser trabeculoplasty in ameliorating glaucoma. 

 Supported by National Eye Institute Grants EY03279, EY08247, and EY10572 and by grants from the Glaucoma Research Foundation, San Francisco, California, Research to Prevent Blindness, and Alcon Laboratories, Fort Worth, Texas.

 Submitted for publication December 3, 2001; revised April 12, 2002; accepted May 21, 2002.

 Commercial relationships policy: F.

 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.

 

Hypothetical signal-transduction pathway involved in TNFα induction of TM MMPs and TIMPs. TNFα binding activates the 55-kDa TNF receptor (TNF-R1), which forms a complex with TNF receptor-associated death domain (TRADD) and TNF receptor-activation factor (TRAF-2) and triggers one or more steps (question marks), which result in the activation of PKC-μ and c-Raf-1. Mek is phosphorylated on S217 and S221 by c-Raf-1. Activated Mek then phosphorylates Erk on Y204 and T202, which activates Mek and Erk and then directly or indirectly phosphorylates (question marks) transcriptional activators, which initiate transcription of the MMPs and TIMPs mRNAs. PKC-μ activity is also necessary in this process. ³¹ PD98059 is a selective Mek activity inhibitor, and Gö 6976 inhibits PKC-μ.

Figure 2.

 

Effects of TPA and TNFα on TM expression of MMP and TIMP. TM cells were serum free for 48 hours before and during treatment with TPA or TNFα, as indicated. Media collected after 72 hours of treatment were subjected to gelatin zymography (first row) or casein zymography (second row) to detect the activity levels of MMP-9, -2, and -3, as indicated. The same media were subjected to Western immunoblot assay and probed with antibodies to MMP-3, MMP-1, TIMP-1, or TIMP-2 (third through sixth rows, respectively) to detect levels of these proteins. Controls were parallel samples that were exposed to vehicle and incubated for the same times.

Figure 3.

 

Effects of TPA and TNFα on phosphorylation of Erk, Mek, and c-Raf-1. TM cells, which had been serum free for 2 days, were treated with 10 ng/mL TPA or TNFα for the indicated times. Western immunoblots of cellular extracts separated on SDS-polyacrylamide gels were probed with antibodies specific for phospho-Erk (A), Mek or phospho-Mek (B) or c-Raf-1 (C). Arrows: expected migration position of the protein. Controls gels were exposed to TPA or TNFα for 0 minutes.

Figure 4.

 

Effects of PD 98059, the Mek inhibitor, on expression of MMP-TIMP induced by treatment with TPA or TNFα. TM cells were maintained serum free for 2 days before the indicated doses of the Mek inhibitor PD 98059 were added. One hour later, the indicated doses of TPA or TNFα were added. After 72 hours, the culture media were removed for zymography or Western blot analysis as indicated. The (−) or (+) over the lanes denote, respectively, the absence or presence of TPA or of TNFα at the indicated doses. PD, samples exposed to the inhibitor PD 98059 at the indicated doses. Gelatin and β-casein zymograms were used to evaluate the activity of MMP-9, -2, and -3, respectively as labeled, and Western immunoblots were used to determine levels of MMP-3, MMP-1, and TIMP-1 proteins.

Figure 5.

 

Effects of PD 98059, the Mek inhibitor, on phosphorylation of Erk and Raf. TM cells, which had been serum free for 2 days, were pretreated for 1 hour with 50 μM PD 98059 as indicated (+PD) and then treated with 10 ng/mL TPA or TNFα added as indicated. The effect on phosphorylation of Erk-1 and -2 was evaluated at several times, as indicated on Western blots with phosphospecific antibodies (A). The effects of similar treatments on the c-Raf-1 bandshift induced by TPA and TNFα was also evaluated with a c-Raf-1 specific antibody (B). Control lanes were serum free for 2 days and then exposed to TPA or TNFα for 0 minutes. For groups labeled +PD, PD 98059 was added 1 hour before other treatments. For comparison purposes, the same actual data for the treatment-without-inhibitor lanes that were shown in Figure 3 are shown in (A).

Figure 6.

 

Effects of PKC inhibitors on phosphorylation of Erk and Mek. TM cells, which had been serum free for 2 days, were pretreated for 1 hour with 200 nM Bis I (Bis) or Gö 6976 (Go) as indicated and then with 10 ng/mL TPA or TNFα for the indicated times. The phosphorylation levels of Erk and Mek were then determined after the indicated treatment times on Western blots with phosphospecific antibodies. Controls cells (−) were treated with inhibitors and/or TPA or TNFα for 0 minutes. For comparison purposes, the same actual data for the effects on phosphorylation of Erk of treatment without inhibitor (leftmost set of lanes) that were shown in Figure 3 are included.