SKIP Downregulation Increases TGF-β1-Induced Matrix Metalloproteinase-9 Production in Transformed Keratinocytes

Transforming growth factor-beta (TGF-β1) is a potent inductor of matrix metalloproteinase-9 (MMP-9) in transformed cells. Recently, Ski-interacting protein (SKIP) has been described as a regulator of TGF-β1 signal transduction, but its role in the induction of cell malignance by TGF-β1 has not been fully elucidated so far. In the present study, we analyzed the role of SKIP on TGF-β1-induced MMP-9 production. Mouse transformed keratinocytes (PDV) were stably transfected with SKIP antisense construct. We observed that SKIP depletion provoked an enhancement in the expression of MMP-9 in response to TGF-β1 treatment. The downregulation of SKIP produced an enhancement in TGF-β1-activated ERK1,2 MAP kinase as well as increased transactivation of downstream Elk1 transcription factor. The increased MMP-9 production in response to TGF-β1 was dependent of MAPK activation as PD98059, an MEK inhibitor, reduced MMP-9 expression in SKIP antisense transfected cells. Thus, we propose SKIP as a regulatory protein in TGF-β1-induced MMP-9 expression acting by controlling ERK1,2 signaling in transformed cells.


Introduction
e transforming growth factor-beta (TGF-) superfamily of factors is implicated in the regulation of cell proliferation, differentiation, migration, extracellular matrix production, apoptosis, and tumorigenesis [1]. TGF-binds to the functional complex of TGF-family of receptors (T Rs) at the cell surface, which consist of two type II and two type I transmembrane serine/threonine kinase receptors [1,2]. Receptors, in turn, activate downstream cellular components, including Smads and members of the Ras/MAP kinase pathways [3]. TGF-1 has been postulated to have a dual role in tumour progression, by acting as a tumour suppressor in early stages of carcinogenesis and exerting a prooncogenic role in the last steps of metastatic disease contributing to tumour cell invasion and metastasis [4,5].
Degradation of the basal membrane and the collagenous extracellular matrix (ECM) presents a critical step in the tumor invasion [6]. Members of the matrix metalloproteinase (MMP) family are involved in the degradation of ECM and implicated in malignancy [7]. Among the human MMPs reported to date, MMP-2 and 9 are the key enzymes involved in degradation of type-I and IV collagen in ECM [8]. Both MMP-2 and 9, which are abundantly expressed in various malignant tumors, contribute to cancer invasion and metastasis [9]. MMP-9 can be stimulated by the in�ammatory cytokine tumor necrosis factor (TNF)-∝, by the epidermal growth factor, and by TGF-, through the activation of different intracellular signaling pathways [10]. TGF-1 regulates MMP-9 expression, either by inducing its expression through Ras-ERK MAPK or by the negative regulation of MMP-9 expression related to Smad3 [11,12]. In fact, it was reported that Smad3 null mice show deregulated expression of MMP-9 [13] while higher expression of this enzyme was detected in serum of osteoarthritis patients with Smad3 mutation [14]. ese data suggest a �ne regulation of MMP-9 expression by TGF-1 either by Ras-ERKs or Smad3.
Recently, it has been reported that Ski-interacting protein (SKIP) interacts with Smad2,3 to augment TGF--dependent transcription [15], suggesting that SKIP may play a role RT-PCR analysis of MMP-9 expression. PDV cells were treated for 24 h with 5 ng/ml of TGF-1, and mRNA obtained was reverse transcribed to cDNA. PCR products were separated in agarose gels. GADPH was used as housekeeper gene to verify the equal amount of cDNA in samples. (c) Western blot analysis for TGF-1-induced activation of Smad3 and ERK1,2 MAPK. Cells were treated with 5 ng/ml of TGF-1 at times indicated. e samples were assayed for phosphorylated forms of Smad3 and ERK1,2. e detection of total forms of Smad3 and ERK1,2 were used to con�rm the amount of each protein loaded.
in regulating cell behavior through the TGF-pathway. SKIP was originally isolated as a putative coactivator protein which interacted with VDR and v-Ski protooncogene, using yeast 2-hybrid screening strategies [16,17]. SKIP is a wellconserved transcriptional adaptor protein that, depending on the cellular context, functions to recruit either activation or repression complexes to mediate multiple signaling pathways involved in the control of cell proliferation and differentiation [18]. However, its precise role in stimulation of tumorigenesis by TGF-1 is poorly understood.
In the present work, we aimed to evaluate the role of SKIP on TGF-1-induced MMP-9 expression and their implications in the regulation of intracellular signal transduction involved. Results obtained revealed that SKIP participates in TGF-1-induced MMP-9 expression in transformed cells. SKIP depletion enhances TGF-1-induced MMP-9 production concomitantly with an increment of TGF-1activated ERK1,2 signaling, implying the role of SKIP in TGF-1-induced expression of MMP-9 in mouse transformed keratinocytes.

Cell Culture and Transfection
Procedures. e transformed mouse keratinocyte cell line (PDV) was kindly provided by Dr. M. Quintanilla (IIB-CSIC, Spain) and cultured as described previously [11]. For stable transfections, PDV cells (∼10 6 ) seeded in 60 mm plates were transfected with 5 g of either antisense mouse-SKIP plasmid or empty pCDNA3 vector using Superfect (Qiagen, Hilden, Germany) following the manufacturer's instructions. Transfected cells were selected by growing in medium containing 10% fetal bovine serum and 400 g/mL of G418 for two weeks. Individual clones were isolated by cloning rings. Transient transfections to analyze MMP-9 promoter activity, Gal4-Elk1 and SRE-luc, were performed as previously described [11,19]. Fire�y luciferase activity (Promega, Adison, WI, USA) was standardized for -galactosidase activity (Tropix, Bedford, MA, USA).

Western Blot, Zymography, and RT-PCR Assays.
Western blots were performed as described elsewhere [11]. Gelatinase activity was assayed in serum-free medium conditioned for 24 h in cell cultures treated or not with TGF-1. Conditioned media were subjected to SDS-PAGE zymography in gels containing 1 mg/ml gelatine, as reported by Santibáñez et al., 2002 [11].

Statistics.
Data are given as means (±SEM) from at least three independent experiments. Asterisks ( * ) denote signi�cant differences at a value of < 0.05 for experimental groups being compared with control in the absence of TGF-1, while ��� denotes signi�cant differences at value of < 0.05 for experimental groups being compared with cells in the presence of TGF-1 only, as determined by Student's t-test.

Effect of TGF-1 on MMP-9 Expression and Smad3/ ERK1,2 Activation in PDV Cells.
To determine the MMP-9 activity aer TGF-1 treatment, serum-free conditioned media from PDV cells were assayed by gelatin zymography. As shown in Figure 1(a), zymography revealed that TGF-1 stimulates the production of a 92-kDa MMP-9 protein in a dose-dependent manner in PDV cells treated with increasing concentrations of TGF-1 (1, 5, and 10 ng/ml) for 24 hours. Semiquantitative RT-PCR analysis revealed that in PDV cells MMP-9 mRNA was upregulated aer 24 hours ( Figure  1(b)) of treatment with TGF-1 (5 ng/ml) compared with constantly expressed levels of GAPDH mRNA, which served as an internal control.
Further on, signaling pathways activated by TGF-1 were investigated. Phosphorylated ERK1,2 and Smad3 proteins were detected in PDV cells by western blot using phospho-speci�c polyclonal antibodies. As shown in Figure 1(c), western blot revealed that 5 ng/ml of TGF-1 activated the phosphorylation of ERK1,2 as early as 5 minutes and peaking at 30 minutes during the time course tested. Also, TGF-1 induced the phosphorylation of Smad3 aer 5 minutes with maximum activation at 120 minutes.

SKIP Depletion Enhances TGF-1-Induced MMP-9
Expression/Secretion. Given that we have previously shown that TGF-1 promotes expression of MMP-9 in PDV cells [11], our next goal was to analyze whether SKIP modulates TGF-1-induced MMP-9 production. erefore, PDV cells were stably transfected with a SKIP antisense construction, and the secretion and expression of MMP-9 in these cells were studied. Figure 2(a) shows a PDV clone with reduced expression of SKIP, designated as AS-S, in comparison with control PDV/Mock (PDV/M) cells, transfected with control vector. As shown in Figure 2(b), and as detected by zymography, stimulation of MMP-9 production by TGF-1 was strongly enhanced in AS-S cells relative to stimulated control cells. is result paralleled with the result obtained by RT-PCR analysis, where the expression of the MMP-9 transcript was enhanced in AS-S cells under TGF-1 treatment compared to PDV/Mock cells (Figure 2(c)). Also, the effect of SKIP depletion on the ability of TGF-1 to transcriptionally stimulate MMP-9 expression was evaluated by expressing a reporter construct containing the luciferase gene under the control of the MMP-9 promoter. It was observed that the lower expression of SKIP in AS-S is correlated with the enhancement of TGF-1-induced transactivation of MMP-9 promoter, whereas the ectopic overexpression of SKIP in control cells inhibited the stimuli of TGF-1 on MMP-9 promoter activity. Moreover, the transfection with SKIP in AS-S also reduced the response of cells to TGF-1 (Figure 3).  activation of signal transduction by TGF-1. Additionally, since TGF-1 induces MMP-9 through ERK1,2 in PDV cells [11], we investigated whether SKIP depletion affects ERK1,2 activation. As expected, the depletion of SKIP in PDV cells (AS-S) decreased TGF-1-mediated induction of Smad3 phosphorylation (Figure 4(a)) Also, signi�cant changes were observed in the TGF-1-induced ERK1,2 phosphorylation levels in AS-S cells in respect to the control. A noteworthy enhancement of downstream ERK1,2 signalling activation by TGF-1 was also noticed, since both transactivation of Elk-1 transcription factor and serum response elements (SRE) was greatly incremented in AS-S cells in response to TGF-1. is result was obtained aer cell transfection using the hybrid Elk1-Gal4 and serum response element (SRE-luc) reporter systems (Figures 4(b) and 4(c)).

e TGF-1-Increased MMP-9 Expression in Downregulated SKIP Cells is Mediated by ERK1,2 Signaling.
In order to determine the role of ERK1,2 signalling on the enhancement of TGF-1 induced MMP-9 by SKIP downregulation, experiments using MEK1,2 chemical inhibitor PD98059 were performed. As shown in Figure 5(a), the treatment of AS-S cells with PD98059 drastically inhibited SRE transactivation by TGF-1. e inhibition of ERK1,2 signalling strongly decreased the capacity of TGF-1 to increase MMP-9 production in AS-S determined either by RT-PCR or Zymography analysis (Figure 5(b)). Furthermore, PD98059 inhibited the TGF-1-stimulated MMP-9 in PDV/M control cells ( Figure 5(c)).

Discussion
TGF-1 is a potent inductor of MMP-9 production in transformed cells, and although the participation of Smad3 and ERK1,2 in the regulation of TGF-1-induced MMP-9 expression has been reported [11,12], the modulation mechanisms of this signaling pathway are not well elucidated.
In the present work, we analyzed the role of SKIP in the production of MMP-9 induced by TGF-1. We observed that the depletion of SKIP, by antisense strategy, provoked an increment in the production of MMP-9 in response to TGF-1 (Figures 2 and 3). Even though it was reported that SKIP modulates Smad3 in TGF-1 signal transduction [15], there is no data about TGF-1 response genes, such as MMP-9. e results shown here represent new �ndings, which imply that SKIP is required for the regulation of MMP-9 induction by TGF-1. SKIP is postulated as a negative and/or positive transcriptional regulator and appears to modulate a number of key signaling pathways involved in the control of cell proliferation and differentiation, and as such may play a role in oncogenesis and development. In that way, the regulatory role of SKIP on Smad3 signaling was recently reported [15], and as we also noticed, the down-regulated expression of SKIP provokes low activation of Smad3 aer TGF-1 treatment (Figure 4(a)). is is in agreement with this report, even though this study was performed by overexpressing SKIP.
In addition to the known fact that TGF-1 activates a plethoric signal transduction beyond Smads, including ERK1,2 MAPK [3,20], we previously reported that this growth factor induces the expression of MMP-9 by activating ERK1,2 signaling [11]. e results presented here demonstrate that SKIP reduction produced signi�cant changes in TGF-1-induced ERK1,2 phosphorylation (Figure 4(a)) and also incremented the ERK1,2-dependent activation of downstream signaling initiated by TGF-1. Enhanced activation of the transcription factor Elk-1, also named ETS domaincontaining protein Elk-1, as well as enhanced transactivation of the cis-serum response elements (SRE) was produced in AS-S cells in response to TGF-1 (Figures 4(b) and 4(c)). It has been demonstrated that the active phospho-Elk1 binds to the MMP-9 promoter and induces the transcription of MMP-9, as well as that MMP-9 promoter contains SRE Ets box binding sites that are essential for MMP-9 expression [21]. us, we may speculate that the depletion of SKIP enhances the capacity of TGF-1-activated ERK1,2 to induce the transactivation of Elk-1-SRE axis, and thereby increase the expression of MMP-9 in cells following TGF-1 treatment. e activation of ERK1,2 is crucial for the MMP-9 production in AS-S cells, as demonstrated by MEK1,2 inhibition with its chemical inhibitor PD98059, which in turn inhibited the transactivation of SRE ( Figure 5). Even though SKIP depletion produces a reduction in Smad3 signaling (Figure 4(a)), we did not perform experiments to determine if the solely Smad3 inhibition can mimic SKIP depletion. Interestingly, other authors have reported that both Smad3 and ERK1,2 mediate TGF-1-induced MMP-9 expression [22,23]. It is possible that the participation of both Smad3 and ERK1,2 in the effect of TGF-1 on MMP-9 expression may be tissue or cell-type speci�c. In addition, missense mutations of the Smad3 gene and knockout Smad3 mice show an increase in the MMP-9 production [13,14], implying that in some cases Smad3 may be a negative regulator of MMP-9 expression, although no other signal transduction pathways further than Smad3 were analyzed in these reports. Interestingly, the knockdown of SKIP with iRNA enhances TGF-1-induced MMP-9 in the prostate carcinoma cells PC-3 (Villar et al. unpublished results), suggesting that SKIP is also able to regulate MMP-9 expression in other cancer cells.
Intriguingly, we have reported that SKIP is necessary for the expression of the urokinase type plasminogen activator, uPA, a key proteinase involved in cancer cell invasion and metastasis [24], but conversely to MMP-9, SKIP depletion decreased the induction of uPA by TGF-1, which suggested a differential involvement of SKIP in TGF--induced cancer proteinases. Additionally, recently we have also reported that Smad3 activity was highly required in the induction of uPA expression by TGF-1 [25], which might explain the divergent role of SKIP on uPA and MMP-9 expression, since in our cell model Smad3 does not play an important role in TGF-1 induced MMP-9 (data not shown). Further studies are necessary to elucidate in deep this differential effect of SKIP.
Given that SKIP in PDV cells appears to be necessary for the regulation of ERK1,2 signaling, the mechanism underlying the increase of ERK1,2 activation remains to be resolved. Nonetheless, SKIP might participate as a control switch in the �ne tuning of signal transduction of TGF-1 in cancer cells, acting as a regulator of Smad3 and/or ERK1,2 signals, by coordinating the magnitude and duration of these signals according to the cellular requirements. In conclusion, our results provide evidence that SKIP mediates TGF-1-induced activation of ERK1,2 downstream signaling, resulting in an enhancement of TGF-1-induced MMP-9 expression in epidermal transformed keratinocytes.