Epidermal growth factor receptor transactivation is implicated in IL-6-induced proliferation and ERK1/2 activation in non-transformed prostate epithelial cells

Abstract

Epidermal growth factor receptor (EGF-R) is a receptor tyrosine kinase that can be activated by molecules other than its cognate ligands. This form of crosstalk called transactivation is frequently observed in both physiological and pathological cellular responses, yet it involves various mechanisms. Using the RWPE-1 cell line as a model of non-transformed prostate epithelial progenitor cells, we observed that interleukin-6 (IL-6) is able to promote cell proliferation and ERK1/2 activation provided that EGF-R kinase activity is not impaired. Treatment with GM6001, a general matrix metalloprotease inhibitor, indicated that IL-6 activates EGF-R through cleavage and release of membrane-anchored EGF-R ligands. Several inhibitors were used to test implication of “a disintegrin and metalloprotease” ADAM10 and ADAM17. GW280264X that targets both ADAM10 and ADAM17 blocked IL-6-induced proliferation and ERK1/2 phosphorylation with same potency as GM6001. However, ADAM10 inhibitor GI254023X and ADAM17 inhibitor TAPI-2 were less efficient in inhibiting response of RWPE-1 cells to IL-6, indicating possible cooperation of ADAM17 with ADAM10 or other metalloproteases. Accordingly, our findings suggest that IL-6 stimulates shedding of EGF-R ligands and transactivation of EGF-R in normal prostate epithelial cells, which may be an important mechanism to promote cell proliferation in inflammatory prostate.

1. Introduction

Growth factors, their receptors and downstream kinases are often over-expressed or mutated in cancer, thereby triggering deregulation of several signal transduction pathways. As a consequence, signaling cascades may be altered in different ways, leading to their constitutive activation by autocrine stimulation or receptor transactivation [1,2]. Significant advances have been recently made in the field of prostate cancer development, pointing to key roles of epidermal growth factor receptor (EGF-R) family and cytokine signaling, in particular interleukin-6 (IL-6), in androgen-independent stages and metastasis [1,3,4]. In contrast, signaling events that might affect early steps of prostate cancer remain poorly understood. Paracrine stimulation of EGF-R has been suspected to participate to proliferation of androgen- dependent prostatic tumors, whereas IL-6 provided by inflammatory cells or reactive stroma might stimulate growth of epithelial cells primed to tumor conversion by initiating mutations [5,6]. Better knowledge of the function and regulation of these signaling cascades in normal prostate progenitor cells should help understand which molecular mechanisms are involved in prostate carcinogenesis. In this context, the major purpose of the present study was to investigate on signaling mechanisms implicating EGF-R and IL-6 in the proliferation of normal prostate epithelial cells.

EGF-R belongs to the ErbB receptor family that consists of four closely related receptors: ErbB1/EGF-R, ErbB2/neu/HER2, ErbB3, and ErbB4. Eleven ligands have been identified that show overlapping but specific binding to ErbB proteins [7,8]. In addition, there is increasing evidence that ErbB receptors can induce intracellular signals not only in response to cognate ligands but also following transactivation by cross-talking receptors and downstream signaling pathways [9]. All four ErbB receptors are expressed in the mouse [10] or human prostate [11,12], as well as most of their cognate ligands [11,13]. Consequently, ErbB receptors are physiologically activated in the normal prostate [10]. In the human prostate, EGF-R is mainly expressed by proliferative epithelium, whereas expression of its ligands, EGF and TGF-α, is found in both epithelial and stromal cells and increases in hyperplastic prostate [14–16]. Interestingly, mice deficient for EGF or TGF-α show defects in the formation of prostatic epithelial ducts during fetal development [17]. Moreover, sustained EGF treatment promotes physiological prostate development in newborn rats [18]. Collectively, data support an important role for EGF-R and its ligands in prostate epithelium growth.

IL-6 is a pleiotropic cytokine that is strongly implicated in regulating immune and inflammatory responses. IL-6 mediates its activity through a receptor complex composed of two subunits: the α-receptor subunit that specifically binds IL-6 (IL-6-Rα), and the signal transducing subunit gp130 [19]. Besides its role as an autocrine proliferation and survival factor in advanced prostate tumor cells [20], IL-6 is gaining interest as a pro-inflammatory cytokine since a wealth of studies highlighted links between prostate inflammation and tumorigenesis [21,22]. In normal prostate, IL-6 and IL-6-Rα are expressed in basal epithelium, whereas gp130 expression is increasing in hyperplasic basal epithelium [23,24]. Finally, IL-6 activates Stat3 and stimulates proliferation in primary epithelial cells from human prostate [20].

In the normal prostate, proliferation compartment is restricted to the basal epithelium where a small population of stem cells can give rise to transient/amplifying (TA) cells. These progenitor cells have strong proliferation potential, but limited self-renewal capabilities, and ultimately differentiate into secretory cells [25]. Here we used the immortalized human prostate cell line RWPE-1 as a model of prostate TA cells [26]. RWPE-1 cells express TA cell markers, undergo in vitro acinar differentiation in response to androgen and do not form tumors when injected into nude mice [27–29]. Interestingly, they show dose- dependent proliferation in response to EGF [26], and IL-6 can stimulate RWPE-1 cell proliferation, albeit with lower efficiency than EGF [30]. Therefore, we used this cell line to study IL-6 response and possible link with EGF signaling in non-transformed prostatic epithelial cells.

2. Materials and methods

2.1. Reagents

EGF and IL-6 (Peprotech) were dissolved to a concentration of 10 μg/ml in Keratinocyte-Serum Free Medium (K-SFM, Invitrogen) and stored in aliquots at −20 ° C. WST-1 cell proliferation reagent was obtained from Roche Diagnostics. TAPI-2 and Insolution inhibitors U0126, PP2, PD153035, Jak Inhibitor I and GM6001 were purchased from Calbiochem. ADAM inhibitors GI254023X and GW280264X were kindly provided of Dr. A. Ludwig (Institute for Molecular Cardiovascular Research, University Hospital Aachen, Germany).

2.2. Cell culture and proliferation assays

RWPE-1 cell line, a gracious gift from Dr. N. Allioli (Institute of Functional Genomics, Lyon), was maintained in K-SFM supplemented with 5 ng/ml EGF and 50 μg/ml bovine pituitary extract (complete K- SFM, Invitrogen). Dose–response tests of cytokines or inhibitors were performed in miniaturised proliferation assay as follows: 5000 cells were seeded in 96-well culture plates in K-SFM supplemented as specified in the text (200 μl per well). After 5 days of culture, WST-1 reagent was added into each well for a 4-hour incubation at 37 °C according to manufacturer’s instructions and optical density at 450 nm was measured using a Victor X5 plate reader (Perkin– Elmer). All cultures were performed in a humidified incubator at 37° in an atmosphere of 5% CO2.

2.3. Cell stimulation and signaling studies

Cells were washed free of growth factors then incubated in basal K-SFM. After 3 h of starvation, medium was replaced by fresh K-SFM supplemented by EGF (10 ng/ml) or IL-6 (10 ng/ml) for various times, as specified in the text. To stop stimulation, medium was removed and changed with cold PBS supplemented by 1 mM Na3VO4 (Sigma- Aldrich). Inhibitors were added to cultures 30 min before EGF or IL-6; other conditions (inhibitor concentrations and stimulation times) are specified in the text. TAPI-2 was dissolved in ethanol whereas other inhibitors were dissolved in DMSO. Therefore, control cultures were performed in the presence of ethanol or DMSO at final concentrations equal to that in corresponding experimental points.

2.4. Cell lysis and immunoblotting

Cells were removed from culture surface by using a cell scrapper, centrifuged and lysed in cold radio immunoprecipitation assay buffer (1% deoxycholic acid, 1% Triton X-100, 0.1% SDS, 50 mM Tris- base, 150 mM NaCl, 20 mM EDTA, pH 7.4) containing protease inhibitor cocktail (Roche Applied Science) and 2 mM Na3VO4. Insoluble material was removed by centrifugation, and the protein concentration was determined by using the Protein Assay solution (Bio-Rad). Proteins from equalized cell lysates were separated on a SDS-polyacrylamide gel, transferred to nitrocellulose membrane, and blotted with specific primary antibodies: anti-phospho-Stat3, anti-Stat3, anti-phospho-EGF-R (Tyr845) or anti-EGF-R (Cell Signal- ing Technologies); anti-phospho-ERK or anti-ERK1 (Santa Cruz). Antibody binding was visualized using HRP-conjugated secondary antibodies (Sigma-Aldrich) and ECL+ reagent (GE Healthcare).

2.5. Statistical analysis

Results were expressed as the mean±standard error of mean for at least 3 independent experiments. Statistical significance was determined by Student’s t-test (paired-data analysis). P valuesb 0.05 were considered as statistically significant.

3. Results

3.1. ERK1/2 activation drives proliferative response of RWPE-1 cells to EGF or IL-6

Cells were cultured in 96-well plates in K-SFM supplemented with BPE for 5 days in the presence of increasing concentrations of EGF or IL-6, then proliferation was assessed using WST-1 reagent. Consistent with previous data [26,30,31], EGF and IL-6 stimulated RWPE-1 cell proliferation in a dose-dependent manner, with a plateau level at 5 ng/ml for each cytokine (242 ± 28% and 139 ± 8%, as compared to control, respectively) (Fig. 1A). Viable cell numeration using Trypan Blue dye was also performed after 4 or 7 days of cultivation in culture flasks, and confirmed higher proliferation rates of RWPE-1 cells in response to EGF than in response to IL-6 (Fig. 1B).

Since persistent ERK1/2 activation was associated to cell prolifer- ation in various models [32], the effects of EGF and IL-6 on ERK1/2 phosphorylation were assessed in time-course studies. For that, exponentially growing RWPE-1 cells were starved of EGF and BPE for 3 h, then stimulated with 10 ng/ml EGF or 10 ng/ml IL-6 for various times, as indicated. Western blotting analysis of cell lysates showed that EGF readily stimulated ERK1/2 phosphorylation reaching a peak at 5 min of EGF stimulation, but persisting for at least 4 h of EGF stimulation (Fig. 1C). In contrast, IL-6 that proved to be a weak mitogen for RWPE-1 cells could only transiently stimulate ERK1/2 phosphorylation (Fig. 1C). Interestingly, IL-6 induced strong and sustained Stat3 phosphorylation under the same conditions, which indicated efficient IL-6 signal transduction (Fig. 2). Therefore, data strongly suggested correlation between ERK1/2 activation and proliferative response in RWPE-1 cells. To test the role of ERK1/2 in mitogenic signaling induced by EGF and IL-6, we used the specific MEK inhibitor U0126. Fig. 1D shows that U0126 inhibited prolifera- tion of RWPE-1 cells in response to EGF or IL-6 in a dose-dependent manner, confirming essential role of ERK1/2 activation in prostate epithelial cell growth.

3.2. ERK1/2 activation and proliferation in response to IL-6 requires EGF-R kinase activity

IL-6 can induce ERK activation through recruitment of the tyrosine phosphatase SHP2 on specific phosphotyrosines of the gp130 subunits constituting the IL-6 receptor. In turn, SHP2 becomes phosphorylated by JAK2 and is able to recruit Grb2-SOS complexes, thereby triggering Ras activation and downstream events leading to ERK1/2 phosphor- ylation and activation [19]. Thus, ERK1/2 phosphorylation in response to IL-6 depends on Jak2 activity, which we verified in RWPE-1 cells by using Jak Inhibitor I (data not shown). However, as shown in Fig. 1C, IL-6 transiently stimulated ERK1/2 phosphorylation while it induced sustained Stat3 phosphorylation. These different kinetics might indicate either that negative regulation occurs downstream of JAK2/ SHP2 to modulate ERK1/2 phosphorylation, or that IL-6 stimulates a JAK2-dependent alternative pathway for activating ERK1/2. Since cytokine receptors have been shown to transactivate EGF-R [33,34], we first asked whether stimulation of ERK1/2 phosphorylation by IL-6 in RWPE-1 cells would require EGF-R activity. For that, 3-hour starved RWPE-1 cells were stimulated by IL-6 for various times in the presence or not of PD153035, a potent inhibitor of EGF-R kinase activity. Fig. 2A clearly shows that blocking EGF-R kinase activity prevented IL-6 to induce ERK1/2 phosphorylation. Moreover, PD153035 similarly inhibited proliferation induced by either EGF or IL-6 in RWPE-1 cells (Fig. 2B), pointing to possible role of EGF-R in IL- 6-induced ERK1/2 activation. Alternately, data might implicate ErbB2 rather than EGF-R since PD153035 can also inhibit other EGF-R family members in intact cells [35] and, in prostate cancer cells, IL-6 stimulation results in association of gp130 subunit of IL-6-R with Erb2, which is essential to ERK1/2 activation [34]. However, we considered this possibility very unlikely since RWPE-1 cells do not express detectable levels of ErbB2 protein [36]. In conclusion, data strongly suggest that IL-6 is able to transactivate EGF-R in non-transformed prostate epithelial cells.

3.3. Src family kinases indirectly contribute to EGF-R transactivation

We first asked whether IL-6 receptor could directly transactivate EGF-R. As previously shown for growth hormone, JAK2 might phosphorylate EGF-R, creating binding site for Grb2/Sos and thereby activating ERK1/2, but this mechanism does not require EGF-R kinase activity [33]. Therefore, our observation that ERK1/2 activation by IL-6 in RWPE-1 cells necessitates EGF-R kinase activity (Fig. 2A) apparently rules out the possibility that JAK2 mediates EGF- R transactivation in response to IL-6. Another mechanism might implicate phosphorylation and subsequent activation of EGF-R by Src family kinases (SFK; reviewed by [37]). To test this possibility, RWPE-1 cells were starved of growth factors for 3 h, then stimulated for various times with EGF or IL-6. Corresponding cell lysates were subjected to western blotting with an antibody specifically detecting phosphorylation of Tyr845, a major SFK substrate on EGF-R [38]. EGF induced strong and sustained Tyr845 phosphorylation, correlating with ERK1/2 activation (Fig. 3A). In contrast, no phosphorylation of Tyr845 could be detected following IL-6 stimulation, while the transient activation of ERK1/2 could be observed. Thus, we concluded that IL-6 did not induce SFK to phosphorylate EGF-R in RWPE-1 cells. Interestingly, the SFK inhibitor PP2 strongly inhibited ERK1/2 phosphorylation induced by IL-6, not that induced by EGF (Fig. 3B), indicating that SFK might be indirectly implicated in EGF-R transactivation.

3.4. Metalloproteases mediate EGF-R transactivation by IL-6

Next, we asked whether EGF-R transactivation by IL-6 would implicate proteolytic release of EGF-R ligands, as previously shown in response to various agonists [39]. EGF-R ligands can be cleaved and released from the cell surface by activation of membrane-bound metalloproteases [40]. This mechanism was tested here by using the general metalloprotease inhibitor GM6001. For that purpose, 3-hour starved RWPE-1 cells were stimulated with 10 ng/ml IL-6 for 5, 10 or 30 min in the presence or absence of 10 μM GM6001. As shown in Fig. 4A, we observed strong inhibition of ERK1/2 phosphorylation in the presence of GM6001, whereas the inhibitor had no effect on EGF induced ERK1/2 phosphorylation (data not shown). Thus, activation of ERK1/2 in response to IL-6 is dependent on both EGF-R activity and proteolytic activity in RWPE-1 cells. Next, we sought to identify which metalloproteases could be implicated in IL-6-induced transactivation of EGF-R. GM6001 has a broad specificity, targeting both matrix metalloproteases and members of ADAM (a disintegrin and metallo- protease) family. ADAM proteases have been implicated in EGF-R ligand shedding, in particular ADAM10 and ADAM17 [9,41]. So, we used 2 molecules that differed in their capacity to block ADAM10 and ADAM17 activities to further characterize IL-6-mediated ERK1/2 phosphorylation in RWPE-1 cells. GW280264X has been shown to preferentially block ADAM17 and to a lesser extent ADAM10, while GI254023X preferentially blocked ADAM10 [42]. Here, using the same protocol as for GM6001, we could observe that GW280264X, but not GI254023X, could strongly impair ERK1/2 phosphorylation upon IL-6 stimulation of RWPE-1 cells, pointing to a role of ADAM17 in this process (Fig. 4B and C, respectively). Consistent with this possibility, the ADAM17-specific inhibitor TAPI-2 was also able to inhibit IL-6- induced ERK1/2 phosphorylation under the same conditions (Fig. 4D). However, TAPI-2 effects were weaker than those of GM6001 or GW280264X, suggesting that ADAM17 might not be the sole protease involved in EGF-R transactivation (Fig. 4E). Finally, the 3 compounds that inhibited IL-6 stimulation of ERK1/2 activation also inhibited IL-6 dependent proliferation, with most prominent effect of GW280264X (Fig. 4F).

4. Discussion

In primary cultures of normal human prostate epithelial cells, IL-6 stimulates proliferation [20], which we confirmed here using RWPE-1 cell line as a model of non-transformed prostate epithelial cells. Proliferation of RWPE-1 cells in response to IL-6 was lower than in response to EGF, correlating with the capabilities of these cytokines to stimulate transient and sustained ERK1/2 phosphorylation in starved cells, respectively. In contrast, IL-6 induces sustained ERK1/2 phosphorylation in vitro or in vivo in advanced prostate cancer cells [34,43]. Therefore, IL-6 stimulation of prostate epithelial cells may result in different outcomes under physiological or pathological conditions. Transactivation of ErbB family receptors has been implicated in both situations, yet distinct mechanisms have been described. Indeed, we have reported here EGF-R transactivation by IL- 6 in non-transformed prostate epithelial cells whereas Qiu et al. have previously shown that IL-6 could induce ErbB2 transactivation in advanced prostate cancer cell lines [34]. It is thus possible that ErbB2 overexpression in prostate cancer cells could be responsible for sustained ERK1/2 activation in response to IL-6 [36,44].

Here, we have shown that the effects of IL-6 on proliferation and ERK1/2 activation in RWPE-1 cells required kinase activity of EGF-R, based on the effects of PD153035. The rapidity at which IL-6 stimulates ERK1/2 phosphorylation was not compatible with tran- scriptional regulation of EGF-R ligand expression. Having also excluded direct crosstalk between IL-6-R and EGF-R, we then focused our study on possible implication of EGF-R ligand shedding from the cell surface following IL-6 stimulation of RWPE-1 cells. Experiments with the general inhibitor of metalloproteases GM6001 confirmed our hypothesis that IL-6 could induce the release of EGF-R ligands from the cell surface, resulting in autocrine/paracrine stimulation of ERK1/2 activity and subsequent cell proliferation. In line with this finding, it was recently shown in endothelial cells that IL-6 causes rapid cleavage of the membrane-anchored pro-form of neuregulin-1 and subsequent ERK1/2 activation through ErbB family members other than EGF-R [45]. Nevertheless, to our knowledge, our study is the first to describe metalloprotease-dependent EGF-R activation induced by IL-6 in non- transformed epithelial cells.
Members of the ADAM family of membrane proteases are responsible for transactivation of EGF-R by proteolytic cleavage of pro-forms of receptor ligands [40]. Depending on cell type, several ADAM proteases share EGF-R ligands and vice versa, yet ADAM10 and ADAM17 are most often implicated in the cleavage of TGF-α, HB-EGF and amphiregulin [41,46]. Our experiments indicated that GW280264X was as potent inhibitor of IL-6-induced ERK1/2 phos- phorylation as GM6001, pointing to possible role of ADAM10 and ADAM17. However, the ADAM10 specific compound GI254023 had no significant effect in our model, ruling out a role of ADAM10, and the ADAM17 inhibitor TAPI-2 had only a mild inhibitory effect on both IL- 6-induced proliferation and ERK1/2 phosphorylation. This suggested that other metalloproteases might act together with ADAM17 upon IL-6 stimulation. Consistent with this possibility, GW280264X also target matrix metalloproteases (A. Ludwig, personal communication). Transcriptome analysis of RWPE-1 cells indicated that amphir- egulin transcripts are much more abundant than those encoding TGF- α or HB-EGF [47]. Amphiregulin is a key factor in epithelial development and may play a role in prostate epithelium proliferation [48,49]. However, although amphiregulin could stimulate EGF-R signaling, anti-amphiregulin neutralizing antibody did not inhibit IL- 6-induced ERK1/2 phosphorylation in RWPE-1 cells (data not shwon). This indicated that either amphiregulin is not implicated in EGF-R transactivation by IL-6 or additional EGF-R ligands are released in response to IL-6. The latter possibility would be consistent with our conclusion that multiple metalloproteases might be activated by IL-6 in RWPE-1 cells.

5. Conclusions

Our data show that IL-6 can stimulate proliferation in non- transformed prostate epithelial cells through metalloprotease activation and subsequent EGF-R transactivation, yet mechanisms underlying this cross-talk remain to be detailed. Nevertheless, importance of IL-6 in prostate epithelial homeostasis remains to be determined. Inflamma- tory microenvironment may result in hyperplastic prostate epithelium, which is now considered to be an early step in prostate cancer progression [50]. Sharp increase in IL-6 expression has been described in hyperplastic prostate, correlating with strong ERK1/2 activation [23,24,51]. Therefore, EGF-R transactivation by IL-6 described here may contribute to excessive epithelial cell proliferation in inflammatory prostate.