Interleukin-6 drives melanoma cell motility through p38α-MAPK-dependent up-regulation of WNT5A expression

Abstract

Extensive research has demonstrated a tumor-promoting role of increased WNT5A expression in malignant melanoma. However, very little light has been shed upon how WNT5A expression is up-regulated in melanoma. A potential regulator of WNT5A expression is the pro-inflammatory cytokine Interleukin (IL)-6, which shares the ability of WNT5A to increase melanoma cell invasion. Here, we investigate whether IL-6 can promote melanoma cell motility through an increased expression of WNT5A. We clearly demonstrate that the WNT5A-antagonistic peptide Box5 could inhibit IL-6-induced melanoma cell migration and invasion. Furthermore, IL-6 stimulation of the human melanoma cell lines HTB63 and A375 increased the expression of WNT5A in a dose-dependent manner. To identify the signaling mechanism responsible for this up-regulation, we explored the involvement of the three main signals induced by IL-6; STAT3, Akt and ERK 1/2. Of these, only STAT3 was activated by IL-6 in the melanoma cell lines tested. However, the STAT3 inhibitor S3I-201 failed to inhibit IL-6-induced WNT5A up-regulation in HTB63 and A375 cells. Nor did STAT3 siRNA silencing affect the expression of WNT5A. In search of an alternative signaling mechanism, we detected IL-6-induced activation of p38-MAPK in HTB63 and A375 cells. The p38-MAPK inhibitor SB203580 abolished the IL-6-induced WNT5A up-regulation and blocked IL-6-induced melanoma cell invasion. The latter effect could be rescued by the addition of recombinant WNT5A. Notably, immunoprecipitation analysis revealed that only the p38α-MAPK isoform was activated by IL-6, and subsequent siRNA silencing of p38α-MAPK abolished the IL-6-induced up-regulation of WNT5A. Taken together, we demonstrate a novel link between the two melanoma pro-metastatic agents IL-6 and WNT5A explaining how IL-6 can increase melanoma cell invasion and thus promote the metastatic process. This finding provides a basis for future therapeutic intervention of melanoma progression.

Keywords: Cell motility; Interleukin-6; Melanoma; STAT3; WNT5A; p38 MAPK.

Figure 1

The WNT5A‐antagonistic peptide Box5 inhibits IL‐6‐induced melanoma cell migration. (A–D) The panels show transwell migration assay results of human WM852, HTB63, A375 and A2058 melanoma cells treated with either Carrier (0.1% BSA in PBS) alone or recombinant IL‐6 (20 ng/ml) in the absence or presence of Box5 (100 μM). The migration assays were performed over 24 h (n = 4) and analyzed as described in Materials and methods. The results are given as means and S.E.M; *, p < 0.05; **, p < 0.01.

Figure 2

Analysis of endogenous WNT5A expression levels in melanoma cell lines. (A) Western blot analysis showing siRNA silencing of endogenous WNT5A in HTB63 cells. Cells were transfected with either negative control siRNA (NC siRNA; 100 nM), anti‐WNT5A‐siRNA #1 (WNT5A siRNA #1; 100 nM) or anti‐WNT5A‐siRNA #2 (WNT5A siRNA #2; 100 nM) and incubated for 48 or 72 h. Two protein bands in the presumed WNT5A region were clearly detected in NC siRNA transfected cells, however only the intensity of the upper band was reduced following transfection with either WNT5A siRNA #1 or WNT5A siRNA #2. Recombinant WNT5A (rW5A) was used as a positive control and was resuspended in the lysate from the WNT5A‐negative human breast cancer cell line MDA‐468. The lower panel shows densitometric analyses of the siRNA effects on WNT5A protein expression normalized against β‐actin (n = 4). (B) qPCR analysis of WNT5A mRNA expression in HTB63 cells treated as described in (A). The relative WNT5A mRNA expression level was normalized against TATA‐binding protein (TBP) mRNA expression (n = 4). (C) Western blot analysis showing the levels of endogenous WNT5A protein expression in the human melanoma cell lines WM852, HTB63, A375 and A2058. α‐tubulin was used to ensure equal loading of all samples. A representative blot from four separate experiments is shown. (D) qPCR analysis of endogenous WNT5A mRNA expression levels in human WM852, HTB63, A375 and A2058 cells. Samples were normalized against the mRNA of the housekeeping gene TBP. The results are presented as relative to WNT5A mRNA expression of WM852 cells (n = 3). The results are given as means and S.E.M; ***, p < 0.001.

Figure 3

Effects of recombinant IL‐6 stimulation on WNT5A protein and WNT5A mRNA expression. (A) Western blot analysis of WNT5A expression in the human melanoma cell lines WM852, HTB63, A375 and A2058, which were treated as indicated in the figure with Carrier (0.1% BSA in PBS) alone or recombinant IL‐6 (20 ng/ml) for 24 h α‐tubulin was used to ensure equal loading of all samples. A representative blot from 4 separate experiments is shown. (B) qPCR analysis of the WNT5A mRNA expression in WM852, HTB63, A375 and A2058 cells treated as described in (A). The samples were normalized according to the expression of the housekeeping gene TATA‐binding protein (TBP). The IL‐6‐induced effects were presented as relative to the Carrier treated cells in the same experiment (n = 4). (C–D), Western blot analysis of WNT5A expression in HTB63 and A375 cells treated with either Carrier or recombinant IL‐6 (2 or 20 ng/ml). Densitometric analyses were performed, and the data were then normalized against α‐tubulin used as a loading control (HTB63, n = 5; A375, n = 4). The results are given as means and S.E.M; *, p < 0.05; **, p < 0.01 ***, p < 0.001.

Figure 4

Recombinant IL‐6 stimulation triggers STAT3 but not Akt or ERK1/2 activation in melanoma cells. (A) Western blot analysis of tyrosine‐705 phosphorylation of STAT3 (p‐STAT3, Tyr705) in the human melanoma cell lines WM852, HTB63, A375 and A2058 treated with Carrier (0.1% BSA in PBS) alone or recombinant IL‐6 (20 ng/ml) for 60 min. Densitometric analyses were performed, and the data were then normalized against total STAT3 (t‐STAT3) (n = 3). (B–C) Western blot analysis of Akt phosphorylation (p‐Akt) and ERK1/2 phosphorylation (p‐ERK1/2) in the human melanoma cell lines WM852, HTB63, A375 and A2058 treated as described in (A). Total Akt and total ERK1/2 were used to ensure equal loading of all samples. Representative blots from three separate experiments are shown in (B) and (C). The results are given as means and S.E.M.

Figure 5

The recombinant IL‐6‐induced expression of WNT5A is not regulated by STAT3 activity. Western blot analysis of tyrosine‐705 phosphorylation of STAT3 (p‐STAT3, Tyr705) in human HTB63 (A) and A375 (C) melanoma cells treated with either Carrier (0.1% BSA in PBS) alone or recombinant IL‐6 (20 ng/ml) for 24 h in the absence (only DMSO) or presence of S3I‐201 (100 μM dissolved in DMSO). Densitometric analyses were performed, and the data were then normalized against total STAT3 (t‐STAT3; n = 4). Western blot analysis of WNT5A expression in HTB63 (B) and A375 (D) cells treated as described in A and C, respectively. Densitometric analyses were performed, and the data were then normalized against α‐tubulin (n = 4). The results are given as means and S.E.M; *, p < 0.05; **, p < 0.01; ***, p < 0.001.

Figure 6

Recombinant IL‐6‐induced p38‐MAPK activation mediates elevated WNT5A expression and increased melanoma cell invasion. (A) Western blot analysis of p38‐MAPK phosphorylation (p‐p38) in the human melanoma cell lines WM852, HTB63, A375 and A2058 treated with Carrier (0.1% BSA in PBS) alone or recombinant IL‐6 (20 ng/ml) for 60 min. Densitometric analyses were performed, and the data were then normalized against total p38‐MAPK (p38; n = 3). (B) Western blot analysis of p38‐MAPK phosphorylation (p‐p38) in HTB63 cells incubated in the absence or presence of recombinant IL‐6 (20 ng/ml) for 30 min, 1 h, 2 h, 3 h or 4 h. Densitometric analyses were performed, and the data were then normalized against total p38‐MAPK (p38; n = 3). (C–D) Western blot analysis of WNT5A expression in HTB63 and A375 cells treated with either Carrier (0.1% BSA in PBS) alone or recombinant IL‐6 (20 ng/ml) for 24 h in the absence (only DMSO) or presence of SB203580 (10 μM dissolved in DMSO). Densitometric analyses were performed, and the data were then normalized against α‐tubulin (n = 4). (E) Matrigel invasion of human HTB63 melanoma cells treated with recombinant IL‐6 (20 ng/ml, dissolved in 0.1% BSA in PBS) in the absence or presence of SB203580 (10 μM dissolved in DMSO) and recombinant WNT5A (rW5A, 0.2 μg/ml, dissolved in 0.1 mM EDTA, 0.5% CHAPS, 0.1% BSA in PBS) or their respective controls, Carrier (0.1% BSA in PBS), DMSO and Vehicle (0.1 mM EDTA, 0.5% CHAPS, 0.1% BSA in PBS). The invasion assays were performed over 24 h (n = 4) and analyzed as described in Materials and methods. The results are given as means and S.E.M; ns, not significant; *, p < 0.05; **, p < 0.01; ***, p < 0.001.

Figure 7

p38α‐MAPK is the only isoform activated by recombinant IL‐6 stimulation and shown to be responsible for the IL‐6‐induced increase in WNT5A expression. (A) Western blot analysis of activated p38‐MAPK isoforms in human HTB63 melanoma cells treated with either Carrier (0.1% BSA in PBS) alone or recombinant IL‐6 (20 ng/ml) for 30 min. After treatment, the cells were lysed and either directly analyzed for their contents of total p38‐MAPK (p38) and α‐tubulin or used for control IgG (IP:IgG) or phospho‐p38‐MAPK (IP:p‐p38) immunoprecipitation. These immunoprecipitates were subsequently analyzed for their contents of the p38‐MAPK isoforms (p38α, p38β, p38γ or p38δ). Representative blots from three separate experiments are shown. (B) Western blot analysis of p38α‐MAPK (p38α), p38β‐MAPK (p38β) and WNT5A expression in HTB63 cells transfected with either negative control siRNA (NC siRNA; 50 nM), anti‐p38α‐siRNA (p38α siRNA; 50 nM) or anti‐p38β‐siRNA (p38β siRNA; 50 nM) and incubated for 48 h prior to treatment with either Carrier (0.1% BSA in PBS) alone or recombinant IL‐6 (20 ng/ml) for 24 h. Densitometric analyses of WNT5A expression were performed, and the data were then normalized against α‐tubulin (n = 4). The results are given as means and S.E.M; ns, not significant; ***, p < 0.001.

Figure 8

Schematic overview of the signaling events that link the two pro‐metastatic agents IL‐6 and WNT5A in melanoma. The broken arrow indicates a mechanism demonstrated by other researchers that together with the present findings suggests a positive feed back‐loop whereby WNT5A indirectly could regulate its own expression level.