Deciphering growth factor signalling in melanoma cells

Melanomas are derived from epidermal melanocytes and are highly metastatic. Although they are the least common type of skin cancer, they account for ~75% of skin cancer-related deaths. The discovery of aberrant v-Raf murine sarcoma viral oncogene homolog B1 (BRAF) in ~50% of melanomas prompted the development of selective inhibitor(s) towards this signalling intermediate. Of all BRAF mutations, BRAFV600E is the most prevalent. Following targeted therapy with vemurafenib, an ATP-competitive inhibitor of BRAFV600E, BRAFV600E melanomas regressed and the patient had an improved quality of life. However, these tumours developed resistance, leading to the eventual death of the patient. Since then, overcoming the development of acquired resistance in response to BRAFV600E inhibition has become one of the major challenges in melanoma research.

One of the mechanisms of resistance involves hepatocyte growth factor (HGF), which is secreted by surrounding stromal cells in BRAFV600E-inhibitor (BRAFi) resistant melanomas, resulting in the reactivation of the mitogen-activated protein kinase (MAPK) and phosphatidylinositol 3-kinase-Ak strain transforming (PI3K-AKT) pathways. Furthermore, epidermal growth factor receptor (EGFR) was shown to be upregulated in BRAF-inhibitor resistant melanoma cells, leading to increased cell proliferation, invasion and metastasis. EGFR ligands include epidermal growth factor and transforming growth factor alpha (TGF-alpha). TGF-alpha is a powerful mitogenic growth factor (GF) that is formed in the skin. Therefore, this study first investigated the ability of GFs (HGF and TGF-alpha) to stimulate MAPK and PI3K-AKT signalling pathways in four distinct melanoma cell lines (MM418-C1, C32, MM329 and D24).

These melanoma cell lines were chosen based on their BRAF mutational status and origin of tumour. While both MM418-C1 and C32 cells harbours the BRAFV600E mutation, the other two cell lines (MM329 and D24) do not. In addition, both MM418-C1 and MM329 come from primary tumours, while C32 and D24 were derived from secondary tumours. While HGF signalled through BRAF in both BRAFV600E melanoma cells (MM418-C1 and C32), only TGF-alpha did in D24 cells (BRAFWT melanoma cells). However, these GF-activated BRAF signals were not related to the increased expression levels of the downstream effector extracellular-signal regulated kinase (ERK). Instead, we observed that the PI3K-AKT pathway can activate ERK. Furthermore, HGF or TGF-alpha also stimulated the activity of both the stress-mediated MAPK pathways (p38 MAPK and c-Jun N-terminal kinase (JNK)) in both MM418-C1 and D24 cells. Of interest was that neither HGF nor TGF-alpha stimulated the MAPK signalling pathway in MM329 cells.

Cytotoxicity of the inhibitors was determined using the cell counting kit-8 (CCK-8) method prior to inhibition studies. This was to ensure that the inhibitor doses used did not cause >50% cell death and hence, any changes observed were due to inhibitory effect of the inhibitors and not because of the triggering of cell death pathways. We showed that PLX4032 (100-400 nM) (BRAFV600E inhibitor) caused significant cell death of BRAFV600E (MM418-C1 and C32), while enhancing the proliferation of BRAFWT melanoma cells (MM329 and D24). On the contrary, U0126 (MAPK/ERK kinase (MEK) inhibitor) at >5 microM caused significant cell death in all four melanoma cell lines. Interestingly, SB202190 (p38 MAPK inhibitor) and SP600125 (JNK inhibitor) triggered a hormetic effect in BRAFWT melanoma cells, but not in BRAFV600E cells.

The effect of specific inhibitors on the GF-induced signalling pathways in these cells were also examined. Inhibition studies showed that pathway crosstalk occurred between the signalling pathways in these cells. In BRAFV600E melanoma cells, PLX4032 or PF-04691502 not only reduced GF-stimulated p-BRAF or p-AKT levels, respectively, but also decreased p-ERK2 levels. Meanwhile, p38 MAPK inhibition led to increased p-ERK2 levels in BRAFWT melanoma cells. Furthermore, both GF-stimulated p-p38 and/or p-JNK levels rose when PI3K was inhibited in all four melanoma cell lines.

In addition, both GFs were tested for their abilities to enhance the migration of these melanoma cell lines. Both GFs stimulated the BRAFV600E and BRAFWT melanoma cells differently. However, PI3K inhibition caused a greater impairment of migration in these cells when compared to that of BRAF or MEK inhibition regardless of the cell's BRAF mutational status and tumour origin. This suggests that the PI3K-AKT pathway plays a crucial role in the migration of melanoma cells, which occurred via the activation of AKT1.

Silencing RNA (siRNA) experiments were performed to confirm the results from the inhibition studies. While p-BRAF levels were reduced in BRAF-siRNA-transfected cells, both p-AKT and p-ERK2 levels were increased in the presence of HGF in the C32 and D24 melanoma cells. This suggests that HGF can stimulate the PI3K-AKT pathway, which in turn activates ERK when its upstream activator BRAF is suppressed, and therefore contributing to the mechanism of vemurafenib drug resistance.

Taken together, our studies suggest that the GF-activated MAPK-ERK pathway is the major pathway involved in the proliferation and survival of melanoma cells. However, when BRAF is inhibited, GFs can activate the PI3K-AKT pathway which can in turn stimulate ERK, contributing to the mechanism of resistance. While this PI3K-AKT-ERK axis is important for providing proliferative and survival signals to these melanoma cells, the other axis of this pathway is involved in cell migration. Although the stress-mediated MAPK pathways (p38 MAPK and JNK) were deemed insignificant in two of the melanoma cell lines, their importance was more pronounce when both MAPK-ERK and PI3K-AKT pathways were blocked. Our studies demonstrated the patterns of GF signalling in four distinct cell lines and would add to the available knowledge of how melanoma cells behave under the influence of GFs and/or specific pathway inhibitors at the molecular level. Exploiting these similarities and differences in melanoma signalling may give rise to the development of novel therapeutic targets.