M-CSF as a therapeutic target in BRAFV600E melanoma resistant to BRAF inhibitors

Abstract

Background: Disseminated BRAF^V600E melanoma responds to BRAF inhibitors (BRAFi) but easily develops resistance with poor prognosis. Secretome plays a pivotal role during tumour progression causing profound effects on therapeutic efficacy. Secreted M-CSF is involved in both cytotoxicity suppression and tumour progression in melanoma. We aimed to analyse the M-CSF contribution in resistant metastatic melanoma to BRAF-targeted therapies.

Methods: Conditioned media from melanoma cells were analysed by citoarray. Viability and migration/invasion assays were performed with paired melanoma cells and tumour growth in xenografted SCID mice. We evaluated the impact of M-CSF plasma levels with clinical prognosis from 35 metastatic BRAF^V600E-mutant melanoma patients.

Results: BRAFi-resistant melanoma cells secretome is rich in pro-tumour cytokines. M-CSF secretion is essential to induce a Vemurafenib-resistant phenotype in melanoma cells. Further, we demonstrated that M-CSF mAb in combination with Vemurafenib and autophagy blockers synergistically induce apoptosis, impair migration and reduce tumour growth in BRAFi-resistant melanoma cells. Interestingly, lower M-CSF plasma levels are associated with better prognosis in metastatic melanoma patients.

Conclusions: Secreted M-CSF induces a BRAFi-resistant phenotype and means worse prognosis in BRAF^V600E metastatic melanoma patients. These results identify secreted M-CSF as a promising therapeutic target toward BRAFi-resistant melanomas.

Fig. 1. Secreted media from Vemurafenib-resistant melanoma cell lines induces a resistance phenotype in sensitive melanoma cells.

a Schematic representation of the in vitro assays. b Cell viability assay or c Wound healing of M3 and Sk-Mel-28 cells after being cultivated with their own conditioned media (CM) and from the corresponding resistant CM treated with Vemurafenib (1 μM, 48 h). d Representative images of Hoechst nuclear staining of the M3 cell line (x10 microscopic field) and percentage (%) of invasive cells into the matrigel after transwell assay with CM (Vemurafenib 1 μM, 48 h). e (Left) Schematic representation of the in vivo assay. (Right) Wound-healing assay of M3 cells cultivated with 10% of mice plasma from SCID mice with xenografts of A375, M238, A375R, M238R and Sk-Mel-28R melanoma cells treated with Vemurafenib (1 μM, 48 h). Graphs show mean + SEM. n > 3 independent experiments. Statistical analysis was performed using ANOVA and Bonferroni tests (*p < 0.05; **p < 0.01; ***p < 0.001; n.s., non-significant).

Fig. 2. Cytokine secretome profile of Vemurafenib-resistant melanoma cell lines exhibit an increase in pro-tumour cytokines.

a GO analysis based on differentially regulated cytokines on resistant melanoma cells. Each node shows an enriched GO term. Heatmap (OD mean) and cytokine expression graphic of b pro-tumour cytokine profile and c pro-angiogenic cytokine profile. Statistical analysis was performed using the Wilcoxson multiple comparison test.

Fig. 3. M-CSF promotes the acquisition of Vemurafenib-resistance in BRAFV600E melanoma cells.

a Levels of M-CSF in conditioned media of Vem-S and Vem-R cell lines determined by ELISA assay. b Viability and wound-healing assay of sensitive melanoma cells treated with Vemurafenib (1 μM, 48 h) and/or rhM-CSF (200 ng/mL). c Schematic representation of the in vitro assays. d MTT assay or e wound-healing assay of sensitive M3 and SK-Mel-28 cells after exposure to 48 h of their corresponding resistant CM treated with Vemurafenib (1 μM) and/or M-CSF mAb (20 ng/mL). f Wound-healing assay of Sk-Mel-28 cells cultivated with 10% of mice plasma from SCID mice with Sk-Mel-28R xenografts and treated with Vemurafenib (1 μM, 48 h) and/or M-CSF mAb (20 ng/mL). Graphs show mean + SEM. n = 3 independent experiments. Statistical analysis was performed using the ANOVA test, followed by Bonferroni’s multiple comparisons test (*p < 0.05; **p < 0.01; ***p < 0.001; ns, non-significant).

Fig. 4. Autophagic blockers Mibefradil/Chloroquine, Vemurafenib and M-CSF mAb-combined therapy induce apoptotic cell death and impair migration of BRAFV600E-resistant melanoma cells.

a Monitoring autophagic flux by mRFP-EGFP-LC3B assay in resistant melanoma cells after M-CSF mAb, Mibefradil or Vemurafenib treatment. b WB and quantification of LC3II protein levels from Sk-Mel-28R melanoma cell line exposed to CQ for 1 h and M-CSF mAb (20 ng/mL). c Cell viability assay, d apoptotic Annexin V assay and e wound-healing assay of Sk-Mel-28R cells treated with Vemurafenib (1 μM) and/or M-CSF mAb (20 ng/mL) and/or combined with Mibefradil (5 μM) or CQ (12.5 μM). Graphs show mean + SEM. n > 3 independent experiments. Statistical analysis was performed using the ANOVA test, followed by Bonferroni’s multiple comparisons test (*p < 0.05; **p < 0.01; ***p < 0.001; ns, non-significant).

Fig. 5. Autophagic blockers Mibefradil/Chloroquine, Vemurafenib and M-CSF mAb-combined therapy reduce tumour growth and vascularisation of BRAF^V600E-resistant melanoma cells

a Workflow of the in vivo experiment. (o.p. = orally; i.p. = intraperitoneal). b Tumour growth curves under treatment with Vemurafenib (50 mg/kg), CQ (30 mg/kg), M-CSF mAb (Lacnotuzumab) (4 mg/kg) or the combination therapy for 2 weeks in mice with SK-Mel-28R xenografts (n = 6). c Weight monitoring during all the following treatments. d Ki-67 Immunochemistry detection in xenografts (left) % of positive Ki-67 (right) and representative images of the Ki-67 immunostaining. e (left) % of positive immunostaining of ENG and (right) images of vascularisation represented by ENG immunostaining. Statistical analysis was performed using the ANOVA test, followed by Bonferroni’s multiple comparisons test (*p < 0.05; **p < 0.01; ***p < 0.001; ns, non-significant).

Fig. 6. M-CSF plasma levels determine melanoma OS and DFS in metastatic melanoma patients.

a Timeline of plasma collection and patient treatment. b Clinical parameters from 35 patients with BRAFV600E advanced-stage melanoma. MAP kinase-targeted therapy: BRAFi (Vemurafenib, Dabrafenib or LGX818) in monotherapy or in combination with MEKi (Dabrafenib/Trametinib, LGX818/MEK162, Vemurafenib/Cobimetinib). Response to the treatment: complete response (CR), partial responders (PR), and non-responders (stable disease (SD) and progressive disease (PD)). c Kaplan–Meier of melanoma overall survival (mOS) curve of metastatic melanoma patients (n = 17) with <5 ng/mL of M-CSF vs. >5 ng/mL. d Contingency of alive/dead patients depending on their M-CSF levels and e analysis of M-CSF plasma levels depending on the status of the patients (alive/dead). f Kaplan–Meier of melanoma disease-free survival (%DFS) of metastatic melanoma patients with <5 ng/mL vs. > 5 ng/mL of M-CSF plasma levels. g M-CSF levels depending on time to relapse after treatment (<12 vs. >12 months). Fisher exact tests were used to examine clinical variables and the ANOVA test for the statistical analysis of M-CSF plasma levels. Melanoma overall survival (mOS) and disease-free survival (DFS) curves were calculated using the Kaplan–Meier method with a log-rank test. p-values are indicated by asterisks *p < 0.05; **p < 0.01; ***p < 0.001.