Extracellular vesicles derived from melanoma cells induce carcinoma-associated fibroblasts via miR-92b-3p mediated downregulation of PTEN

Abstract

In melanoma, carcinoma-associated fibroblasts (CAFs) are important cellular components in the tumour microenvironment due to their potential to promote tumour growth and metastatic spread of malignant cells. Melanoma cells have the ability to affect non-tumour cells in the microenvironment by releasing extracellular vesicles (EVs). The mechanisms responsible for reprogramming normal dermal fibroblasts (NHDFs) into CAFs remain incompletely understood. However, it is likely thought to be mediated by melanoma-specific miRNAs, which are transported by EVs derived from melanoma cells. Therefore, we wondered if one of the most enriched miRNAs in EVs secreted by melanoma cells, miR-92b-3p, is involved in the conversion of normal fibroblasts into CAFs. We observed that melanoma cell-derived EVs indeed delivered miR-92b-3p into NHDFs and that its accumulation correlated with CAF formation, as demonstrated by enhanced expression of CAF marker genes and increased proliferation and migration. Overexpression of miR-92b-3p in NHDFs revealed similar results, while EVs deficient of miR-92b-3p did not induce a CAF phenotype. As a target we identified PTEN, whose repression led to increased expression of CAF markers. We thus provide a novel pathway of intercellular communication by which melanoma cells control the transformation of CAFs by virtue of EV-transported miRNAs.

Keywords: carcinoma‐associated fibroblasts; extracellular vesicles; melanoma; miRNA.

FIGURE 1

Characterization of extracellular vesicles derived from BLM and MV3. (a) TEM studies of EVs isolated from cell culture supernatants. The size scale corresponds to 100 nm. (b) Measurement of particle concentration or particle size of isolated EVs by nanoparticle tracking analysis (NTA). (c) Western blot analysis of EVs and corresponding cells analysed for CD63, CD81, CD9, ALIX and CANX.

FIGURE 2

Melanoma cell‐derived EVs induce a CAF phenotype in normal human dermal fibroblasts (NHDFs). NHDFs were incubated with or without melanoma cell‐derived EVs (10 µg/mL BLM and MV3) for 48 h and the resulting phenotype was analysed. (a) Representative images of actin filament staining (red) with Alexa Fluor™ 594 phalloidin and nuclei staining with DAPI (blue) (scale bar represents 50 µm). CAF marker genes were analysed by (b) Western blot for αSMA and qRT‐PCRs for (c) IL‐6, (d) IL‐8 and (e) FAP. In addition, (f) proliferation and (g) migration was investigated. (h) Proliferation and (i/j) migration of MV3 co‐cultured with NHDFs pre‐incubated with (10 µg/mL) or without MV3‐derived EVs. (k) Representative images of the wound closure assay 6 h after the wound was inflicted. (Bars represent mean ± standard deviation of at least three individual experiments on NHDFs from different donors; * p ≤ 0.05; ** p ≤ 0.01; *** p ≤ 0.001; **** p ≤ 0,0001; n.s., not significant).

FIGURE 3

Overexpression of miR‐92b‐3p mimics CAF phenotype in NHDFs. (a) qRT‐PCRs for miR‐92b‐3p in NHDFs after incubation with 10 µg/mL EVs derived from melanoma cells. EVs from NHEMs or normal medium were used as control. NHDFs were transfected with 100 nM miR‐92b‐3p mimic or control mimic (ctrl). The resulting phenotype was analysed 48 h after transfection. (b) qRT‐PCRs to validate miR‐92b‐3p overexpression in NHDFs after transfection. (c) Representative images of actin filament staining (red) with Alexa Fluor™ 594 phalloidin and nuclei staining with DAPI (blue). The scale bar represents 50 µm. CAF marker genes were analysed by d) Western blot analysis for αSMA and qRT‐PCRs for (e) IL‐6, (f) IL‐8 and (g) FAP. (h) Proliferation and (i) migration assays of NHDF cells transfected with miR‐92b‐3p mimic or control. (j) Proliferation and (k,l) migration of MV3 co‐cultured with NHDFs transfected with 100 nM miR‐92b‐3p mimic or control. (l) Representative images of the wound closure assay 6 h after the wound was inflicted (Bars represent mean ± standard deviation of at least three individual experiments on NHDFs from different donors; * p ≤ 0.05; ** p ≤ 0.01; **** p ≤ 0.0001; n.s., not significant).

FIGURE 4

Locked nucleic acids (LNA) mediated miR‐92b‐3p inhibition prevents EV‐induced CAF formation. Melanoma cells were transfected with 100 nM LNA‐92b‐3p or LNA ctrl. Non‐transfected cells were used as control (ctrl). (a) 48 h after transfection, EVs were isolated and analysed by qRT‐PCR for miR‐92b‐3p enrichment. NHDFs were incubated with or without the indicated EVs (10 µg/mL) and analysed for (b) miR‐92b‐3p accumulation. (c) Representative images of actin filament staining (red) with Alexa Fluor™ 594 phalloidin and nuclei staining with DAPI (blue). The scale bar represents 50 µm. CAF marker genes were analysed by (d) Western blot analysis for αSMA and qRT‐PCRs for (e) IL‐6, (f) IL‐8 and (g) FAP. Functional assays examined (h) proliferation and (i) migration. J) Proliferation and K//L migration of MV3 co‐cultured with NHDFs pre‐incubated with 10 µg/mL EVs lacking miR‐92b‐3p cargo or control EVs. (j) Representative images of the wound closure assay 6 h after the wound was inflicted (Bars represent mean ± standard deviation of at least three individual experiments on NHDFs from different donors; * p ≤ 0.05; ** p ≤ 0.01; *** p ≤ 0.001).

FIGURE 5

MiR‐92b‐3p targets PTEN in NHDFs. (a) Scheme of the predicted miR‐92b‐3p binding site in the PTEN 3′UTR. Western blot analysis for PTEN of (b) NHDFs incubated with 10 µg/mL melanoma cell‐derived EVs or without (ctrl) and (c) NHDFs transfected with 100 nM miR‐92b‐3p mimic or control mimic (ctrl). (d) Single cell sequencing analysis (GSE254918) of PTEN expression of in vivo Fibroblasts derived from normal skin of healthy donors (HD) or malignant melanoma tumours (MM). (e) Western blot analysis for PTEN of NHDFs transfected with 100 nM PTEN siRNA. (f) Representative images of actin filament staining (red) with Alexa Fluor™ 594 phalloidin and nuclei staining with DAPI (blue). (Scale bar represents 50 µm). Analysis of CAF marker genes by qRT‐PCR for (g) IL‐6, (h) IL‐8 and (i) FAP. (j) Functional assays examined migration. (k) Proliferation and (l) migration of MV3 co‐cultured with NHDFs transfected with 100 nM PTEN siRNA or control. (m) Representative images of the wound closure assay 6 h after the wound was inflicted. (Bars represent mean ± standard deviation of at least 3 individual experiments on NHDFs from different donors; **** p ≤ 0.0001; * p ≤ 0.05; n.s.: not significant).