Use of non-small cell lung cancer multicellular tumor spheroids to study the impact of chemotherapy.

BACKGROUND: Lung cancers represent the main cause of cancer related-death worldwide. Recently, immunotherapy alone or in combination with chemotherapy has deeply impacted the therapeutic care leading to an improved overall survival. However, relapse will finally occur, with no efficient second line treatment so far. New therapies development based on the comprehension of resistance mechanisms is necessary. However, the difficulties to obtain tumor samples before and after first line treatment hamper to clearly understand the consequence of these molecules on tumor cells and also to identify adapted second line therapies. METHODS: To overcome this difficulty, we developed multicellular tumor spheroids (MCTS) using characterized Non-Small Cell Lung Cancer (NSCLC) cell lines, monocytes from healthy donors and fibroblasts. MCTS were treated with carboplatin-paclitaxel or -gemcitabine combinations according to clinical administration schedules. The treatments impact was studied using cell viability assay, histological analyses, 3'RNA sequencing, real-time PCR, flow cytometry and confocal microscopy. RESULTS: We showed that treatments induced a decrease in cell viability and strong modifications in the transcriptomic profile notably at the level of pathways involved in DNA damage repair and cell cycle. Interestingly, we also observed a modification of genes expression considered as hallmarks of response to immune check point inhibitors and immunogenicity, particularly an increase in CD274 gene expression, coding for PD-L1. This result was validated at the protein level and shown to be restricted to tumor cells on MCTS containing fibroblasts and macrophages. This increase was also observed in an additional cell line, expressing low basal CD274 level. CONCLUSIONS: This study shows that MCTS are interesting models to study the impact of first line therapies using conditions close to clinical practice and also to identify more adapted second line or concomitant therapies for lung cancer treatment.

Oncolytic attenuated measles virus encoding NY-ESO-1 induces HLA I and II presentation of this tumor antigen by melanoma and dendritic cells.

Antitumor virotherapy stimulates the antitumor immune response during tumor cell lysis induced by oncolytic viruses (OVs). OV can be modified to express additional transgenes that enhance their therapeutic potential. In this study, we armed the spontaneously oncolytic Schwarz strain of measles viruses (MVs) with the gene encoding the cancer/testis antigen NY-ESO-1 to obtain MVny. We compared MV and MVny oncolytic activity and ability to induce NY-ESO-1 expression in six human melanoma cell lines. After MVny infection, we measured the capacity of melanoma cells to present NY-ESO-1 peptides to CD4 + and CD8 + T cell clones specific for this antigen. We assessed the ability of MVny to induce NY-ESO-1 expression and presentation in monocyte-derived dendritic cells (DCs). Our results show that MVny and MV oncolytic activity are similar with a faster cell lysis induced by MVny. We also observed that melanoma cell lines and DC expressed the NY-ESO-1 protein after MVny infection. In addition, MVny-infected melanoma cells and DCs were able to stimulate NY-ESO-1-specific CD4 + and CD8 + T cells. Finally, MVny was able to induce DC maturation. Altogether, these results show that MVny could be an interesting candidate to stimulate NY-ESO-1-specific T cells in melanoma patients with NY-ESO-1-expressing tumor cells.

The Tumor Antigen NY-ESO-1 Mediates Direct Recognition of Melanoma Cells by CD4+ T Cells after Intercellular Antigen Transfer.

NY-ESO-1-specific CD4(+) T cells are of interest for immune therapy against tumors, because it has been shown that their transfer into a patient with melanoma resulted in tumor regression. Therefore, we investigated how NY-ESO-1 is processed onto MHC class II molecules for direct CD4(+) T cell recognition of melanoma cells. We could rule out proteasome and autophagy-dependent endogenous Ag processing for MHC class II presentation. In contrast, intercellular Ag transfer, followed by classical MHC class II Ag processing via endocytosis, sensitized neighboring melanoma cells for CD4(+) T cell recognition. However, macroautophagy targeting of NY-ESO-1 enhanced MHC class II presentation. Therefore, both elevated NY-ESO-1 release and macroautophagy targeting could improve melanoma cell recognition by CD4(+) T cells and should be explored during immunotherapy of melanoma.

Inhibition of effector antigen-specific T cells by intradermal administration of heme oxygenase-1 inducers.

Developing protocols aimed at inhibiting effector T cells would be key for the treatment of T cell-dependent autoimmune diseases including type 1 autoimmune diabetes (T1D) and multiple sclerosis (MS). While heme oxygenase-1 (HO-1) inducers are clinically approved drugs for non-immune-related diseases, they do have immunosuppressive properties when administered systemically in rodents. Here we show that HO-1 inducers inhibit antigen-specific effector T cells when injected intradermally together with the T cell cognate antigens in mice. This phenomenon was observed in both a CD8+ T cell-mediated model of T1D and in a CD4+ T cell-dependent MS model. Intradermal injection of HO-1 inducers induced the recruitment of HO-1+ monocyte-derived dendritic cell (MoDCs) exclusively to the lymph nodes (LN) draining the site of intradermal injection. After encountering HO-1+MoDCs, effector T-cells exhibited a lower velocity and a reduced ability to migrate towards chemokine gradients resulting in impaired accumulation to the inflamed organ. Intradermal co-injection of a clinically approved HO-1 inducer and a specific antigen to non-human primates also induced HO-1+ MoDCs to accumulate in dermal draining LN and to suppress delayed-type hypersensitivity. Therefore, in both mice and non-human primates, HO-1 inducers delivered locally inhibited effector T-cells in an antigen-specific manner, paving the way for repositioning these drugs for the treatment of immune-mediated diseases.

A spliced antigenic peptide comprising a single spliced amino acid is produced in the proteasome by reverse splicing of a longer peptide fragment followed by trimming.

Peptide splicing is a novel mechanism of production of peptides relying on the proteasome and involving the linkage of fragments originally distant in the parental protein. Peptides produced by splicing can be presented on class I molecules of the MHC and recognized by CTLs. In this study, we describe a new antigenic peptide, which is presented by HLA-A3 and comprises two noncontiguous fragments of the melanoma differentiation Ag gp100(PMEL17) spliced together in the reverse order to that in which they appear in the parental protein. Contrary to the previously described spliced peptides, which are produced by the association of fragments of 3-6 aa, the peptide described in this work results from the ultimate association of an 8-aa fragment with a single arginine residue. As described before, peptide splicing takes place in the proteasome by transpeptidation involving an acyl-enzyme intermediate linking one of the peptide fragment to a catalytic subunit of the proteasome. Interestingly, we observe that the peptide causing the nucleophilic attack on the acyl-enzyme intermediate must be at least 3 aa long to give rise to a spliced peptide. The spliced peptide produced from this reaction therefore bears an extended C terminus that needs to be further trimmed to produce the final antigenic peptide. We show that the proteasome is able to perform the final trimming step required to produce the antigenic peptide described in this work.

Human natural killer cells promote cross-presentation of tumor cell-derived antigens by dendritic cells.

Dendritic cells (DCs) cross-present antigen (Ag) to initiate T-cell immunity against most infections and tumors. Natural killer (NK) cells are innate cytolytic lymphocytes that have emerged as key modulators of multiple DC functions. Here, we show that human NK cells promote cross-presentation of tumor cell-derived Ag by DC leading to Ag-specific CD8(+) T-cell activation. Surprisingly, cytotoxic function of NK cells was not required. Instead, we highlight a critical and nonredundant role for IFN-γ and TNF-α production by NK cells to enhance cross-presentation by DC using two different Ag models. Importantly, we observed that NK cells promote cell-associated Ag cross-presentation selectively by monocytes-derived DC (Mo-DC) and CD34-derived CD11b(neg) CD141(high) DC subsets but not by myeloid CD11b(+) DC. Moreover, we demonstrate that triggering NK cell activation by monoclonal antibodies (mAbs)-coated tumor cells leads to efficient DC cross-presentation, supporting the concept that NK cells can contribute to therapeutic mAbs efficiency by inducing downstream adaptive immunity. Taken together, our findings point toward a novel role of human NK cells bridging innate and adaptive immunity through selective induction of cell-associated Ag cross-presentation by CD141(high) DC, a process that could be exploited to better harness Ag-specific cellular immunity in immunotherapy.

Interplay between oncolytic measles virus, macrophages and cancer cells induces a proinflammatory tumor microenvironment.

Attenuated measles virus (MV) exerts its oncolytic activity in malignant pleural mesothelioma (MPM) cells that lack type-I interferon (IFN-I) production or responsiveness. However, other cells in the tumor microenvironment (TME), such as myeloid cells, possess functional antiviral pathways. In this study, we aimed to characterize the interplay between MV and the myeloid cells in human MPM. We cocultured MPM cell lines with monocytes or macrophages and infected them with MV. We analyzed the transcriptome of each cell type and studied their secretion and phenotypes by high-dimensional flow cytometry. We also measured transgene expression using an MV encoding GFP (MV-GFP). We show that MPM cells drive the differentiation of monocytes into M2-like macrophages. These macrophages inhibit GFP expression in tumor cells harboring a defect in IFN-I production and a functional signaling downstream of the IFN-I receptor, while having minimal effects on GFP expression in tumor cells with defect of responsiveness to IFN-I. Interestingly, inhibition of the IFN-I signaling by ruxolitinib restores GFP expression in tumor cells. Upon MV infection, cocultured macrophages express antiviral pro-inflammatory genes and induce the expression of IFN-stimulated genes in tumor cells. MV also increases the expression of HLA and costimulatory molecules on macrophages and their phagocytic activity. Finally, MV induces the secretion of inflammatory cytokines, especially IFN-I, and PD-L1 expression in tumor cells and macrophages. These results show that macrophages reduce viral proteins expression in some MPM cell lines through their IFN-I production and generate a pro-inflammatory interplay that may stimulate the patient's anti-tumor immune response.

Downregulation of MUC1 expression and its recognition by CD8⁺ T cells on the surface of malignant pleural mesothelioma cells treated with HDACi.

Research into new treatments against malignant pleural mesothelioma (MPM) is of great interest, as this aggressive cancer is often resistant to conventional therapies. One potential strategy is the use of epigenetic drugs, such as 5-aza-2'-deoxycytidine (5-azaCdR), a DNA-hypomethylating drug, and valproate (VPA), a histone deacetylase inhibitor (HDACi). Indeed, these drugs not only trigger MPM cell death, but also induce the expression of cancer testis antigens recognized by CD8(+) T cells, such as New York-esophageal cancer-1 (NY-ESO-1). The objective of this study was to assess effects of these drugs on the expression and recognition by CD8(+) T cells of Mucin1 (MUC1), a tumor-associated antigen that is overexpressed by MPM. MPM tumor cell lines were treated with epigenetic drugs, alone or in combination. MUC1 expression by MPM cells, and its recognition by a MUC1-specific CD8(+) T-cell clone, was downregulated by HDACi when used alone or in combination with 5-azaCdR. This effect was not due to a blocking of the HLA class I presentation pathway in treated MPM cells, as NY-ESO-1 induced by 5-azaCdR alone, or with VPA, was recognized by a NY-ESO-1-specific T-cell clone. This study suggests that the choice of tumor antigens could be critical for strategies combining epigenetic drugs with immunotherapy.

TP53 mutations correlate with the non-coding RNA content of small extracellular vesicles in melanoma.

Non-coding RNAs (ncRNAs) are important regulators of gene expression. They are expressed not only in cells, but also in cell-derived extracellular vesicles (EVs). The mechanisms controlling their loading and sorting remain poorly understood. Here, we investigated the impact of TP53 mutations on the non-coding RNA content of small melanoma EVs. After purification of small EVs from six different patient-derived melanoma cell lines, we characterized them by small RNA sequencing and lncRNA microarray analysis. We found that TP53 mutations are associated with a specific micro and long non-coding RNA content in small EVs. Then, we showed that long and small non-coding RNAs enriched in TP53 mutant small EVs share a common sequence motif, highly similar to the RNA-binding motif of Sam68, a protein interacting with hnRNP proteins. This protein thus may be an interesting partner of p53, involved in the expression and loading of the ncRNAs. To conclude, our data support the existence of cellular mechanisms associate with TP53 mutations which control the ncRNA content of small EVs in melanoma.

Purification of circulating plasmacytoid dendritic cells using counterflow centrifugal elutriation and immunomagnetic beads.

BACKGROUND AIMS: Plasmacytoid dendritic cells (pDC) are a dendritic cell (DC) subset specialized in the production of high amounts of interferon (IFN) type I (IFN-α, -ß) in response to viruses. They can be purified from peripheral blood mononuclear cells (PBMC), usually using magnetic bead sorting. METHODS: In this study, we set up a counterflow centrifugal elutriation (CCE) procedure to enrich pDC from PBMC. We first analyzed each CCE fraction for the presence of pDC using CD123 and BDCA-2 as markers. We then purified pDC using CCE and magnetic beads and verified that their functions were not affected by this procedure. RESULTS: pDC were sorted by CCE into intermediate fractions between those containing lymphocytes and monocytes. The pDC frequency in these intermediate fractions was 3-fold that in PBMC. Using negative-magnetic bead sorting, starting with the same number of cells and beads, we obtained more than twice as many pDC from intermediate fractions as from PBMC. The phenotypes and IFN-α production capacities of sorted pDC from PBMC and from intermediate fractions were similar, both immediately after sorting and after stimulation with CpG-A oligodeoxynucleotides. In addition, we showed that intermediate fractions could be cryopreserved and that magnetic bead sorting could be performed with the same efficiency after thawing. CONCLUSIONS: Altogether, our results show that CCE can be used to enrich lymphocytes, monocytes and pDC from the same donor, without magnetic beads on their surface. Our method should be useful for the purification of these cells for experimental research and may also be adaptable for clinical use in immunotherapy.

Human dendritic cells sequentially matured with CD4(+) T cells as a secondary signal favor CTL and long-term T memory cell responses.

Dendritic cells (DCs) are professional antigen-presenting cells involved in the control and initiation of immune responses. In vivo, DCs exposed at the periphery to maturation stimuli migrate to lymph nodes, where they receive secondary signals from CD4+ T helper cells. These DCs become able to initiate CD8+ cytotoxic T lymphocyte (CTL) responses. However, in vitro investigations concerning human monocyte-derived DCs have never focused on their functional properties after such sequential maturation. Here, we studied human DC phenotypes and functions according to this sequential exposure to maturation stimuli. As first signals, we used TNF-α/polyI:C mimicking inflammatory and pathogen stimuli and, as second signals, we compared activated CD4+ T helper cells to a combination of CD40-L/ IFN-γ. Our results show that a sequential activation with activated CD4+ T cells dramatically increased the maturation of DCs in terms of their phenotype and cytokine secretion compared to DCs activated with maturation stimuli delivered simultaneously. Furthermore, this sequential maturation led to the induction of CTL with a long-term effector and central memory phenotypes. Thus, sequential delivery of maturation stimuli, which includes CD4+ T cells, should be considered in the future to improve the induction of long-term CTL memory in DC-based immunotherapy.

The DCMU Herbicide Shapes T-cell Functions By Modulating Micro-RNA Expression Profiles.

DCMU [N-(3,4-dichlorophenyl)-N-dimethylurea] or diuron is a widely used herbicide, which can cause adverse effects on human, especially on immune cells, due to their intrinsic properties and wide distribution. These cells are important for fighting not only against virus or bacteria but also against neoplastic cell development. We developed an approach that combines functional studies and miRNA and RNA sequencing data to evaluate the effects of DCMU on the human immune response against cancer, particularly the one carried out by CD8+ T cells. We found that DCMU modulates the expression of miRNA in a dose-dependent manner, leading to a specific pattern of gene expression and consequently to a diminished cytokine and granzyme B secretions. Using mimics or anti-miRs, we identified several miRNA, such as hsa-miR-3135b and hsa-miR-21-5p, that regulate these secretions. All these changes reduce the CD8+ T cells' cytotoxic activity directed against cancer cells, in vitro and in vivo in a zebrafish model. To conclude, our study suggests that DCMU reduces T-cell abilities, participating thus to the establishment of an environment conducive to cancer development.

Viral Mimicry Response Is Associated With Clinical Outcome in Pleural Mesothelioma.

Introduction: The aim of this study was to investigate endogenous retrovirus (ERV) expression and type I interferon (IFN) activation in human pleural mesothelioma (PM) and their association with clinical outcome. Methods: The expression of ERV was determined from PM cohorts and mesothelial precursor RNA sequencing data. The expression of ERV was confirmed by quantitative polymerase chain reaction (qPCR). Methylation of genomic DNA was assessed by quantitative methylation-specific PCR. DNA demethylation was induced in cells by demethylating agent 5-Aza-2'-deoxycytidine (5-Aza-CdR) treatment. To block type I IFN signaling, the cells were treated with ruxolitinib or MAVS silencing. The expression of IFN-stimulated genes (ISGs) was determined by qPCR and Western blot. Circulating ERVs were detected by qPCR. Results: Long terminal repeats (LTRs) represent the most abundant transposable elements up-regulated in PM. Within the LTR, ERVmap_1248 and LTR7Y, which are specifically enriched in PM, were further analyzed. The 5-Aza-CdR treatment increased the levels of ERVmap_1248 expression and induced ERVmap_1248 promoter demethylation in mesothelial cells. In addition, ERVmap_1248 promoter was more demethylated in the mesothelioma tissue compared with nontumor tissue. The 5-Aza-CdR treatment of the mesothelial cells also increased the levels of ISGs. Basal ISG expression was higher in the mesothelioma cells compared with the mesothelial cells, and it was significantly decreased by ruxolitinib treatment or MAVS silencing. Furthermore, ISG expression was higher in the tumor tissue with high expression levels of ERVmap_1248. High expression of ERVmap_1248 was associated with longer overall survival and BAP1 mutations. ERVmap_1248 and LTR7Y can be detected in the PM plasma. Conclusions: We provide clues for patient stratification especially for immunotherapy where best clinical responses are associated with an activated basal immune response.

Human MuStem cells repress T-cell proliferation and cytotoxicity through both paracrine and contact-dependent pathways.

BACKGROUND: Muscular dystrophies (MDs) are inherited diseases in which a dysregulation of the immune response exacerbates disease severity and are characterized by infiltration of various immune cell types leading to muscle inflammation, fiber necrosis and fibrosis. Immunosuppressive properties have been attributed to mesenchymal stem cells (MSCs) that regulate the phenotype and function of different immune cells. However, such properties were poorly considered until now for adult stem cells with myogenic potential and advanced as possible therapeutic candidates for MDs. In the present study, we investigated the immunoregulatory potential of human MuStem (hMuStem) cells, for which we previously demonstrated that they can survive in injured muscle and robustly counteract adverse tissue remodeling. METHODS: The impact of hMuStem cells or their secretome on the proliferative and phenotypic properties of T-cells was explored by co-culture experiments with either peripheral blood mononucleated cells or CD3-sorted T-cells. A comparative study was produced with the bone marrow (BM)-MSCs. The expression profile of immune cell-related markers on hMuStem cells was determined by flow cytometry while their secretory profile was examined by ELISA assays. Finally, the paracrine and cell contact-dependent effects of hMuStem cells on the T-cell-mediated cytotoxic response were analyzed through IFN-γ expression and lysis activity. RESULTS: Here, we show that hMuStem cells have an immunosuppressive phenotype and can inhibit the proliferation and the cytotoxic response of T-cells as well as promote the generation of regulatory T-cells through direct contact and via soluble factors. These effects are associated, in part, with the production of mediators including heme-oxygenase-1, leukemia inhibitory factor and intracellular cell adhesion molecule-1, all of which are produced at significantly higher levels by hMuStem cells than BM-MSCs. While the production of prostaglandin E2 is involved in the suppression of T-cell proliferation by both hMuStem cells and BM-MSCs, the participation of inducible nitric oxide synthase activity appears to be specific to hMuStem cell-mediated one. CONCLUSIONS: Together, our findings demonstrate that hMuStem cells are potent immunoregulatory cells. Combined with their myogenic potential, the attribution of these properties reinforces the positioning of hMuStem cells as candidate therapeutic agents for the treatment of MDs.

Heterogeneous RNA editing and influence of ADAR2 on mesothelioma chemoresistance and the tumor microenvironment.

We previously observed increased levels of adenosine-deaminase-acting-on-dsRNA (Adar)-dependent RNA editing during mesothelioma development in mice exposed to asbestos. The aim of this study was to characterize and assess the role of ADAR-dependent RNA editing in mesothelioma. We found that tumors and mesothelioma primary cultures have higher ADAR-mediated RNA editing compared to mesothelial cells. Unsupervised clustering of editing in different genomic regions revealed heterogeneity between tumor samples as well as mesothelioma primary cultures. ADAR2 expression levels are higher in BRCA1-associated protein 1 wild-type tumors, with corresponding changes in RNA editing in transcripts and 3'UTR. ADAR2 knockdown and rescue models indicated a role in cell proliferation, altered cell cycle, increased sensitivity to antifolate treatment, and type-1 interferon signaling upregulation, leading to changes in the microenvironment in vivo. Our data indicate that RNA editing contributes to mesothelioma heterogeneity and highlights an important role of ADAR2 not only in growth regulation in mesothelioma but also in chemotherapy response, in addition to regulating inflammatory response downstream of sensing nucleic acid structures.

Homozygous Co-Deletion of Type I Interferons and CDKN2A Genes in Thoracic Cancers: Potential Consequences for Therapy.

Homozygous deletion (HD) of the tumor suppressor gene CDKN2A is the most frequent genetic alteration in malignant pleural mesothelioma and is also frequent in non-small cell lung cancers. This HD is often accompanied by the HD of the type I interferons (IFN I) genes that are located closed to the CDKN2A gene on the p21.3 region of chromosome 9. IFN I genes encode sixteen cytokines (IFN-α, IFN-߅) that are implicated in cellular antiviral and antitumor defense and in the induction of the immune response. In this review, we discuss the potential influence of IFN I genes HD on thoracic cancers therapy and speak in favor of better taking these HD into account in patients monitoring.

Endogenous retrovirus expression activates type-I interferon signaling in an experimental mouse model of mesothelioma development.

Early events in an experimental model of mesothelioma development include increased levels of editing in double-stranded RNA (dsRNA). We hypothesised that expression of endogenous retroviruses (ERV) contributes to dsRNA formation and type-I interferon signaling. ERV and interferon stimulated genes (ISGs) expression were significantly higher in tumor compared to non-tumor samples. 12 tumor specific ERV ("MesoERV1-12") were identified and verified by qPCR in mouse tissues. "MesoERV1-12" expression was lower in mouse embryonic fibroblasts (MEF) compared to mesothelioma cells. "MesoERV1-12" levels were significantly increased by demethylating agent 5-Aza-2'-deoxycytidine treatment and were accompanied by increased levels of dsRNA and ISGs. Basal ISGs expression was higher in mesothelioma cells compared to MEF and was significantly decreased by JAK inhibitor Ruxolitinib, by blocking Ifnar1 and by silencing Mavs. "MesoERV7" promoter was demethylated in asbestos-exposed compared to sham mice tissue as well as in mesothelioma cells and MEF upon 5-Aza-CdR treatment. These observations uncover novel aspects of asbestos-induced mesothelioma whereby ERV expression increases due to promoter demethylation and is paralleled by increased levels of dsRNA and activation of type-I IFN signaling. These features are important for early diagnosis and therapy.

A Functional Assay to Determine the Capacity of Oncolytic Viruses to Induce Immunogenic Tumor Cell Death.

Oncolytic immunotherapy efficacy relies partially on the induction of immunogenic tumor cell death following infection with oncolytic viruses (OV) to induce an antitumor immune response. Here, we describe a method to determine if an OV is able to induce such an immunogenic tumor cell death. This method consists in testing whether tumor cells lysed by an OV are able to induce the maturation of human monocyte-derived immature dendritic cells (Mo-iDC).

MicroRNAs in Tumor Exosomes Drive Immune Escape in Melanoma.

MicroRNAs (miRNA), small noncoding RNAs that regulate gene expression, exist not only in cells but also in a variety of body fluids. These circulating miRNAs could enable intercellular communication. miRNAs are packaged in membrane-encapsulated vesicles, such as exosomes, or protected by RNA-binding proteins. Here, we report that miRNAs included in human melanoma exosomes regulate the tumor immune response. Using microscopy and flow cytometry, we demonstrate that CD8+ T cells internalize exosomes from different tumor types even if these cells do not internalize vesicles as readily as other immune cells. We explored the function of melanoma-derived exosomes in CD8+ T cells and showed that these exosomes downregulate T-cell responses through decreased T-cell receptor (TCR) signaling and diminished cytokine and granzyme B secretions. The result reduces the cells' cytotoxic activity. Using mimics, we found that miRNAs enriched in exosomes-such as Homo sapiens (hsa)-miR-3187-3p, hsa-miR-498, hsa-miR-122, hsa-miR149, and hsa-miR-181a/b-regulate TCR signaling and TNFα secretion. Our observations suggest that miRNAs in melanoma-derived exosomes aid tumor immune evasion and could be a therapeutic target.