In melanoma cells, TRAF6 promotes nuclear factor kappa-light-chain-enhancer of activated B cells (NFkB)-dependent release of fibroblast growth factor 19 (FGF19), implicated in the transformation and activation of fibroblasts. FGF19-mediated CAFs’ activation supports, in turn, the malignant and invasive phenotype of melanoma cells and their drugs resistance. On the other hand, TRAF6 upregulation in fibroblasts results in ECM remodeling through the release of matrix metalloproteinases (MMPs) 2 and 9 [91].

The mutual interaction between melanoma and CAFs can promote drug resistance in different ways. Straussman and collaborators highlighted the role of hepatocyte growth factor (HGF) in the development of acquired resistance to BRAF inhibitors (Figure 3B). Co-culture systems and proteomics analysis showed that HGF secreted by fibroblasts, by interacting with its mesenchymal ephitelial transition (MET) receptor on melanoma, induced MAPK and PI3K pathways’ activation, thereby promoting resistance to RAF inhibition. Simultaneous downregulation of both RAF and MET reverted resistance in vitro, and it has been proposed as a possible therapeutic approach for the treatment of BRAF-mutant melanomas. Most importantly, the authors confirmed increased HGF expression in stromal cells of BRAF-mutant melanoma patients undergoing BRAF-targeted treatment in vivo, which resulted in poor prognosis and decreased response to treatments [92]. Neuregulin 1 (NRG1) is another paracrine factor through which CAFs may influence melanoma response to MAPK inhibitors (Figure 3C). NRG1 is the ligand of v-erb-b2 avian erythroblastic leukemia viral oncogene homolog3 (ErbB3), which is upregulated in melanoma cells after treatment with BRAF inhibitors. The use of ErbB3/ErbB2 antibodies restores the cytotoxic activity of these drugs in BRAF-mutant melanoma cell lines [93]. Furthermore, vemurafenib treatment increases the production of transforming growth factor β (TGF-β) by melanoma cells; TGF-β, in turn, causes CAFs’ activation and increased fibronectin production, involved in BRAF inhibitors’ resistance (Figure 3D) [94].

Moreover, paradoxical MAPK activation, induced by BRAF inhibitors in genetically “normal” stromal cells, promotes a “therapy-resistant” microenvironment: Intravital imaging analyses conducted in melanoma have shown major paradox MAPK reactivation, especially in areas with high stromal density. Such activated CAFs, in turn, promoted matrix remodeling and ERK reactivation in melanoma, through integrin β1/focal adhesion kinase (FAK)/v-src sarcoma (Schmidt–Ruppin A-2) viral oncogene homolog avian (Src) signaling [95].

Uncontrolled production of reactive oxygen species (ROS) by TME fibroblasts is also associated with resistance to BRAF-targeted agents in melanoma (Figure 3E). Aging fibroblasts tend to release high levels of ROS in the TME, thereby modulating MAPK and PI3K pathways’ activation in tumor cells and promoting cells’ growth and drug resistance [96]. Such phenomenon could potentially be reversed by treating melanoma cells with antioxidants, thereby restoring drug responsiveness [97]. Along these lines, a recent preclinical study analyzed the role of secreted frizzled-related protein 2 (sFRP2), a wingless type MMTV integration site family member (Wnt) antagonist, in vemurafenib resistance: The sFRP2, produced and released in the TME by aged fibroblasts, actives a cascade of events in melanoma cells, ultimately leading to the loss of the redox effector apurinic/apyrimidinic endonuclease 1 (APE1). This condition significantly reduces the ability of melanoma cells to overcome ROS-induced DNA damage. Moreover, sFRP2-mediated inhibition of β-catenin leads to reduced melanoma response to vemurafenib [97].

It has recently been shown that CAFs are involved in the induction of a protumor immune microenvironment in many cancer models, favoring tumor growth and pharmacological resistance [98,99,100]. CAFs-mediated CXC chemokine ligand CXCL-2 production promotes regulatory T cells’ (Tregs) growth and recruitment in the tumor stroma [100,101]. Melanoma-associated fibroblasts also directly influence tumor cells’ ability to adapt and modify the response to immunotherapy. Experiments conducted on melanoma cell cultures have shown that CAFs release CXC motif chemokine 5 (CXCL5), which, in turn, induces PI3K/ protein kinase B (AKT)-dependent PDL-1 expression and resistance to immunotherapy in melanoma cells (Figure 3E) [102]. In addition, TGF-β secreted by CAFs also promotes resistance to PD-1 inhibitors: Transcriptomic and flow cytometric analysis, conducted on biopsies from 94 melanoma patients at various treatment stages, revealed a subset of patients characterized by loss of major histocompatibility complex 1 (MCH-I) and disease progression: Such phenomenon, induced by TGF-β released by CAFs, promotes a microphthalmia-associated transcription factor (MITF)_low/AXL_high phenotype in melanoma cells, associated with resistance to MAPK pathway and PD-1 inhibitors [103].

Finally, CAFs’ involvement in drug resistance is not limited to the production of paracrine factors but is also associated with cell-to-cell contact with cancer cells. Several studies have shown that fibroblasts create a physical barrier around the tumor mass and directly activate survival pathways in cancer cells [104,105,106]. In particular, these interactions are promoted by N-cadherin, expressed by both melanoma cells and fibroblasts, and are involved in tumor activation of the survival pathway PI3K/AKT/ BCL2 associated agonist of cell death (BAD) [104].

The immuno-microenvironment is characterized by T lymphocytes that recognize antigenic peptides presented by other components of the immune system [107]. CD4⁺ T cells act as immune response “adjuvants” through the secretion of specific cytokines. CD8⁺ T lymphocytes, on the other hand, are responsible for direct antigen/tumor cell individuation/elimination and are considered the most important mediators of tumor immune surveillance [82,108].

Depending on their genetic background, melanoma cells can influence the development of an immunosuppressive microenvironment. Phosphatase and tensin homolog deleted on chromosome 10 (PTEN), for example, is an important tumor suppressor gene often mutated/deleted in several cancer types, including melanoma; indeed, PTEN loss is present and concomitant with BRAF mutations in about 44% of melanomas and is associated with reduced OS [109]. PTEN loss promotes the formation of TME with low levels of cytotoxic T and natural killer (NK) cells and high concentrations of immunosuppressive elements, such as myeloid-derived suppressor (MDSCs) Tregs cells [110]. As described in in vitro and in vivo studies, PTEN-null melanoma cells inhibit antitumor activity of T cells and, consequently, response to immunotherapy. Through its negative regulation of PI3K and signal transducer and activator of transcription (STAT) 3 pathways, PTEN inhibits the production of immunosuppressive cytokines, such as interleukins (IL) 6 and 10 and vascular endothelial growth factor (VEGF). In melanoma, PTEN loss promotes STAT3 activation and, consequently, overproduction of these cytokines [111]. Moreover, PTEN loss is associated with reduced T cells’ recruitment to the tumor site and cytotoxic activity [112].

An immunosuppressive TME influences the differentiation of dysfunctional CD8⁺ T lymphocytes, i.e., T cells with reduced growth and effectors’ cell recognition capacity and high concentrations of PD-1 and CTLA-4 receptors. If physiological conditions such as this status are necessary for immune homeostasis and to avoid self-reactive phenomena, in tumor contexts it may be an escape route that cancer cells use to evade immune response and promote resistance to immunotherapy [113]. A study conducted in patients with advanced melanoma demonstrated the presence of a subpopulation of T cells with high levels of PD-1 and immunoglobulin and mucin domain-containing molecule 3 (Tim3), another inhibitory receptor. Tim3 inhibition partially reverted the dysfunctional condition of T cells and increased their antitumor abilities. These results form the rationale for simultaneous blockade of PD-1 and Tim3 as a possible therapeutic approach to restore CD8⁺ T lymphocytes’ functionality in context of melanoma [114]. The same research group identified an additional inhibitory receptor, called T cell immunoglobulin (Ig) and immunoreceptor tyrosine-based inhibition motif (ITIM) domain (TIGIT). Inhibition of this receptor together with PD-1 may counteract dysregulated T cells’ activity in a manner similar to Tim3 inhibition [115]. Extensive transcriptional profiling of the tumor infiltrate in 25 melanoma patients recently showed clonal expansion of dysfunctional CD8⁺ T cell subset. The authors highlighted the reactivity and differentiation of these cells, which are likely involved in the regulation of antitumor activity and resistance to immunotherapeutic agents, making them an attractive target for more targeted and effective immunotherapeutic treatments in melanoma [116].

B lymphocytes are the cells responsible for humoral and acquired immunity. Their main function is to produce specific antibodies against foreign antigens, but they are also involved in maintenance of immune memory [117]. In melanoma, tumor-associated B cells (TAB) account for up to 33% of TME immune cells and are involved in resistance to targeted therapy by promoting angiogenesis and chronic inflammation. In addition, the presence of B cells in the tumor infiltrate is associated with increased metastatic capacity of melanoma cells and reduced patients’ OS [118]. Recently, an interesting study analyzed the cross talk between melanoma cells and TAB and identified specific stimulating factors involved in the modulation of tumor response to different drugs. Melanoma secretes fibroblast growth factor 2 (FGF2), which actives B cells through its binding to fibroblast growth factor receptor 3 (FGFR-3) and promotes the release of insulin-like growth factor 1 (IGF-1). This factor, on tumor cells, induces proliferation and drug resistance. IGF-1, in turn, induces tumor cell proliferation and drug resistance. High levels of IGF-1 and FGFR-3 have been found in biopsies of melanoma patients treated with BRAF inhibitors in monotherapy or in combination with MEK inhibitors and IGF-1, and its receptor (IGF-1R) are associated with resistance to MAPK inhibitors [118]. However, TABs may have an opposite function in response to immunotherapy in melanoma. Indeed, a particular subtype of TABs can instead promote melanoma response to ICIs, by promoting the recruitment of CD8⁺ T cells in the tumor compartment. The authors observed that the presence of higher concentrations of these B cells, in pretreated melanoma patients, is associated with a better response to future immunotherapy treatments [119]. More recently, analysis of metastatic melanoma samples showed that the co-occurrence of tumor- associated CD8⁺ T cells and CD20⁺ B cells is associated with improved survival [120]. The formation of tertiary lymphoid structures in these CD8⁺/CD20⁺ tumors is associated with a gene signature, which predicts clinical outcomes in melanoma patients treated with ICIs. Moreover, B cell-rich melanomas displayed increased levels of transcription factor 7 (TCF7)+ naive and/or memory T cells, whereas T cells in tumors without tertiary lymphoid structures had a dysfunctional molecular phenotype. In another study, it was shown that B cell signatures are enriched in human melanoma samples from patients who responded to neoadjuvant ICI treatment [121]. B cell markers were, indeed, the most differentially expressed genes in the tumors of responders versus non responders [122]. Histological evaluation again highlighted the localization of B cells within tertiary lymphoid structures, while RNA sequencing demonstrated clonal expansion and unique functional states of B cells (switched memory B cells) in responder.

NKs are an important subclass of granular lymphocytes, involved in the recognition and elimination of virus-infected and transformed cells [123,124]. In general, cancer promotes several mechanisms that destabilize the functionality of NKs, determining immune evasion: (1) Hyperproduction of activating ligands that paradoxically block NKs’ receptors and (2) release of immunosuppressive factors, such as TGF-β and prostaglandin E [125]. Moreover, vemurafenib treatment of melanoma cells induces suppression of NKs activity in vitro, through downregulation of natural killer group 2D (NKG2D) and DNAX accessory molecule-1 (DNAM-1) activating receptors and simultaneous upregulation of major histocompatibility complex (MHC) I, which plays an inhibitory effect on NK cells [126].

Tregs represent a CD4⁺ T cell’s subpopulation with immunosuppressive properties [127,128]. In different cancer types, including melanoma, Tregs are able to promote immune evasion and cancer progression and are associated with poor prognosis [129,130,131]. In an analysis conducted on peripheral blood mononuclear cells (PBMCs) collected from healthy volunteers, Baumgartner and collaborators observed that melanoma evades the immune system by activation of Treg cells. Indeed, PBMCs exposed to melanoma-conditioned medium for a week presented an increase in Tregs’ induction and a major presence of IL-10 and TGF-β in the supernatant, as compared to the same PBMCs grown in control medium [132]. In BRAF-mutant melanomas, uncontrolled MAPK activation leads to an increased production of different ILsand VEGF that influence the activity of the immune system toward a protumor condition. Sumimoto and collaborators showed that in BRAF-mutant melanomas Tregs are activated and suppress the antitumor function of T lymphocytes. Moreover, pharmacological blockades or genetic manipulation of key components of MAPK pathway drastically decrease tumor production of immunosuppressive cytokines, allowing for the development of an immune microenvironment favorable to tumor suppression [133].

Regulation of Tregs’ differentiation and function could, therefore, be considered a valid therapeutic target for many cancers, including melanoma. Tregs are characterized by constitutive upregulation of PD-1 and CTLA-4 receptors and this condition leads to the hypothesis that Tregs could be the actual targets of ICI-based immunotherapy [134]. Unfortunately, results obtained in different studies are conflicting. Indeed, some studies have confirmed the inhibitory action of ICI on Tregs’ functionality, while others have reported opposite results that could support the hypothesis of an involvement of ICI-mediated activation of these cells in immune-resistance [135,136,137]. Analysis conducted on murine models of autoimmune pancreatitis have partly elucidated the suppressive role of PD-1 on Treg cells activity. Indeed, mice characterized by PD1-deficient Tregs showed greater immunosuppressive capacities and rapid development of autoimmune disease [135]. On the basis of these results, it can be speculated that, physiologically, the PD-1 axis plays an important role in the regulation of Tregs’ functionality and its inhibition may result in their increased activity. In vitro and in vivo experiments showed that, after treatments with nivolumab, Tregs proliferate and are functionally activated, resulting in the inhibition of antitumor activity [137]. Although with somewhat conflicting results, anti-CTLA-4 therapy would seem to bring more favorable effects on Tregs inhibition. Melanoma patients treated with ipilimumab showed a reduction in Tregs’ levels and major benefits in terms of decreased tumor growth and survival [138]. In mice models of melanoma, CTLA-4 blockade increases the intratumor effector T cells/Tregs ratio, through fragment crystallizable (Fc)-gamma receptor (FcγR)-dependent mechanism. FcγR is expressed by several immune cells, such as macrophages, neutrophils and NK cells and, therefore, TME composition may influence the response to CTLA-4 inhibitors. Melanomas presenting low concentrations of macrophages or immune cells deficient for FcγR tend to respond less to therapy [139].

MDSCs are immature myeloid cells, with immunosuppressive functions, associated to tumor progression, metastasis, angiogenesis, and drug resistance. Absent or present in small concentrations in physiological conditions, they are recruited by tumor cells or by inflammatory stimuli and are responsible for the production of several factors involved in tumor growth and immune evasion, such as IL-10, TGF-β, and VEGF [140,141]. In melanoma, chronic inflammation promotes MSDCs’ accumulation and activation in TME; thus, these cells are considered a possible therapeutic target in melanoma treatment [142]. A recent study identified a set of microRNAs (miRNAs) that regulate the differentiation and polarization of MDSCs in melanoma [143]. The authors found a significant association between the levels of these circulating miRNAs and reduced PFS and OS for melanoma patients treated with PD-1 and CTLA-4 inhibitors. This relationship was not reproduced in liquid biopsies of patients treated with MAPK pathway inhibitors, indicating a possible specific correlation with resistance to immunotherapy, as opposed to targeted therapy. These data suggest the possibility that combined treatments to inhibit myeloid dysfunctions could be able to overcome resistance to ICI in melanoma [143].

As reported by Gebhardt C. and collaborators, high levels of MDSCs in the TME are associated with ipilimumab resistance. Analysis of peripheral blood of 59 metastatic melanoma patients showed increased levels of MDSCs and their chemoattractant factors in patients poorly responsive or resistant to treatment with the CTLA-4 inhibitor. In addition, MDSCs exhibited a higher production of nitric oxide and were characterized by a higher expression of PDL-1, as compared to those isolated from responsive patients [144].

In melanoma, high levels of MDSCs are also associated with resistance to BRAF inhibitors. A recent preclinical study conducted in BRAF-inhibitor resistant mouse models showed that, after an initial response to treatment, these mice develop acquired resistance associated to an increase of MDSCs in TME. MAPK signaling reactivation in BRAF-resistant mice promotes the release of a complex system of stimulating cytokines, including C-C Motif Chemokine Ligand 2 (CCL2), that attract MDSCs and suppress the immune response [145].

Macrophages are immune cells involved in phagocytosis, pro-inflammatory cytokines’ production, and specific immunity. The tumor-associated macrophages (TAMs), under the influence of cancer cells and the other microenvironment components, may be promoters or repressors of the tumorigenic process [146]. They are divided into two categories. M1-like macrophages (M1-TAMs), with antitumor activity, are important for the early stages of the inflammatory response. M2-like macrophages (M2-TAMs), predominant in TME, are correlated with tumor progression [82,147,148]. Different studies confirmed the key role of TAMs in tumor progression and have highlighted the significant correlation between the high levels of TAMs in the TME and poor prognosis for the patients [149,150]. In particular, melanoma cells, releasing miRNA-125b-5p in the microenvironment, inhibit the lysosomal acid lipase A (LIPA) and promote M2-macrophages’ phenotype and their survival [151].

TAMs induce resistance to MAPK inhibitors by favoring the expression of tumor resistance factors or through their direct paradoxical activation of the pathway, driven by BRAF inhibitors. In in vivo melanoma models, MAPK-targeted agents induce tumor necrosis factor α (TNFα) production by macrophages, which promotes NFkB pathway activation and higher MITF expression. To overcome the TNFα- and MITF-mediated resistance to MAPK inhibitors, the authors proposed a selective inhibition of NFkB pathway that, in in vitro and in vivo analyses, synergized with MEK blockade and decreased the TNFα production [152]. Moreover, TAMs suffer the paradoxical reactivation of the MAPK pathway under the influence of BRAF inhibitors, with increased production of pro-angiogenic factors, such as VEGF and IL-8 and resistance to treatments [153].

TAMs express high levels of V-domain Ig suppressor of T cell activation (VISTA), another negative immune checkpoint, which is correlated with resistance to immunotherapy [154,155]. VISTA, associated with a significant decrease of survival in primary melanomas, in vivo, promotes a protumoral microenvironment mediated by upregulation of Tregs’ levels and PDL-1 expression on macrophages’ surface [155,156,157,158].

The switch of TAMs to an antitumor phenotype is considered an alternative approach to evade TAM-mediated resistance and to reconstitute the response to PD-1 axis inhibitors. TAMs’ transformation process, mediated by STAT6 inactivation and NFkB phosphorylation, results in an increase of interleukin 12 (IL-12) production by TAMs and a reduction in inhibitory cytokine levels, such as IL-10 and C-C motif chemokine 22 (CCL22) [159]. TAMs induce immunotherapy resistance also by inhibiting the recruitment of CD8⁺ T lymphocytes in the tumor site. Indeed, TAMs, by stable and durable interactions with T cells, lead to the maintenance of an immunosuppressive microenvironment not responsive to the inhibitory activity of anti-PD1 molecules [160].

DCs are immune cells derived from myeloid precursors and are implicated in recognition and capture of antigens considered "foreign", such as pathogens or cancer cells. They are antigen presenting cells (APCs) and interact with T lymphocytes through MHC present on their surface. DCs are involved in the production of cytokines and chemokines with anti- or pro-inflammatory function according to the stimuli received from the surrounding environment [161,162,163].

Unlike other immune cells, the involvement of DC in resistance to targeted and immunotherapy is mainly associated with the absence in tumor infiltrate of this type of cell. Melanoma is able to elude the complex mechanism of T cell activation by influencing DCs’ maturation. Tumor cells produce inhibitory cytokines such as IL-8, IL-10, and VEGF and create an unfavorable environment for DCs’ maturation. This condition negatively affects the DCs’ ability to present antigen to T cells and, therefore, determines a reduced immune response [164]. Moreover, López González and collaborators demonstrated that, if inhibition of glycogen synthase kinase 3 beta (GSK3β) obstructs DC differentiation, a constitutively active GSK3β overcomes the IL-10 inhibition, leading to DC maturation [165]. In addition, melanoma promotes the switch of myeloid cells through immuno-suppressive macrophage-like cells rather than DCs [166]. The use of oncolytic virus (i.e., ORCA-010) could stimulate a specific differentiation of DCs and T cell priming by producing tumor-associated neo-antigen in order to increase the response to ICIs [167]. The therapeutic potential of DC vaccines was, recently, supported in an interesting preclinical work by Zhou and collaborators. The authors produced in vitro CD103+ murine and evaluated its activity in murine models of melanoma and osteosarcoma. CD103+ stimulated a favorable environment to the action of T lymphocytes, resulting in a reduced primary and metastatic tumor growth [168]. Further and recent results have also been obtained by direct DC targeting with molecular inhibitors. A recent study demonstrated that dasatinib (TK inhibitor) induces the activation of allogenic T cells by impairing the phosphorylation and metabolism of tryptophan induced by Indoleamine-2,3-dioxygenase (IDO), one of the most important intermediary cancer tolerants [169]. In addition, the same RAF kinase inhibitors could induce acquired resistance by influencing the differentiation and activation of DCs. Preclinical experiments, carried out on human and mouse DC cells, have detected a reduced or lack of DCs’ ability to recruit T cells after treatment with RAF kinase inhibitors. These experiments, therefore, open possible new therapeutic scenarios, not only in melanoma, considering the negative effects of pan-RAF inhibitors on the immune response modulation [170].