Immune surveillance in melanoma: From immune attack to melanoma escape and even counterattack

ABSTRACT

Pharmacologic inhibition of the cytotoxic T lymphocyte antigen 4 (CTLA4) and the programmed death receptor-1 (PD1) has resulted in unprecedented durable responses in metastatic melanoma. However, resistance to immunotherapy remains a major challenge. Effective immune surveillance against melanoma requires 4 essential steps: activation of the T lymphocytes, homing of the activated T lymphocytes to the melanoma microenvironment, identification and episode of melanoma cells by activated T lymphocytes, and the sensitivity of melanoma cells to apoptosis. At each of these steps, there are multiple factors that may interfere with the immune surveillance machinery, thus allowing melanoma cells to escape immune attack and develop resistance to immunotherapy. We provide a comprehensive review of the complex immune surveillance mechanisms at play in melanoma, and a detailed discussion of how these mechanisms may allow for the development of intrinsic or acquired resistance to immunotherapeutic modalities, and potential avenues for overcoming this resistance.

KEYWORDS:

• Apoptosis
• CTLA4
• immune surveillance
• immunotherapy resistance
• melanoma
• PD1

Introduction

The concept of immunotherapy in metastatic melanoma is not new. Scientists and physicians used vaccines, bacterial toxins, and T cell growth factors to stimulate the immune system to fight cancer.¹ The success rate has been variable but these different techniques have helped 5 to 10 percent of patients with metastatic melanoma achieve long-term remission. The recent introduction of checkpoint inhibitors has revolutionized the treatment of metastatic melanoma.¹

Unquestionably, recent advances in melanoma drugs that target the host immune response have led to growing interest in the role of immunotherapy. Ipilimumab, pembrolizumab, and nivolumab, are immune checkpoint inhibitors approved by the FDA to treat metastatic melanoma.^1,2 Ipilimumab is a fully humanized monoclonal antibody directed against CTLA-4. Patients with metastatic melanoma treated with this drug have a 2 y overall survival rate of 24%.^2,3 The 10-year overall survival rate of patients treated with ipilimumab is approximately 22% in a pooled analysis of overall survival data from 12 clinical trials, suggesting a sustained response rate.⁴ Pembrolizumab is a selective anti-PD-1 humanized monoclonal antibody. It received an accelerated FDA approval based on the results of a phase I study in which it was used in the setting of ipilimumab-resistant melanoma with a median follow-up of 18 months.³ The response rate of 34% was maintained in 81% of patients who responded to therapy with a median overall survival was 25.9 months.³ Nivolumab is another selective anti-PD-1 humanized monoclonal antibody. In a phase III trial, the objective response to nivolumab was significantly better than chemotherapy (32% vs 10%) in patients with metastatic melanoma who progressed after ipilimumab and BRAF inhibitor therapy (if BRAF was mutated).⁵ The combination of ipilimumab and nivolumab significantly improved progression free survival compared with ipilimumab alone [HR 0.40 (95% CI: 0.22, 0.71); p-value < 0.002] in a subsequent study.⁶

Unfortunately, not all patients with metastatic melanoma respond to immunotherapy. Many patients fail to respond to immunotherapy (i.e. primary resistant) while others develop initial responses which can be later lost due to development of acquired resistance (i.e., secondary resistant) leading to disease relapse.^7,8 Primary resistance to anti-PD1 therapies is common, affecting up to 60% of patients in some cancer types.⁹ In primary resistance there are intrinsic- host, tumor, or microenvironment-related factors that prevent the mounting of an effective T cell response. We speculate that one or more of these factors will prevent patients with metastatic melanoma from responding to immunotherapy upfront. In secondary resistance, patients initially respond to immunotherapy; however, this response is lost due to acquired tumor or microenvironment-related changes. We speculate that the T cell activation process begins intact in cases of secondary resistance; however, acquired changes at the level of the melanoma cell (epigenetic or oncogenic changes) or its microenvironment result in the formation of a new clone that is resistant to immunotherapy. It is essential to have an awareness of the different potential mechanisms of resistance so that we can identify strategies to target these tumors and enhance the efficacy of immunotherapy as an effective treatment modality in metastatic melanoma. This article will review the concepts of immune surveillance in melanoma and potential mechanisms adopted by melanoma tumor cells to escape it.

Immune surveillance in melanoma

In order for the immune system to mount an effective immune response against melanoma several steps must occur (Fig. 1). These include 1) activation of T cells, 2) homing of the activated T cells to the melanoma microenvironment, 3) identification and episode of melanoma cells by activated T lymphocytes, and 4) sensitivity of melanoma cells to programmed cell death (apoptosis).

Figure 1. The 4 steps of immune surveillance in black boxes; factors that enhance the efficacy and contribute to an effective immune surveillance in green boxes; factors that impair an effective immune surveillance and help melanoma escape immune attack in red boxes. DC: dendritic cells, TVEC: Talimogene laherparepvec, IFN-y: interferon gamma, CCL4: chemokine CCL4, TAP: Transporter associated with antigen processing, PD1: programmed death receptor 1, CTLA4: cytotoxic lymphocyte antigen 4, LAG3: lymphocyte-activation gene-3, TIM3: T-cell immunoglobulin mucin-3, BTLA: B- and T-lymphocyte attenuator, FLIP: FLICE-inhibitory proteins, IAPs: inhibitor of apoptosis proteins, CD95: death receptor also known as Fas, TRAIL/Apo2L: TNF-related apoptosis-inducing ligand, DISC: death-inducing signaling complex.

Activation of the T cells

There are 2 types of T cells involved in adaptive immunity; CD4+ T helper and cytotoxic CD8+ T cells. The CD4+ T helper cell type 1 (Th1) cells provide “help” for the activation of CD8+ T cells while CD4+ Th2 cells stimulate a humoral immune response and suppress the development of a Th1 response. CD4+ T cells can also display cytotoxic activity in some situations.¹⁰ Both cell types are derived from the thymocytes (bone marrow-derived progenitors) in the thymus where they undergo a process of rearrangement and selection.¹¹ This process is important in selecting appropriate T cell subsets that can detect the major histocompatibility complex (MHC) and react to non-self0peptides rather than self-peptides.¹¹

The process of T cell activation starts with the recruitment of dendritic cells (DCs), the most specialized of the antigen presenting cells (APCs), to the tumor microenvironment where they initiate phagocytosis (Fig. 2), a process by which tumor antigens are processed into peptides and loaded onto the MHC.¹² Danger signals, originating from the tumor microenvironment, stimulate Toll-like receptors (TLRs) on DCs, leading to their maturation and migration to regional lymph nodes, where they present the MHC-bound tumor-specific antigens to the T-cell receptors on lymphocytes –– a process called T cell priming.¹³ The binding of the CD28 receptors on the T cell with B7 molecules on the APCs generates a stimulatory signal from the APCs assuring the T cells that the presented tumor antigen is non-self and an appropriate target for destruction. If co-stimulation does not occur, the primed T cells will become hypoactive (i.e., T cell anergy) (Fig. 3).

Figure 2. Phagocytosis of tumor antigens by dendritic cells (DCs) and the antigen-presenting machinery that helps load the tumor peptides onto MHC-II molecules on the DC surface; 1) Endocytosis of tumor specific antigen (TSA), 2) Formation of endocytic vesicle that transports the tumor specific antigens into the cytosol, 3) the tumor specific antigens are degraded in the cytosol by proteasomes that degrades the proteins into peptides, 4) The tumor peptides are transported to the endoplasmic reticulum (ER) by transporter associated with antigen processing (TAP), 5) In the ER the tumor peptides are loaded on the MHC I or MHC II molecules to form complexes that (6) that moves through the Golgi apparatus and got displayed on the dendritic cell wall where they interact with the T cell receptors.

Figure 3. T cell activation and the mechanism of actions of both PD1 and CTLA4 inhibitors; PD1: programmed death 1 receptor, CTLA4: cytotoxic lymphocyte antigen-4, MHC: major histocompatibility complex, TCR: T cell receptor, TSA: tumor specific antigen, APS: antigen presenting cell, PDL1: programmed death ligand-1.

Homing of the activated T cells

Once activated, the T cells are now prepared to recognize and “licensed” to attack melanoma cells, but they first need to home to the melanoma microenvironment. This “homing” mechanism by which activated T-cells return to the tumor microenvironment, may be variably successful. We know that patients with high CD8+ T cell infiltration within their melanoma have the best clinical outcomes,¹⁴ however, diffuse CD8+ T cell infiltration into the melanoma microenvironment is observed in less than 10% of melanoma metastases.¹⁴ Several mechanisms facilitate homing of activated T-cells to the tumor microenvironment. The expression of cell adhesion molecules by the activated T cells and both adhesion molecules and chemokines by the vascular endothelium will enable homing of the activated T cells to the tumor microenvironment.^15,16 Murine models of vaccination for melanoma have shown that homing of the activated T cells appears to be integrin-dependent.^17,18 In humans, chemokine immune signatures of metastatic melanoma have correlated with CD8+ T cell infiltration.¹⁵ Specifically, Salerno et al. revealed a significant role for retention integrins α1β1, α2β1, and αEβ7 in maintaining intratumoral T cells. A site-specific homing to bowel metastases was mediated by CCR9 and increased expression of retention integrins.¹⁹ Hence, successful T cell infiltration is associated with immune signatures that include many genes associated with T cell homing such as adhesion molecules, retention integrins, and the chemokine ligands for CCR5, CCR9, and CXCR3.^20,21 These immune signatures may identify patients who are likely to respond to immunotherapy.

Identification and episode of melanoma cells

Once the activated T cells are in the tumor microenvironment, they will scan the area searching for melanoma cells that express and share the same tumor-specific antigens that initiated the T cells activation. The tumor specific-antigen processing machinery includes molecules such as the transporter protein TAP, the linker protein tapasin, and β2-microglobulin. These molecules ensure that the tumor peptides are processed and loaded onto the MHC-I on the melanoma cell surface.^22,23 Once the activated T cells are in contact with the tumor antigen-MHC I complex that is expressed on melanoma cells, they initiate a cytotoxic signal, leading to destruction and lysis of the tumor cells.^22,23

Sensitivity of melanoma cells to programmed cell death

There are 2 major mechanisms by which T cells and natural killer (NK) cells kill melanoma: the apoptosis pathway and the granule exocytosis pathway.²⁴

Apoptosis

Cytotoxic ligands, which generally belong to the tumor necrosis factor (TNF) family of ligands, include tumor necrosis factor α (TNF-α), Fas ligand (FasL or CD95L) and TRAIL/Apo2L. These ligands transmit a death signal upon binding to their cognate receptors.²⁵ The binding of the death ligand FasL (CD95L on lymphocytes) to the death receptor Fas (CD95 on melanoma cells) attracts adaptor proteins to the intracellular domain of the death receptor, which in turn recruits procaspase-8 (Fig. 4).^26-29 The resulting protein complex is called death-inducing signaling complex (DISC), in which procaspase-8 gets activated to “the initiator” caspase-8. In some cells, the amount of active caspase-8 is sufficient to initiate apoptosis directly, while in others the amount is too small and requires mitochondrial activation to amplify the apoptotic signal.³⁰ The Bcl⁻2 family member Bid, which is cleaved by active caspase-8 and translocated to the mitochondria, mediates mitochondrial activation. After activation, mitochondria release cytochrome c, an apoptosis-inducing factor, and other apoptogenic factors into the cytosol.^31,32 Cytochrome c, Apaf-1, adenosine 5′-triphosphate, and procaspase-9 form a complex called the apoptosome, which activates caspase-9. Activated “initiator” caspases (caspase-8 and −9) activate “executioner” caspases (caspase-3, −6, and −7), which cleave cellular substrates, known as the “death substrates,” leading to the biochemical and morphological changes characteristic of apoptosis.^31,32 (Fig. 4)

Figure 4. Apoptosis.

The granule exocytosis pathway

Perforin and granzymes are located within cytotoxic lymphocyte (CTL) granules. In the presence of calcium, perforin polymerization induces changes in the melanoma cell membrane which allow the passing of granzymes from the CTL into the melanoma cells.^33,34 Granzymes are neutral serine proteases that can activate caspases in the target cell.^35,36 In addition, granzyme B may directly cleave the Bcl⁻2 family member Bid, initiating the mitochondrial death pathway as described above.³⁷

Mechanisms of immune escape in melanoma

As we have now described, there are 4 complex steps that are required to ensure effective immune surveillance against melanoma (Fig. 1). Moving forward, we discuss factors that may interfere with each step of this immune surveillance and may thus contribute to melanoma escaping from the immune attack.

Factors that interfere with T cell activation

The immunosuppressed host

Recent studies have documented poor survival of melanoma among immunocompromised individuals, such as those with HIV or those who have received an organ transplant, presumably because these individuals are unable to elicit sufficient immune anti-tumor responses. However, further investigation is needed as there is a substantial lack of immunotherapy-outcome data among these groups of patients as they are almost always excluded from clinical trials of immunotherapy in metastatic melanoma.³⁸

Elderly individuals are another population with decreased diversity of the total T cell subset pool, and in theory, might be less able to mount an effective immune response against melanoma.^39,40 However, this may not be an accurate assumption as many of our elderly patients with metastatic melanoma respond to immunotherapy as well, if not better, than younger patients and with fewer adverse events

Mutational load

Genomic and RNA-based studies in melanoma suggest that tumor-specific mutational load and neoantigen signature as well as cytolytic activity are significantly associated with clinical benefit and increased overall survival.⁴¹ Tumors with a high mutational load respond better to immunotherapy because these mutations give rise to a variety of neoantigens that are truly non-self and can more easily be recognized by the immune system.⁴² Mutations or translocations of normal cellular genes in a tumor (i.e., melanoma) generate tumor-specific antigens that are identified by T cells as non-self-antigens.^10,43,44 Tumor-associated antigens, on the other hand, are expressed by tumor cells as well as normal cells.¹⁰ Differentiation antigens (e.g. MART-1/Melan-A, tyrosinase, gp100) are examples of tumor-associated antigens present on both normal/benign melanocytes and melanoma cells, against which reactive T lymphocytes have been identified.⁴⁵ Melanoma, lung cancer, and renal cell carcinoma are examples of tumors with high mutational loads that derive benefit from checkpoint inhibitors.^46,47 Conversely, tumors with low mutational loads such as pediatric tumors are more likely to evade immune surveillance.⁴⁸

As discussed earlier, once the T cells are activated, they home to the tumor microenvironment and eliminate melanoma cells, which share the same tumor-specific antigens that caused the T cell activation. Melanoma recurrences seen in some patients after an initial response phase to immunotherapy could be related to the development of further mutations that result in the formation of new and/or the loss of original tumor-specific antigens to which the T cells were activated.^49-51 Thus acquiring new mutations help melanoma cells escape immune surveillance and could contribute to disease progression.^52,53

Maturation and migration of dendritic cells

Cytokines, present in the tumor microenvironment, may help melanoma cells evade immune surveillance.⁸ The vascular endothelial growth factor (VEGF), a cytokine commonly produced by both tumor and associated immune cells, via stimulation of VEGF receptor 1 (VEGFR1), can suppress dendritic cells maturation and migration via a nuclear factor kB (NFкB)-dependent pathway.^8,54 Elevated tumor-derived transforming growth factor β (TGFβ) and IL-10 produced by immunosuppressive cells such as regulatory T-cells (Tregs) can suppress dendritic cells maturation and migration to lymph node basin by reducing expression of class II MHC and costimulatory machinery. Moreover, VEGF and IL-10 can affect antigen presentation and prevents T-cell activation.⁵⁵

The activated WNT/β-catenin pathway in melanoma cells impairs maturation of DCs at least in part through increased production of IL-10, and it inhibited IFN-α production by melanoma-specific CTLs.⁵⁶ This activated pathway also results in failure of CD103+ DC recruitment due to defective CCL4 chemokine production.⁵⁷ Failure of DCs to be recruited and migrate to lymph nodes prevents T cell activation and hence is one method that can lead to melanoma immune escape. Thus, patients with metastatic melanoma and an activated WNT/β-catenin pathway may represent one group of patients who will not respond up front to immunotherapy.

Factors that interfere with homing of activated T cells

Oncogenic and epigenetic changes

Chemokines play an important role in the homing of activated T cells to the tumor microenvironment. Specific oncogenic or epigenetic alterations may result in defective production of chemokines needed for T cell homing.⁵⁰ IFN-y increases the production of important chemokines CXCL9, 10, and 11 important for T cell migration and infiltration into the tumor microenvironment.⁵⁸ The activated WNT/β-catenin pathway in melanoma cells impairs homing of T cells to the tumor microenvironment by inhibiting IFN-y production by melanoma-specific CTLs.⁵⁶ Moreover, lack of interferon gamma signaling due to Jak1/2 loss of function mutations or perhaps epigenetic silencing of the Jak1/2, results in decreased production of chemokines required for T cell homing and may represent a mechanism of resistant to immune surveillance.^58-60 It is entirely possible and even likely that other oncogenic pathways in melanoma also mediate immune suppression.⁶¹

Epigenetic alterations including DNA methylation patterns, post-translational modification of histones, and chromatin remodeling play a large role in melanoma progression and metastasis.^62,63 There is growing interest in epigenetic changes and their role in the formation of resistance to immunotherapy in melanoma.⁶³ EZH2 is a histone modifier that functions as the catalytic component of the Polycomb Repressive Complex 2 (PRC2).^62,63 EZH2 mediated H3K27 hypermethylation has been shown to repress the production of Th1- type chemokines CXCL9 and CXCL10 in the tumor resulting in low tumor infiltrating CD8+ cells and poor patient outcomes in ovarian cancer, but this deserves further research in melanoma.⁶⁴ EZH2 mediated H3K27 hypermethylation may also silence important immune-protective genes allowing melanoma cells to evade immune surveillance.⁶³ EZH2 inhibitors are already entering clinical trials in lymphoma and sarcoma, thus future studies should look into combining an EZH2 inhibitor with a checkpoint inhibitor in melanoma tumor models to overcome a range of resistant mechanisms and enhance the clinical efficacy of immunotherapy.⁶³

Lack of T cell infiltration

IHC-based studies also support the notion that CD8⁺, CD4⁺, PD-1⁺, and programmed death-ligand 1–positive (PD-L1⁺) cell densities in pretreatment biopsies can predict response to therapy.^14,41,65 The absolute numbers of CD8+ T-cells infused during adoptive cell therapy for patients with metastatic melanoma strongly correlates with response to treatment, which suggests an important role for MHC class I restricted CTL-mediated tumor killing.^66,67 Oncogenic and epigenetic changes, as discussed earlier, may alter the function of certain immune-protective genes, such as with defective chemokine production, leading to a decreased homing of T cells to the tumor microenvironment and a decreased response to immunotherapy.⁶⁸ Down-regulation of adhesion molecules in melanoma may also inhibit homing of activated T cells to the melanoma microenvironment leading to continued tumor progression despite appropriate activation of T cells.^43,69

Intralesional injection of the oncolytic Newcastle disease virus (NDV) resulted in increased T cell infiltration locally at the injection site and distally at distant sites in a murine B16 melanoma model. NDV enhanced the efficacy of cytotoxic T lymphocyte 4 (CTLA4) inhibitor, perhaps due to the increased presence of tumor infiltrating lymphocytes, and resulted in rejection of pre-established distant tumors. These observations, though limited to a single model, provide a rationale for exploring the potential of oncolytic viruses such as NDV to enhance T-cell homing to the microenvironment, potentially in combination with immune checkpoint antibodies.⁷⁰

Factors that interfere with the function of activated T cells

The checkpoints

T cells possess inhibitory molecules, namely CTLA4 and the programmed T cell death 1 (PD1) receptor that, when coupled with their ligand, negatively regulate T cell activation and help prevent autoimmunity.⁷¹ Normally, CTLA4 expressed on T cells competes with the CD28 receptor on the T cell, for binding to B7 molecules on APCs, causing T cell anergy by interruption of the co-stimulatory signal.⁷² CTLA4 expression on T cells is upregulated in melanoma; hence the T cells are primed but turned off. Inhibiting CTLA-4 by targeted ipilimumab and thereby interrupting the negative inhibitory signal helps restore T cell function against melanoma. In addition, CTLA4 is expressed on regulatory T-cells (Tregs), and the immune effects observed with anti-CTLA4 antibodies may be a combination of enhancing effector T-cell function while blocking Tregs activity.⁷³ CTLA4 is also expressed on tumor cell lines⁷⁴ and in human melanoma.⁷⁵ Blockade of CTLA4 in vitro induces apoptosis of melanoma cells, indicating that CTLA4 might have additional, non-immune functions as well. CTLA-4 blockade enhances T cell responses, including production of IFN-y.⁷⁶ Patient identified as non-responders to CTLA-4 blockade have tumors with genomic defects in IFN-y pathway genes. Mice bearing melanoma tumors with knockdown of IFN-y receptor (IFNGR1) have impaired tumor rejection upon anti-CTLA-4 therapy. Loss of the IFN-y signaling pathway is associated with primary resistance to anti CTLA-4. Loss of IFNGR1, IFNGR2, Jak2, and IRF1 genes would impair IFN-y signaling pathway.^59,76 Other genomic losses in these non-responders include downstream IFN-y pathway genes such as the family of IFIT1, IFIT2, and IFIT3.⁷⁶

Melanoma cells can also evade immune surveillance via upregulation of the programmed death-ligand 1 (PDL1) which binds to PD1 receptors on T cells, thereby suppressing their function.¹ Anti PD1 agents that block the interaction between the PD1 receptors on T cells and the PDL1 ligand on tumor cells can restore T cell function against melanoma as evidenced by the success of the targeted drugs pembroluzimab and nivolumab.¹

Likewise, after chronic exposure to tumor antigens, CD8+ T cells begin to further express immune inhibitory receptors beyond PD-1 such as lymphocyte-activation gene-3 (LAG3), T-cell immunoglobulin mucin-3 (TIM3), and B- and T-lymphocyte attenuator (BTLA); the exact mechanisms and roles of these receptors in melanoma are yet to be determined.⁷⁷

Another mechanism of melanoma immune evasion is the production of IFN-γ by the effector T cells. This results in induction of PD-L1 expression on tumor cells that, in turn, deactivate T cell function by binding to the PD1 receptors.⁷⁹ However, IFNγ has been shown to enhance the MHC-I antigen processing and presentation pathway.⁵⁹ Buferne et al. have showed that CD8+ T cells specific for tumor antigen/MHC-I were efficient at inducing regression of the MHC-I-deficient melanoma, provided that the T cells were endowed with properties permitting their migration into the tumor and their efficient production of IFNγ.⁷⁹ The amount of IFNγ produced ex vivo from T cells correlated with an increase in surface expression of MHC-I molecules by the melanoma cells. These CD8+ T cells expressed PD-1 and upregulated its PDL-1 ligand on melanoma cells within the tumor. Despite upregulation of this immunosuppressive pathway, efficient IFNγ production in the melanoma microenvironment was actually found to be associated with resistance of the CD8+ T cells to inhibition both by PD-1/PDL-1 engagement and by TGFβ1, 2 main immune regulatory mechanisms hampering the efficiency of immunotherapy in patients.⁷⁹

An adaptive immune resistant process occurs in response to interferon gamma (IFN-γ) signaling in which melanoma cells would reactively overexpress PDL1 that binds to PD1 receptors on cytotoxic T cells causing its deactivation and allowing melanoma cells to escape immune attack.⁵⁸ Lack of INF y signaling due to loss-of-function mutations in Jak1/2 has been linked to an acquired resistant to PD1 inhibitors in patients with metastatic melanoma.^58,80

Gene expression profiling (GEP), in longitudinal tumor biopsies, predicted response in patients with metastatic melanoma treated with sequential CTLA4 and PD-1 blockade.⁴¹ While no significant GEP differences were found at pretreatment CTLA4 blockade, on-treatment CTLA4 blockade, and pretreatment PD-1 blockade, between responders and nonresponders; there was an early on treatment difference in samples of patients on anti–PD-1 therapy.⁴¹ The latter showed 411 significantly differentially expressed genes (DEG) in responders (P < 0.05), mostly upregulated as compared with nonresponders, including cytolytic markers, HLA molecules, IFNγ pathway effectors, chemokines and select adhesion molecules.⁴¹ Notably, vascular endothelial growth factor (VEGFA), was lower in responders than in nonresponders on PD-1 blockade⁴¹ suggesting a possible role of angiogenesis in resistant to immunotherapy and could potentially be a target of therapy.^81-83

The secretion of inhibitory molecules

Many types of molecules and cells in the melanoma microenvironment have been shown to suppress immune responses via cell-cell interactions and/or the production of immune suppressing molecules. Transforming growth factor β (TGF-β), interleukin-10 (IL-10) and enzyme indoleamine-2, 3-dioxygenase (IDO) may have a direct negative effect on the function of T-cells in the tumor microenvironment,^84-86 or an indirect one via the recruitment of different immunosuppressive cells that could be used by the tumor cells to evade immune surveillance. Tolerogenic immature DCs, myeloid derived-suppressor cells (MDSCs), immunosuppressive macrophages, regulatory B cells, or regulatory CD4+ T-cells are all examples of immunosuppressive cells that can be deregulated by melanoma to facilitate its progression and evasion of the immune surveillance.^87-89 Murine models have demonstrated that CD4+, CD25+, FOXP3+ Tregs hamper the efficacy of CTLs. Therefore, selective depletion of Tregs could be beneficial.²⁵ A recent study found that the T cell subset most induced by IL-2 treatment in melanoma patients is upregulation of Tregs expressing CD4, CD25, Foxp3 and the inducible T cell co-stimulator (ICOS). High levels of ICOS-expressing peripheral Tregs were a strong predictor of the lack of response to IL-2.⁹⁰ MDSCs have been implicated as an important component in suppression of immune responses in melanoma. A higher numbers of these cells in serum samples of patients with advanced melanoma is associated with worse prognoses.^91,92 Strategies targeting these cells are being developed to improve the efficiency of adoptive cell transfer therapy in melanoma.⁹³

The infiltrating T cells in the tumor microenvironment are highly regulated by the amino acid tryptophan levels. The enzyme indoleamine-2,3-dioxygenase (IDO), upregulated by CD8+ cells in the tumor microenvironment, converts tryptophan into the immunosuppressive molecule kynurenine.^94-96 Inhibitors of IDO may thus provide a targetable strategy for blocking this innate immunosuppression. INCB024360 and 1-methyl-D-tryptophan are in several clinical trials for different malignancies including metastatic melanoma.⁶¹

TGF-β is a cytokine that affects proliferation, activation, and differentiation of cells of innate and adaptive immunity and thus inhibits the anti-tumor immune response. The vascular endothelial growth factor (VEGF)⁹⁷ inhibits the differentiation of progenitors into DCs. Prostaglandins,⁹⁸ interleukin-10,⁹⁹ and soluble tumor gangliosides¹⁰⁰ are all immunosuppressive factors that may contribute to melanoma immune escape. Although the etiology is unclear, a subset of CD4+ T cells seems to suppress the T cell responses against melanoma.¹⁰¹

Factors that interfere with the ability of activated T cells to identify melanoma cells

Defects in the melanoma-antigen processing machinery

In order for them to be recognized by the activated T cells, melanoma-specific antigens need to be loaded onto the MHC-I molecules. Defects in this antigen-processing machinery can thus lead to immune escape.^22,23 The transporter associated with antigen processing (TAP) and tapasin are 2 proteins involved in loading antigenic peptides onto MHC-I molecules.^102,103 By forming physical links between class I heavy chains and TAP molecules, tapasin plays an important role in the assembly of MHC class I molecules with peptides in the endoplasmic reticulum.¹⁰⁴ Tapasin downregulation has been significantly associated with melanoma progression, perhaps due to immune escape.¹⁰⁴ Chang et al. reported an irreversible tapasin mutation that was associated with loss of the HLA haplotype and selective epigenetic silencing of HLA-A3 gene responsiveness to IFN-y in a melanoma cell line obtained from an axillary lymph node metastasis removed from a patient with a progressive disease despite surgery, vaccination with C-Vax/BCG, and chemotherapy combined with high-dose IL2.¹⁰⁵

Total loss of MHC-I by melanoma cells results in their complete resistance to recognition and destruction by CTLs.^106-108 The mechanism underlying this abnormal tumor phenotype has frequently been the loss of the β2-microglobulin (β2m), which is required for the formation of HLA class I H chain-β2m-peptide complexes and their transport to the cell surface.¹⁰⁹ Two genetic events – namely, mutations of one copy of the β2m gene and loss of the other copy [i.e., loss of heterozygosity (LOH)] – have been found to underlie β2m loss in malignant cells, although the chronological sequence of the 2 events is not known.^22,102 It has been demonstrated that in melanoma patients undergoing immunotherapy, a lack of response to immunotherapy and generation of progressing metastases appears to be associated with immune selection of MHC-I negative melanoma cell variants.¹¹⁰ Porgador et al described a very high prevalence (5 in 13 cases) of irreversible, complete loss of MHC-I expression in patients with metastatic melanoma who were treated with various immunotherapeutic regimens.¹¹¹ Although the melanoma cells in these 5 patients were very sensitive to autologous natural killer cells in vitro, clinical outcomes were very poor.¹¹¹ Studies have shown that loss of MHC-I expression may target melanoma cells for NK cell attack; however, the expression of MHC-I surrogates may help melanoma cells resist NK cell-mediated lysis.^112,113 Del Campo et al. revealed heterogeneous loss of β2m, caused by either mutation or loss of heterozygosity (LOH), in nests of a patient's primary melanoma; 10 months later the metastasis displayed homogenous loss of β2m- expression.¹¹⁴ Reconstitution of β2m expression has also been shown to restore HLA class I antigen expression in the 5 distinct melanoma cell lines. Immunotherapy imposes a selective pressure, perhaps, leading to an increased survival advantage for melanoma cells lacking the antigen-presenting MHC-I molecule or the tumor-specific antigen that contains the immunogenic peptide epitope (or both).^22,23 This selective pressure has the propensity to result in the emergence of immunotherapy resistant clone.^105,115 Restoring β2m function and HLA class I expression should be considered therapeutic targets in the development of cancer immunotherapy protocols.^114,116,117

Finally, the interferon gamma signaling pathway plays an important role in the function of activated T cells. Upon recognizing their tumor specific antigens on melanoma cell surface, the cytotoxic CD8+ T cells secrets interferon gamma⁵⁸ which acts via the interferon gamma receptors (the Janus kinases Jak1/2 and the signal transducers and activators of transcription STAT) to stimulate the expression of genes that exhibit anti-tumor effects.⁵⁸ Interferon gamma increases the antigen presentation machinery by melanoma cells so tumor specific antigens can be easily be recognized by the cytotoxic T cells. Moreover, interferon gamma exhibits direct tumor growth arrest and enhance apoptosis.⁵⁸ Loss-of-function mutations in Jak1/2 or in β−2-microglobulin resulted in acquired resistance to PD1 blockade in patients with metastatic melanoma.⁸⁰

Epithelial Mesenchymal Transition

Over time, some melanoma cells switch phenotype from proliferative, or epithelial “E-like,” to more invasive, or mesenchymal “M-like,” state in a process similar to classical epithelial-mesenchymal transition (EMT).¹¹⁸ In a panel of 57 melanoma cell lines derived from patient samples,¹¹⁹ the expression levels of E-cadherin or N-cadherin characterized these cell lines as E-like or mesenchymal-like (M-like), respectively.¹²⁰ Gene expression profiling enabled a detailed molecular classification of both melanoma cell types.^121-123 The proliferative state is characterized by high expression of microphthalmia-associated transcription factor (MITF),¹²⁴ while the invasive state expresses low or no levels of MITF with increased expression of EMT-related transcription factors.^121,122 Expression of Melan-A, tyrosinase, and MITF are significantly downregulated in M-like, compared with E-like, melanoma cells.¹²⁰ As such, immune responses induced to tumor-associated antigens such as Melan-A and tyrosinase will potentially only be effective against the E-like melanoma cells. EMT like phenotypic switch in melanoma was confirmed as a method of escape from adoptive T cell therapy with T-lymphocytes targeting Melan-A.¹²⁵ Of interest, there was an upregulation of the SPANX family of cancer testis antigens (CTAgs), which has demonstrated immunogenicity¹²⁶ in the M-like melanoma cells.^118,120 Moreover, high levels of SPANX family antigen expression was found in more aggressive melanoma cell lines.¹²⁷ Hence, it might be feasible to target the M-type melanoma cells with adoptive cell therapy using T-lymphocytes specific for SPANX family antigens.

Proinflammatory cytokines released in response to therapy-induced tumor tissue injury are critical signals that induce phenotypic plasticity of melanoma and immune cells.¹²⁸ Proinflammatory signals in the tumor microenvironment may induce EMT and dedifferentiation of melanoma cells. TNF-α resulted in dedifferentiation of melanoma cells, leading to loss of gp 100 (HMB-45) expression and it abrogated immune recognition by gp100-specific adoptive T cell transfer.¹²⁸ Neutrophil infiltration into the melanoma microenvironment is gaining interest as a therapeutic target because it is a sign of proinflammation that may drive melanoma plasticity toward the dedifferentiated mesenchymal phenotype that easily escapes immune surveillance. Ulceration of primary melanoma promotes neutrophil infiltration and correlates with increased angiotropsim and a higher risk of distant metastasis with poor prognosis.¹²⁹ Finally, resistance to BRAF inhibitors in both cultured cells and patient samples has been associated with an invasive, M-type phenotype switch with loss of MITF and SOX10 expression (116–118). This change was accompanied by a reduction in CTL infiltrates and an emergence of a CD8+ T cell exhaustion signature in the tumor microenvironment suggesting an acquired lowered state of immune surveillance.

Resistance of melanoma cells to apoptosis

Overexpression of antiapoptotic molecules

Changes in the expression of molecules involved in apoptosis signaling may contribute to melanoma resistance to immunotherapy.¹⁰ Several proteins regulate the apoptotic pathway.

FLICE-inhibitory proteins (FLIPs)

FLIPs interfere with the initiation of apoptosis directly at the level of death receptors but do not affect perforin/granzyme induced apoptosis (Fig. 4).^130-132 High levels of FLIP expression were reported in tissues from melanoma skin lesions.^130,133,134 Overexpression of FLIP by transfection into a Fas- and TRAIL-sensitive human melanoma cell line rendered this cell line more resistant to death mediated by both TRAIL and FasL.¹³⁵ Selective expression of FLIP by human melanomas may represent a mechanism by which melanoma cells escape immune destruction.¹³⁵

BCL2 family of antiapoptotic proteins

The Bcl⁻2 family has both pro- and anti-apoptotic proteins. Pro-apoptotic Bcl⁻2 proteins include Bak and Bax which oligomerize in the mitochondrial membrane to allow release of cytochrome c and other apoptotic effectors into the cytoplasm (Fig. 4). Bak and Bax activity are facilitated by other pro-apoptotis Bcl2 family proteins such as Bim, Bid, Bad, Noxa, and Puma. The anti-apoptotic Bcl⁻2 proteins include Bcl⁻2, Bcl⁻xL, Mcl⁻1, Bcl⁻w and A1.¹³⁶ Overexpression of the antiapoptotic molecules of the Bcl⁻2 family, which regulate apoptosis at the mitochondrial level, is associated with poor outcomes in melanoma.^{10,32,131,137} Bcl⁻2 expression, both in in vitro and in vivo models, confers resistance to CD95L and other apoptosis stimuli.⁶¹ Activating ras mutations, seen in 15% of melanomas, are associated with more progressive disease. N-ras mutation in particular causes resistant to apoptosis due to increased levels of Bcl⁻2 in melanoma.¹³⁸

The BH3 mimetic ABT-737 (inhibiting both Bcl2 and Bcl⁻xL) sensitizes human melanoma cells to marked BIM (Bcl⁻2-interacting mediator of cell death) -mediated apoptosis to BRAF inhibitors or dacarbazine in vitro.^61,139 ABT-737 deserves further study in combination with immunotherapy to enhance efficacy and overcome resistance.¹⁴⁰ Attempts to target BCL2 with antisense RNA (in effect inhibiting its action) in melanoma patients have not yet been successful. An international randomized controlled trial of 771 melanoma patients comparing DTIC and oblimersen (an antisense agent that targets Bcl⁻2) with DTIC alone resulted in an increased median PFS, but no significant difference in OS.¹⁴¹ It was never adequately established that this agent modulated Bcl⁻2 sufficiently to render cells more susceptible to cytotoxicity.¹⁴² A phase I-II clinical study investigated the combination of antisense oligonucleotides (ASO) targeted against BCL2 and dacarbazine in 14 patients with advanced malignant melanoma expressing BCL2.¹⁴² 40% exhibited decreased Bcl⁻2 expression, increased tumor cell apoptosis, and increased antitumor responses.¹⁴² Obatoclax, a drug inhibiting interaction of BCL2 with the pro-apoptotic proteins Bax and Bak, is being investigated in clinical trials for other malignancies.⁶¹

Mcl⁻1, a structurally distinct member of the anti-apoptotic Bcl⁻2 family, is strongly expressed at all stages of melanoma,¹⁴³ and is highly selective for BAK inhibition.¹⁴⁴ The combination of DTIC plus antisense oligonucleotide against Mcl⁻1-sensitized melanomas to DTIC in a mouse model.¹⁴⁵ Preclinical data suggest that taxane resistance may be due to increased MEK signaling resulting in anti-apoptotic changes.¹⁴⁶ MEK inhibitors achieve much of their apoptotic effect through Mcl⁻1 suppression.¹⁴⁷ These data suggest that dual MEK/Mcl⁻1 inhibition could be an effective means of improving clinical response.

The inhibitors of apoptosis proteins (IAP)

The IAPs bind to and inhibit caspases (Fig. 4).^134,148,149 In melanoma, 2 members of the IAP family (survivin and ML-IAP/Livin) and FLIP have been associated with melanomagenesis, as they are strongly expressed in melanocytic nevi and further overexpressed in invasive and metastatic melanomas.^150,151

A key antiapoptotic function of survivin relates to inhibition of mitochondrial and apoptosis inducing factor-dependent apoptotic pathways. Its expression in melanoma likely protects against both caspase-independent and -dependent apoptosis.¹⁵² Blocking the expression of survivin triggers spontaneous apoptosis in melanoma cell lines.¹⁵³ Livin selectively binds to SMAC (IAP antagonist) and caspase-3, caspase-7, and caspase-9, resulting in blockade of both the death receptor and mitochondria-based apoptotic pathways.^154,155 Smac/DIABLO (Fig. 4) released from the mitochondria during apoptosis promotes caspase activation by binding to IAPs.^134,148 In the cytoplasm, SMAC binds and inhibits IAPs and thereby acts as a proapoptotic protein.¹⁵⁶ Livin binds and sequesters SMAC, thereby permitting other IAPs, such as XIAP, to function uninhibited.¹⁵⁷ Hence livin is considered a sink for SMAC.¹⁵⁷

Given its role as a downstream inhibitor of apoptosis and due to its limited expression in normal adult tissues, ML-IAP/Livin has been identified as a potential selective target for immune-mediated tumor destruction in melanoma vaccine trials.^149,158 The presence of anti-livin antibodies in the serum of patients with GI, breast, and lung malignancies support anti-livin based immunotherapy as a therapeutic approach.^159-161 Spontaneous CTL responses to ML-IAP/Livin epitopes have been found in melanoma patients and a limited number of healthy controls.¹⁶² Spontaneous immune responses to ML-IAP/Livin in melanoma patients demonstrate both specific cellular and humoral activity as well as correlations with favorable clinical outcomes.¹⁴⁹ Therefore, inhibitors of livin could be useful adjuncts in the treatment of melanoma.¹⁶³

Expression of the proteinase inhibitor 9 PI-9

The caspase 8 and caspase-10 dependent Fas Ligand pathway becomes activated when tumor specific antigens, displayed on the surface of melanoma cells, are engaged with the cytotoxic T lymphocyte receptors on the immune cells. Activation of the Fas receptor pathway leads to, a cascade of caspase activation ending with apoptotic death of melanoma cells. Granzyme proteases are released by this process. ³³ After entry into melanoma cells, granzyme B cleaves the precursor forms of several caspases and induces apoptosis as well.¹⁶⁶ In humans, Proteinase inhibitor 9 (PtdIns-9) is the only intracellular inhibitor of granzyme B. Expression of PtdIns-9 in metastatic melanoma cells is associated with unfavorable overall and disease-free survival.³³ Expression of PtdIns-9 could be an important immune escape mechanism, and its inhibition may enhance the efficacy of immunotherapy.^164,165

Loss of Apoptotic protease activating factor 1 (Apaf-1)

p53 is a tumor suppresser gene that promotes apoptosis and cell cycle arrest in response to DNA damage due to radiation or chemotherapy. Mutation of the p53, though rare in melanoma, has been associated with chemoresistance (3). Apaf-1, a critical downstream mediator of p53-dependent apoptosis, and an integral part of the apoptosome, is frequently lost in metastatic melanoma.¹⁶⁷ A high rate of allelic loss of the Apaf-1 locus is seen, with the remaining allele transcriptionally inactivated by gene methylation. Apaf-1-negative melanomas are invariably resistant to treatment and are unable to execute a p53 induced apoptosis. Restoring physiologic levels of Apaf-1 through gene transfer or 5aza2dC treatment markedly enhances chemosensitivity and rescues the apoptotic defects associated with Apaf-1 loss.¹⁶⁸ Restoring Apaf-1 regulation to some melanomas may have therapeutic benefit, though; it is unclear if restoring Apaf-1 could enhance the effects of immunotherapy in melanoma.

Downregulation of FAS/CD95 signaling

While the Fas-apoptotic pathway is active in early stages of melanomas, it is often impaired in metastatic advanced-phase melanomas.^169,170 Functional impairment of the Fas/CD95 receptor-ligand interaction is associated with the development and progression of malignancies.¹⁷¹ One possible cause might be the expression of soluble receptors such as soluble CD95 (sCD95) and decoy receptor 3 (DcR3) that act as decoys for the death ligands, thus interfering with apoptosis via competitive inhibition of Fas/CD95 signaling.¹⁷¹ Indeed, increased soluble CD95 serum levels correlate with poor prognoses in melanoma patients.^171-173 Elevated sFas/CD95 serum levels were associated with poor overall (P < 0.005) and low progression-free (P < 0.0005) survival.¹⁷¹ Multivariate analysis revealed sFas/CD95 serum concentration as an independent predictive factor for progression-free survival (P = 0.011), but not overall survival (P = 0.078), in patients with melanoma.¹⁷¹

Downregulation of Fas signaling can also be attributed to changes at the genomic, transcriptional and post-translational levels.¹⁷⁴ Mutations in the death domain of Fas have been reported to take place in 7% of metastatic melanomas.¹⁷⁵ These mutated forms of CD95 may interfere in a dominant-negative way with apoptosis induction via CD95. Hypermethylation of Fas receptor promoter may represent a more common event in melanoma, resulting in transcriptional downregulation of Fas expression.¹⁷⁴

High levels of FasL expression have been implicated in suppression of the immune response in metastatic melanoma.^176,177 An inverse correlation between expression of FasL and that of Fas has been found during melanoma progression.^178,179 Transfection and overexpression of exogenous Fas in melanoma cells with high levels of endogenous FasL induce efficient cell death.¹⁸⁰ Similarly, expression of exogenous FasL in tumor cells that express endogenous Fas receptor has often been tested as a possible anticancer treatment.¹⁸¹

Downregulation of TRAIL signaling

TNF-related apoptosis-inducing ligand (TRAIL/Apo2L) is another major cytotoxic mechanism used by CTLs and NK cells to eradicate melanoma cells. TRAIL/Apo2L interacts with its cognate receptors (DR4, DR5, DcR1, DcR2, and osteoprotegerin) located on the melanoma cell surface, leading to melanoma cell death by apoptosis. Downregulation or malfunction of DR4 and the initiator caspase-8,¹⁸⁴ decreased release of Smac/DIABLO and thereby insufficient caspase-3 activation,¹⁸³ NF-κB mediated survival,¹⁸⁵ high expression of cFLIP and anti-apoptotic molecules such as Bcl⁻2, Bcl⁻XL, Mcl⁻1 and inhibitors of apoptosis (IAPs) as well as loss of Bax and Bak function have been reported to be of importance in TRAIL and apoptosis resistance in melanoma.¹⁸⁶

Cultured melanoma cells that express high levels of TRAIL receptors are highly sensitive to TRAIL-mediated apoptosis.¹³³ Decreased expression or function of TRAIL in melanoma cells resulted in attenuation of TRAIL-dependent apoptosis.^182,183 Moreover, some melanoma cells have completely lost expression of TRAIL receptors, presumably as a result of chromosome 8p22–21 loss^182,183 or through an epigenetic silencing mechanism.¹⁷⁴ Furthermore, at least 2 decoy receptors bind TRAIL without the ability to induce apoptosis, thereby pointing to an additional layer of regulation for the TRAIL apoptosis cascade.¹⁸⁷

Due to its apoptotic activity in a wide range of cancer cell types and low toxicity in most normal cells, TRAIL/Apo2L has become a promising anticancer agent.¹⁸⁸ TRAIL induces apoptosis in melanoma cell lines resistant to TNF-α, FASL and CD40L.¹⁸⁹ Agonistic monoclonal antibodies directed against TRAIL/Apo2L death-inducing receptors have become an attractive immunological therapeutic tool in clinical oncology due to their selective killing of tumors and lack of affinity toward healthy cells. Downregulation of Bcl⁻2, FLIP, or IAPs (XIAP and survivin) by siRNAs sensitizes resistant melanoma cells to Apo2L/TRAIL-induced apoptosis.¹⁹⁰ Full-length TRAIL, delivered from an adenoviral vector (Ad-hTRAIL), was studied in combination with DTIC or the histone deacetylase inhibitor suberoylanilide hydroxamic acid (SAHA) in human melanoma cell lines.¹⁹¹ Administration of Ad-hTRAIL in combination with DTIC or SAHA enhances apoptosis in human melanoma cell lines, and suggests that the therapeutic potential of such treatment strategies should be further evaluated for possible clinical use^191,192 HDACIs, inhibitors of histone deacetylation; a process essential in the regulation of gene expression, have proved to induce apoptosis, differentiation and cell cycle arrest in a variety of tumor cell lines, and to have anticancer potential in malignant melanoma.¹⁹³

The tumor counterattack

Melanomas with low T cell infiltration have poor prognoses and do not respond well to immunotherapy. The acquisition of death receptor ligand expression (FasL or CD95L) enables melanoma cells to deliver death signals to activate Fas (CD95)-positive CTLs leading to the death of CTLs. This mechanism is called the “tumor counterattack.”^{10,187,194-197} Immune-privileged sites such as the eye, the brain, and the testis have a high level of CD95L expression in the cells native to these sites, which explains why corneal transplants are accepted without tissue matching or immunosuppressive therapy in humans.^{10,133,197-199} Moreover, this may explain why immunotherapy may not be effective in melanoma involving the brain or the eye.

Melanoma cells not only can acquire resistance to Fas-induced apoptosis, which preserves them from apoptosis-mediated killing by cytotoxic T cells, but also can overexpress FasL which they use to deliver death signals and counterattack cytotoxic T lymphocytes in vitro.^10,176,200 The growth of a subcutaneously injected CD95L-positive murine melanoma cell line was slightly faster in wild-type compared with mutated CD95 receptor mice, suggesting impaired cell death in the faster growing tumor.¹⁷⁶ Another study hypothesized that CD95L is expressed intracellularly in melanoma cells. Melanosome preparations purified from melanoma cells induced dose-dependent Fas-mediated apoptosis of a cell line susceptible to Fas-dependent apoptosis.^200,201 This data suggested a new mechanism whereby FasL may be secreted by FasL-bearing vesicles inside the melanoma cells to mediate Fas-dependent apoptosis in the tumor microenvironment that does not necessarily imply cell-to-cell contact.^200,201 Contrary to the studies that support the counterattack phenomenon, others have shown that CD95L expression has a proinflammatory function leading to rejection of CD95L-expressing tumors by neutrophils.^10,181,202 To date, it is still unclear whether CD95L expression by tumor cells will provide them with an immune escape mechanism or will make them prone to rejection by the CD95L induced neutrophilic infiltration in vivo.¹⁰

Similar to CD95L, the death ligand TRAIL has been suggested to suppress tumor-specific T cells.²⁰³ Moreover, the membrane protein RCAS1²⁰⁴ not only inhibits proliferation, but also induces apoptosis in T cells in vitro. Chemokine production by tumor cells may also sensitize T cells to apoptosis.²⁰⁵ Additionally, the combination with treatments aimed at down-modulating FasL expression or at reducing microvesicle release by tumor cells may represent a strategy for improving clinical efficacy of active or adoptive immunotherapy in cancer patients.²⁰⁰

Conclusions

Understanding the mechanisms of resistance to immunotherapy will help us find new therapies and candidate targets in melanoma to prevent or overcome resistance. Mutational load is important for T lymphocyte activation, and we have learned that immunotherapy works best in tumors with a high, as opposed to low, number of mutations. We have also learned that activated T lymphocytes must need to successfully return home to the melanoma microenvironment for optimal activity. An immune cell gene signature for T cell homing may include expression of adhesion molecules, retention integrins, and the chemokine ligands for CCR5, CCR9, and CXCR3. Defects in expression of these proteins, through genetic mutations or epigenetic silencing, may interfere with homing of activated T cells and contribute to melanoma-immune escape. The acquisition of new mutations in melanoma cells may give rise to new and/or the loss of original tumor-specific antigens to which the T cells were activated. Epigenetic silencing or genetic mutations may contribute to downregulation of death ligands and/or their receptors leading to failure of melanoma cells to undergo apoptosis. Future studies should look into combining an EZH2 inhibitor with a checkpoint inhibitor in melanoma tumor models to overcome a variety of resistance mechanisms.

The Intralesional injections of oncolytic viruses, such as the FDA approved agent Talimogen Laherparepvec or the NDV (currently being studied as an investigational new drug) resulted in increased T cell infiltration in the tumor microenvironment. Future studies should look into combining these viruses with anti PD1 agents to enhance the clinical efficacy of immunotherapy. The production of IFN-γ by the effector T cells induce PD-L1 expression on tumor cells that in turn deactivate T cell function by binding to the PD1. In that regard, IFN-γ may enhance the clinical efficacy of anti PD1 therapies. Moreover, IFN-γ enhances the MHC-I antigen processing and presentation pathway. Hence, the combination of IFN-γ with anti PD1 inhibitors may be an alternative approach and deserve further investigation.

The recruitment of different immunosuppressive cells can help melanoma cells evade immune surveillance; hence, strategies targeting these cells may be of benefit. Likewise, finding new therapies that target immunosuppressive molecules in the melanoma microenvironment such as IDO, TGF-B, IL-10 may enhance the efficacy of immunotherapy. Inhibitors of IDO are in several clinical trials for different malignancies, including a randomized phase I-II trial combining ipilimumab with INCB024360 (NCT01604889).⁶¹ It appears that treatment with BRAF inhibitors provides a great landscape for the checkpoints inhibitors to work. BRAF or BRAF+MEK inhibitors enhance melanoma- antigen expression, increase CD8+ T lymphocyte infiltration, and decrease the immunosuppressive cytokines IL-6 and IL-8. Hence, it makes sense to start patients on BRAF inhibitors first to enhance the efficacy and benefit of checkpoint inhibitors later. However, the sequencing of current FDA approved agents is an area of active research and ongoing trials. Interestingly, inhibitory PD-1 was increased on T cells and its immunosuppressive ligand PDL1 was also increased with BRAF inhibition. This supports the hypothesis that targeting inhibitory immune interactions may be critical in augmenting responses to BRAF-targeted therapy in patients with melanoma.²⁰⁶

Even if we succeed in all the above, resistance to immunotherapy may still emerge if the apoptotic machinery is dysfunctional. Targeting apoptotic inhibitors such as survivin and ML-IAP/Livin and using AKT inhibitors to reduce the anti-apoptotic Bcl2 family proteins and Mcl⁻1 are all attractive therapeutic strategies to enhance apoptosis. These strategies deserve further investigation combined with the current available anti PD1 and CTLA4 inhibitors to overcome resistance. p53 is a tumor suppresser gene that promotes apoptosis and cell cycle arrest in response to DNA damage due to radiation or chemotherapy. Hence, combining immunotherapy with checkpoint inhibitors with either stereotactic radiation therapy or chemotherapy may enhance an effective immune response by providing more local inflammation that attracts inflammatory cells, and more tumor specific antigens to the cytotoxic T cells, in addition to their role in enhancing p53-induced apoptosis. These areas of translational research are urgently needed as there currently are no available treatments of patients with metastatic melanoma who have failed checkpoint inhibitors, other than adoptive cell transfer which is done only in major academic centers. The future advances in immunotherapy in the treatment of melanoma will likely involve combinatorial therapy to that allows different therapeutic agents to synergistically work together to optimize tumor antigenicity and the host immune anti-tumor response.

Disclosure of potential conflicts of interest

No potential conflicts of interest were disclosed.

Acknowledgment

Dr. Mahmoud wishes to acknowledge Dorothy Graves for her critical review of this manuscript.