Immune checkpoint inhibitors for the treatment of melanoma

Abstract

Introduction:

Immune checkpoint inhibitor (ICI) based immunotherapy is dramatically changing the management of many types of cancers including melanoma. In this malignancy, ICIs have been shown to prolong disease and progression free survival as well as overall survival of a percentage of treated patients, becoming the cornerstone of melanoma treatment.

Areas covered:

In this review, first, we will describe the mechanisms of immune checkpoint activation and inhibition, second, we will summarize the results obtained with ICIs in melanoma treatment in terms of efficacy as well as toxicity, third, we will discuss the potential mechanisms of immune escape from ICI, and lastly, we will review the potential predictive biomarkers of clinical efficacy of ICI-based immunotherapy in melanoma.

Expert opinion:

ICIs represent one of the pillars of melanoma treatment. The success of ICI-based therapy is limited by the development of escape mechanisms which allow melanoma cells to avoid recognition and destruction by immune cells. These results emphasize the need of additional studies to confirm the efficacy of therapies which combine different classes of ICIs as well as ICIs with other types of therapies. Furthermore, novel and more effective predictive biomarkers are needed to better stratify melanoma patients in order to define more precisely the therapeutic algorithms.

1. Introduction

Melanoma is a type of skin cancer characterized by poor prognosis and drug-resistance, especially when at a metastatic stage [1]. Until few years ago, standard chemotherapy was the only therapeutic option for metastatic melanoma patients. However the 5-year overall survival (OS) of treated patients was less than 5% [2]. During the past 10 years, immune checkpoint inhibitor (ICI) based immunotherapy has dramatically changed the clinical outcomes of melanoma patients [3]. ICI based immunotherapy represents an innovative oncological therapeutic approach since it enhances the host’s immune response against his/her own tumor by unleashing cognate immune cells [4]. Targeting immune checkpoint molecules by specific monoclonal antibodies (mAbs) has been developed and tested for the treatment of metastatic melanoma patients. At present, Cytotoxic T-Lymphocyte Antigen 4 (CTLA-4)-specific mAb ipilimumab and Programmed cell death protein-1 (PD-1)-specific mAbs pembrolizumab and nivolumab are utilized in melanoma treatment clinical practice [5]. In addition, several clinical trials, testing the efficacy and tolerability of other ICIs as well as ICI combination therapies, are still ongoing.

2. Mechanisms of action of immune checkpoints

2.1. CTLA-4

The interaction between T cell receptor (TCR) of T cells and an antigen-derived peptide presented by a major histocompatibility complex (MHC) allele on the surface of antigen presenting cells (APCs) is the most important event for the activation and regulation of T cells and as a result for the development of a host’ s immune response [6,7]. Several co-stimulatory or co-inhibitory signals are also needed to modulate T cell activation as well as cytokine production, cellular cycle and metabolism [8–10]. CTLA-4 (also known as CD152) is a receptor that plays a crucial role in the regulation of T cell activity [11,12] by mediating a co-inhibitory signal through its interaction with B7 molecules on APC surface [11]. Specifically, CTLA-4, by acting as a competitive antagonist receptor with a higher affinity binding, inhibits the interaction between the co-stimulatory signal of CD28 with its ligands B7.1 and B7.2 (CD80 and CD86) [13]. As a result, CTLA-4 down-regulates the activation of T cells, counteracting the stimulatory signals generated by CD28:B7 interactions and by TCR:MHC class I antigen binding [12]. An additional CTLA-4 mediated regulatory mechanism is represented by the regulatory T (Treg) cell activation [14]. This possibility is suggested by the high CTL4 expression level on Treg cells [15] and by the suppression of their regulatory function by mAbs blocking CTLA-4 [16–18]. Moreover, CTLA-4 on Treg cells might orchestrate memory CD8+ T cell quiescence by suppressing effector and proliferation programs [16,19].

On the other hand, some lines of evidence showed that T cell response is enhanced by CTLA-4 blockade even following Treg cell depletion. As a result CTLA-4 blockade can function independently of Treg activity [16,20]. Although inhibitory functions of CTLA-4 are well shown, the mechanisms they use to regulate the immune system remain to be defined [16,21].

Several lines of evidence have showed the crucial role of CTLA-4 in the immune system homeostasis as demonstrated by the fact that CTLA-4 knockout mice die because of the development of lymphoproliferative disease [22–24]. In addition in vitro and in vivo studies have also shown that CTLA-4 blockade has anti-tumor activity [25]. CTLA-4 is believed to inhibit potentially autoreactive T cells at the initial stage of naive T cell activation, typically in lymph nodes. As a result, CTLA-4 blockade affects the immune priming phase of T cell activation by supporting the activation and proliferation of a higher number of effector T cells, regardless of TCR specificity, and by reducing Treg-mediated suppression of T cell response [25].

2.2. PD-1/PD-L1

The receptor PD-1 (also known as CD279) interacts with its ligands Programmed death-ligand 1 (PD-L1) (also known as CD274) and Programmed death-ligand 2 (PD-L2 (also known as CD273) [26,27]. PD-1 is expressed on the surface of immune cells [28]. In contrast its ligands are more widely expressed. Specifically, PD-L1 is expressed on leukocytes and non-hematopoietic cells as well as in non-lymphoid tissues [27,29,30]. It can be induced on parenchymal cells by inflammatory cytokines (IFN-γ) or tumorigenic signaling pathways [31]. PD-L1 expression is also found on many tumor types and is associated with an increased number of tumor-infiltrating lymphocytes (TILs) and poorer prognosis [32,33]. In contrast, PD-L2 is primarily expressed on dendritic cells (DCs) and monocytes but can be induced on a wide variety of other immune cells and non-immune cells, depending on the local microenvironment [27]. PD-1 has a higher binding affinity for PD-L2 than for PD-L1, and this difference may be responsible for differential contributions of these ligands to immune responses [34]. Like CTLA-4, PD-1 mediates a co-inhibitory signal that is crucial to mitigate T cell effector activities by rendering these cells anergic when they are activated [35]. Because of PD-L1 and PD-L2 expression in peripheral tissues, the PD-1 axis maintains tolerance within locally infiltrated tissues. In the case of malignancies, PD-L1 or PD-L2 expression by cancer cells in the infiltrated tumor tissue inhibits cognate T cell activation and leads to T cell exhaustion. As a result cancer cells escape from T cell mediated immune destruction [29,36,37]. Blockade of PD-1/PD-L1 axis releases T cell anergy, reactivating T cell mediated immune destruction [24].

Moreover, as mentioned above for CTLA-4, also PD-1/PD-L1 axis has a key role in regulating the induction of Treg cell development and function. Specifically, some lines of evidence showed that PD-L1 mediated Treg cell development is caused by AKT, ERK2, mTOR and S6 down regulation as well as PTEN up regulation [38].

However, further studies are needed to explain all the mechanisms underlying PD-1/PD-L1 axis mediated interactions between T cells and Treg cells or other immune cellular subtypes.

2.3. Other immune checkpoints

Besides CTLA-4, PD-1 and PD-L1, other immune checkpoints have been identified and consequently new ICIs have been developed. To date, none of the latter ICIs is used in clinical practice although several clinical trials are evaluating their safety and efficacy. Inhibitors of the Lymphocyte activation gene-3 (LAG-3), T cell immunoglobulin and mucin domain 3 (Tim-3) and Indoleamine 2,3-dioxygenase (IDO) are at the most advanced stage of development [39]. LAG-3 (CD223) is an inhibitory receptor expressed on T cell surface. This receptor binds HLA class II antigens (HLA-II) expressed by DCs [40]. HLA-II are crucial for the recognition and subsequent T cell mediated destruction of cancer cells. LAG-3: HLA-II interactions lead to reduction of cancer cell recognition by T cells [41,42], providing cancer cells, including melanoma cells, with an immune escape mechanism [43]. Furthermore, LAG-3 inhibits T cell activation and generally promotes a more suppressive immune response by reducing cytokine and granzyme production and promoting Treg cell proliferation [44,45]. Several lines of evidence have shown that LAG-3 blockade by mAbs mitigates LAG-3 mediated suppressive immune response and re-activate cancer cell recognition and destruction [41,42].

Tim-3 is a receptor expressed on various types of immune cells and especially on T Helper-1 cells [46]. Binding its ligand (C-type lectin galectin-9), Tim-3 promotes down-regulation of interferon-ɣ-producing cells [47]. It is also a marker of exhausted CD8+ T cells [48]. Sakuishi et al have also shown that co-inhibition of both Tim-3 and PD-1 induces tumor regression [48]. Tim-3-specific mAbs are currently tested in several trials for the treatment of many cancer types, both as a single agent or in combination with other ICIs. Lastly, IDO is a rate-limiting enzyme involved in tryptophan catabolism. The latter plays a key role in immune tolerance since tryptophan depletion and storage of its catabolites (especially kynurenines) have immunosuppressive and inflammatory activity [49]. In addition, tryptophan catabolism mediated immunosuppression and inflammation promote carcinogenesis [50,51]. Blockade of IDO activity has been shown both in vivo and in vitro to display antitumor activity [52]. IDO inhibitors are currently being evaluated for their tolerability and antitumor activity in many types of cancer including melanoma.

3. ICIs in melanoma

3.1. CTLA-4-specific mAbs

CTLA-4-specific mAbs, tremelimumab and ipilimumab, have been the first ICIs to be tested for the treatment of melanoma patients. Tremelimumab is a fully human IgG2 mAb. Ipilimumab is a fully human IgG1 mAb. Administration of tremelimumab failed to demonstrate an OS benefit as compared to standard chemotherapy (dacarbazine 1000 mg/m² every 3 weeks or oral temozolomide 200 mg/m² once daily for 5 days every 4 weeks) in unresectable stage IIIc or IV (according to AJCC 6^th edition) treatment-naive melanoma patients. Furthermore, its administration was associated with a high rate of immune-related adverse events (irAEs) [53]. On the other hand, two phase 3 clinical trials (NCT00094653 and NCT00324155) have clearly shown a significant OS benefit of ipilimumab administration in unresectable stage III or IV (according to AJCC 6^th edition) melanoma patients, leading to its clinical approval. In the first trial a total of 676 pretreated melanoma patients was randomized, in a 3:1:1 ratio, to receive ipilimumab (3 mg/kg) plus gp100 (melanoma peptide vaccine) or ipilimumab (3 mg/kg) alone or gp100 alone. Other crucial inclusion criteria included HLA-A*0201 expression and no systemic treatment in the previous 28 days. Each agent was administered four times, once every 3 weeks. In the vaccine groups, patients received two modified HLA-A⋆0201–restricted peptides, injected subcutaneously as an emulsion with incomplete Freund’s adjuvant (Montanide ISA-51): a gp100:209–217(210M) peptide, 1 mg injected in the right anterior thigh, and a gp100:280–288(288V) peptide, 1 mg injected in the left anterior thigh. Median OS was 10.0, 10.1 and 6.4 months in ipilimumab plus gp100, ipilimumab alone and gp100 alone arm, respectively [54]. No difference was found between ipilimumab plus gp100 and ipilimumab alone. The rate of irAEs of grade 3 or 4 was 10–15% and 3% for ipilimumab and gp100 administration, respectively [54]. In the second trial a total of 502 untreated unresectable stage III or IV (according to AJCC 6^th edition) melanoma patients were randomized to receive ipilimumab (10 mg/kg) plus dacarbazine (850 mg/m²) or dacarbazine plus placebo. Ipilimumab was administered four times, once every 3 weeks, followed by dacarbazine alone every 3 weeks. The addition of ipilimumab to dacarbazine significantly prolonged median OS (11.2 vs 9.1 months) and increased survival rate at 1 (47.3% vs 36.3%), 2 (28.5% vs 17.9%) and 3 (20.8% vs 12.2%) years as compared to dacarbazine alone [55]. About 56.3% of ipilimumab treated patients experienced AEs of grade 3–4 [55]. An updated 5-year survival analysis has recently confirmed the OS benefit of ipilimumab over dacarbazine (18.2% vs 8.8%) [56].

In order to decrease the high rate of irAEs (especially colitis, rashes and endocrine disorders) an additional phase 3 clinical trial was conducted utilizing different doses of ipilimumab. Patients were randomized to receive ipilimumab at the dose of 10 or 3 mg/kg every 3 weeks or four doses. The results showed a longer median OS for the 10 mg/kg arm as compared to the 3 mg/kg arm (15.7 vs 11.5 months). As expected the highest dose of ipilimumab was associated with the highest rate of irAEs of grade 3–4 (37.0% vs 18.0% for the 10mg/Kg arm and the 3 mg/kg arm, respectively) [57]. An update 5-year survival analysis has confirmed the OS benefit for patients who receive the 10 mg/kg arm as compared to those who receive the 3 mg/kg arm. Moreover, these results suggest the emergence of a plateau in OS curve, consistent with previous ipilimumab trials [58].

Lastly, based on the results obtained in the metastatic setting, the clinical efficacy of ipilimumab was also tested in the adjuvant setting. Complete resected melanoma patients with high risk stage 3 were randomized to receive ipilimumab 10 mg/kg or placebo every 3 weeks for four doses. At a median follow-up of 5.3 years, the 5-year rate of recurrence-free survival (RFS) was 40.8% in the ipilimumab group, as compared to 30.3% in the placebo group. The OS rate at 5 years was 65.4% in the ipilimumab group, as compared to 54.4% in the placebo group. The rate of distant metastasis–free survival at 5 years was 48.3% in the ipilimumab group, as compared to 38.9% in the placebo group. irAEs of grade 3 or 4 occurred in 41.6% of the patients in the ipilimumab group and in 2.7% of those in the placebo group [59,60].

3.2. PD-1-specific mAbs

To date PD-1-specific mAbs are the cornerstone of treatment of metastatic melanoma patients. So far, two mAbs have clearly shown a clinical efficacy in the treatment of melanoma patients: nivolumab and pembrolizumab. Both have been approved by the FDA for the treatment of both metastatic and completely resected high risk melanoma patients.

Nivolumab is a fully human IgG4. Several clinical trials have shown the efficacy of nivolumab for the treatment of melanoma patients. In the phase 3 clinical trial NCT01721746 nivolumab administration at the dose of 3 mg/kg every 2 weeks was compared to investigator’s choice chemotherapy (dacarbazine 1000 mg/m² every 3 weeks or paclitaxel 175 mg/m² combined with carboplatin area under Curve (AUC) 6, every 3 weeks), all until progression or unacceptable toxicity, in unresectable stage III or IV (according to AJCC 7^th edition) melanoma patients who progressed following ipilimumab or BRAF inhibitor administration (in presence of BRAF mutations). Nivolumab administration demonstrated a significant clinical benefit. In fact, the objective response rate (ORR) and median OS was higher in the nivolumab arm as compared to investigator’s choice chemotherapy arm (31.7% vs 10.6% and 16 months vs 14 months, respectively) [61,62]. In a second phase 3 clinical trial (NCT01721772), 418 treatment naïve patients affected by unresectable stage III or IV (according to AJCC 7^th edition) melanoma without BRAF mutations were randomized to receive nivolumab (3 mg/kg) every 2 weeks or dacarbazine (1000 mg/m²) every 3 weeks, all until progression or unacceptable toxicity. Nivolumab administration was associated to a significant longer 1 year-OS rate (72.9% vs 42.1%) and median progression free survival (PFS) (5.1 vs 2.2 months) as compared to dacarbazine alone [63]. Furthermore, patients treated with nivolumab experienced less grade 3–4 AEs (11.7% vs 17.6%) [63]. These results were further confirmed following a longer follow-up [64,65]. Lastly, nivolumab administration was also tested as an adjuvant treatment for completely resected stage IIIB, IIIC and IV (according to AJCC 7^th edition) melanoma patients. Specifically, in the phase 3 clinical trial NCT02388906 nivolumab at the dose of 3 mg/kg every 2 weeks was compared to ipilimumab administration at the dose of 10 mg/kg every 3 weeks for four doses and then every 12 weeks. Treatment was administered for up to 1 year or until disease recurrence or a report of unacceptable toxic effects. In this trial nivolumab administration was associated to a significant clinical benefit in both 1-year RFS rate (70.5% vs 60.8%) and rate of grade 3–4 AEs (14.4% vs 45.9%) as compared to ipilimumab [66]. This result was confirmed following a 4 year follow-up update [67].

Pembrolizumab, another PD-1 specific mAb, is a humanized IgG4. It has been first tested in ipilimumab-refractory melanoma patients. Specifically, in the phase 3 clinical trial NCT01704287 540 unresectable stage III or IV (according to AJCC 7^th edition) patients were randomized, in a 1:1:1 ratio, to receive pembrolizumab at the dose of 2 or 10 mg/kg every 3 weeks or investigator’s choice chemotherapy (paclitaxel 225 mg/m² combined with carboplatin area under Curve (AUC) 6, every 3 weeks, or dacarbazine 1000 mg/m² every 3 weeks, or paclitaxel 175 mg/m² every 3 weeks, or oral temozolomide 200 mg/m² once daily for 5 days every 4 weeks). Both doses of pembrolizumab significantly prolonged PFS for the treated melanoma patients as compared to chemotherapy [68]. However at the final analysis of the study pembrolizumab administration did not significantly prolong the OS as compared to chemotherapy [69]. In a second phase 3 clinical trial (NCT01866319) unresectable stage III or IV (according to AJCC 7^th edition) melanoma patients were randomized, in a 1:1:1 ratio, to receive pembrolizumab at the dose of 10 mg/kg every 2 or 3 weeks or ipilimumab at 3 mg/kg every three weeks for four doses. The study demonstrated a significant longer PFS and OS for pembrolizumab-treated patients as compared to ipilimumab-treated patients. These results were confirmed in 3- and 5-year follow-up updates. In addition, pembrolizumab administration was associated to less grade 3–4 AEs [70–72]. Lastly, based on its clinical activity in the metastatic setting, pembrolizumab was also tested in an adjuvant setting for stage 3 melanoma patients following a complete resection of their primary tumor. In the phase 3 clinical trial NCT02362594, completely resected stage IIIA (patients with stage N1a melanoma had to have at least one micro-metastasis measuring >1 mm in greatest diameter) or stage IIIB or IIIC disease with no in-transit metastases (according to AJCC 7^th edition) melanoma patients were randomized to receive pembrolizumab at the dose of 200 mg or placebo every 3 weeks for a total of 18 doses. Patients treated with pembrolizumab had a significantly longer 1-year RFS rate as compared to patients treated with placebo (75.4% vs 61.0%) [73]. This result was confirmed following a longer follow-up [74,75].

4. ICI and toxicities

Despite the improved survival benefit associated with ICIs, irAEs might develop because of their pharmacological mechanisms [76,77]. By blocking the pathways that regulate the immune response, ICIs can promote the immune system’s activity inducing organ inflammation and thus increasing the risk of irAEs [77,78]. These toxicities can potentially involve multiple tissues or systems, including the skin (rash, pruritus), gastrointestinal tract (diarrhea, colitis), endocrine system (hypothyroidism, hypophysitis, hypocortisolism), liver (hepatitis), and lung (pneumonitis) [77]. Globally the occurrence of any grade irAEs is reported from 54% to 96% among patients with advanced melanoma receiving ICI-based immunotherapy [57,62,69,77,79]. irAEs can be severe and life-threatening resulting in treatment discontinuation or failure. Then, a proper management of irAEs is crucial to avoid serious sequalae [76]. Spectrum and grade of irAE might depend by the ICI utilized, dosage, schedule, regimen of monotherapy or combinatorial therapy, and clinical setting [77]. Chang et al, in their network analysis yielded three important findings about irAEs in ICI-treated melanoma patients. First, the three ICI regimens associated with the lowest risk of any or severe irAEs were nivolumab 3 mg/kg every 2 weeks, pembrolizumab 2 mg/kg every 3 weeks, and pembrolizumab 10 mg/kg every 3 weeks. In contrast, ipilimumab 10 mg/kg every 3 weeks or nivolumab 1 mg/kg every 3 weeks combined with ipilimumab 3 mg/kg every 3 weeks were associated with higher risks of irAEs as compared with the other ICI regimens. Second, within the various organ systems and irAE severities, different ICI regimens were ranked as being associated with the lowest risk of irAEs. Nivolumab 3mg/kg every 2 weeks was associated with a lower risk of dermatologic, gastrointestinal, and endocrine irAEs (regardless of AE severity) as compared with other ICI regimens. Ipilimumab 3mg/kg every 3 weeks was associated with a lower risk of any or severe liver irAEs, and pembrolizumab 10mg/kg every 2 weeks was associated with a lower risk of any or severe pulmonary irAEs. Third, ICI regimens were associated with a higher risk of any pruritus and diarrhea as compared with traditional chemotherapy regimens [77].

Several studies suggest that patients who experience irAEs during ICI therapy show favorable outcomes in cases of melanoma, non-small cell lung cancer, and urothelial carcinoma [80–84]. However, as compared with other cancer types, the association between irAE occurrence and ICI efficacy is not well established in melanoma patients [80]. It is unclear whether specific irAEs are more strongly related with favorable outcomes than other types of irAEs. To date, the development of vitiligo is of particular interest as an increasing number of studies have shown that its development is associated with a better clinical outcome in patients with stages III and IV melanoma treated with anti-PD-1 therapies [85–88].

Nevertheless, further studies are needed to clarify the potential role of development of vitiligo and other irAEs in ICI treated melanoma patients as a potential prognostic and/or predictive biomarker.

5. Mechanisms of immune escape to ICIs in melanoma

Although ICI based immunotherapy has improved clinical outcomes in melanoma patients, not all patients benefit from this type of therapy. A proportion of patients doesn’t have any benefit; another proportion has clinical benefit, but only for a short period. The latter group is characterized by acquired or secondary ICI resistance, while the former one is characterized by innate or primary ICI resistance [89–91]. Primary resistance to anti PD-1 and anti-CTLA-4 based immunotherapy occurs in roughly 40–65% and in more than 70% of treated patients, respectively [54,55,70]. ICI resistance can be linked to tumor (tumor-intrinsic) as well as to extrinsic mechanisms (tumor-extrinsic) [92]. In tumor-intrinsic resistance, cancer cells modify processes involved in DNA damage response, cell signaling pathways and immune recognition. In contrast, tumor-extrinsic resistance depends by the interactions among different immune cells into the tumor microenvironment [92–94]. So far, several mechanisms of innate and acquired ICI resistance have been identified [95–97]. It is important to underline that many mechanisms might overlap between innate and acquired ICI resistance [95].

The most notable trigger of innate ICI resistance depends by genetic as well as epigenetic alterations in tumor cells that influence tumor neoantigen presentation [92,98,99]. Tumor neoantigens are peptides produced from somatic mutations occurring in cancer cells [92,98,99]. The tumor neoantigen repertoire is crucial for the activation of T-cell mediated immune responses against cancer cells [92,98]. Emerging evidence suggests that some malignancies including melanoma lose or down regulate neoantigen generation allowing cancer cells to escape from T-cell recognition and cytotoxicity [92,95,100]. In addition to tumor neoantigen down regulation, melanomas are able to escape from T-cell cytotoxicity through alterations of the antigen presentation by both cancer cells and DCs [92,95,101]. DCs are crucial in antigen presentation. They promote immune response through uptake and presentation of tumor neoantigens in order to activate naïve CD4+ and CD8+ T-cells [102,103]. However, before presenting neoantigens, DCs must undergo a complex developmental program called “maturation” which makes them efficient T cell inducers [95,104]. Consequently, any factor that impairs DC maturation or function also contributes to immune escape [95]. For instance, tumor cells as well as myeloid derived suppressor cells (MDSCs), tumor-associated macrophages (TAMs) and Treg cells might impair DC maturation by maintaining high levels of IL-6 and IL-10 in tumor microenvironment or directing their maturation into regulatory phenotypes [92,95,105]. In cancer cells antigen presentation can be impaired by alterations in the structure of MHC or antigen presentation machinery (APM) components. As a result cancer cells escape from immune cells recognition [92,101,106]. Specifically the MHC class I pathway play a crucial role in antigen presentation and any defects in the genes associated with MHC-1 pathways such as the HLA class I and the beta 2 microglobulin (B2M) can affect antigen presentation and therefore immune response [92,107]. These modifications have been observed in several types of malignancy and are a frequent event in melanoma [92,101,108–111]. Composition of the tumor immune microenvironment (TIME) plays also a key role in innate ICI resistance. TIME contains several types of cells which promote or inhibit tumor growth as well as immune response [92]. High number of FOXP3+ cells and low number of CD8+ T cells in the TIME have been correlated with a worse prognosis as well as with more aggressive clinicopathological characteristics in melanoma [112–117]. TIME cell types include Treg, MDSCs, TAMs, cancer-associated adipocytes, fibroblasts and endothelial cells [118]. These cells together with tumor-associated stromal cells can produce molecules that inhibit T-cell response [119,120] as well as T cell migration to tumor core through the release of molecules or chemokines such as nitration, CCL2 [121,122] or overexpression of the endothelin B receptor [123].

Secondary or acquired resistance develops after a period of clinical response to treatment following which the tumor progression occurs [95]. There are specific changes observed in melanoma patients that play a potential role in acquired resistance [95]. As described in innate ICI resistance, defects in antigen presentation is also one of most important cause of acquired resistance. Several lines of evidence demonstrated that B2M mutations associated with the loss of HLA class I expression develop in melanoma patients who progress on ICI treatment [95,124]. Besides antigen presentation alteration acquired ICI resistance in melanoma can be mediated by loss of function of Janus Kinase 1 and 2/Signal transducer and activator of transcription (JAK1,2/STAT) signaling pathway. The latter is involved in many biological processes including cell differentiation, growth, apoptosis, and plays a crucial role in adaptative immune response [125,126]. Emerging evidence showed that JAK1,2/STAT pathway mutations have been reported in various cases of secondary resistant melanoma patients treated with anti-PD-1 [95,108]. Lastly, mechanisms of acquired ICI resistance include up-regulation alternative immune checkpoint molecules or their ligands [95,96]. Among them, PD-L1 overexpression leads to cytotoxic CD8+ T cell exhaustion and contributes to tumor immune evasion by shifting the maturation of naïve CD4+ T cells into Treg [97]. Besides PD-L1, up-regulation of LAG-3, Tim-3 and Fc receptor-like 6 protein (FCRL6) and IDO have been also described [95]. The latter is a rate-limiting enzyme involved in the catabolism of tryptophan. Its up-regulation leads to an over production of tryptophan catabolites (especially kynurenines) that are characterized by immunosuppressive activities.

6. Immunotherapy based on ICI combinations

In order to increase the therapeutic efficacy of ICIs and overcome intrinsic or acquired resistance mechanism(s) to ICI based immunotherapy a number of clinical studies (Table 1) is examining whether the combination of ICIs with conventional treatments, targeted therapy or other immunotherapeutic agents could improve patient outcomes.

Table 1.

Overview of main clinical trials of anti-PD-1 and anti-CTLA-4 agents in melanoma.

Trial	Phase	Treatment	Reference
NCT00257205	3	Tremelimumab vs dacarbazine or temozolomide	[53]
NCT00094653	3	Ipilimumab plus gp100 vs ipilimumab vs gp100	[54]
NCT00324155	3	Ipilimumab plus dacarbazine vs dacarbazine	[55]
NCT01515189	3	Ipilimumab 10 mg/kg vs ipilimumab 3 mg/kg	[57]
NCT00636168	3	Ipilimumab vs placebo	[59]
NCT01721746	3	Nivolumab vs dacarbazine or carboplatin plus paclitaxel	[62]
NCT01721772	3	Nivolumab plus dacarbazine vs dacarbazine	[63]
NCT02388906	3	Nivolumab vs ipilimumab	[66]
NCT01704287	2	Pembrolizumab vs dacarbazine or carboplatin plus paclitaxel or paclitaxel or temozolomide	[68]
NCT01866319	3	Pembrolizumab vs ipilimumab	[70]
NCT02362594	3	Pembrolizumab vs placebo	[75]
NCT01024231	1	Ipilimumab plus nivolumab	[79]
NCT01927419	2	Ipilimumab plus nivolumab vs ipilimumab	[128]
NCT01844505	2	Ipilimumab vs nivolumab vs ipilimumab plus nivolumab	[130]
NCT02714218	3/4	Ipilimumab 3 mg/kg plus nivolumab 1 mg/kg vs ipilimumab 1 mg/kg plus nivolumab 3 mg/kg	[135]
NCT03068455	3	Ipilimumab plus nivolumab vs nivolumab	[136]
NCT02437279	1B	Ipilimumab plus nivolumab	[137]
NCT02519322	2	Ipilimumab plus nivolumab vs nivolumab	[138]
NCT02977052	2	Ipilimumab 3 mg/kg plus nivolumab 1 mg/kg vs ipilimumab 1 mg/kg plus nivolumab 3 mg/kg vs ipilimumab 3 mg/kg followed by nivolumab 3 mg/kg	[139]
NCT03470922	2/3	Relatlimab plus nivolumab vs nivolumab	[142]
NCT02130466	1/2	Dabrafenib plus trametinib plus pembrolizumab vs dabrafenib plus trametinib plus placebo	[147]
NCT02908672	3	Vemurafenib plus cobimetinib plus atezolizumab vs vemurafenib plus cobimetinib	[148]

One of the approaches utilized is represented by the combination of both CTLA-4- and PD-1-specific mAbs. The combination of ipilimumab and nivolumab was the first to be evaluated in melanoma patients. Its manageable safety profile was firstly demonstrated in the phase 1 clinical trial NCT01024231 [79,127]. Its clinical efficacy was secondly evaluated in the phase 2 clinical trial NCT01927419 (CheckMate 069). In this trial, treatment-naive unresectable stage III or IV (according to AJCC 7^th edition) melanoma patients were randomized, in a 2:1 ratio, to receive nivolumab (1 mg/kg) plus ipilimumab (3 mg/kg) or ipilimumab (3 mg/kg) plus placebo, every 3 weeks for four doses. Subsequently, patients treated with nivolumab plus ipilimumab received nivolumab (3 mg/kg) every 2 weeks until disease progression or unacceptable toxicity, whereas patients treated with ipilimumab alone received placebo every 2 weeks during this phase. Both BRAF mutated and wild type patients were included in the study. Analysis of the results demonstrated a higher clinical efficacy of the combination as compared to that of ipilimumab alone, regardless of BRAF mutations [128,129]. Lastly, in the phase 3 clinical trial NCT01844505 (CheckMate 067), 945 previously untreated unresectable stage III or IV (according to AJCC 7^th edition) melanoma patients were randomized, in a 1:1:1 ratio, to receive nivolumab (1 mg/kg) plus ipilimumab (3 mg/kg) every 3 weeks for four doses, followed by nivolumab (3 mg/kg) every 2 weeks, or nivolumab (3 mg/kg) every 2 weeks plus placebo, or ipilimumab (3 mg/kg) every 3 weeks for four doses plus placebo. Median PFS was 11.5, 6.9 and 2.9 months in combination, nivolumab alone and ipilimumab alone arms, respectively. The subgroup analysis showed that in patients with PD-L1 expressing tumors the PFS was not significantly different between those treated with the combination and those treated with nivolumab alone. PFS of patients in both arms was significantly longer than that of patients treated with ipilimumab alone. In contrast, in patients with PD-L1 negative tumors the PFS of patients treated with the combination was significantly longer than that of patients treated with nivolumab alone or ipilimumab alone (11.2 vs 5.3 vs 2.8 months, respectively) [130]. Three-year OS rate was 58%, 52% and 34% in combination, nivolumab alone and ipilimumab alone arms, respectively. These results were also confirmed in a longer follow-up [131–133]. Grade 3–4 AE rates were 59, 21 and 28% in combination, nivolumab alone and ipilimumab alone arms, respectively [134]. The combination of nivolumab and ipilimumab was also tested utilizing different doses of both nivolumab and ipilimumab in order to decrease the rate of AEs. In the phase 3–4 clinical trial NCT02714218 the combination of nivolumab and ipilimumab at the dose of 3 and 1 mg/kg, respectively, every 3 weeks was compared to standard doses of nivolumab and ipilimumab combination of 1 and 3 mg/kg every 3 weeks, respectively. Following combination therapy, patients in both groups received nivolumab at the dose of 480 mg every 4 weeks until progression or unacceptable toxicity. Severe AEs rate was 34% for the new tested combination as compared to 48% for the standard combination [135]. In terms of efficacy there were no significant differences between the groups for any efficacy end point, although the standard combination displayed the highest rate of ORR (50.6 vs 45.6%) as well as the longest median PFS (9.9 vs 8.9 months)[135].

Moreover, ipilimumab plus nivolumab combination was tested also in adjuvant as well as neoadjuvant setting. In NCT03068455, patients with resected stage IIIB, IIIC, IIID, and IV (according to AJCC 8^th edition) melanoma were randomized 1:1 to receive either nivolumab (240 mg every 2 weeks) plus ipilimumab (1 mg/kg every 6 weeks) or nivolumab (480 mg every 4 weeks) for up to approximately 1 year therapy. The trial showed no benefit in terms of recurrence-free survival between the two arms [136].

On the other hand, other trials evaluated the role of ipilimumab plus nivolumab as an integral treatment for resectable stage III melanoma patients. Specifically, in the OpACIN trial, 20 patients with resectable stage IIIB (according to AJCC 7^th edition) were randomized to receive ipilimumab (3 mg/kg) and nivolumab (1 mg/kg) either as an adjuvant (four courses after surgery) or neoadjuvant-adjuvant treatment (two courses before surgery and two courses post-surgery). Among 9 patients treated in the neoadjuvant arm, 7 achieved a pathological response, and no patient relapsed at the time of the analysis (median follow-up: 25.6 months). However, 9/10 patients experienced one or more severe AEs [137]. In another randomized phase 2 trial, twenty-three stage III or oligometastatic stage IV (according to AJCC 7^th edition) melanoma patients received either neoadjuvant nivolumab 3 mg/kg for up to 4 doses or ipilimumab 3 mg/kg plus nivolumab 1 mg/kg for up to 3 doses. Combination treatment achieved high response rates but also a high incidence of severe toxicities, while treatment with nivolumab as single agent was safe but had a low clinical activity [138]. Lastly, in the OpACIN-neo trial 89 patients with resectable stage III melanoma were randomized 1:1:1 to receive one of the following neoadjuvant regimens: (Arm A) two cycles of ipilimumab (3 mg/kg) plus nivolumab (1 mg/kg) every three weeks; (Arm B) two cycles of ipilimumab (1 mg/kg) plus nivolumab (3 mg/kg) every three weeks; (Arm C) two cycles of ipilimumab (3 mg/kg) once every 3 weeks followed by two cycles of nivolumab (3 mg/kg) every two weeks. Arm B was associated with lower incidence of irAEs with 20% of patients experiencing a grade 3–4, as compared to 40% and 50% of patients in arm A and C, respectively. Despite the lower toxicity, the regimen with low dose ipilimumab achieved a similar rate of radiological and pathological response compared with arm A, while arm C achieved the worst results. In particular, 57% of patients in arm B achieved a radiological response, and 77% a pathological response [139,140].

The combination of pembrolizumab and ipilimumab has been also tested. In the NCT02089685 phase 1b trial, the standard dose (2 mg/kg) of pembrolizumab in combination with the low dose of ipilimumab (1 mg/kg) every 3 weeks for 4 doses followed by pembrolizumab (2 mg/kg) every 3 weeks for up to 2 years or disease progression was administered to unresectable stage III or IV (according to AJCC 7^th edition) melanoma patients. The estimated 1-year PFS was 69%. The estimated 1-year OS was 89%. Moreover, 45% of patients had grade 3–4 AEs although no treatment-related deaths occurred [141].

Besides anti-CTLA-4 and anti-PD-1 combination many other clinical trials are currently testing combinations of anti-CTLA-4 and anti-PD-1 mAbs with different ICIs. Table 2 summarizes some of these ICI combinations being currently tested in ongoing clinical trials. Among them, NCT03470922 trial also known as Relativity-047 is one of the most promising. Indeed, in this trial 714 patients with previously untreated, unresectable stage III or IV (according to AJCC 8^th edition) melanoma were 1:1 randomized to receive nivolumab at the dose of 480 mg plus placebo every 4 weeks or nivolumab at the dose of 480 mg plus relatlimab (anti-LAG3 mAb) at the dose 160 mg every 4 weeks. At a median follow-up of 13.2 months, the addition of relatlimab is associated with a longer median PFS (10.12 vs 4.6 months) as compared to placebo. On the other hand, the incidence of grade 3 or 4 treatment-related AEs was higher with the combination (18.9% vs 9.7%) [142]. To date OS results are still pending.

Table 2.

Novel clinical trials of ICI combination.

Agents	Phase	Status	ClinicalTrials.gov ID
Anti-LAG-3 (Relatlimab) +/− Anti-PD-1 (Nivolumab)	1/2a	Recruiting	NCT01968109
Anti-LAG-3 (Relatlimab) + Anti-PD-1 (Nivoluamb) vs Nivolumab	2/3	Active, not recruiting	NCT03470922
Anti-LAG-3 (Relatlimab) + Anti-PD-1 (Nivoluamb) + Anti-IDO1 (BMS-986205) vs Anti-LAG-3 (Relatlimab) + Anti-PD-1 (Nivoluamb) + Anti-CTLA-4 (Ipilimumab)	1/2	Recruiting	NCT03459222
Anti-IDO1 (Epacadostat) + Anti-PD-1 (Pembrolizumab)	1/2	Completed	NCT02178722
Anti-TIM3 (MBG453) +/− Anti-PD-1 (Spartalizumab)	1b/2	Active, not recruiting	NCT02608268
Anti-TIM3 (LY3321367) +/− Anti-PD-L1 (LY3300054)	1a/1b	Active, not recruiting	NCT03099109

CTLA-4: Cytotoxic T-Lymphocyte Antigen 4, IDO1: Indoleamine 2,3-dioxygenase-1, LAG-3: Lymphocyte activation gene-3, PD-1: Programmed cell death protein-1, PD-L1: Programmed death-ligand 1, TIM-3: T cell immunoglobulin and mucin domain 3.

Lastly, ICIs are also currently tested in combination with targeted agents, especially for melanoma patients harboring a BRAF V600 mutation. The latter is detected in about 50% of melanoma tumors and sensitizes the patient population to BRAF and MEK inhibitor combination. This combination of targeted agents represents a cornerstone of BRAF mutated melanoma treatment [143,144] since its administration is associated to both high rate of ORR and longer survival outcomes as compared to standard therapies. However as previously described for ICI-based immunotherapy the development of intrinsic and acquired drug resistance also limits the efficacy of BRAF and MEK inhibitor combination [145]. As a result, to date, it appears difficult to establish the right therapeutic algorithm for BRAF mutated melanoma patients, since both ICI-based immunotherapy and BRAF/MEK inhibitors are effective, but to a limited extent. Thus, several ongoing trials are evaluating not only the best treatment choice, but also potential combinations or sequential therapies including targeted therapy and ICI-based immunotherapy in BRAF mutated melanoma patients. Targeted therapy such as BRAF inhibitor can be combined to ICI-based immunotherapy such as anti-PD-1/PD-L1 mAbs. Indeed, BRAF pathway inhibition (with or without MEK inhibition) is associated to the up-regulation of melanoma antigen and HLA class I antigen expression as well as to a reduced production of immunosuppressive cytokines (interleukin-6 and interleukin-8). Furthermore, BRAF inhibition and the combination of BRAF and MEK inhibition are associated with PD-1 and PD-L1 up-regulation [146]. As a results BRAF inhibitor enhance the anti-tumor activity of PD-1/PD-L1 inhibitor by facilitating tumor cell recognition and destruction by cognate T cells. [146].

On this basis, the combination of targeted therapy and ICIs has been tested in a phase 2 clinical trial (NCT02130466). In this study BRAF V600E/K mutated unresectable stage III or IV (according to AJCC 8^th edition) melanoma patients were randomized to receive dabrafenib (a BRAF inhibitor) plus trametinib (a MEK inhibitor) plus pembrolizumab or dabrafenib plus trametinib and placebo. Median PFS was longer in patients treated with the combination of immunotherapy and targeted therapy as compared to those treated with targeted therapy alone (16.0 vs 10.3 months). However, the study did not reach the planned benefit for the statistically significant improvement of combinatorial treatment as compared to targeted therapy alone [147]. Two important ongoing phase 3 clinical trials are the NCT02908672 and the NCT02967692. In the NCT02908672, treatment-naive unresectable stage III or IV (according to AJCC 7^th edition) melanoma patients carrying a BRAF V600 mutation are randomized following a run-in period (28 days) with targeted therapy alone to receive atezolizumab (a fully engineered humanized IgG1 anti-PD-L1 mAb) plus cobimetinib (a MEK inhibitor) plus vemurafenib (a BRAF inhibitor) or cobimetinib plus vemurafenib and placebo. At a median follow-up of 18.9 months, PFS was significantly prolonged with atezolizumab versus control (15.1 vs 10.6 months). However, a longer follow-up is needed to evaluate benefits in terms of OS [148]. In the NCT02967692, unresectable stage III or IV (according to AJCC 7^th edition) melanoma patients carrying a BRAF V600 mutation are randomized to receive spartalizumab (a humanized IgG4 anti-PD-1 mAb) plus dabrafenib and trametinib or placebo plus dabrafenib and trametinib. To date survival outcome results are still pending.

So far, two main ongoing clinical trials are evaluating a sequential therapeutic approach. In the phase 2 clinical trial NCT02631447 BRAF V600 mutated metastatic melanoma patients are randomized, in a 1:1:1 ratio, to i) encorafenib (a BRAF inhibitor) and binimetinib (a MEK inhibitor) until disease progression followed by ipilimumab and nivolumab; ii) ipilimumab and nivolumab until disease progression followed by encorafenib and binimetinib; and iii) run-in period (8 weeks) of encorafenib and binimetinib followed by ipilimumab and nivolumab until disease progression followed by encorafenib and binimetinib. In the phase 2 clinical trial NCT03235245 unresectable and metastatic melanoma patients carrying a BRAF V600 mutation are randomized to receive ipilimumab and nivolumab with or without encorafenib and binimetinib following a previously run-in period (12 weeks) of targeted therapy.

7. Predictive biomarkers of response to ICIs in melanoma

The impressive but limited clinical efficacy of ICI therapy underlies the need to identify robust biomarkers to select patients who are likely to benefit from this type of immunotherapy. Several in vitro and in vivo studies have tried to identify a strong predictive biomarker of ICI based immunotherapy in melanoma patients. So far no one is effective. PD-L1 expression has been investigated as a predictive biomarker of clinical response to ICI-based therapy in various types of cancer including melanoma [149–151]. PD-L1 expression by malignant cells has been shown to correlate with a better clinical response to ICI-based immunotherapy in many types of solid cancer including triple negative breast cancer, NSCLC and urothelial cancer [152–154]. In contrast, in melanoma patients conflicting data have been reported [150] since also PD-L1 negative tumors may respond to ICI treatment [130]. The latter conflicting results might reflect the different characteristics of the antibodies utilized as a companion test to detect PD-L1 expression, the different cut-off utilized for immunohistochemical detection of PD-L1 expression, the different site of tissue sample biopsy analyzed and/or the different score in PD-L1 expression generated by the evaluation of its expression by cancer cells or by both cancer and immune cells [155]. Besides PD-L1, tumor mutational burden (TMB) is also currently under investigation as a potential biomarker. TMB is defined as the number of somatic mutations per megabase of tumor DNA. It has been hypothesized that a high tumor mutation load may affect the probability of generating immunogenic neoantigens, representing the mutations effectively targeted by activated TILs [98,156]. Several lines of evidence have shown that a high TMB is correlated with a better response rate as well as a longer survival outcome as compared with a low TMB in patients who have undergone ICI-based immunotherapy. These results have been validated not only in melanoma but also in other types of cancer [157–159]. Although it seems to be a reliable predictive biomarker, it is very difficult to stratify melanoma patients based on TMB, since not all the treated patients have a clinical benefit from ICI therapy, although most of them display a high TMB. Thus, to date, TMB is not considered an effective biomarker in melanoma for ICI-treated patients. Since neoantigen presentation to host’s immune system requires a fully functional HLA class I APM, one wonders whether the TMB predictive value would be increased by taking into account also the patient’s HLA class I APM component expression/function.

Recent technological innovations have allowed to delineate gene signatures that can be correlated with clinical outcome and predict response to immunotherapy in various malignancies. In a phase 2 study evaluating ipilimumab in patients with melanoma, a gene expression profiling was performed on tumor biopsies before and after treatment. Patients with high immune-related gene expression levels responded better to ipilimumab than those with lower basal expression [160]. Similar results have been shown utilizing an interferon gamma (IFNγ)- related gene signature in melanoma patients treated with pembrolizumab [161]. However, this data needs to be validated in a large cohort of melanoma patients.

As we have described above, HLA class I APM component expression/function is frequently defective in melanoma cells. One would expect that HLA class I antigen expression might be a reliable predictive biomarker, because of its crucial role in the interactions of cancer cells with host’s immune system. However so far this topic has been neglected and only a few studies have investigated this possibility. A retrospective analysis of two clinical trials (CheckMate 064 and CheckMate 069) has shown that primary response to anti–CTLA-4 requires a high HLA class I expression level by melanoma cells. In contrast, primary response to anti–PD-1 is associated with preexisting IFNɣ-mediated immune activation with HLA class II expression and components of innate immunity when HLA class I is compromised [162].

Several lines of evidence have also proposed the use of TIME characteristics as a predictive biomarker of clinical response to ICI-based therapy [112–115,163]. To date, the density, spatial distribution and localization of TILs within TIME are parameters considered important predictors of response to ICI-based immunotherapy for many types of cancer [164–166]. In patients with melanoma, tumor associated macrophages and myeloid-derived suppressor cells (MDSC) were found to correlate with poor anti-PD-1 [167] and anti-CTLA-4 [168] responses, respectively. However, the predictive value of TIME characteristics in melanoma patients treated with ICI still remains to be validated.

Lastly, growing evidence suggests the pivotal role of commensal microbiota in response to ICI therapy [169]. Several studies have identified strains of bacteria associated with response or resistance to ICI therapy in melanoma [170–173]. Moreover, patients with high microbiota diversity, have been shown to have a better ICI response, regardless of the bacterial species [172,174]. Noteworthy patients treated with antibiotics, during ICI therapy had a decreased tumor antigen-specific immune response [174] thus suggesting that the gut microbiota plays a key role in the response to ICI therapy. On the other hand, the underlying mechanisms by which the gut microbiota influences ICI therapy are unknown. In a recent study, He et al, suggest that the gut microbial metabolite butyrate could promote antitumor therapeutic efficacy through the ID2-dependent regulation of CD8⁺ T cell immunity [175]. Moreover, Baruch et al, have demonstrated in a phase 1 study the safety and feasibility of fecal microbiota transplantation and reinduction of anti PD-1 therapy in 10 patients with anti-PD-1-refractory metastatic melanoma. They observed two partial responses and one complete response [176]. However, additional studies are needed to clarify the mechanisms underlying microbiota and immune response interactions and how they influence the ICI treatments.

8. Conclusion

ICI-based immunotherapy has completely revolutionized the melanoma treatment both in metastatic and in adjuvant settings. Several studies are evaluating the potential efficacy and tolerability of combination and sequential therapies. Drugs under investigations include not only the “classic” ICIs but also the novel generation ICIs. Moreover, search for predictive biomarkers as well as characterization of immune escape mechanisms are areas of active research.

9. Expert opinion

ICI-based immunotherapy has revolutionized the treatment of melanoma both in metastatic and in adjuvant settings, significantly prolonging the DFS, PFS and OS of treated patients. However, not all patients benefit from ICI-based immunotherapy while some others achieve long lasting clinical responses. Results from clinical trials testing the efficacy of CTLA- or PD-1-specific mAbs have demonstrated that dosage and schedule of ICIs play a crucial role in their efficacy and tolerability. Indeed, different doses of ipilimumab or pembrolizumab displayed a different efficacy and tolerability profiles. The PD-1-specific mAbs, nivolumab and pembrolizumab, are more effective as compared to the CTLA-4-specific mAb ipilimumab in prolonging the DFS and OS of treated melanoma patients both in adjuvant and metastatic settings. Both are currently approved for the treatment of melanoma patients and display a similar toxic profile. However, it still remains to be defined which of the two PD-1-specific mAbs can be considered standard of care. A novel clinical trial comparing their respective clinical activities still needs to be designed and performed in melanoma patients. One should take into account that in a metastatic setting both nivolumab and pembrolizumab have been compared to standard chemotherapy or to ipilimumab. In contrast in an adjuvant setting nivolumab has been compared to ipilimumab while pembrolizumab has been compared to placebo.

Treatment of metastatic melanoma patients with ipilimumab includes four doses of drug administration. In contrast, both nivolumab and pembrolizumab based treatments are administered until disease progression including those patients who achieve long lasting complete clinical responses. For those patients the treatment duration remains still to be determined. One would expect that patients treated with PD-1-specific mAbs need to continue the treatment as compared to those treated with ipilimumab since CTLA-4 is thought to regulate T-cell proliferation early in an immune response whereas PD-1 suppresses T cells later in an immune response, primarily in peripheral tissues. As a result, a T cell memory response is expected to be more stimulated by ipilimumab rather than PD-1-specific mAbs. Clinical trials are currently evaluating the possibility of therapeutic withdrawal following two years of PD-1-specific mAb treatment. The results obtained from these studies will be crucial to reduce the occurrence of AEs as well as to increase patient compliance.

In order to define the best therapeutic algorithm for melanoma treatment it should be taken into account that about 50% of melanoma carries a BRAF V600 mutation. Both targeted therapy and ICI-based immunotherapy are currently clinically approved for those patients. To date, no trial was designed to test directly whether ICI-based immunotherapy is effective in BRAF mutant melanoma patients. The clinical evidence of the therapeutic efficacy of ipilimumab, nivolumab or pembrolizumab in BRAF mutant melanoma patients is derived from subgroup analysis of phase 3 clinical trials including both BRAF mutated and wild type patients. As a result a novel clinical trial should be designed and implemented in order to compare the clinical efficacy of targeted therapy and PD-1-specific mAb-based monotherapy in BRAF V600 melanoma patients.

Many clinical studies are currently evaluating the potential role of the combination of different ICIs or the combination of ICIs and targeted therapies in melanoma patients in order to increase their therapeutic efficacy. Targeted therapy may augment the efficacy of ICI-based immunotherapy, improving tumor antigen presentation by tumor cells as well as T cell infiltration and decreasing tumor burden. On the other hand, as shown by ipilimumab and nivolumab combination, use of different ICIs may overcome some types of ICI resistance. However, in view of the expected high rate of AEs, this type of therapies might be limited only to cancer centers with high levels of clinical expertise. As a result, not all patients will benefit from this type of therapies.

As we have previously described, identification of effective predictive biomarkers represents an unmet clinical need in order to identify patients who may benefit from ICI-based immunotherapy. Although useful, both PD-L1 and TMB do not appear reliable predictive biomarkers. In addition, the role of IFNγ-gene signature appears to be limited by its complexity and high related costs. On the other hand, the role of HLA class I expression on tumor cells as a predictive biomarker has been widely investigated only in preclinical models while, few clinical evidence have been produced so far. The resulting paucity of this information is surprising, since i) HLA class I antigens play a crucial role in cancer cell recognition by host’s immune system, ii) HLA class I antigen expression has been shown to be defective in a high percentage of cases in most if not all the malignancies investigated and iii) HLA class I defects have been shown to have clinical significance. As we have described above, Rodig SJ et al. showed that HLA class I expression plays a crucial role in predicting clinical response to CTLA-4-specific mAb but not to PD-1-specific mAbs. In addition, clinical benefit from PD-1 specific mAbs are predicted by HLA class II expression in the same population of treated patients. The former results may reflect the cancer immune editing process and therefore the development of ICI resistance following chronic exposure to PD-1-specific mAbs. Lack and down-regulation of HLA class I expression can be caused by development of mutations in some of the antigen processing machinery components such as β2-microglobulin and attenuation of IFNγ-induced HLA class I expression elements, respectively. However, further clinical studies are needed to define the role of HLA class I and II expression as well as other antigen processing machinery components in order to implement them as ICI predictive biomarkers in clinical practice.

Lastly, the potential role of the commensal microbiota in predicting ICI-based immunotherapy appears as a novel field of investigation. So far it has been investigated as a potential modulator of tumor-immune response. However, one should take into account that commensal microbiota may also exert a direct effect on tumor cell proliferation.

Identification of predictive biomarkers will also be useful to design novel clinical trials based on personalized immunotherapy. In contrast the lack of selection of treated patients might cause treatment failure as the administration of pembrolizumab and epacadostat (IDO inhibitor) in combination in the NCT02752074 clinical trial failed to increase the PFS of metastatic melanoma patients as compared to pembrolizumab administration alone.

Lastly in order to define an effective therapeutic algorithm, we should take into account that some subgroups of melanoma patients do not benefit from ICI monotherapy because of presence of brain metastasis, high levels of LDH, high tumor burden and requirement of the administration of high doses of corticosteroids. In these patients, because of their poor prognosis, the combination of different ICIs or of ICIs and targeted therapies should be a better therapeutic strategy.

Article highlights.

• The ICI based immunotherapy has dramatically changed the management of metastatic melanoma as well as of complete resected high risk melanoma patients.

• So far, the main drugs used in melanoma treatment are anti-CTLA-4 mAbs and anti-PD-1 mAbs.

• Several ongoing clinical trials are evaluating the role of novel ICIs in melanoma treatment.

• Other trials are currently evaluating ICIs in combination or in different sequencing approaches with other type of therapies.

• A lot of patients didn’t achieve any benefits from this therapy. Then, the study of immune escape mechanisms is crucial to better understand while these patients didn’t respond to ICIs. Moreover, is essential to determine ICI acquired resistance mechanisms.

• In order to define the most correct therapeutic algorithm, it is needed to determine effective predictive biomarkers.