The first step in the initiation of the host immune response in primary cutaneous melanoma is recognition of the antigens that will induce inflammatory intratumoral infiltrate (19).

Tumor antigens in melanoma are of two types: specific antigens that are expressed by tumor cells (tumor-specific antigens, TSAs) and tumor associated antigens (TAAs), which are predominantly found in the tumor cells. TSA are also called ‘neoantigens’ because they are newly formed antigens and they are not found in the normal human genome. TSA exert high immunogenicity and induce T-lymphocytes that are not affected by central tolerance and maintain for a period of time an efficient adaptative anti-tumor activity. There is an extremely large number of possible mutations, so that each patient tumor is unique, tumors from the same patient can be different, and there is also intratumoral heterogeneity (20). On the other hand, TAA are mainly found in tumor cells but can also be expressed by normal melanocytes and most of them are of intracellular protein origin. Most often melanoma cells express only MHC class I and do not express MHC class II molecules. TAA do not induce an ideal cytotoxic T-cell reaction and generate reduced formation of tumor-specific CD4⁺ and antibodies (4,11,21).

Malignant tumor cells can employ different stratagem to avoid T-cell intervention. In some cases, tumor cells can reduce their expression of TAA or they can produce chemokines and cytokines with immunosuppressive consequences (22–25).

DC are among the first cells to reach the tumor, recognize tumor antigens and play a pivotal role in the initiation and regulation of both innate and adaptive immunity (17). They are the most potent antigen-presenting cells (APCs) and after processing of tumor-associated antigens, they induce a specific antitumor response by cross-presentation of antigens to CD8⁺ T lymphocytes by MHC class I molecules and to CD4⁺ by MHC class II molecules (26). In order to obtain a potent anti-tumoral immune response of T lymphocytes, the antigen presentation must be efficiently done by mature dendritic cells. DC can additionally contribute to the cytotoxic immune reaction directly and by activating NK cells (27).

DC from the skin, Langerhans cells, dermal DC and plasmacytoid DC may exhibit different phenotypes, with dualistic functions (28). In the tumor microenvironment, mediators released by tumor cells or tumor-associated macrophages (TAMs) such as IL-8, IL-10, TGF-β1 and VEGF limit normal DC maturation, in an attempt to evade host immune response (20,29,30).

DCs in the immature state fail to properly stimulate T-cells. On the other hand, by favoring proliferation of regulatory T-cells, they block T-cell responses. In this way they are mediators of immune tolerance despite immune activation, a phenomenon described in patients with melanoma in 1997 by Enk et al (31). Immature/tolerogenic DC regulate tumor angiogenesis and favor an active tumor growth (32). This proangiogenic effect is no longer operative with DC maturation. The maturation process of dendritic cells is dependent on factors encountered in the tumor microenvironment (33).

Mature DC are distributed mainly peritumorally and their density, together with the activation status of T lymphocyte, correlated with melanoma tumor thickness and patient's survival, making Kobayashi et al and Simonetti et al to recommend the use of these parameters as a predictor of treatment response in patients treated with immunotherapy (34,35).

An adequate number of mature DC in tumor area can eliminate malignant cells (36). Lotze showed that infiltration with mature DC in primary tumors is associated with fewer metastases and better prognosis (37).

In an ideal situation, antigen-presenting cells lead to activation of effector memory T-cells in the lymph nodes that mediate antitumor effects at tumor site, producing new antigens from destroyed tumor cell and creating a tumor-immunity cycle (2,38).

After recognizing the antigen presented by APC, T-cells require several signals for full activation. The first signal depends on the antigen and is represented by MHC I or MHC II complexed with the presented peptide, binding to the T-cell receptor (TCR) (39). Insufficiently presented antigen on APC cannot activate T-cells and induce immunologic ignorance (40).

Cell adhesion molecules are responsible for maintaining the connection between the two cells and allow as many TCRs on the T-cell as possible to become activated (41). In metastatic disease a diversity of TCR was observed between different sites that can have different clinical evolution, explained by the selection of T-cells directed against different tumor antigens or against different epitopes from the same antigen (38).

The second signal of activation requires specific interaction between T-cell receptors from their ligands and APC (from CD28/B7 family), and is antigen-independent (39). These receptors are of two types: co-stimulators or co-inhibitors. Many ligands can bind to multiple receptors, of each type (42).

The interplay between costimulatory receptors and their ligand completely activate T-cells and induce production of tumor-specific T-cells (43).

In normal conditions, co-inhibitor receptors [programmed cell death-1 (PD-1), cytotoxic T lymphocyte antigen-4 (CTLA-4), B and T lymphocyte attenuator (BTLA) and T-cell immunoglobulin and mucin domain-3 (TIM-3), known as ‘immune checkpoint molecules’, and their ligands (respectively, programmed death-ligand 1 (PD-L1) or PD-L2, CD80 or CD86, herpesvirus entry mediator (HVEM) and galectin 9], inhibit T-cell activity and are involved in maintaining self-tolerance and modulating the intensity and duration of an anti-infectious immune response (30,42). Activation of inhibitory immune checkpoints by cancer cells shield them from the immune attack and allow them to proliferate further (38,44).

APC T-cell interaction can be mediated also by tumor necrosis factor (TNF) family members and their receptors, by the soluble cytokines from the microenvironment (TGF-β, IL-1, IL-10, IL-12, IL-18) and by immune-inhibitory molecules, metabolic enzymes that deprive lymphocytes of necessary aminoacids [indoleamine 2,3-dioxygenase (IDO)] that can be new therapeutic targets (42,45,46).

Thus, it has been observed that the immune system has a dual role, by modifying the interactions between co-stimulatory and co-inhibitory signals. On the one hand, it can control malignant cells but on the other hand, it may favor tumor progression (47).

TIL include different subsets of lymphocytes, in different proportions, adjusted by cytokines secreted by tumor cells or other residents: CD8⁺ T-cells, CD4⁺ T-cells, NK cells and B-cells (48,49). T lymphocytes have a major role in the antitumor immune responses and are the dominant elements in the tumor microenvironment.

TSA generate highly sensitive and specific CD8⁺ T recognition and data suggest that TIL may specifically target TSA (50,51). TAA can induce tumor-specific CD8⁺, CD4⁺ T-cells and antibodies against TAA (30).

CD8 effector T-cells (or cytotoxic T-cells) inhibit tumor proliferation, either through direct cytolytic action on tumor cells or by releasing interferon (IFN)-γ and TNF-α (50,52). Therefore, a large infiltration of CD8⁺ T-cells in the tumor is related to a good prognosis in patients with melanoma (53). Protracted antigen exposure can cause tumor antigen-specific T-cells ‘exhaustion’. PD-1 and TIM-3 have been considered immunohistochemical markers for ‘exhaustion’ (4). Based on the balance between co-stimulatory and co-inhibitory signals from the micro-environment, CD8 T-cells may exert different functional states (54).

There are four main types of CD4⁺ T-cells (T helper cells), with distinct properties and in variable percentage: Th1, Th2 Th17, and Treg.

Th1, Th2 help antitumor fight by stimulating the activity of CD8⁺ T-cells through mediators such as IFN-γ, TGF-β or IL-2 (55,56). Th1 can also favor activation of macrophages and maturation of dendritic cells, while Th2 can use cytotoxic weapons of eosinophils (57,58). A consistent intratumoral infiltration of CD8⁺ T-cells and Th1 cells is correlated with favorable prognosis and better survival in most human cancers (59).

Th17 cells have two antagonist forms: in certain cytokine milieu can convert to Th1 characteristics and exert antitumor effects, while in other conditions can switch to regulatory T-cells characteristics and induce tumor progression (60).

Regulatory T-cells (CD25/FoxP3 suppressive T-cells, Treg) are a subtype of T-cells with a key role in preventing autoimmune diseases. Their presence in the tumor microenvironment inhibits the antitumor immune responses. They are attracted to the tumor microenvironment by chemokines secreted by tumor cells and macrophages. Tregs are activated after recognizing TAA released from destroyed tumor cells and then specifically suppress the activation of TAA-specific effector T-cells and the efficient tumor cell destruction by various mechanisms (IL-10, TGF-β) (41,61). Moreover, Treg can suppress the action of several types of immune system cells such as CD8⁺ T-cells, NK cells, B cells and APC (62).

There are few studies done on the prognostic role of Treg in melanoma. Some have noted the association between the presence of a high Tregs infiltrate with the local recurrence of the disease, fast tumor progression, a higher metastasis rate in the sentinel lymph node, but without any association with tumor thickness or patient survival (63–66).

It is considered that ratios between different subsets of T-cells can provide more accurate information on the local immune balance and CD8/FoxP3 (effector/regulatory) ratio and CD8/CD4 (effector/helper) ratio are the most commonly used (67). To skip from immune attack, tumor cells suppress tumor-specific T effector and induce immunosuppressive Treg cells, thus reducing CD8/Treg ratios. In patients treated with combination checkpoint therapy, CD8/Treg ratios increased and was associated with improving survival in melanoma (68).

NK are effector cells of the innate immune system and play an essential role in cancer immune surveillance due to their ability to destroy tumor cells independent of MHC or previous activation (69).

NK also participate in regulation of adaptive immune reactions through interactions with DCs, that activate NK cells and determine a potent cytotoxic immune response against tumor cells (27).

Tumor cells may defeat NK by releasing TGF-β, by low antigenicity expression or by increased MHC I expression (33,69–71). Tregs can also contribute by rivaling with NK for IL-2 (30). The activation state of NK cells is modulated by activating and inhibitory receptors.

The prognostic role of NK cell infiltration in melanoma has not been evaluated, but they seem to limit the development of hematogenous metastases (28).

B cells represent 15–20% of all infiltrating lymphocytes (16,30,66). The exact roles of tumor-infiltrating B cells in antitumoral immune response are not defined. A trend for their higher prevalence was observed in thicker tumors and an increased density of B lymphocytes infiltrating primary cutaneous melanomas was associated with reduced risk of distant metastases and longer survival (28,72) The roles of plasma cells are even less understood. Bosisio et al recently reported a significantly worse survival in primary melanomas with clusters of plasma cells compared to a better outcome in the cases with sparse plasma cell infiltrate. Most plasma cells were polyclonal, expressing predominantly IgG and IgA (73).

TAMs are a heterogeneous group of cells with antigen-presenting capacity and represent the predominant inflammatory cells of the tumor infiltrating lymphocytes. They are found in all stages of the tumor progression (30). Based on the signals from microenvironment, macrophages can render antagonist functional characteristics.

M1 macrophages, the classical type, have high antigen-presenting action and also can produce Th1 cytokines and exert antitumor effect (69).

M2 macrophages have a low antigen-presenting activity, inhibit CD8⁺ T-cell and NK cell activity, stimulate switch to Th2 and Treg predominance, stimulate antibody production, angiogenetic effects and favor tumor cell migration (74,75). It is the predominant profile in TAM, induced by products secreted by dendritic cells, Treg lymphocytes and tumor microenvironment. During tumor growth and progression, macrophages convert from M1 to M2 phenotype (29).

Studies have shown correlation between TAM density and tumor thickness and ulceration, but no significant correlation with survival in patients with melanoma (76).