CD8 T cells expressing NK associated receptors are increased in melanoma patients and display an effector phenotype

Introduction

Inhibitory and activating MHC-class I specific receptors were initially described as receptors present on NK cells that regulate lysis of target cells expressing the appropriate ligand [7, 27, 31, 38]. Several studies have demonstrated that the expression of these receptors is not exclusive to NK cells and it has been made clear that different subsets of T lymphocytes can express NK-associated receptors (NKRs) [35, 55]. NKRs expressed on T cells are functionally active and may act as regulators of T cell-mediated cytotoxicity [35, 64]. In humans, three families of HLA class-I specific NKRs have been identified including activating and inhibitory isoforms: Killer Immunoglobulin-like Receptors (KIRs), leukocyte immunoglobulin-like receptors (LILR; also termed Immunoglobulin Like Transcripts, ILT, or CD85) and C-type lectin receptors [27, 31, 37]. Together with these HLA class-I specific receptors, studies have described the expression on CD8⁺ T lymphocytes of non-HLA class I specific NKRs such as CD56, CD57 or CD244 that have been correlated with cell activation and may have functional implications in the immune response against tumour antigens [42, 45, 57, 59].

Tumour-specific CTLs can express HLA class-I specific NKRs and signalling via inhibitory receptors interferes with CTL effector function upon TCR engagement by MHC-peptide complexes [26, 54]. Several studies have investigated whether signalling via KIR can affect melanoma recognition by CTLs. Thus, infection of gp100-specific CTLs with a vaccine virus encoding human KIR3DL1 receptor [4] or the natural expression of CD94/NKG2A on Melan-A specific CTLs [54] resulted in inhibition of lysis of gp100 and Melan-A positive melanomas respectively. Consequently, the expression of the appropriate HLA class-I ligand by melanoma cells may regulate T cell function by crosslinking MHC-specific inhibitory receptors. In addition, melanoma cells that have HLA class-I loss of selective alleles can elude the inhibitory effect of HLA-specific NKRs. Thus, KIRD2L2 expression on an HLA-A24-restricted CTL clone from a melanoma patient inhibited lysis of autologous melanoma cells but did not affect lysis of a melanoma variant that had lost the expression of all HLA alleles except for HLA-A24, the restriction element [26]. Total or partial loss of HLA class I antigens is a frequent event in different tumours [1, 12, 32, 46, 47, 50]. In an extensive study of 61 melanoma cell lines analyzed by flow cytometry, it has been reported that 51.7% of melanoma cell lines had altered HLA class I phenotypes, including 3% of HLA class I total loss, 3% of HLA-ABC downregulation, 19.7% loss of heterozygosity and 26% of HLA-B downregulation [48]. Although inhibitory HLA class-I specific NKRs on CTLs can regulate melanoma cell lysis in vitro [54], their role in the control of the immune response in vivo has not yet been demonstrated. On the other hand, activating NKRs are also present in a subset of CD8⁺ T cells and can amplify signals mediated through the TCR acting as costimulatory molecules [52]. Thus, changes in the HLA class-I profile in melanoma cells can have profound effects on both CTL and NK cell responses [1, 12, 32, 46, 47, 50].

It has been recently demonstrated that, the analysis of KIR gene distribution and the frequency of inhibitory and activating KIR genes in melanoma patients and health control did not show significant differences in the Bulgarian population. However, an association of the inhibitory KIR2DL2/2DL3 genes with their HLA class I ligands was found more frequently in melanoma patients suggesting that the prevalence of inhibitory signals could be relevant for the outcome and progression of melanoma [40].

NKR⁺ T cells are almost totally absent in cord blood but have been found in increased quantity with age [17, 58]. NKR expression on CTLs has been found altered in several clinical conditions that involve chronic activation of the immune system such as HIV infection, tumors, rheumatoid arthritis or aging [2, 4, 6, 23, 59]. Thus, the expression of CD56 on CD8⁺ T cells has been correlated with the cytolytic effector function [45] and has been found in decreased quantities in HIV infected individuals [59]. In addition, the expression of CD57 on CD8⁺ T cells can define replicative senescence and antigen-induced apoptotic death [11].

Several data indicate that NKRs are preferentially expressed on differentiated CD8⁺ CD28⁻ T cells [39, 55, 59]. An increased proportion of CD28⁻ T cells has been reported in aging and other situations of chronic immune stimulation, suggesting that NKR expression on differentiated CD8⁺ CD28⁻ T cells constitutes a regulatory mechanism of effector function [20, 55, 59]. In addition, it has recently been demonstrated that the differential expression of HLA class-I specific NKRs on CTLs is correlated with transition from effector to memory T cells suggesting a role for NKRs in the survival of CD8⁺ memory T cells [46, 63, 69].

This work characterizes peripheral blood CD8⁺ T cells from melanoma patients in relation to the expression of several NKRs and costimulatory molecules. The analysis performed included the study of HLA class-I specific receptors (KIRs, LILRB1 and CD94/NKG2) and non-HLA class-I specific receptors (CD56, CD57, CD161 and CD16) related to the immune response against melanoma. As the expression of NKRs on CD8⁺ T cells is preferential on CD8⁺ CD28⁻ T cells, we have also analyzed the expression of these NKRs on CD8⁺ T cells according to the CD28 phenotype.

Materials and methods

Results

Discussion

Previous studies have shown that NKRs can be expressed on several subsets of human peripheral blood CD8⁺ T cells [34, 35, 37, 55, 58, 64]. Further characterization of these subsets showed that NKR⁺ T cells express intracellular and surface markers characteristic of antigen-experienced T cells. The expression of inhibitory HLA class-I specific NKRs may govern TCR-mediated CD8⁺ T cell activation [35, 58, 67]. In contrast, activating NKR expressed on T cells can act as co-stimulatory molecules that augment TCR-mediated activation [52]. The expression of NKRs on CD8⁺ T cells has been found altered in several pathologies including infections and tumours and several changes in the expression of NKRs have been correlated with disease progression [13, 14, 19, 23, 34, 59].

We have analyzed the expression of HLA-specific NKRs on CD8⁺ T cells from melanoma patients. The analysis of the CD94 C-type lectin receptor shows that this molecule is expressed in a subset of CD8⁺ T cells and that CD94⁺ T cells express the inhibitory NKG2A molecule at higher levels than the activating NKG2C molecule both in melanoma and healthy donors. Thus, both forms of the receptor, inhibitory CD94-NKG2A/B and costimulator CD94-NKG2C/E receptors [52] are present in circulating T cells [30]. Differential expression of CD94-NKG2 isoforms has been observed on MART-1 specific T cells in different areas of melanoma. Whereas the inhibitory receptors were found within the vitiligo areas, both activating and inhibitory isoforms were found within the tumor [44]. Vetter et al., by RT-PCR, showed a predominant expression of CD94 associated with NKG2C/E molecules in tumour infiltrating lymphocytes (TIL). The reversal in the ratio of inhibitory to activating receptors in TIL versus peripheral blood lymphocytes could suggest either the presence of specific factors in the tumor microenvironment or the preferential homing of T cells with distinct stages of activation [66]. Furthermore, it has previously been demonstrated that melanoma cell lysis mediated by CTLs can be inhibited by CD94/NKG2A heterodimer [54] but the possible role of CD94-NKG2C/E as regulators of immune response against melanoma should be further analyzed. The ligand of CD94/NKG2 receptors is the nonclassical class I molecule HLA-E [8, 10, 28]. The expression of HLA-E on the cell surface of tumour cell lines (including melanoma cell lines) has been correlated inversely with the expression of other HLA class I molecules suggesting that HLA-E expression on tumour cells could regulate specific CTL-mediated cytotoxicity by interaction with inhibitory or costimulatory CD94/NKG2 receptors [32].

The percentage of KIR⁺ CD8⁺ T cells was very low both in healthy donors and melanoma patients, with values below 5% in most individuals. The possible functional relevance of these results are difficult to evaluate since the mAbs used in our study recognize both the inhibitory and activating forms of the receptors and also because the KIR expression depends on the individual genotype. In this respect it is interesting that, the recent report by Naumova et al. [40] showing no significant differences among KIR genotypes between melanoma patients and healthy persons. The expression of the LILRB1 receptor, CD85j, was low in the majority of the individuals and no statistically significant differences were found between melanoma patients and healthy donors. However, two melanoma patients showed a high percentage of CD85j⁺ CD8⁺ T cells (63%, 73%). The expression of the inhibitory receptor CD85j on CD8⁺ T cells has been shown increased in some patients with lymphoproliferative disorders [14]. This receptor has been found to interact through its amino-terminal Ig domain with several classical HLA class I molecules and the nonclassical HLA-G and HLA-F molecules [16, 29, 41]. UL18, a human CMV protein homologous to HLA class I, has also been identified as a ligand for CD85j [15, 18].

We have also observed that circulating CD8⁺ T cells from melanoma patients have a significant increase of CD56 and CD57 expression. CD56 surface expression has been correlated with cytolytic effector function of CD8⁺ T cells [45]. Furthermore, it has been described that in vitro stimulation with anti-CD3 leads to an expansion of CD3⁺ CD56⁺ T cells that display cytotoxic ability against several tumours including melanoma [24]. It is interesting to note that the expression of CD57 on CD8⁺ T cells has been associated with chronic antigen stimulation and considered as a marker of proliferative inability [11]. Thus an expansion of CD57⁺ CD28⁻ CD8⁺ T cells was also found in myeloma patients probably as the consequence of persistent antigen stimulation [56]. Altogether, our results suggest that the expansion and accumulation of circulating CD56⁺ and CD57⁺ T cells in melanoma patients is also caused by antigen-driven activation of specific T cells.

The expression of NKRs on CD8⁺ T cells from melanoma patients is highly variable suggesting that lymphocyte activity may differ considerably between melanoma patients, as previously described by Speiser et al. on CD3⁺ T cells [55]. Moreover, the expression of NKRs on CD8⁺ T cells is preferential in the CD28⁻ subset [5, 13, 55, 59]. Our results showing that circulating CD8⁺ T cells from melanoma patients have a decreased expression of the costimulatory molecule CD28 when compared with age-matched healthy donors confirm and extend previous work by Martinez-Escribano et al. [33]. It could be speculated that the increase in the percentage of CD8⁺ CD28⁻ cells is related with the tumour load and that patients with metastatic melanoma should have a greater expansion of the CD8⁺ CD28⁻ T cell subset. However, our preliminary results show that only one out of the five patients with metastatic melanoma has a significant decrease in the expression of CD28 (data not shown) suggesting that metastatic melanoma patients have a deficient CD8 mediated immune response against the tumour. Further studies are necessary to test this possibility. When the analysis of NKRs was performed according to the CD28 phenotype, no statistically significant differences between melanoma patients and healthy donors were found, suggesting that the increase of NKR⁺ CD8⁺ T cells observed in melanoma patients is the consequence of CD8⁺ CD28⁻ T cell expansions and that the high variability observed in the expression of NKRs might be related to the differences in the proportion of CD8⁺ CD28⁻ T cells among melanoma patients.

It has been proposed that the CD8⁺ CD28⁻ T cell subset observed in vivo arose as a consequence of repeated rounds of antigen-driven proliferation [20–22]. In vitro experiments clearly showed that CD28⁻ cells are derived from CD28⁺ precursors and can be considered as terminally differentiated antigen experienced effector cells that emerge and expand in response to chronic antigenic stimulation [43, 62]. In head and neck cancer patients an expanded proportion of CD8⁺ T cells lacking CD28 has been observed. In these patients CD8⁺ CD28⁻ T cells showed an increased apoptosis rate that might suggest that chronic antigenic stimulation drives the recruitment, differentiation and finally death of these cells [62]. After surgical resection of the tumour the proportions of CD28⁻ and CD28⁺ CD8⁺ T cells were found normalized [62]. In addition, the downregulation of CD28 on CD8⁺ T cells has been previously described in many others clinical conditions that involve chronic activation of the immune system as HIV infection or aging [9, 13, 20, 59, 68].

Recent studies show that the simultaneous staining of CD28 and CD27 molecules defines several stages of T cell differentiation [3, 49, 60]. We have found that melanoma patients had a decrease of the naïve CD8⁺ CD28⁺ CD27⁺ T cell subset while the CD28⁻ CD27⁻ CD8⁺ T cell subset was expanded. Recently, this CD28⁻ CD27⁻ CD8⁺ T cell subset has been characterized by the expression of perforin, granzyme B, and CD95L [65] and by an increased cytotoxicity and a decreased proliferative potential [25]. Downregulation of CD28 and CD27 has been correlated to the differentiation status of antigen-specific T cells [60, 61, 65].

Within CD8⁺ CD28⁻ T cells several subpopulations can be distinguished based on the expression of surface and intracellular markers. The increased mean fluorescent channel of perforin within the CD8⁺ CD28⁻ T cell subset in melanoma patients compared to healthy persons may indicate an active immune response against the tumour. It has been previously described that CD244 expression on CD8⁺ T cells is correlated with perforin content [53] indicating that CD28⁻ CD27⁻ CD8⁺ T cells expanded in melanoma patients are terminally differentiated effector cells that are CD244⁺ and express high levels of perforin.

Based on the coexpression of CD28 and CD45RA, different CD8⁺ T cells subsets can be defined: naïve T cells (CD28⁺ CD45RA⁺), memory T cells (CD28⁺ CD45RA⁻) and effector T cells (CD28⁻ CD45RA⁺) [25]. Whereas the percentage of naïve cells was higher than 40% in 75% of healthy persons, only 31% of melanoma patients showed this phenotype. Interestingly, 25% of melanoma patients had less than 20% of naïve CD8⁺ T cells further suggesting that a significant expansion of terminally differentiated CD8 effector cells can be found in a high proportion of melanoma patients.

Once generated, tumour-specific effector T cells must undergo clonal expansion in order to maintain control over the tumour. Chronic exposure to tumour antigens either during tumour development or during immunotherapy may hamper the ability of the immune system to prevent cancer occurance or progression as, after multiple rounds of antigenic stimulation, T cells can undergo replicative senescence, comprising irreversible cell cycle arrest and alterations in effector function. In addition to irreversible growth arrest, senescent cells acquire resistance to apoptotic stimuli and show increased levels of bcl2 expression [21]. Together with CD28 as hallmark of immunosenescence, CD57 has also been associated with senescence and is expressed by the majority of CD28⁻ T cells [11, 36, 51, 56].

In conclusion, we have found that circulating CD8⁺ T cells from melanoma patients display a distinct effector phenotype characterized by downregulation of costimulatory molecules and high expression of NKRs and perforin. The increased percentage of these cells is probably related to the acquisition of an effector phenotype as the consequence of tumour antigen stimulation. We suggest that the increased expression of NKRs on T cells may contribute to the final outcome of the immune response against melanoma, either stimulating or inhibiting the function of CD8 effector cells. Blocking inhibitory receptor function and enhancing activating receptors may represent new strategies with therapeutic potential against melanoma.