NKp30 isoforms and NKp46 transcripts in metastatic melanoma patients: Unique NKp30 pattern in rare melanoma patients with favorable evolution

ABSTRACT

Given the NK cell-based immunosurveillance of melanoma, we investigated the prognostic value of NKp46 transcript and NKp30 isoform (NKp30A, NKp30B and NKp30C) profiling in blood of 187 melanoma patients including 13 long survivors (LS), metastatic patients that have controlled the disease. Compared to healthy volunteers (HV), patients had reduced amounts of transcripts of the three NKp30 isoforms (NKp30 A, B and C) but similar ratios between NKp30 isoforms (ΔAB, ΔAC, ΔBC). Stratification of patients according to disease stage showed higher NKp30C and lower NKp46 transcripts in stage IV patients. Furthermore, patients with previous history of conventional chemotherapy displayed reduced NKp30A transcripts. The expression levels of NKp30 isoforms failed to predict survival from sampling of patients, while NKp46 expression predicted melanoma outcome. LS patients displayed elevated NKp30A levels, accordingly high ΔAB and ΔBC ratios, and a unique pattern of rare allelic variants of NKp30 SNPs. Moreover, NK cells from LS displayed correlated NKp30/NKp46 membrane expression, high spontaneous and NKp30- or NKp46-triggered degranulation. These data outline the impact of NKp30 and NKp46 transcripts on melanoma evolution and identify unique genetic features of NKp30 associated with higher NK activation in rare LS melanoma patients that control a metastatic disease.

Introduction

Melanoma is a severe form of skin cancer with high metastatic potential. Metastatic stage IV melanoma patients have a short median survival (6–9 mo). In some patients, the immune system may control the evolution of the disease. The rare cases of spontaneous regression of metastatic melanoma have prolonged tumor-free survival (>24 mo) and present large particular infiltrate of immune cells comforting the role of the immune system in tumor control.^1,2 Cytotoxic immune cells are key antitumor effectors releasing cytotoxic molecules in tumor cells. Among them, Natural Killer (NK) cells infiltrate several tumors. They are potent cytotoxic effectors, acquiring high lytic potential in response to cytokine activation. Their role was first demonstrated in myeloid leukemia and there is now growing evidence that NK cells are involved in the control of human solid tumors.^3,4

NK cells have the natural ability to distinguish stress-induced molecules and alterations of MHC-I molecules on transformed targets. Activating and inhibitory receptors on NK cells are triggered during target cell recognition and the integration of these opposite signals determines NK activation.^5,6 The main activating NK receptors are NKG2D and the natural cytotoxicity receptors (NCRs), NKp30 (also named NCR3), NKp46 (also named NCR1) expressed by resting NK cells and NKp44 (also named NCR2) induced on cytokine activated NK cells. Activating and co-receptors recognize various ligands on targets that are upregulated upon cellular stress.⁷ Most of the inhibitory receptors recognize different HLA class-I molecules and include the Killer Immunoglobulin Receptors (KIRs), NKG2A, CD85j and LAIRs.^8-10 Accordingly, NK cells can kill target cells that have lost (or express low amounts of) HLA class I molecules which are frequent alterations of tumor cells including melanoma cells.

Recently, polymorphisms of activating NCR3/NKp30 receptor controlling the receptor functionality and involved in the evolution of Gastrointestinal Stromal Tumor (GIST) were identified. Three major NKp30 isoforms were identified, each one with a specific intracytoplasmic domain i.e. NKp30A, NKp30B and NKp30C. NKp30A and NKp30B signaling mediates cytolytic and IFNγ/TNF-α production, respectively, while the NKp30C isoform induces the production of the immunosuppressive cytokine IL-10.⁹

Several recent studies correlated the NKp30 isoform expression pattern with the prognosis and evolution in different malignancies. Hence, a high level of the NKp30C or a low level of NKp30B isoforms was found to correlate with a negative impact on the prognosis of patients with GIST or neuroblastoma, respectively.^11-13 Furthermore, Delahaye et al., have shown that the predominant transcription of the NKp30C isoform determined the prognosis of metastatic GIST patients treated with Imatinib (IM).¹¹ The authors also showed a relation between a NKp30 single-nucleotide polymorphism (SNP) (rs986475) and the distribution of spliced RNA isoforms.

The aim of the present study was to explore further the impact of NCR transcripts and the genetic diversity of NCR3 in metastatic melanoma patients. We studied a series of melanoma patients at different stages of the disease including a small group of 13 LS patients whose metastatic tumor had regressed and who remained tumor free for at least 24 mo without treatment.

We demonstrate that (i) the expression levels of each NKp30 (NCR3) isoforms are lower in melanoma patients at all stages of the disease compared to HV, (ii) NKp46 (NCR1) transcripts were decreased in metastatic patients and predicted survival from sampling and finally we found, (iii) a unique pattern of expression of NKp30 isoforms and frequency of genetic variant in rare melanoma patients, associated with a higher NK cell cytotoxic activity, that controlled the disease.

Materials and methods

Patients and specimens

Patient characteristics are depicted in Table 1. The study was approved by the ethic committee and informed written consent was obtained from patients. Heparinized blood was drawn from patients at the time of clinical consultation from Gustave Roussy Cancer Campus and Cochin hospital. One hundred and six HV (sex and age-matched) were used as controls (EFS, Créteil and Besançon, France). The patients enrolled consisted of stage I to stage IV melanoma patients from Cochin hospital and from Institute Gustave Roussy.

Table 1.

Characteristics of the melanoma patients included in the studies.

	STAGE 0–IIIA (n = 27)	STAGE IIIB–IIIC (n = 40)	STAGE IV (n = 107)	Long survival (LS) (n = 13)
Gander (male/female)	(16/11)	(20/20)	(58/49)	(8/5)
Age at diagnosis	59.7[25.21–88.34]	53.5[28.9–93.5]	49.1[13.22–84.44]	52.68[36.15–73.45]
Tumor thickness
<2 mm	15	19	32	9
>2 mm	7	11	39	1
Unknown	5	10	36	3
Ulceration
Present	7	9	29	2
Absent	14	12	19	3
Unknown	5	19	59	8
Conventional chemotherapy
Yes		12	56	12
No		22	29	1
Unknown		6	22

Peripheral blood mononuclear cells (PBMC) preparations

PBMC were obtained by Ficoll-Hypaque density gradient (PAA Laboratories), washed twice in Dulbecco's phosphate-buffered saline (PBS, GIBCO Invitrogen), re-suspended in RLT buffer containing β-2 mercaptoethanol (Quiagen) to preserve the RNA quality, and stored at −80°C.

Analyses of transcripts by qRT-PCR

Total cellular RNA was isolated from PBMC samples with the RNeasy kit (Qiagen) following the manufacturer's recommendations. First strand cDNA was synthesized from 1 µg of total RNA using SuperScript™ III Reverse Transcriptase (Invitrogen) and random primers (Promega) according to the manufacturer's recommendations. The NKp30 isoforms (NKp30A, B, C) and NKp46 expression was determined by qRT-PCR and normalized on β2M expression as previously described.^11,13 The expression of the gene of interest was calculated with the 2^−ΔCq method ¹⁴ normalized to the expression level of β2M, with the following formula: 2^−ΔCt = 2ˆ(Cq (gene of interest) – Cq (housekeeping gene))*10⁶. The level of expression of the distinct NKp30 isoforms compared with each other (delta, Δ) was determined using the following formula ΔNKp30_x NKp30_y = CqNKp30_y – CqNKp30_x, with Cq = Cycle threshold. The median of each NKp30 isoform pair delta (Δ) was used for the overall survival (OS) curve analysis.

Analysis of SNPs in NKp30 gene

NCR3 SNP genotyping: DNA was isolated from whole blood according to the DNeasy kit protocol (QIAGEN). After dilution, in DNase free water at a concentration of 5 ng/mL, the NCR3 SNPs (rs986475 and rs11575842) were genotyped using a custom TaqMan® SNP Genotyping Assay (pre-designed primers and Taqman genotyping Master Mix) in 96-well plates that were read and analyzed using the 7,300 Real-time PCR system (Applied Biosystems; Life technologies™).

Phenotypic and CD107a degranulation analyses

PBMC were stained with conjugated antibodies (CD3, CD56, NKp30 and NKp46) for 20 min at 4°C, washed and fixed with 1% paraformaldehyde and analyzed by a 9-color flow cytometer (CantoII, BD bioscience). Data were analyzed with DIVA software (BD bioscience). To evaluate the degranulation of NK cells, 10⁶ PBMC were stimulated with 10⁵ K562 targets (10/1effector/target ratio) in V-bottom plates for 5 h. Cells were then labeled for 20 min at 4°C with CD3, CD56 and CD107a (BD Biosciences). Cells were collected on a FACSCanto II flow cytometer. Results are expressed as the percentages of CD107a.

Ex-vivo NKp30 and NKp46 stimulation

1.10⁵ PBMC cells were cultured 6 h in 96-flat bottom plates precoated with anti-NKp30 mAb or anti-NKp46 mAbs (5 µg/mL, R&D systems), in RPMI Medium 1640–GlutaMAX™-I (Life technologies) supplemented with 10% FCS, penicillin/streptomycin (Life technologies) with or without 10ng/mL rhIL-2 (Miltenyi) + PMA/ ionomycin at 37°C and 5% CO2. The control IgG2a isotype (5 µg/mL; clone 133304; R&D systems) for the anti-NKp30 mAb was systematically included in control cultures. Then, the cells were harvested and stained with CD3, CD56 and CD107a (BD bioscience).

Statistical analyses

Data were analyzed with Prism 5 software (GraphPad San Diego, CA, USA). The unpaired t-test and the non-parametric Mann–Whitney test were used for comparison of the different groups, as indicated in the Figure Legends. Correlations were analyzed by Spearman test. Survival curves were plotted according to the Kaplan–Meier method and compared by Firth's penalized-likelihood Cox regression. A p value of < 0.05 was considered statistically significant.

Results

Definition of the cohort of patients

The present study included 187 melanoma patients enrolled from two institutions, Cochin hospital and Gustave Roussy Cancer Center. The cohort included 79 non-treated patients (from stage I to stage IV patients, not treated at the time of sampling), 80 patients that had received conventional chemotherapy or multiple therapy and 28 patients whose treatment was unknown as detailed in Table 1. Characteristics of the primary melanoma as well as disease stage are indicated at the time of sampling: the cohorts gather 27 patients with localized disease (stage 0-IIIA), 40 stage IIIB/C, 107 stage IV patients (Table 1). The OS of the patients enrolled in the studies is as expected, correlated with the Breslow index (Fig. S1). For metastatic Stage IV patients, it is of note that sampling was done at least 8 weeks after the last course of chemotherapy. In addition, 13 patients with exceptionally favorable outcome defined as LS (Table 1) were analyzed separately, due to their particular and unique characteristics.

Reduced expression levels of NKp30 isoforms in melanoma patients

The expression of each NCR3-encoded NKp30 isoforms (NKp30A, NKp30B and NKp30C) normalized to β-2 microglobulin expression was assessed by qRT-PCR in PBMC from 164 melanoma patients. NKp30 transcript analyses of patients were compared with 106 age- and sex-matched HV, a control series included in previous studies performed in GIST patients.¹² The relative expression of the three distinct isoforms (NKp30A, NKp30B and NKp30C) was significantly lower in melanoma patients than in HV (Fig. 1A). We then assessed the relative expression of NKp30 isoforms compared with each other, using the “delta” formula ΔNKp30x NKp30y = NKp30y − NKp30x for ΔAB, ΔAC, ΔBC as described elsewhere.¹³ Importantly, this NKp30 cell profiling is stable over time and was found similar in melanoma patients and in HV (Fig. 1B). There was no correlation between the transcripts levels of each NKp30 isoforms and the percentage of NKp30 expression in circulating NK cells from patients (Fig. 1C). Furthermore, using a nonparametric Spearman test, we found that the relative expression of NKp30 isoforms did not reflect the percentages of circulating NK cells (Fig. 1D). In conclusion, NKp30 isoforms transcripts are lower in NK cells from melanoma patients but the ratios between the isoforms are comparable to those found in NK cells from healthy donors.

Figure 1.

Relative expression levels of NKp30 isoforms in melanoma patients. (A) The expression of NKp30A (left panel), NKp30B (middle panel) and NKp30C (right panel), normalized to β-2microglobulin housekeeping gene, in Healthy Volunteers (HV) and in melanoma patients is shown. (B) NKp30 isoforms ratios (ΔAB, ΔAC, ΔBC) in HV and melanoma patients, non-parametric Mann–Whitney test was used compare groups, *p < or = 0.05, **p < 0.01, ***p < 0.001. ns = not significant. (C–D) Correlation between transcript levels of each NKp30 isoform with membrane expression of NKp30 by blood NK cells (C, n = 63) or with NK cells frequency (D, n = 33) from melanoma patients. Spearman test was used.

Impact of disease stage and of conventional chemotherapies on expression levels of NKp30 isoforms

We next analyzed the NKp30 isoforms expression according to their melanoma stage at sampling. We found that patients at any stage of the disease exhibited low levels of NKp30 isoforms (NKp30A, NKp30B and NKp30C) compared to NK cells from HV (Fig. 2A). Among the patients bearing local, regional or metastatic disease, there was no difference in the levels of NKp30A or NKp30B transcripts whereas NKp30C (the immunosuppressive isoform) transcript levels were higher in NK cells from stage IV compared to stage 0-IIIa melanoma patients (Fig. 2A). In addition, we observed that the ratios between isoforms were not modified when HV and melanoma patients are compared (Fig. 2B), indicating a global decrease of NKp30 transcripts with a maintained hierarchy of each isoform in patients. Interestingly, in agreements with the high NKp30C transcripts in stage IV patients, the ΔBC ratio value was slightly lower in NK cells from these patients compared to NK cells from less advanced patients (Fig. 2B).

Figure 2.

Expression of NCR transcripts and NKp30 isoforms in NK cells from patients at different stages of the disease Expression (A) and ratios (B) of NKp30 isoforms in melanoma patients at different stages of the disease are shown. Non-parametric Mann–Whitney test.

In this study, most metastatic patients analyzed had received several previous courses of conventional chemotherapy followed or not by other agents at the time of sampling (Table 1). However, we also studied 37 stage IV patients that were untreated at the time of sampling, allowing the evaluation of the impact of chemotherapy treatment on NKp30 isoform expression. Thereby, when metastatic melanoma patients were analyzed according to a previous history of chemotherapy, we found lower NKp30A expression levels (Fig. 3A) and decreased ΔAC (Fig. 3B) ratio in 60 chemotherapy treated patients (30 patients received dacarbazine, 26 patients received fotemustine or dacarbazine + fotemustine, three patients received cisplatin or dacarbazine + fotemustine + cisplatin and one patient received endoxan) compared to 37 non-treated patients (Fig. 3B), indicating that these conventional chemotherapies may modulate the relative expression levels of isoforms.

Figure 3.

Impact of Chemotherapy treatment on NKp30 transcripts. (A–B) Expression of NKp30A, B, and C isoforms (A) and NKp30 isoform ratios (B) in non-treated (n = 37) and chemotherapy treated patients (n = 60): non-parametric Mann–Whitney test was used.

Impact of expression of NKp30 isoforms and NKp46 transcripts in melanoma patients with disease evolution

We analyzed the prognostic value of the relative expression of NKp30 isoforms by stratifying patients according to the median of the transcript level considered as a cut-off value. It is of note that expression of transcripts may change over time whereas ratio between isoforms remains constant,¹³ thus correlation with survival was assessed from day of sampling for transcripts levels and isoform ratios. Metastatic patients were included in the analyses.

Kaplan–Meier curves showed no impact of the relative expression of NKp30A, NKp30B and NKp30C on the survival from the day of sampling (Fig. 4A). We then analyzed the prognostic value of the ΔNKp30 ratios according to the median level of the cohort. There was no correlation of the three different ΔNKp30 ratios with better survival from day of sampling (Fig. 4B).

Figure 4.

Prognostic values of NCR transcripts in melanoma patients. Survival from sampling day (OSS) of melanoma patients according to the median value of (A) the levels of NKp30A (left), NKp30B (middle) and NKp30C (right) and of (B) NKp30 ratio ΔAB (left), ΔAC (middle) and ΔBC ratios (right) were assessed by univariate analysis using the Kaplan–Meier method (n = 117). (C) Expression of NKp46/NCR1 transcripts in 47 melanoma patients according to their disease stage at sampling. (D) Kaplan–Meier curves of melanoma patient OSS based on the NCR1/NKp46 transcript levels from the day of sampling (n = 31), Log-rank (Mantel-Cox). *p < ou= 0.05, **p < 0.01, ***p < 0.001. ns = not significant.

In addition, we analyzed the levels of another NK cell activating receptor, NKp46, on a subgroup of 47 melanoma patients including 36 inoperable stage III or stage IV metastatic melanoma patients. When patients were stratified according to the stage at sampling, NKp46 expression was lower in stage IV and certain stage III (three out of seven) metastatic patients (Fig. 4C, left panel). In contrast to NKp30, in 31 stage IV metastatic patients, the expression levels tend to predict the OS from the day of sampling (Fig. 4C right panel). These data are in accordance with our previous work describing a correlation between NKp46 expression and duration of stage IV in metastatic patients ¹⁵ indicating the importance of NKp46 in melanoma evolution.

Long survivor (LS) patients display unique NKp30 transcripts and particular NKp30 polymorphisms

Interestingly, LS patients display a distinct pattern of NKp30 isoforms. We found significantly higher expression of NKp30A transcript in LS patients compared to progressor melanoma patients (metastatic stage IV patients with a progressive disease at the time of sampling) (Fig. 5A). Expression levels of NKp30B and NKp30C were close to that of progressor patients. As a consequence of the high NKp30A transcripts, LS were characterized by high ΔAB and ΔAC ratios compared to HV (Fig. 5B), indicating a unique NKp30 regulation in these patients.

Figure 5.

Expression levels of NKp30 isoforms and NKp30 ratios in Long Survivors melanoma patients. (A–B) NKp30 isoforms transcripts (A) and ΔNKp30 ratios (B) in 106 HV, 12 LS and 91 stage IV melanoma patients. Non-parametric Mann–Whitney test was used to compare groups.

According to the fact that NCR3 genetic variants may inter-individually control (at least partly) NKp30 expression, we analyzed the frequencies of two NCR3 SNPs previously described to be associated with NKp30 function¹¹ (and HJ Garchon unpublished data). Genotyping analyses revealed that the NKp30 SNP rs986475 (NKp30/NCR3*3790: T/C), associated with increased NKp30C transcript levels, was present in 32 of 108 (29.5%) progressor patients and in 25 of 109 HV (23%). In contrast, in the LS patients, the heterozygous CT genotype (associated with high NKp30C) was present in only 1 out of 13 patients and we detected the rare homozygous CC genotype in two patients (Fig. 6A). The SNP rs11575842 T minor allele associated with higher NKp30A isoform in HV (HJ Garchon unpublished data) was present in 7/25 (28%) of HV, 34 of 122 (27.9%) of progressors and 5/13 (46%) of LS melanoma patients (Fig. 6B). These results indicated particular frequencies of the two SNPs related to NKp30 isoforms in LS patients. There was no relation between presence of alleles and the percentages of membrane NKp30 on circulating NK cells (Fig. 6C–E). Of note, as expected from previous studies, the relative expression of NKp30C was significantly higher in patients bearing the rs986475 C/T allele and consequently slightly decreased ΔAC and ΔBC ratios (Fig. 6D). The rs11575842 C allele was not associated with higher NKp30A expression in progressor melanoma patients (Fig. 6F). In LS, the high levels of NKp30A was independent of the rs11575842 C/T allele (Fig. 6F). These data further indicate that higher NKp30A and the particular NKp30 isoform ratio values in LS are independent of the SNP alleles present in the patients (Fig. 6–F).

Figure 6.

NCR3 genetic variants in Long Survivors and Progressors Melanoma patients. (A–B) Frequencies of NCR3 genetic variants rs986475 (A) and rs15575842 (B) in HV and progressor or LS melanoma patients. (C–F) Analysis of membrane expression of NKp30 according to the presence of the two NCR3 SNP, rs986475 (C) and rs11575842 (E) in progressor (n = 37, black dots) and 13 LS (n = 13, white dots) melanoma patients. Transcript levels (left panels) or ratios (right panels) of NKp30 isoforms according to rs986475 (D) and rs11575842 (F) SNPs in progressor (n = 104) or LS (n = 13) melanoma patients. Non-parametric Mann–Whitney test was used to compare groups.

Long survivor (LS) patients with unique NKp30 profile display activated NK cells

Phenotypic analyses of blood NK cells in melanoma patients indicated that percentages of NKp30⁺ NK cells from metastatic progressors and LS patients were comparable to that of NK cells from HV (Fig. 7A, left panel). Percentages of NKp46⁺ NK cells were decreased in melanoma patients (which is significant in LS) compared to blood NK cells from HV (Fig. 7A, right panel). However, we found that in HV and LS patients, there was a correlation between NKp46 and NKp30 expression (Fig. 7B, left and right panels, respectively) whereas the two receptors were not correlated in progressor patients (Fig. 7B, middle panel). In addition, we found that NK cells from LS patients exhibited significantly higher degranulation potential than NK cells from HV or progressors (Fig. 7C).

Figure 7.

Unique NCR expression in LS patients. (A) Analysis of NKp30 (left panel) and NKp46 (right panel) in 10 HV, 17 progressor and 13 LS melanoma patients. (B) Correlation between NKp30 and NK46 percentages of expression in HV (left panel), progressors (middle panel) and LS (right panel) melanoma patients. (C) K562-induced CD107a degranulation levels of NK cells from 15 HV, 21 progressor and 11LS melanoma patients. Non-parametric Mann–Whitney test was used to compare groups.

We assessed the function of NK cells from LS and progressor patients by triggering NKp30 or NKp46 with or without IL-2 in a 5-h degranulation assay. NK cells derived from LS patients had a significantly higher capacity to degranulate upon NKp30 and NKp46 cross-linking than progressor patients or HV (Fig. 8). Thus, the particular NKp30 profile, namely high NKp30A transcripts in LS, conferred a higher capacity to respond to NKp30 stimulation.

Figure 8.

Superior NK cell activity in LS melanoma patients. Analysis of degranulation following mAb mediated triggering of NCR1/NKp46 or NCR3/NKp30 in one representative HV (A), one representative metastatic patients (B), and two LS melanoma patients (C). Resting NK cells (Med) or NK cells activated with (IL−2 + PMA/Iomomycine) were triggered with isotopic control (ISO) anti-NKp30 or anti-NKp46 mAbs and percentages of degranulation (CD107a) were assessed by flow cytometry.

Discussion

There is growing number of reports showing that the immunosurveillance of metastatic tumors involves NK cells.^16,17 Indeed, immunophenotyping studies showed that NK infiltrating tumors cells, exhibited altered effector functions, as shown in breast and lung malignancies.^18,19 NK cell relative anergy correlated with downregulation of their natural cytotoxic receptors.^18-20 In addition to the numbers of tumor infiltrating NK cells, NK cell receptor expressions (NKp30 or KIRs) were found to strongly influence the prognosis in several malignancies such as in renal cell carcinoma,²¹ GIST,¹² neuroblastoma¹³ and acute myeloid leukemia.²²

Our recent studies on NK cells in melanoma patients also indicate the involvement of NK cells in the natural history of the disease. We showed that NK cells infiltrated primary thick melanoma (more numerous in melanomas from aged patients) but their numbers did not impact survival.²³ In blood NK cells, we observed a progressive decrease of NKp46 expression level in patients with advanced melanoma and this alteration correlated with shorter duration of stage IV.¹⁵ In addition, chemotherapy treated patients displayed distinct phenotypic modulations,²⁴ indicating that tumor progression and conventional chemotherapy may interfere with NK cell status in melanoma patients.²⁵

In a recent study, comparing tumor positive and tumor-free Sentinel lymph nodes (SLN), we found that more numerous cytotoxic cells (CD8⁺ cells, Granzyme⁺ cells, NK cells) infiltrated tumor-free SLN. We also found that NK cells in SLN correlated with higher rate of relapses, supporting a tumor-promoting role of NK cells in SLN.²³ However, we also showed that metastatic lymph nodes contained NK subset CD56^brightCD16⁺ endowed with high lytic potential following activation by cytokines.²⁶ Depending of the disease stage localized or metastatic, lymph node NK cells may display different functions likely reflecting their recruitment from the tumor or their local activation.

In the present report, we studied the mRNA expression levels of NKp46 and NKp30 isoforms (NKp30A, NKp30B and NKp30C) receptors and their ratio in PBMC of melanoma patients sampled at the different stages of the disease. Compared to blood NK cells from healthy donors, NK cells from melanoma patients with a cutaneous or metastatic melanoma displayed lower transcripts levels of each isoform of NKp30. Correlation with patient evolution showed that neither levels of NKp30A, NKp30B or NKp30C had a prognostic impact in melanoma patients. Among metastatic patients, the NKp30A transcripts were lower in patients with previous history of treatment with conventional chemotherapy compared to untreated patients, indicating that conventional chemotherapies may modify NK cell activation. It is of note that the splice variants of NCR (NKp30 and NKp44) were controlling the distinct functions of NK cells in response to cytokine environment was recently evidence in NK cells from decidua.²⁷

These results echo to recent report investigating the NKp30/B7H6 axis in other tumor types, gastrointestinal sarcoma (GIST), neuroblastoma and non-small cell lung carcinoma (NSCLC). The presence of high levels of NKp30C isoform in blood NK cells influences GIST patient's prognosis¹¹ and markedly impacted both event-free and OS, in two independent cohorts of metastatic GIST.²⁸ Low ΔBC values and low expression levels of NKp30A were identified in one-third of patients with dismal prognosis across molecular subtypes. This low ΔBC value in PBMC was associated with a pro-inflammatory and immunosuppressive tumor microenvironment. In neuroblastoma patients, the expression levels of the distinct NKp30 isoforms correlated with 10-y event-free survival in three independent cohorts of patients in remission from metastases after induction chemotherapy.¹³

In NSCLC patients, the relative expression of NKp30 was lower than in HV as well as in all other analyzed malignancies and the low relative expression of NKp30 correlated with poor overall and progression-free survival (Fend L and Rusakiewicz S, personal communications). Moreover, a discrete proportion of NSCLC patients exhibited a particular low ratio between NKp30B and NKp30C (ΔBC).

Previous and accompanying reports indicate that ΔBC ratios are important regulator of NCR3 gene in patients bearing GIST, Neuroblastoma and NSCLC. In Melanoma, decreased ΔBC ratios occur with disease progression. In addition, NKp30A expression levels are modulated in chemotherapy treated patients and high levels of NKp30A are present in LS patients. The data altogether outline that the mechanisms controlling total and isoforms of NCR3 mRNA expression are complex, yet not elucidated, and contribute to specific and distinct patterns of NKp30 isoform expression in blood NK cells from different tumor patients. There is no correlation between transcripts levels of isoforms and the membrane expression of NKp30 by blood NK cells, but isoform ratio interferes with functionality of NKp30 and subsequent NK cell activation. In melanoma and NSCLC patients, in contrast to in leukemic (ALM) and breast cancer patients, membrane expression of NKp30 by blood NK cells is close to that of donor NK cells. In NSCLC, NKp30 expression is altered in tumor NK cells, while maintained in NK cells from metastatic lymph nodes of melanoma patients.

Our data also indicated the diminution of NKp46 transcripts in metastatic melanoma and its prognostic value on OS. These results confirmed and further strengthened our previous data showing an impact of NKp46 expression at the surface of blood NK cells on the duration of stage IV in melanoma patients.¹⁵

Interestingly, our data show that NKp30A expression level isoform (NK cell fitness) was elevated in rare LS patients that have controlled a metastatic disease. These LS are characterized by unique polymorphisms and isoforms proportions of NKp30. In addition, strong correlations between NKp46 and NKp30 isoforms expression were found in LS patients (and donor NK cells) but not in progressor melanoma patients.

Despite history of chemotherapy, they express high levels of NKp30A. Blood NK cells from these patients display increased lytic potential in response to NKp30 triggering, favoring the involvement of NK cells in the disease control. In addition, NK cells of LS patients bear the genetic variant rs986475 associated with low levels of NKp30C isoform and are endowed with high lytic potential, suggesting that this rare genetic variant of NCR3 gene may confer a unique immune status in patients developing sporadic melanoma. It is of note that LS patients developed melanomas that were not distinct in terms of oncogenic mutations from those developed by progressor patients, indicating that NKp30 isoforms and SNP are not related to specific oncogenic anomalies in melanoma patients. These data suggest the importance of immunogenetic tests (NKp30/NCR3 transcript and isoforms levels, SNPs) in melanoma patients to distinguish rare patients with particular evolution and to identify relevant immune pathway required for tumor control.

Disclosure of potential conflicts of interest

No potential conflicts of interest were disclosed.

Acknowledgments

We thank Deborah Rechard, Christelle Auger (URC Paris center) and Isabelle Scheer-Senyarich for their contribution to the data management. Drs Nathalie Franck, Nora Kramkimel, Elodie Regnier-Rosencher and Françoise Boitier for their contributions in patients care and sample collections. *co senior authors.

Funding

This work was supported by grants from l'Institut National du Cancer (PLBIO-2011-6) and PAIR Melanome (2013-0662013), l'Association pour la Recherche contre le Cancer (ARC, 3964 to AC), la Ligue Nationale contre le Cancer (Comité Ile de France to AC, PhD grant for MM), l'Assistance Publique des Hôpitaux de Paris (APHP) (DRCD, Immumela program).