CD5 and CD6 as immunoregulatory biomarkers in non-small cell lung cancer.

BACKGROUND: The study of immune surveillance in the tumour microenvironment is leading to the development of new biomarkers and therapies. The present research focuses on the expression of CD5 and CD6-two lymphocyte surface markers involved in the fine tuning of TCR signaling-as potential prognostic biomarkers in resectable stages of non-small cell lung cancer (NSCLC).
METHODS: CD5 and CD6 gene expression was analysed by reverse transcription quantitative polymerase chain reaction (RTqPCR) in 186 paired fresh frozen tumour and normal tissue samples of resected NSCLC.
RESULTS: Patients with higher CD5 expression had significantly increased overall survival (OS, 49.63 vs. 99.90 months, P=0.013). CD5 expression levels were correlated to CD4 infiltration and expression levels, and survival analysis showed that patients with a higher CD5/CD4 + ratio had significantly improved prognosis. Multivariate analysis established CD5 expression as an independent prognostic biomarker for OS in early stages of NSCLC (HR=0.554; 95% CI, 0.360-0.853; P=0.007). Further survival analysis of NSCLC cases (n=97) from The Cancer Genome Atlas (TCGA) database, confirmed the prognostic value of both CD5 and CD6 expression¸ although CD6 expression alone did not reach significant prognostic value in our NSCLC training cohort.
CONCLUSIONS: Our data support further studies on CD5 and CD6 as novel prognostic markers in resectable NSCLC and other cancer types (i.e., melanoma), as well as a role for these receptors in immune surveillance. 2020 Translational Lung Cancer Research. All rights reserved.

Introduction

The immune system’s natural ability to detect and eliminate malignant cells is currently considered the best weapon in the battle against cancer. The study of immune biomarkers is critical to diagnose, prevent and choose the appropriate immunotherapy strategy in cancer treatment. The presence, localization, and proportion of helper and especially cytotoxic T cells (CTLs) in tumour-infiltrating lymphocytes (TILs) from aggressive cancers such as non-small cell lung cancer (NSCLC) and melanoma has been associated with a favorable prognosis (1-3). In recent years, gene expression patterns and single nucleotide polymorphism studies have contributed to cancer prognosis (4). Particularly, inducible immunoregulatory genes like immune checkpoints blockers (CTLA-4 or PD-1) expressed in TILs, are among the most predictive biomarkers for cancer clinical outcome and targets for immunotherapy (5,6). Identification of new immune biomarkers is still needed to fully understand the mechanisms of immune evasion and facilitate subsequent development of novel immunotherapies.

The lymphocyte surface co-receptors CD5 and CD6 are immunomodulators involved in the development, activation, differentiation and survival of lymphocytes (7,8). They are encoded by homologous genes derived from duplication of a common ancestral gene (9), and both are constitutively expressed by all T cells and the small B1a cell subset. CD5 and CD6 are signal-transducing receptors that physically associate with the T and B cell antigen-specific clonotypic receptor (TCR and BCR, respectively) at the centre of the immune synapse (10,11). The endogenous CD6 ligands involve CD166/ALCAM (12), Galectins 1 and 3 (13) and CD318 (14), all broadly distributed on immune, epithelial, mesenchymal and/or cancer cells. In contrast, the nature of the CD5 ligand/s is ill-defined because no reported candidate (CD72, IgVH framework region, gp200, gp150, gp40-80, CD5 itself and IL-6) has been firmly validated by independent groups (7,15). Based on monoclonal antibody in vitro data, CD5 and CD6 were initially considered co-stimulatory molecules (7,8). However, analysis of CD5 and CD6 knockout mice unveiled their negative modulatory role for thymocyte (and B1a) activation and differentiation signals upon clonotypic receptor cross-linking (16-19). Interestingly, this immunomodulatory role occurs even in the absence of ligand interaction (20,21). This implies that lymphocyte function can be up- or down-regulated by CD5 and CD6 expression. Accordingly, anergic T and B cells show upregulated surface CD5 expression (22,23). CD5 and CD6 expression also parallels TCR/CD3 levels and is predictive of TCR avidity and survival of T cells (24-27). Moreover, regulation of CD5 expression by TCR engagement has been reported in peripheral T cells (28). Regarding antitumour responses, in situ regulation of CD5 expression by CTLs is thought to adapt their sensitivity to intra-tumour peptide-major histocompatibility complex (p-MHC) levels (29). Indeed, CTL clones from lung cancer patients show that CD5 expression is inversely proportional to their anti-tumour cytolytic activity, preventing activation-induced cell death (AICD) during T cell overactivation (29).

CD5 and CD6’s modulation of lymphocyte activation and survival supports the hypothesis that their overall intra-tumour expression levels alter the anti-tumour immune response and can be therefore used as prognosis biomarkers in cancer. Accordingly, this study investigated the prognostic value of CD5 and CD6 gene expression in a cohort of 186 patients with resectable NSCLC. The results show that high intra-tumour levels of both CD5 and CD6 associate to better prognosis as measured by overall survival and relapse-free survival. This was validated in silico using NSCLC biopsy information from The Cancer Genome Atlas (TCGA) database.

Methods

NSCLC patient cohort and sample collection

The training cohort included 186 patients with resected and non-pretreated stage I to IIIA NSCLC from Consorcio Hospital General Universitario de Valencia. Between 2004 and 2017, 186 fresh-frozen tumour and normal tissue samples were obtained from surgical resection and preserved in RNAlater® (Applied Biosystems, USA). Patients who had received neoadjuvant treatment and those with a follow-up shorter than 1 month were excluded. This study abides by the Declaration of Helsinki (as revised in 2013) and was approved by the institutional review board. All patients had signed the informed consent prior to sample collection.

Real-time PCR analysis of NSCLC patients

RNA was isolated using TRI Reagent® (Sigma, USA) and retrotranscribed using High Capacity cDNA Reverse Transcription Kit (Applied Biosystems, USA) in a MasterCycler® thermocycler (Eppendorf, Germany) following the manufacturer’s instructions. CD3D, CD3E, CD4, CD5, CD6 and CD8 gene expression was analysed by RTqPCR using Taqman® hydrolysis probes and Taqman® Gene Expression Master Mix (Applied Biosystems, USA) in a LightCycler 480 thermocycler (Roche, Switzerland).

Relative gene expression was calculated by Pfaffl formula, taking into consideration expression differences between normal and tumour tissue, as well as RTqPCR efficiencies of each TaqMan® assay. Reference gene expression corresponds to the geometric mean of ACTB, CDKN1B and GUSB, endogenous controls used after evaluation with GeNorm software (30,31).

Immunohistochemistry (IHC)

CD5 and CD6 relative gene expression data were normalized against CD4+ and CD8+ lymphocyte infiltration in NSCLC samples (n=60). To do so, CD4 and CD8 expression was evaluated in 60 Formalin-Fixed Paraffin Embedded (FFPE) samples using a Dako Autostainer Link 48 and the Dako EnVisionTM FLEX detection system (Dako, Canada). After section drying and antigen dewaxing in a PT Link instrument, the endogenous peroxidase activity was quenched with peroxidase blocking reagent in the Autostainer Link 48 instrument. Immunostaining was carried out with Dako FLEX Ready to-Use format for CD4 (Clone 4B12, Dako) and CD8 (Clone C8/144B, Dako). Briefly, a detection system chromogen (3,3'-diaminobenzidine, DAB) was used after primary antibody incubation, followed by washing and counterstaining of sections with hematoxylin, dehydration and mounting. CD4+ and CD8+ lymphocytes were counted in 10 high power fields (HPF) (magnification ×400) for tumour areas. Negative controls and normal human tonsil positive control tissue were included. IHC staining quantification was performed by two independent evaluators.

TCGA database search for CD5 and CD6 mRNA expression levels in NSCLC biopsies

Online information available at TCGA database (https://cancergenome.nih.gov/) for NSCLC patients was downloaded and used as independent validation cohorts. Patients with resected NSCLC and available gene expression data for CD5 and CD6 in normal and tumour tissue samples were selected (n=97). Patients who had received neoadjuvant treatment or with follow-up shorter than 1 month were excluded.

Statistical analyses

Relative gene expression was dichotomized using median as a cut-off value. Non-parametric tests were used for correlations between clinico-pathological and analytical variables. Survival analyses were performed considering relapse-free survival (RFS) and overall survival (OS). RFS spans from surgery to relapse or exitus dates, and OS from surgery to exitus dates, following the Response Evaluation Criteria in Solid Tumours (RECIST). For patients who neither relapsed nor died, the last recorded follow-up was considered. Gene prognostic value was assessed using Kaplan-Meier curves (Log-rank test) and univariate Cox regression analyses, followed by a multivariate Cox regression analysis, using all significant variables to establish independent prognostic biomarkers. Statistical analyses were performed using the Statistical Package for Social Sciences (SPSS) 15.0 software (Chicago, USA), considering significant P<0.05.

Results

Analysis of CD5 and CD6 expression in NSCLC samples from training cohort

The NSCLC training cohort was composed of 186 patients (see for demographic and clinicopathological data), mainly men (85.0%) and current or former smokers (88.7%). Histology was squamous cell carcinoma in 47.9% and adenocarcinoma in 41.4% of all cases. During the follow-up (median 34.2 months), 85 patients relapsed (45.7%) and 91 died (48.9%). Non-parametric tests were conducted to determine an association of relative gene expression with clinicopathological variables.

Table S1

Clinicopathological characteristics of patients included in NSCLC training Cohort

Characteristics	N, n=186	%
Age at surgery (median, range)	65 [26–85]
Gender
Male	158	84.95
Female	28	15.05
Smoking Status
Current	91	48.92
Former	74	39.79
Never	21	11.29
Performance Status
0	125	67.20
1–2	61	32.80
Stage
I	96	51.61
II	55	29.57
IIIA	35	18.82
Lymph node involvement
Yes	52	27.96
No	134	72.04
Histology
SCC	89	47.85
ADC	77	41.40
Others	20	10.75
Differentiation grade
Poor	47	30.52
Moderate	73	47.40
Well	34	22.08
NS	32
EGFR
Wild type	89	89.00
Mutated	11	11.0
NS	86
KRAS
Wild type	152	83.98
Mutated	29	16.02
NS	5
Relapse
Yes	85	45.70
No	101	54.30
Exitus
Yes	91	48.92
No	95	51.08

NSCLC, non-small cell lung cancer; SCC, squamous cell carcinoma; ADC, adenocarcinoma; NS, not specified.

The survival analyses revealed that high CD5 expression had significant biomarker prognostic value for OS (OS, 49.63 vs. 99.90 months, P=0.013). Furthermore, a statistical trend toward significant RFS was detected (RFS, 29.20 vs. 44.30 months, P=0.076) (; ). In contrast, similar analyses did not provide significant results for CD6 expression (). Multivariate Cox regression analysis revealed high CD5 expression as a potential prognostic biomarker for OS in resected NSCLC patients (HR=0.554; 95% CI, 0.360–0.853; P=0.007) ().

Table 1

Survival analysis of CD5 and CD6 in NSCLC patients of training cohort. Cox univariate analyses were conducted with dichotomized relative expression of CD5 and CD6

Univariate analysis	RFS				OS
	HR	95% CI	P value		HR	95% CI	P value
CD5 (high vs. low)	0.708	0.483–1.038	0.077		0.592	0.390–0.898	0.014*
CD6 (high vs. low)	0.997	0.680–1.460	0.986		0.929	0.614–1.404	0.725

*, P value <0.05. NSCLC, non-small cell lung cancer; RFS, relapse-free survival; OS, overall survival; HR, hazard ratio; CI, confidence interval.

Figure 1

NSCLC survival curves according to CD5 relative expression. Kaplan-Meier survival curves for (A) RFS and (B) OS of CD5 expression levels in NSCLC patients of training cohort. NSCLC, non-small cell lung cancer; RFS, relapse-free survival; OS, overall survival.

Table 2

Multivariate cox regression analysis in NSCLC patients of training cohort. Cox multivariate analysis was conducted with all significant variables resulted from the univariate analyses, that were performance status, stage, tumour size, lymph node involvement, KRAS mutational status and CD5

Multivariate analysis	RFS				OS
	HR	95% CI	P value		HR	95% CI	P value
Stage (IIIA vs. IIA/IIB vs. IA/IB)	1.565	1.219–2.010	0.0004*		1.625	1.234–2.139	0.001*
KRAS mutational status (mutated vs. WT)	2.126	1.280–3.531	0.004*		2.054	1.134–3.723	0.018*
PS (1/2 vs. 0)	1.661	1.102–2.504	0.015*		1.800	1.148–2.823	0.010*
CD5 (high vs. low)					0.554	0.360–0.853	0.007*

*, P value <0.05. NSCLC, non-small cell lung cancer; RFS, relapse-free survival; OS, overall survival; PS, performance status; WT, wild-type; HR, hazard ratio; CI, confidence interval.

CD5 and CD6 are constitutively expressed lymphocytic receptors, whose expression can be regulated during lymphocyte development and activation events (32,33). Thus, mRNA level correlation to lymphocyte infiltration or to up-regulation in infiltrating lymphocytes was assessed. To this end, CD5 and CD6 gene expression was normalized to CD4+ and CD8+ lymphocyte infiltration by IHC, and the CD5/CD4+, CD5/CD8+, CD6/CD4+ and CD6/CD8+ ratios were calculated in 60 samples. Survival analysis indicated that patients with higher CD5/CD4+ ratios had significantly improved prognosis (RFS, 13.33 vs. 66.97 months, P=0.023; OS, 25.73 vs. 73.93 months, P=0.019) (). Furthermore, a tendency for improved RFS and OS was observed for CD5/CD8+, CD6/CD4+ and CD6/CD8+ ratios ().

Figure S1

NSCLC survival curves according to CD5 ratios. (A) Kaplan-Meier survival curve for RFS of CD5 relative expression normalized to CD4 infiltration (CD5/CD4+). (B) Kaplan-Meier survival curve for OS of CD5 relative expression normalized to CD4 infiltration (CD5/CD4+). (C) Kaplan-Meier survival curve for OS of CD5/CD4 expression levels in NSCLC patients of the training cohort. NSCLC, non-small cell lung cancer; RFS, relapse-free survival; OS, overall survival.

Table S2

Survival analysis of CD5 and CD6 ratios in NSCLC patients of training cohort. Cox univariate analyses were conducted with dichotomized relative expression of CD5 and CD6, normalized against CD4+ and CD8+ infiltration measured by IHC in a subset of 60 patients

Univariate analysis	RFS				OS
	HR	95% CI	P value		HR	95% CI	P value
CD5/CD4+ (high vs. low)	0.428	0.202–0.907	0.027*		0.392	0.175–0.879	0.023*
CD5/CD8+ (high vs. low)	0.675	0.360–1.266	0.221		0.584	0.304–1.122	0.107
CD6/CD4+ (high vs. low)	0.644	0.309–1.342	0.240		0.580	0.265–1.268	0.172
CD6/CD8+ (high vs. low)	0.664	0.353–1.247	0.203		0.574	0.299–1.105	0.097

*, P value <0.05. NSCLC, non-small cell lung cancer; RFS, relapse-free survival; OS, overall survival; HR, hazard ratio; CI, confidence interval.

In order to confirm these results, relative CD5 and CD6 expression using RTqPCR was normalized against CD4 and CD8 expression in 186 NSCLC patients from our training cohort. Survival analysis confirmed that higher CD5/CD4 expression was associated to improved OS (OS, 49.80 vs. 99.90 months, P=0.042) (), but no significant results were obtained neither for the rest of the ratios nor for RFS (). Additionally, and taking into account that the Spearman’s test correlated CD5 and CD6 to CD3D and CD3E, the CD5/CD3D, CD5/CD3E, CD6/CD3D and CD6/CD3E ratios were evaluated but showed no significant association to prognosis ().

Table S3

Survival analysis of CD5 and CD6 ratios in NSCLC patients of training cohort. Cox univariate analyses were conducted with dichotomized relative expression of CD5/CD3D, CD5/CD3E, CD5/CD4, CD5/CD8, CD6/CD3D, CD6/CD3E, CD6/CD4 and CD6/CD8 ratios in 186 patients

Univariate analysis	RFS				OS
	HR	95% CI	P value		HR	95% CI	P value
CD5/CD3D (high vs. low)	1.010	0.687–1.485	0.958		0.859	0.567–1.301	0.473
CD5/CD3E (high vs. low)	0.925	0.631–1.357	0.691		0.846	0.559–1.281	0.430
CD5/CD4 (high vs. low)	0.906	0.617–1.330	0.613		0.633	0.406–0.988	0.044*
CD5/CD8 (high vs. low)	0.941	0.641–1.380	0.754		0.833	0.550–1.261	0.387
CD6/CD3D (high vs. low)	1.093	0.744–1.605	0.650		1.081	0.714–1.637	0.712
CD6/CD3E (high vs. low)	0.977	0.664–1.436	0.905		0.908	0.598–1.379	0.650
CD6/CD4 (high vs. low)	0.858	0.585–1.259	0.435		0.798	0.527–1.209	0.287
CD6/CD8 (high vs. low)	0.970	0.661–1.424	0.878		0.835	0.551–1.266	0.396

*, P value <0.05. NSCLC, non-small cell lung cancer; RFS, relapse-free survival; OS, overall survival; HR, hazard ratio; CI, confidence interval.

In silico analysis of CD5 and CD6 expression from NSCLC TCGA database

Online information of 97 NSCLC patients from the TCGA database was used as a validation cohort. Their median follow-up was 27.61 months, time in which 27 (32.1%) patients had relapsed and 46 (47.4%) were exitus. The statistical analysis confirmed that higher expression of both CD5 and CD6 improved prognosis in NSCLC patients, as it was associated with increased RFS (34.98 vs. 75.57 months, P=0.033; 25.31 vs. 75.57 months, P=0.020, respectively) and OS (40.49 vs. 77.97 months, P=0.038; 39.02 vs. 77.97 months, P=0.034, respectively) (; ). Multivariate analysis did not provide significant results on account of the small sample size (n=97). In all, our data supports that higher CD5 expression in early-stage NSCLC patients associates to increased OS. Regarding CD6, the validation cohort indicates a NSCLC biomarker potential that should be explored further.

Table 3

Survival analysis of CD5 and CD6 in NSCLC patients of TCGA database. Cox univariate analyses were conducted with dichotomized relative expression of CD5 and CD6

Variables	RFS				OS
	HR	95% CI	P value		HR	95% CI	P value
CD5 (high vs. low)	0.542	0.306–0.960	0.036*		0.537	0.296–0.975	0.041*
CD6 (high vs. low)	0.513	0.289–0.911	0.023*		0.530	0.292–0.961	0.037*

*, P value <0.05. RFS, relapse-free survival; NSCLC, non-small cell lung cancer; OS, overall survival; HR, hazard ratio; CI, confidence interval.

Figure 2

NSCLC survival curves according to the expression status of CD5 and CD6 of TCGA database patients. Kaplan-Meier survival curves for RFS and OS of CD5 (A and B) CD6 (C and D) expression levels in NSCLC patients of TCGA database. NSCLC, non-small cell lung cancer; RFS, relapse-free survival; OS, overall survival.

Discussion

The presence of immune cells in the tumour microenvironment plays a key role in NSCLC and melanoma prognosis (34,35). Our work provides a prognostic value for intratumour CD5 and CD6 mRNA expression, two lymphocyte co-receptors involved in modulation of T (and B1a) cell development, activation, differentiation and survival (7,8). High CD5 expression associates to favourable prognosis in 186 fresh-frozen samples of a NSCLC training cohort, validated further with online data from the NSCLC TCGA database. The analysis of CD6 expression in NSCLC tumour samples reveals similar associations. Furthermore, CD5 and CD6 expression correlate to the number of CD4+ and CD8+ lymphocytes and these ratios indicate better prognosis. Importantly, better prognosis also correlates with CD5 overexpression by TILs.

In order to explore the prognostic potential of CD5 and CD6 expression in other cancer types, melanoma information from the TCGA database was analyzed. As illustrated in , better survival rate (“alive” group) presented significantly higher CD5 and CD6 expression, together with other T cell (CD3, CD4, CD8) and B cell (CD19) markers. There was also a significant overall difference rate between the “alive” and “dead” groups regarding time to progression (TTP) and OS considering high vs. low CD5 (Figure S3A,B) or CD6 expression (Figure S3C,D).

Figure S2

Retrospective analysis of gene expression in human melanoma samples from TCGA database. (A) Melanoma tumor biopsies (n=409) were divided into “alive” [being alive for more than 4 years (1,460 days) from day of diagnosis; n=74, blue plot] or “dead” [being dead in less than 2 years (730 days) from diagnosis; n=49], black plot] groups. (B) The alive group was subdivided into two other groups; with tumor (information of last follow-up; n=16, black plot) and without tumor (information of last follow-up; n=58, blue plot). Gene expression values of CD3δ, CD3ε, CD4, CD8α, CD8β, CD19, CD6, CD5 and ALCAM are presented in both graphs. Mann-Whitney test was used for statistical comparisons between groups.

In addition, patients with better survival and tumour regression (no tumour in last follow-up) showed higher CD5 and CD6 expression, but lower CD3, CD4, CD8, CD19 levels (Figure S2B), in support of infiltrating lymphocytes over-expressing CD5 and CD6 with better prognosis in melanoma.

Tumour microenvironment genetics has identified inhibitory or stimulatory lymphocyte accessory molecules (e.g., PD-1/CD279, CTLA-4, LAG-3, TIM-3 and 4-1BB/CD137) that can be regulated to modulate T-cell activation and survival. Interestingly, these biomarkers distinguish tumour-specific T cells from unrelated T cells in the tumour infiltrate (36,37). This is illustrated by PD-1—an inhibitory receptor inducibly expressed on activated T cells—as a marker of the tumour-reactive CD8+ T cell fraction in melanoma tumours (37) and of high avidity CD8+ T cells specific for Melan-A (38) or neoantigens (39). PD-1 expression is related to TCR signal strength, and thus to the functional avidity of specific T cells, underlining the complex significance of PD-1 expression on tumour-specific T cells (38).

Similar to PD-1 and CTLA-4, the CD5 and CD6 co-receptors act as negative regulators of T cell activation (8,40). CD5- and CD6-deficient thymocytes are hyper-reactive to TCR/CD3 cross-linking (16,18). Surface CD5 and CD6 expression is set in the thymus, and parallels TCR/CD3, becoming predictive of TCR avidity (16,18,19,24,25). Thus, thymocytes binding self-peptide-MHC (self-pMHC) with high affinity consequently deliver strong TCR-mediated activation signals, and express higher CD5 and CD6 surface levels to overcome negative selection. Surface CD5 and CD6 promote thymocyte survival by different mechanisms (25,41). Post-positively selected (peripheral) CD4+ and CD8+ T cells with high CD5/CD6 expression (CD5/6hi) respond to foreign peptides with increased activation and survival (42-44). In other words, T cells with TCRs of stronger avidity for self-pMHC dominate in responses to foreign antigens and accumulate in aging individuals, revealing that positive selection contributes to effective immunity (42). This would be in line with our finding that high intratumour CD5 (and CD6) expression correlates with better NSCLC and Melanoma prognosis. Accordingly, higher intratumour CD5 (and CD6) levels reflect infiltrating T cells with higher avidity for tumour antigens and more resistant to activation-induced cell death (29).

Self-pMHC contact modulates CD5 expression, survival and homeostatic proliferation of naïve T cells in the periphery (28,45) suggesting modulation of CD5 on TILs at the tumour site. Indeed, in situ CD5 expression adaptation to pMHC levels of TILs from a lung carcinoma patient has been reported (46). The decreased MHC class I expression, observed in tumour escape from CD8+ T-cell killing, would induce TILs down-modulation of CD5 to prevent tumour evasion. In line with this, CD5lo CTL clones from lung carcinoma patients displayed higher tumour-specific cytotoxicity than CD5hi clones and increased susceptibility to tumour-induced AICD (29). This is in agreement with CD5’s prevention of AICD by negatively regulating T-cell activation (47). Higher AICD susceptibility of CD5lo CTLs would explain the transient control of tumour growth observed in CD5-deficient mice challenged with melanoma (48). Our work demonstrates that protection of CD5-deficient T cells from AICD by adenoviral-mediated expression of soluble Fas-Fc results in reduced melanoma growth (48). The latter may reflect a therapeutic strategy for patients showing low intratumour CD5 mRNA levels. In contrast, potentiation of TILs efficiency in tumours with high CD5 expression should include CD5 and Fas-FasL blocking strategies. A necessary CD5 blockade is supported by improved anti-tumour responses in mice expressing the soluble human CD5 (shCD5) transgene or injected with recombinant shCD5 (49,50).

Recent IHC analyses in small series (n=30) of untreated advanced-stage NSCLC patients concluded that intra-tumour high CD3+ and low CD5+ infiltrates associate to poor prognosis (51). Melanoma patient genetics has revealed that the hypofunctional CD5 haplotype (isoform Pro224-Ala471, a poor down-regulator of TCR/CD3-mediated T-cell activation) associates to better survival (29).

This is the first study of CD6 expression in tumour-resected samples from early-stage NSCLC patients. No significance has been obtained in our NSCLC training cohort, but higher CD6 expression correlates with improved RFS and OS in our analyses of NSCLC and melanoma TCGA data. There is little information on the function of CD6. However, the homology between CD5 and CD6, may enable some degree of functional redundancy. Knockout mice show that both receptors share modulatory roles in thymocyte activation (negative) and survival (positive) (16,18,19). Moreover, Nur-77 levels -indicative of TCR signaling strength- are elevated in CD6hi compared to CD6lo peripheral T cells (18). CD6lo/neg peripheral T cell populations are less responsive to T-cell activators, more susceptible to apoptosis and enriched in regulatory T cells (Treg) (27,52). High TCR avidity and survival of CD6hi T cells would be compatible with high intratumour CD6 expression and favourable cancer prognosis, suggesting its biomarker potential only awaits confirmation.

In order to determine if CD5 expression association to favourable prognosis is due to higher lymphocyte infiltration, CD5 and CD6 ratios to CD4 and CD8 were calculated. Our results confirm their prognostic value and support that higher CD5 expression and lymphocyte infiltration associate to increased antitumour immune responses and improved patient prognosis in early-stage NSCLC.

Conclusions

This study points to a positive prognostic role for two lymphocyte inhibitory co-receptors, CD5 and CD6, in early-stage NSCLC (and in Melanoma). This conclusion is compatible with high surface levels of both CD5 and CD6 associated to TCR avidity and resistance to AICD. This evidence suggests that CD5 and CD6, along with other checkpoint inhibitors (e.g., PD-1 and CTLA-4), may be additional markers of tumour-specific T cells. Further studies deciphering the exact role of CD5, CD6 and their ligands in cancer would benefit patient stratification for personalized immunotherapies and development of new and more efficient strategies.

Melanoma-specific survival curves according to the expression status of CD5 and CD6. Kaplan-Meier survival curves for time to progression (TTP) and OS of high vs. low CD5 (A and B) or CD6 (C and D) expression levels, in tumor samples from melanoma patients. OS, overall survival.

The article’s supplementary files as