Low Expression of CD5 and CD6 Is Associated with Poor Overall Survival for Patients with T-Cell Malignancies

Background T-cell malignancies (TCMs), including T-cell acute lymphoblastic leukemia (T-ALL) and T-cell lymphoma (TCL), are highly aggressive and have a poor prognosis. To further understand prognostic stratifications and to design targeted therapies, this study aims to explore novel, potential biomarkers based on alterations in immune costimulatory molecules (CMs) for TCMs. Methods Peripheral blood from 25 de novo T-ALL patients in our clinical center and transcriptome data from 131 to 162 patients with peripheral TCL (PTCL) from the GSE19069 and GSE58445 dataset, respectively, were obtained to assess the expression levels of CMs and their prognostic significance. Results Seven CMs were associated with overall survival (OS). Among these CMs, CD5 and CD6 had the highest pairwise positive correlation (R = 0.69). CD5 and CD6 were significantly down-regulated in TCM patients compared with healthy individuals (HIs), and lower CD5 and CD6 expression was associated with poor OS for both T-ALL and TCL patients, particularly for patients greater than 60 years old. Furthermore, CD5 was positively correlated with CD6 in TCM patients. Compared with patients who were CD5highCD6high, T-ALL and TCL patients who were CD5lowCD6low had poor OS. Importantly, CD5highCD6high was an independent prognostic predictor for OS in T-ALL (HR = 0.39, 95% CI: 0.23–0.65, P < 0.001) and TCL (HR = 0.35, 95% CI: 0.19–0.62, P < 0.001) patients. Conclusions Low expression of CD5 and CD6 was associated with poor OS for TCM patients, and this may be a potential immune biomarker panel for prognostic stratification of TCM patients.


Introduction
T-cell malignancies (TCMs), including T-cell acute lymphoblastic leukemia (T-ALL) and T-cell lymphoma (TCL), are a group of heterogeneous diseases with high relapse and mortality rates. T-ALL is a malignant disease caused by clonal proliferation of precursor T lymphocytes, and it accounts for 10-15% of pediatric and 20-25% of adult ALL [1,2]. Adult T-ALL patients have adverse clinical outcomes due to chemotherapy resistance, relapse, and no effective targeted drugs [3]. Moreover, TCL accounts for approximately 2% of non-Hodgkin's lymphoma and has a poor prognosis compared with B-cell lymphoma [4]. Gene mutations and abnormal expression are often used as prognostic biomarkers for risk stratification of cancer patients [5][6][7]. Recently, abnormal expressions of costimulatory and inhibitory molecules were also used as prognostic biomarkers for hematological malignancies [8,9]. Hence, novel biomarkers derived from genetics for prognostic stratification and the development of targeted therapy in TCM are urgently needed to solve this dilemma.
Costimulatory receptor-mediated signaling has a strong impact on T-cell responses. For example, CD28, ICOS, CD40, and CD58 are potential biomarkers for prognosis and immunotherapy for hematological malignancies [10][11][12][13]. Our previous publication demonstrated that low CD58 expression is associated with poor clinical outcomes for cytogenetically normal acute myeloid leukemia (AML) patients [11]. CD5 and CD6 are class I scavenger receptors that have highly homologous extracellular regions but little conserved cytoplasmic tails [14,15]. CD5 and CD6 are transmembrane glycoproteins that are highly similar in structure and function [16]. ese proteins are expressed on the surface of the same lymphocyte populations, including mature T cells and B1 cells [17,18]. Both are involved in the development, activation, differentiation, and survival of lymphocytes [19,20]. Multiple studies have demonstrated that CD5 and CD6 act as costimulatory molecules for lymphocyte activation and proliferation based on monoclonal antibody experiments [20,21]. ese findings were subsequently disputed by reports of their negative modulatory effects on activation signals in CD5 and CD6 deficient mice [22][23][24]. Currently, there are bidirectional roles of CD5 and CD6 in cancer immunity. CD5 and CD6 have been demonstrated to affect the immune response to cancers. Furthermore, higher CD5 and CD6 expression predicts favorable outcomes for patients with nonsmall cell lung cancer (NSCLC) and plays positive roles in immune surveillance [25]. In contrast, down-regulation of CD5 expression in tumor-infiltrating lymphocytes was reported to improve the anti-tumor response in lung cancer patients [26]. Additionally, different functional variations of CD5 are related to either favorable or poor prognosis in patients with chronic lymphocytic leukemia [27]. However, the prognostic importance of CD5 and CD6 expression in TCM patients remains unclear.
In this study, peripheral blood (PB) samples from 25 de novo patients with T-ALL in our clinical center and transcriptome sequencing data from 131 to 162 patients with peripheral TCL (PTCL) from the GSE19069 and GSE58445 dataset of Gene Expression Omnibus (GEO) database, respectively, were obtained to investigate the prognostic value and expression levels of costimulatory molecules (CMs) in TCM.

Methods and Materials
2.1. T-ALL Samples. PB samples were collected from 25 de novo patients with T-ALL in our clinical center (JNU) from July 2009 to August 2016, and this was designated as a training cohort (Figure 1). e median follow-up time for surviving T-ALL patients was 7.3 years, and their clinical information is summarized in Table S1. In addition, PB samples from 9 healthy individuals (HIs) were obtained for controls. All participants provided written informed consent, and this study was conducted according to the Declaration of Helsinki principles and approved by the Ethical Committee of Jinan University.

Publicly Available Datasets.
e GSE19069 dataset, including 10 normal T-cell samples and 131 PTCL samples, was downloaded from the GEO database (https://www.ncbi. nlm.nih.gov/geo/) to analyze the expression levels of the CMs [28]. Moreover, the GSE58445 dataset, including transcriptome data and clinical information from 162 PTCL patients, was also obtained from the GEO database [29]. Data from the GSE58445 dataset were used as the validation cohort. e clinical characteristics, including overall survival (OS) time, event, age, and gender, are listed in Table S1.
e qRT-PCR reaction procedures were as follows: preincubation, 95°C for 3 min, and amplification, 95°C for 5 sec and 60°C for 15 sec for a total of 45 cycles. mRNA expression levels were normalized to β-actin using the 2 −ΔΔCT method. e sequences of the primers are presented in Table S2. e prognostic cut-off values for quantitative variables were calculated by X-tile software (version3.6.1, Yale University, New Haven, CT, USA) [8,31]. Kaplan-Meier curves were plotted using the R package "survival," and differences between subgroups were compared by the log-rank test [5]. Univariate and multivariate Cox regression models were constructed by SPSS. Wilcoxon test (two-tailed) or Spearman method were used to evaluate the difference or correlation between two groups of quantitative data, respectively. P value <0.05 was regarded as statistically significant.

Low CD5 and CD6 Expression Is Associated with Poor OS for Patients with TCM.
To identify prognostic CM predictors of OS in TCM patients, we first performed a Kaplan-Meier survival analysis using the GSE58445 dataset. Interestingly, a total of seven CMs including CD5, CD6, CD2, CD40, CD80, CD86, and ICOS were significantly associated with the OS for TCL patients (P < 0.05, Figure S1(a)). Among these CMs, CD5 and CD6 had the highest pairwise correlation; thus, CD5 and CD6 were targeted for subsequent analysis in this study ( Figure S1(b)). Compared with HIs, both CD5 and CD6 were significantly down-regulated in T-ALL patients, which was confirmed in TCL patients (P < 0. CD5 and CD6 are constitutively expressed lymphocyte receptors whose expression can be regulated during lymphocyte development and activation events. erefore, correlations with the up-regulated CD3, CD4, and CD8 expression levels in T cells were evaluated, which would relatively exclude the effects of T-cell counts on CD5 and CD6 expression. e gene expression of CD5 and CD6 was normalized to that of CD3E, CD3G, CD3D, CD247, CD4, CD8A, and CD8B. Interestingly, higher CD5/CD3G (HR � 0.220, P � 0.008), CD5/CD3D (HR � 0.294, P � 0.025), and CD5/CD247 (HR � 0.346, P � 0.057) expression was associated with improved OS in T-ALL patients (Figure 3(a), left panel).

Co-Expression of CD5/CD6 for Prognostic Stratification in TCM Patients
. Because CD5 and CD6 are co-receptors on the surface of lymphocytes, their correlation was investigated. We demonstrated a strong positive relationship between CD5 and CD6 in TCL patients, and this was also found in T-ALL patients (R = 0.41, P � 0.044) (Figure 4(a)). Combinations of genes may be better than a single gene in predicting prognoses and performing risk stratification for cancer patients. Interestingly, T-ALL patients who were CD5 low and CD6 low had poor OS (HR = 0.214, 95% CI: 0.046-0.995, P � 0.032; 5-year OS: 9.1% vs. 62.5%) and a shorter median OS (1.24 vs. 6.56 years) than those who were CD5 high and CD6 high (Figures 4(b) (Figures 4(b)-4(c), right panel). In addition, when sex, age, and CD5/CD6 ratio were included in univariate and multivariate COX regression models for survival analysis, the results indicated that CD5 high CD6 high was an independent prognostic predictor of OS in T-ALL patients (HR = 0.39, 95% CI: 0.23-0.65, P < 0.001).

Discussion
Risk stratification based on the International Prognostic Index (IPI) (including age, stage, performance status, serum lactate dehydrogenase level, and extranodal involvement) has made great progress in predicting the prognosis of patients with TCL, who can be divided into four groups: low, low-intermediate, high-intermediate, and high-risk [32].
is precise risk stratification can provide important references for the management of TCL patients and clinical decision-making, thereby improving patient outcomes [32,33]. However, IPI-based risk stratification cannot accurately predict a prognosis for all TCL patients [33,34]. e reasons for this heterogeneity may be due to clinical characteristics, morphology, genetics, and immunophenotype. Notably, gene alterations play an important role in constructing risk stratification for hematological malignancies, particularly acute myeloid leukemia, but there is a lack of information on the genetic alterations that complement risk stratification to more accurately predict clinical outcomes for TCL patients [32,35,36]. In addition, although multiple studies have been actively exploring the role of genetic alterations combined with minimal residual disease (MRD) and clinical information in risk stratification for T-ALL patients, high heterogeneity makes it difficult to accurately stratify all patients [37]. erefore, further exploration of novel biomarkers to improve risk stratification for TCM patients is needed.
Previous studies have reported that some CMs can be used as prognostic biomarkers for hematological malignancies. AML patients with B7-2 positivity shared a poorer prognosis compared to AML patients who were B7-2 negative [38]. In this study, eleven CMs were analyzed, and seven were associated with the prognosis of TCM patients. However, different hematological malignancies may have different immune receptor abnormalities, and the combination of two related immune receptors has greater advantages compared with a single molecule in predicting the prognosis of patients [5,6,8,11]. Interestingly, CD5 had the strongest correlation with CD6, and low expression of CD5 and CD6 was significantly associated with adverse outcomes in TCM patients. ese findings were consistent with the results that low CD5 and CD6 expression predicts poor prognosis in patients with NSCLC or melanoma [25]. TCL patients older than 60 years of age have adverse clinical outcomes and higher risk stratification compared with those younger than 60 [32]. us, more precise stratification is required for TCL patients older than 60 years for rational decision-making. Interestingly, our study suggests that low expression of CD5 and CD6 could predict poor OS in TCL patients older than 60 years. However, due to the small T-ALL sample size, subgroup analysis could not be performed for validation.
CD5 and CD6 are transmembrane glycoproteins expressed on the surface of T cells that act as costimulatory molecules in the TCR signaling pathway [39,40]. TCMs are transformed by the malignant proliferation of T cells and an increase in T-cell counts. To this end, we normalized the CD5 and CD6 levels to that of the mRNA expression of CD3, CD4, and CD8 [25]. Interestingly, even after normalization, the expression of CD5 and CD6 predicted the prognosis of TCM patients but was not affected by the T-cell counts. Altogether, CD5 and CD6 might be immune biomarkers for the prognostic stratification of TCM patients. We attempted to validate our findings with additional T-ALL and TCL publicly available datasets, but we could not obtain the complete prognostic information and transcriptome data for necessary analysis; thus, more T-ALL or TCL samples are needed to further validate our results in the future.
In conclusion, we observed that lower expression of CD5 and CD6 was associated with poor OS for patients with TCM, and co-expression of CD5 and CD6 was an independent prognostic predictor of OS in TCM patients. ese findings provided deep insight that CD5 and CD6 might be immune biomarkers for prognostic stratification and the development of targeted therapies for TCM patients.  Data Availability e GSE58445 and GSE19069 datasets were downloaded from the Gene Expression Omnibus (GEO) dataset (https:// www.ncbi.nlm.nih.gov/geo/). e datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.

Conflicts of Interest
e authors declare that there are no conflicts of interest.

Authors' Contributions
YQL and CTC contributed to the concept development, study design, and editing of the manuscript. SNS and ZYL performed the experiments, interpreted the data, and wrote the manuscript. CWZ and OJHL reviewed the manuscript. GXL, LW, and JXT provided the clinical samples and treated the patients. All authors read and approved the final manuscript.
Acknowledgments is work was supported by grants from the Intergovernmental International Cooperation on Scientific and Technological Innovation Project of Chinese Ministry of Science and Technology (No. 2017YFE0131600) and the Guangdong Science and Technology Project (No. 2020A0505100042). Table S1: clinical characteristics of patients with TCM. Table  S2: primers for qRT-PCR. Figure S1: prognostic (A) and correlation (B) analysis of costimulatory molecules in TCL patients. Figure S2