SLAMF8, a potential new immune checkpoint molecule, is associated with the prognosis of colorectal cancer

Highlights

• •SLAMF8 is a potential new immune checkpoint molecule.
• •High SLAMF8 expression is associated with poor prognosis of colorectal cancer.
• •The expression of SLAMF8 corelates with the expression of classical immune checkpoint molecules.
• •Prediction of the potential function of SLAMF8 in the tumor microenvironment in CRC.

Abstract

Recently, immune checkpoint inhibitors (ICIs), such as programmed cell death 1 (PD-1) monoclonal antibodies (mAbs), have revolutionized the treatment of malignant tumors. Therefore, the number of studies aiming to screen and identify new immune checkpoint molecules for antitumor immunotherapy is increasing. Signaling lymphocytic activation molecule (SLAM) family members are mainly expressed by and regulate the functions of immune cells. Recent studies have shown that several SLAM family members are involved in the regulation of the tumor immune microenvironment and are promising targets for antitumor immunotherapy. Signaling lymphocytic activation molecule family member 8 (SLAMF8) is a type I cell surface glycoprotein and is encoded on chromosome 1q21. To further illustrate the clinical value of SLAMF8 in colorectal cancer (CRC), we retrospectively analyzed the relationship between SLAMF8 expression and the prognosis of CRC patients and the associations between SLAMF8 expression and the expression levels of other SLAM family members and other classic immune checkpoint molecules using The Cancer Genome Atlas (TCGA) data, RNA sequencing data, tissue immunohistochemistry staining, and systematic follow-up analysis. Here, high SLAMF8 expression was associated with poor overall survival (OS) in CRC. The mRNA expression level of SLAMF8 was positively correlated with the expression levels of multiple classic immune checkpoints and other SLAM family members. Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis suggested that the pathways enriched in CRC tissues with high SLAMF8 expression were associated with the regulation of the tumor immune microenvironment.

Introduction

Colorectal cancer (CRC) is the third most commonly diagnosed cancer and the second leading cause of cancer-related death [1], [2], [3]. Although surgery, chemoradiotherapy, and targeted therapy have greatly improved the prognosis of CRC patients, the prognosis of advanced CRC is still unsatisfactory [4,5]. Thus, a major breakthrough is urgently needed for the treatment of such cases. In recent years, the results of studies of immune checkpoint inhibitors (ICIs), such as anti-programmed cell death 1 (PD-1) monoclonal antibodies (mAbs), have enabled continuous updating of guidelines of antitumor therapy [6]. However, this striking efficacy has been limited to microsatellite instability-high (MSI-H) or mismatch repair-deficient (dMMR) CRC [7,8], and how to ensure that most patients with CRC benefit from immunotherapy a focus of immunotherapy research. Studies have shown that multiple immunosuppressive immune checkpoint molecules in the tumor microenvironment jointly participate in the negative regulation of T-cell function and affect the efficacy of antitumor immunotherapy [9,10]. Therefore, screening to identify additional immune checkpoints with immunosuppressive effects in the tumor immune microenvironment will enable the identification of new immunotherapy targets. The signaling lymphocytic activation molecule (SLAM) family is part of the larger CD2 protein family, and SLAM family members are differentially expressed on the surface of immune cells [11]. Recent studies have indicated that signaling lymphocytic activation molecule family (SLAMF) members are involved in regulating adaptive and innate immune responses [12], [13], [14]. SLAMF7 signaling reprograms T cells toward exhaustion in the tumor microenvironment [15], [16], [17]. Lewinsky H et al. reported that SLAMF5 is a regulator of the immunosuppressive microenvironment in multiple myeloma and regulates PD-1/PD-L1 expression and function in chronic lymphocytic leukemia [18]. In head and neck squamous cell carcinoma, the expression of SLAMF4 was significantly increased in tumor-infiltrating CD8+ T cells and was significantly correlated with the expression of PD-1 and PD-L1, which was related to the formation of an immunosuppressive tumor microenvironment [19]. SLAM family member 8 (SLAMF8; also known as B-lymphocyte activator macrophage expressed (BLAME)/CD353) is a cell-surface protein and a member of the SLAM family [11]. Few studies have investigated the role of SLAMF8 in the tumor microenvironment. Our previous report confirmed that SLAMF8 is mainly expressed on the surface of tumor‐associated macrophages (TAMs), which contribute to the immunosuppressive tumor microenvironment [12]. Accordingly, we hypothesized that SLAMF8 may be involved in the regulation of an immunosuppressive tumor microenvironment in CRC. We explored the clinical and functional effects of SLAMF8 in CRC, and our results may provide novel insights into the role of immune checkpoints in CRC.

Materials and methods

Clinical specimen collection

We retrospectively collected data from 118 CRC patients who received treatment in Nanjing Drum Tower Hospital from January 2017 to May 2022. All patients were diagnosed pathologically and did not undergo chemotherapy or related treatment before surgery. All clinicopathological cancer tissues were collected. We collected the clinicopathological features of the patients, including age, sex, tumor site, pathological TNM stage, and several tumor markers.

Survival analysis

Patients were followed up until death or until 25 May 2020. Overall survival (OS) was defined as the time interval from surgery to death from any cause. This study was approved by the Institutional Review Committee of Nanjing Drum Tower Hospital.

Cell culture and stimulation

The human monocyte cell line THP-1 was provided by the Cell Bank of the Chinese Academy of Sciences (Shanghai, China). THP-1 monocytes were cultured at 37 °C in 5% CO2 a RPMI 1640 medium (Corning, New York, USA) supplemented with 10% heat-inactivated fetal bovine serum (FBS; ExCell Bio, Jiangsu, China) and 1% penicillin‒streptomycin (Beyotime Biotechnology, Shanghai, China). THP-1 macrophages were subsequently stimulated with human IFN-γ (20 ng/ml) and LPS (100 ng/ml) for 48 h to attain the M1 phenotype, and human IL-4 (20 ng/ml) and IL-13 (20 ng/ml) were used to attain the M2 phenotype. THP-1 cells stimulated with PMA (150 ng/mL) were used as the M0 phenotype. PMA, LPS, IFN-γ, IL-4 and IL-13 were purchased from Beyotime (Shanghai, China).

Immunohistochemistry (IHC) staining

IHC assays were performed according to the standard protocol. Paraffin tissue samples containing tumor tissue were used. After being dewaxed and rehydrated, the 2 μm paraffin-embedded sections were incubated with primary mouse anti-human PD-L1 mAb (clone number 22C3) and then the secondary antibody. The antibodies were purchased from the Dako Corporation. Additionally, a rabbit polyclonal anti-human SLAMF8 antibody (1:200, ab221703, Abcam, Cambridgeshire, England) and a primary anti-human PD-L2 mAb (1:750, a b288298, Abcam, Cambridgeshire, England) were used. For mismatch repair protein staining, antibodies against MLH1 (clone number ES05), PMS2 (clone number EP51), MSH2 (clone number FE11), and MSH6 (clone number Pu29) (all from Dako, Denmark) were used at a dilution of 1:200. Negative and positive control staining was performed for each section. The assays were performed strictly according to the kit instructions.

IHC scoring

The results of PD-L1 and PD-L2 staining were obtained by using the combined positive score (CPS), calculated as follows: the number of tumor cells with a given intensity of membrane staining/ the number of lymphocytes/macrophages directly associated with tumor cells with the same intensity of membrane/cytoplasm staining, and the ratio to was multiplied by 100. A CPS of 10 was defined as positive for PD-L1 and PD-L2, and a CPS ≤10 was considered negative. SLAMF8 expression was quantitatively analyzed by using Image-Pro Plus 6.2 version software. All specimens were counted by two independent blinded pathologists. The dMMR criteria were assessed with an internal control (normal intestinal glandular epithelial and mesenchymal cells) with well-colored nuclei. Tumor cells with nuclei with brownish-yellow granular staining were considered positive, while tumor cells without nuclear staining were considered negative. pMMR positivity was defined as positivity for all four proteins. IHC results were interpreted in a double-blinded manner by two or more pathologists independently, and if needed, a consensus was obtained.

qRT–PCR and RNA‐sequencing analysis

For qRT‒PCR, total RNA was extracted and solubilized into cDNA using a transcribed first-strand cDNA synthesis kit (Roche, Basel, Switzerland). qRT–PCR was performed using an ABI 7900 System (Thermo Fisher Scientific, Massachusetts, USA). Primers were synthesized by Tsingke Biological Technology (Beijing, China), and the sequences of the primers were as follows: GAPDH: forward 5′‐GGAGCGAGATCCCTCCAAAAT‐3′, reverse: 5′‐GGCTGTTGTCATACTTCTCATGG‐3′; SLAMF8: forward5′‐CTGATGGTGGATACAAGGG‐3′, reverse: 5′‐GGAAATGGACGTAACGGA‐3′; CD80: forward: 5′‐GCAGGGAACATCACCATCCA‐3′, reverse: 5′‐TCACGTGGATAACACCTGAACA‐3′; CD163:forward:5′‐CCGGGAGATGAATTCTTGCCT‐3′,reverse:5′‐AGACACAGAAATTAGTTCAGCAGCAGCA‐3′. GAPDH was used as the endogenous control. A total of 20 CRC tissues were successfully subjected to RNA‐sequencing analysis. The 2^−ΔΔCt method was used to calculate the relative expression of RNA. RNA isolation and RNA-sequencing library construction and sequencing were carried out by Shanghai Bioengineering Technology (Shanghai, China) according to the manufacturer's protocol. We compared the enrichment results of SLAMF8 expression in CRC tissues (high expression vs. low expression) using GSEA software (v4.0.3, UC San Diego and Broad Institute, USA).

Statistical analysis

The relationships between the expression of PD-L1, PD-L2, and SLAMF8 and clinicopathological variables were statistically assessed by χ² analysis, Fisher's exact test or the Kruskal-Wallis test, as appropriate. The staining area and relative staining intensity of SLAMF8 were calculated using Image-Pro Plus 6.2 version software for grouping, and univariate and multivariate survival analyses were performed using a Cox proportional hazards regression model to assess independent prognostic factors. Survival curves were generated using the Kaplan‒Meier method, and differences in survival curves were determined by the log-rank test. All continuous data are presented as the mean ± standard deviation. A two-tailed P < 0.05 was considered to be statistically significant. SPSS 23.0 software and R4.1.0 software were used for analyses.

Result

Clinicopathological parameters

As mentioned above, we retrospectively collected data from one hundred eighteen patients who were diagnosed with CRC, and the clinicopathologic parameters of the study group are depicted in Table 1. Of these 118 patients, 79 were male, and 39 were female. The mean age was 59 years (26–84 years). In terms of the degree of tumor differentiation, 4 patients had highly differentiated disease, 79 patients had moderately differentiated disease, and 35 patients had poorly differentiated disease. Forty-four patients had vascular invasion, 67 patients had nerve invasion, and 63 patients had lymph node metastasis. Regarding tumor stage, 47 patients had stage 1 or 2 disease, and 71 patients had stage 3 or 4 disease (Table 1).

Table 1.

Patient characteristics.

Characteristics	Number of patients	Percentage (%)
Gender
Male	39	33
Female	79	69.9
Age (year)
<60	56	47.4
≥60	62	52.5
Diameter of tumor (cm)
≤5	22	18.6
>5	96	81.3
Nerve infiltration
No	51	43.2
Yes	67	56.7
Vascular invasion
No	74	62.7
Yes	44	37.2
Lymphatic metastasis
No	55	46.6
Yes	63	53.3
pTNM stage
I + II	47	39.8
III + IV	71	60.1
Differentiation
Well	4	3.3
Moderate	79	66.9
Low	35	29.6
Histological type
Adenocarcinoma	106	89.8
Mucinous adenocarcinoma or others	12	10.1
Tumor location
Left half	50	42.3
Right half	27	22.8
Rectum	41	34.7
MMR
pMMR	96	81.3
dMMR	22	18.6

SLAMF8 expression and clinicopathologic variables

The associations between SLAMF8 protein expression and clinicopathologic variables are shown in Table 2. Patients with high SLAMF8 expression were more likely to have lymphatic metastasis. There was a borderline significant correlation between high SLAMF8 expression and advanced TNM stage. In addition, we found no significant correlation between SLAMF8 expression levels and patient age, sex, tumor size, vascular invasion, or MMR status (P > 0.05 for all) (Table 2).

Table 2.

The relationship between the expressions of SLAMF8 and clinicopathological features in colorectal cancer.

	SLAMF8		X2	P
	Neg	Pos
Gender
Male	18	21	0.3452	>0.05
Female	41	38
Age (year)
<60	31	25	1.224	>0.05
≥60	28	34
Diameter of tumor (cm)
≤5	12	10	0.223	>0.05
>5	47	49
Nerve infiltration
No	25	26	0.035	>0.05
Yes	34	33
Vascular invasion
No	37	37	0.00	>0.05
Yes	22	22
Lymphatic metastasis
No	33	22	4.121	0.042
Yes	26	37
pTNM stage
I + II	28	19	2.8646	0.091
III + IV	31	40
Differentiation
Well	38	48	4.663	0.097
Moderate	15	9
Low	6	2
Histological type
Adenocarcinoma	51	55	1.484	>0.05
Mucinous adenocarcinoma or others	8	4
Tumor location
Left half	25	25	1.536	>0.05
Right half	16	11
Rectum	18	23
MMR
pMMR	49	10	0.223	>0.05
dMMR	47	12

High SLAMF8 expression was associated with poor prognosis

Cun‐Yi Zou et al. reported that higher SLAMF8 expression levels were correlated with shorter survival in glioma [20]. To further evaluate the relationship between SLAMF8 expression and the prognosis of CRC patients, we downloaded a CRC dataset from the TCGA database and included the random tumor tissues of 118 CRC patients from our center for IHC analysis. We analyzed the association between SLAMF8 expression at the mRNA and protein levels and the prognosis of CRC patients. SLAMF8 expression was divided into low expression and high expression based on the median value. In the TCGA cohort, high SLAMF8 mRNA expression was associated with poor prognosis (P = 0.035) (Fig. 1(A)). Kaplan‒Meier survival analysis showed that high SLAMF8 protein expression was significantly associated with poor OS (P = 0. 0039) (Fig. 1(B)).

Fig. 1.

Higher SLAMF8 expression was correlated with shorter survival in CRC.

(A) Kaplan‒Meier survival analysis of patients with CRC with high SLAMF8 mRNA expression or low SLAMF8 mRNA expression using the TCGA CRC dataset. (B) Kaplan‒Meier survival analysis of patients with CRC with high protein expression of SLAMF8 or low protein expression of SLAMF8 based on IHC of tumor tissues from 118 CRC patients from our center.

Correlation between the expression of SLAMF8 and the expression of other SLAM family members

As mentioned previously, multiple SLAM family members have been confirmed to be involved in the regulation of the tumor immune microenvironment and are potential immunotherapy targets. Previous studies by our team also preliminarily confirmed that SLAMF8 expression was associated with CD8-positive T cells in CRC. These findings suggest that SLAMF8 may be similar to other SLAM family members and participate in the regulation of the tumor immune microenvironment. We used the TCGA dataset and our RNA sequencing dataset for analysis, and we found that SLAMF8 was significantly correlated with the expression levels of multiple SLAM family members (Fig. 2(A)), especially SLAMF7 (Fig. 2(B) and (C)).

Fig. 2.

The relationship between the expression of SLAMF8 and the expression of SLAM family members.

(A) SLAMF8 expression was significantly correlated with the expression levels of multiple SLAM family members according to analysis of the TCGA CRC dataset. Both the TCGA CRC dataset (B) and our RNA sequencing dataset (C) showed a relationship between the expression of SLAMF8 and SLAMF7.

Correlations between the expression of SLAMF8 and the expression of classical immune checkpoint molecules

It has been reported that increased expression of multiple immune checkpoints in tumor tissue is correlated with the immunosuppressive tumor microenvironment. To demonstrate the role of SLAMF8 in the CRC tumor immune microenvironment, we analyzed the mRNA expression levels of SLAMF8 and several classic immune checkpoint molecules using public datasets. We observed a positive correlation between SLAMF8 and these molecules in CRC (Fig. 3(A)). Furthermore, we validated the associations between the expression of SLAMF8 and the expression of classic immune checkpoint molecules using our RNA sequencing dataset and obtained similar results to those from the public datasets (Fig. 3(B)).

Fig. 3.

The relationships between SLAMF8 expression and the expression of classical immune checkpoint molecules.

Analyses of the TCGA CRC dataset (A) and our RNA sequencing dataset (B) indicated that SLAMF8 expression was positively related to the expression levels of CD274, PDCD1LG2, PDCD1, HAVCR2, CTLA4, and LILRB4.

We also assessed the association between the protein expression level of SLAMF8 and PD-L1/PD-L2 protein expression using IHC analysis of CRC tissue samples (n = 118) (Fig. 4(A)-(F)). We validated the significant association between the protein expression of SLAMF8 and PD-L1 protein expression in our CRC dataset with 118 tissue samples (P = 0.0349) (Fig. 4(H)), but we observed no significant association between the protein expression of SLAMF8 and PD-L2 protein expression (P > 0.05) (Fig. 4(I)).

Fig. 4.

Correlation of SLAMF8 protein expression with PDL1 protein expression.

Tissues from our center were used. Representative images of SLAMF8 ((A) and (D)), PDL1 ((B) and (E)), and PDL2 ((C) and (F)) IHC staining in CRC (n = 118) tissues are shown, with scale bars of 200 µm. Quantification of the IHC analysis showed a correlation between SLAMF8 expression and PDL1 expression (H) and between SLAMF8 expression and PDL2 expression (I).

Prediction of the functions of SLAMF8 in the CRC tumor microenvironment

Our previous study investigated the utility of SLAMF8 for predicting the efficacy of anti-PD1 mAb therapy in gastrointestinal tumors; the results suggested that SLAMF8 may be involved in the regulation of the tumor immune microenvironment of gastrointestinal tumors. Other previous studies have suggested that SLAMF8 is mainly expressed on the surface of macrophages. In this study, we found for the first time that SLAMF8 was expressed at a higher level in M1 macrophages than in M2 macrophages (Fig. 5(A)). To further reveal the role that SLAMF8 plays in the tumor immune microenvironment in CRC, we performed KEGG analysis using TCGA datasets. The results revealed that the pathways enriched in CRC tissues with high SLAMF8 expression were associated with the regulation of the tumor immune microenvironment, including the chemokine signaling pathway, cytokine‒cytokine receptor interaction, antigen processing and presentation, Fc gamma R-mediated phagocytosis and pathways in cancer (Fig. 5(A)-(F)).

Fig. 5.

The pathways enriched in CRC tissues with high SLAMF8 expression were associated with the regulation of the tumor immune microenvironment.

(A) SLAMF8 was expressed at a higher level in M1 macrophages than in M2 macrophages. KEGG pathway analysis was used to explore the pathways enriched in CRC tissues with high SLAMF8 expression: antigen processing and presentation (B), cytokine‒cytokine receptor interaction, (C) chemokine signaling pathway (D), Fc gamma R-mediated phagocytosis (E), and pathways in cancer (F).

Discussion

Immunotherapy is a promising treatment, but challenges remain, especially for CRC [[7], [8], [21], [22], [23]]. In CRC, most patients who receive immunotherapy have MSI-H CRC [24], [25], [26]. There are still great challenges for immunotherapy for CRC patients with MSS disease, which represents approximately 90% of CRC cases; thus, targeted therapies involving new immune checkpoint molecules are urgently needed to complement immunotherapy [27,28]. Recently, several novel immune checkpoints and their functions have been explored, such as T-cell immunoglobulin and mucin domain-containing protein 3 (TIM3), V-domain Ig suppressor of T-cell activation (VISTA), lymphocyte activation gene-3 (LAG-3), T-cell immunoreceptor with immunoglobulin and ITIM domain (TIGIT) and leukocyte immunoglobulin (Ig)-like receptor B4 (LILRB4) [29], [30], [31], [32], [33], [34], [35], [36], [37], [38].

The SLAM family is reported to modulate both adaptive and innate immune responses, and several SLAM family members have been confirmed as new target candidates for immunotherapy [39]. In this retrospective study of 118 CRC patients, we demonstrated that SLAMF8 expression was significantly higher in the lymphatic metastasis group, and high SLAMF8 expression was associated with poor prognosis. These findings support the prognostic value of SLAMF8 and suggest that SLAMF8 plays a role in supporting the formation of an immunosuppressive tumor microenvironment in CRC. However, the function of SLAMF8 in CRC still needs to be studied. Moreover, we found that SLAMF8 expression was positively associated with the expression levels of several classic immune checkpoint molecules, such as PD-L1 and PD-L2.

Lewinsky, H. et al. reported that CD84 (SLAMF5) regulates PD-L1 expression in vivo [40]. We also further explored the relationship between SLAMF8 and PD-L1. We found that the mRNA expression of SLAMF8 was positively correlated with that of other SLAM family members, such as SLAMF7. SLAMF7 is an effective immunotherapy target in multiple myeloma [41], and SLAMF7 signaling reprograms T cells toward exhaustion in the tumor microenvironment [14], which indicates that SLAMF8 may be involved in the response to antitumor immunotherapy. Similar to other SLAM family members, SLAMF8 is mainly expressed on the surface of macrophages, and it is expressed at a low level on the surface of T cells. Previous reports have confirmed that SLAMF8 is involved in the regulation of the function of macrophages, and TAMs are important components in the tumor microenvironment and participate in the formation of an immunosuppressive tumor microenvironment [42], [43], [44]. In this study, we found that SLAMF8 expression was higher in M1 macrophages than in M2 macrophages. The innate immune response can activate the adaptive immune response, and bridging these two responses is the key to a lasting antitumor immune response. Studies have shown that ICI therapy may be hindered by the polarization of macrophages in the tumor microenvironment to M2 TAMs. TAMs can inhibit the antitumor immune response and promote tumor growth by releasing anti-inflammatory cytokines and angiogenic factors [45,46]. The plasticity of macrophages makes it possible to polarize M2-type macrophages back into M1-type macrophages. This process can also enhance the vitality of T cells and improve the antitumor immune response [47]. In addition, macrophages can affect not only the innate immune response but also the adaptive immune response by phagocytizing tumor cells and promoting antigen presentation. Therefore, we speculate that SLAMF8 has effects in CRC by regulating the plasticity of macrophages. Previous studies have shown that when T cells are activated by IL-2, the expression of SLAMF8 is significantly upregulated [12]. SLAMF8 is a self-ligand. SLAMF8 is expressed at low levels on the surface of resting T cells, but the expression of SLAMF8 is significantly upregulated on activated T cells and M1 macrophages. We speculate that the interaction between SLAMF8 on the surface of activated T cells and SLAMF8 on the surface of TAMs forms a SLAMF8-SLAMF8 signaling axis, which may promote the formation of an immunosuppressive microenvironment. Furthermore, in this study, we found that the enriched signaling pathways in CRC tissues with high SLAMF8 expression were significantly related to the regulation of the tumor immune microenvironment, indicating that SLAMF8 may play a role in the regulation of the tumor immune microenvironment in CRC.

Conclusion

In summary, SLAMF8 expression was associated with malignant progression, unfavorable prognosis, and the expression of several immune checkpoint molecules in CRC. Further studies are needed to determine whether SLAMF8 can be used as an immunotherapy target.

Author contributions

Xiaoping Qian conceived the idea of the article. Yaping Zhang and Qun Zhang composed the manuscript and figure-making. Yaping Zhang and Qun Zhang have contributed equally to this work. Xingzhi Han and Lu Han and Xiao Shi Provided data statistics. Ting Wang provided the IHC test. Jing Hu and Li Li and Zhou Ding supported the clinical data. All authors contributed to the article and approved the submitted version.

Funding

This study was supported by the Jiangsu Provincial Natural Science Foundation (SBK2021021543), the Nanjing Science and Technology Development Project (ZKX21028), the High-level Innovation and Entrepreneurship Talent Introduction Plan of Jiangsu Province (JSSCBS20211532), and the Nanjing Drum Tower Hospital National Youth Training Foundation (2022-JCYJ-QP-94).

CRediT authorship contribution statement

Yaping Zhang: Formal analysis, Writing – original draft, Methodology. Qun Zhang: Formal analysis, Writing – original draft, Methodology, Funding acquisition. Xingzhi Han: Methodology. Lu Han: Resources. Ting Wang: Resources. Jing Hu: Resources. Li Li: Formal analysis. Zhou Ding: Visualization. Xiao Shi: Resources. Xiaoping Qian: Visualization, Funding acquisition, Formal analysis, Writing – original draft.

Declaration of Competing Interest

Authors declare that they have no conflict of interest.