Identification and characterization of CLEC11A and its derived immune signature in gastric cancer

Abstract

Introduction: C-type lectin domain family 11 member A (CLEC11A) was characterized as a growth factor that mainly regulates hematopoietic function and differentiation of bone cells. However, the involvement of CLEC11A in gastric cancer (GC) is not well understood.

Methods: Transcriptomic data and clinical information pertaining to GC were obtained and analyzed from publicly available databases. The relationships between CLEC11A and prognoses, genetic alterations, tumor microenvironment (TME), and therapeutic responses in GC patients were analyzed by bioinformatics methods. A CLEC11A-derived immune signature was developed and validated, and its mutational landscapes, immunological characteristics as well as drug sensitivities were explored. A nomogram was established by combining CLEC11A-derived immune signature and clinical factors. The expression and carcinogenic effects of CLEC11A in GC were verified by qRT-PCR, cell migration, invasion, cell cycle analysis, and in vivo model analysis. Myeloid-derived suppressor cells (MDSCs), regulatory T cells (Tregs), M2 macrophages, and T cells in tumor samples extracted from mice were analyzed utilizing flow cytometry analysis.

Results: CLEC11A was over-expressed in GC, and the elevated CLEC11A expression indicated an unfavorable prognosis in GC patients. CLEC11A was involved in genomic alterations and associated with the TME in GC. Moreover, elevated CLEC11A was found to reduce the benefit of immunotherapy according to immunophenoscore (IPS) and the tumor immune dysfunction, exclusion (TIDE). After validation, the CLEC11A-derived immune signature demonstrated a consistent ability to predict the survival outcomes in GC patients. A nomogram that quantifies survival probability was constructed to improve the accuracy of prognosis prediction in GC patients. Using shRNA to suppress the expression of CLEC11A led to significant inhibitions of cell cycle progression, migration, and invasion, as well as a marked reduction of in vivo tumor growth. Moreover, the flow cytometry assay showed that the knock-down of CLEC11A increased the infiltration of cytotoxic CD8+ T cells and helper CD4+ T into tumors while decreasing the percentage of M2 macrophages, MDSCs, and Tregs.

Conclusion: Collectively, our findings revealed that CLEC11A could be a prognostic and immunological biomarker in GC, and CLEC11A-derived immune signature might serve as a new option for clinicians to predict outcomes and formulate personalized treatment plans for GC patients.

Keywords: Clec11a; gastric cancer; immunotherapy; prognosis; tumor microenvironment.

Figure 1

A flowchart of the study design. The mRNA and protein expressions of CLEC11A in GC were investigated using TCGA, GEO, and HPA databases. Kaplan-Meier curves were used to assess the overall survival in CLEC11A subgroups. CLEC11A was involved in genomic instability. The functional enrichment analyses identified CLEC11A’s relevance with cancer immunity. The associations between CLEC11A and ESTIMATE, immune checkpoints, and immunocyte infiltration were further explored. A 6-gene CLEC11A-derived immune signature was constructed to predict prognosis and immune therapy response, and guide precision therapy. The mRNA expression of CLEC11A in GC was verified by qRT−PCR, and the carcinogenic effects of CLEC11A were examined by cell migration, invasion, cell cycle analysis, and in vivo analysis. The associations between CLEC11A and immunocyte infiltration were analyzed utilizing flow cytometry.

Figure 2

CLEC11A was over-expressed in GC. (A) CLEC11A mRNA expression across TCGA pan-cancer. (B) CLEC11A mRNA expression in GSE13861, GSE13911, GSE26899, GSE29272, GSE54129, and GSE66229. (C) The CLEC11A mRNA expression in GC was verified by PCR analysis. (D) Immunohistochemical analysis of CLEC11A in GC and normal stomach tissues by HPA database. * P < 0.05, ** P < 0.01, *** P < 0.001, **** P < 0.0001, ns: not significant.

Figure 3

Prognosis analyses of CLEC11A in GC. (A) Cox regression of CLEC11A across TCGA cancer types. Overall survival analyses of CLEC11A in (B) TCGA-STAD, (C) GSE26899, (D) GSE13861, (E) GSE26901, (F) GSE15459, (G) GSE29272, (H) GSE26253, (I) GSE62254, (J) GSE84426, (K) GSE84433, and (L) GSE84437.

Figure 4

CLEC11A was related to epigenetic modulations and genomic instability in GC. (A) The genomic alterations of CLEC11A in GC were explored by the cBioPortal online web tool, including mutation and amplification. (B) The correlation between CLEC11A and genomic heterogeneity. The correlation between CLEC11A and (C) 4 methyltransferases, (D) 5 MMR genes, and (E) 44 RNA modulations. (F) The gene mutation frequency and chromosomal gain/loss were analyzed between CLEC11A subgroups in TCGA-STAD. * P < 0.05, ** P < 0.01, *** P < 0.001.

Figure 5

Functional analysis of CLEC11A in GC. (A) Top 20 differential genes between different CLEC11A expression subgroups. (B) GO analysis and (C) KEGG analysis of differential genes of CLEC11A.

Figure 6

The oncogenic effect of CLEC11A in GC. (A) Migration and invasion assays. (B) Cell cycle assay. (C) Reduced CLEC11A expression inhibited tumor growth in vivo. * P < 0.05, ** P < 0.01, *** P < 0.001.

Figure 7

CLEC11A mediated immune lymphocytes in TME. (A) Distributions of TME scores between CLEC11A subgroups. (B) Correlations between CLEC11A and immune lymphocytes in TISIDB. Flow cytometry immunophenotyping analysis of the populations of (C) cytotoxic CD8+ and helper CD4+ T cells, (D) Tregs, (E) M2 macrophages, and (F) MDSCs in MFC tumor-bearing mice after reducing CLEC11A expression.

Figure 8

Immune-related genes and immune response analyses. (A) Expression correlations between CLEC11A and immunoinhibitors, immunostimulators according to TISIDB database. (B) The expression patterns of CLEC11A in four TCGA molecular subtypes of GC. (C) TIDE and (D) IPS between CLEC11A subgroups. * P < 0.05, ** P < 0.01, *** P < 0.001.

Figure 9

A CLEC11A-derived immune signature was developed and validated. (A) The results of univariate Cox regression were presented by forest plot. (B) The importance of CLEC11A-related immune genes was calculated by random survival forest analysis. (C) Overall survival analysis between risk score subgroups in TCGA-STAD cohort, GSE26899, and GSE15459. (D) ROC curves of CLEC11A-derived immune signature in predicting one, three, and five-year overall survival in the TCGA training set, GSE26899 and GSE15459. (E) Five genes (CSF1R, CXCR4, TGFB1, TGFBR1, and TNFSF18) included in the signature showed associations with the overall survival of patients in the TCGA-STAD cohort. Independent prognostic analyses of the clinical features and CLEC11A-derived immune signature in (F) TCGA-STAD cohort, (G) GSE26899, and (H) GSE15459.

Figure 10

A nomogram was developed and validated. (A) The distributions of risk scores in clinical features. (B) Nomogram construction based on the CLEC11A-derived immune signature and clinical characteristics, including stage, N, TMB, and MSI. (C, D) Independent prognostic analyses of the CLEC11A-derived immune signature and clinical features in the TCGA-STAD cohort. (E) ROC curves showed the prediction performances of the nomogram in one-, three-, and five-year overall survival. (F) Calibration curves of the nomogram for one, three, and five-year overall survival.

Figure 11

Somatic mutations and CNVs analysis in GC patients. (A, B) Waterfall graphs illustrated the mutational landscape in risk score subgroups. (C) Genetic alteration of 6 signature genes. (D) CNV frequencies of 6 signature genes. (E) Genomic positions of 6 signature genes. The bands on the inner circle represented the corresponding expression levels.

Figure 12

The TME characteristics of CLEC11A-derived immune signature. (A, B) GO terms enriched in risk score subgroups were determined by GSEA analysis. (C) Immune fractions between the risk score subgroups were quantified by TME scores. (D) Differences of immune-related pathways between risk score subgroups. (E) The relationships between TME infiltrated cells and genes included in CLEC11A-derived immune signature. (F) CIBERSORT algorithm quantified the TME infiltrated cells between risk score subgroups. * P < 0.05, ** P < 0.01, *** P < 0.001.

Figure 13

Immunotherapy response and drug sensitivity. (A) TMB in risk score subgroups. (B) The differences of risk scores in MSI-H/MSS/MSI-L. (C) The TIDE scores and (D) IPS in risk score subgroups. (E) Drug sensitivities of Pazopanib, Dasatinib, Sunitinib, Vinorelbine, Sorafenib, Doxorubicin, Pyrimethamine, and Etoposide in risk score subgroups. ** P < 0.01, *** P < 0.001.