Pan-cancer analysis identifies PD-L2 as a tumor promotor in the tumor microenvironment

Abstract

Background: Programmed cell death protein 1 (PD-1) receptor has two ligands,programmed death-ligand 1 (PD-L1) and PD-L2. When compared with PD-L1, PD-L2 has not received much attention, and its role remains unclear.

Methods: The expression profiles of pdcd1lg2 (PD-L2-encoding gene) mRNA and PD-L2 protein were analyzed using TCGA, ICGC, and HPA databases. Kaplan-Meier and Cox regression analyses were used to assess the prognostic significance of PD-L2. We used GSEA, Spearman's correlation analysis and PPI network to explore the biological functions of PD-L2. PD-L2-associated immune cell infiltration was evaluated using the ESTIMATE algorithm and TIMER 2.0. The expressions of PD-L2 in tumor-associated macrophages (TAMs) in human colon cancer samples, and in mice in an immunocompetent syngeneic setting were verified using scRNA-seq datasets, multiplex immunofluorescence staining, and flow cytometry. After fluorescence-activated cell sorting, flow cytometry and qRT-PCR and transwell and colony formation assays were used to evaluate the phenotype and functions of PD-L2⁺TAMs. Immune checkpoint inhibitors (ICIs) therapy prediction analysis was performed using TIDE and TISMO. Last, a series of targeted small-molecule drugs with promising therapeutic effects were predicted using the GSCA platform.

Results: PD-L2 was expressed in all the common human cancer types and deteriorated outcomes in multiple cancers. PPI network and Spearman's correlation analysis revealed that PD-L2 was closely associated with many immune molecules. Moreover, both GSEA results of KEGG pathways and GSEA results for Reactome analysis indicated that PD-L2 expression played an important role in cancer immune response. Further analysis showed that PD-L2 expression was strongly associated with the infiltration of immune cells in tumor tissue in almost all cancer types, among which macrophages were the most positively associated with PD-L2 in colon cancer. According to the results mentioned above, we verified the expression of PD-L2 in TAMs in colon cancer and found that PD-L2⁺TAMs population was not static. Additionally, PD-L2⁺TAMs exhibited protumor M2 phenotype and increased the migration, invasion, and proliferative capacity of colon cancer cells. Furthermore, PD-L2 had a substantial predictive value for ICIs therapy cohorts.

Conclusion: PD-L2 in the TME, especially expressed on TAMs, could be applied as a potential therapeutic target.

Keywords: PD-L2; colon cancer; immune checkpoint; pan-cancer analysis; tumor associated macrophages.

Figure 1

The flow chart of the entire study. PD-L2, programmed cell death 1 ligand 2; GSEA, gene set enrichment analysis; KEGG, Kyoto Encyclopedia of genes and genomes; PPI, protein-protein interaction; scRNA-seq, single-cell RNA sequencing; TAMs, tumor associated macrophages; qRT-PCR, quantitative reverse transcription-polymerase chain reaction; TIDE, Tumor Immune Dysfunction and Exclusion; TISMO, Tumor Immune Syngeneic MOse; GDSC, Genomics of Drug Sensitivity in Cancer; CTRP, Cancer Therapeutics Response Portal.

Figure 2

Pdcd1lg2 expression in The Cancer Genome Atlas (TCGA) database.

Figure 3

Prognostic values of pdcd1lg2. (A) The prognostic value of pdcd1lg2 on overall survival (OS) or progression-free survival (PFS) displayed by the Kaplan-Meier method. (B) Cox regression analysis of pdcd1lg2 on OS or PFS in pan-cancer described by the forest plot. (C) Based on the Kaplan-Meier models and Cox regression, summary of the correlation between pdcd1lg2 expression and OS or PFS.

Figure 4

Signaling pathways associated with pdcd1lg2 expression according to Kyoto Encyclopedia of genes and genomes (KEGG) analyzed by the Gene Set Enrichment Analysis (GSEA) in pan-cancer. FDR, false discovery rate; NES, normalized enrichment score.

Figure 5

Relationship between pdcd1lg2/PD-L2 expression and immune-related genesin pan-cancer. (A) The Spearman correlation heatmap between pdcd1lg2 expression levels and immune-related genes. (B) The protein-protein interaction (PPI) network presents the proteins interacting with PD-L2. *p<0.05, **p<0.01, ***p<0.001.

Figure 6

The correlation between the expression of pdcd1lg2 and various immune cells infiltration levels in cancers. DC, dentritic cell; NK cell, natural killer cell; Treg, T cell regulatory; CAF, cancer-associated fibroblast; Endo, endothelial cell; Eos, eosinophil; γ/δ T cell, T cell gamma delta; Tfh, T cell follicular helper; NKT, nature killer T cell; HSC, hematopoietic stem cell; MDSC, myeloid-derived suppressor cell.

Figure 7

The phenotype and functions of PD-L2⁺ tumor associated macrophages (TAMs). (A) The localization of pdcd1lg2 expression analyzing in two scRNA-seq datasets. (B) The multiplex immunofluorescence images of PD-L2⁺TAMs in human colon cancer tissues. (C) Representative flow cytometry plots and analysis of the expression of PD-L2 on TAMs in MC38 tumors at different days after engraftment. (D) Flow cytometric analysis of the expression of CD206 on PD-L2⁺TAMs and PD-L2^-TAMs. (E) M1 macrophage and M2 macrophage marker gene expressions in macrophages. (F) Schematic overview of the strategy for identification the functions of PD-L2⁺TAMs. (G) Tranwell assays and colony formation to detect the role of PD-L2⁺TAMs in the migration, invasion, and proliferation of MC38 cells. *p<0.05, **p<0.01, ***p<0.001.

Figure 8

Immunotherapy response prediction, biomarker relevance and sensitive drug prediction of PD-L2. (A) Immunotherapy response of pdcd1lg2 in murine immune checkpoint blockade (ICB) therapy cohorts analyzed by TISMO database. (B) Biomarker relevance of pdcd1lg2 compared to standardized biomarkers with consistent evidence on cancer immune evasion in ICB therapy cohorts. The area under the receiver operating characteristic curve (AUC) is applied to evaluate the prediction performance of the biomarker on ICIs response status. (C) Predictive drugs based on the pdcd1lg2 expression from the GDSC (left) and CTRP (right) databases.