Increased co-expression of 4-1BB with PD-1 on CD8+ tumor-infiltrating lymphocytes is associated with improved prognosis and immunotherapy response in cervical cancer

Abstract

Background: The combination of agonistic antibodies with immune checkpoint inhibitors presents a promising avenue for cancer immunotherapy. Our objective is to explore the co-expression of 4-1BB, ICOS, CD28, with PD-1 on CD8+ T cells in the peripheral blood and tumor tissue of cervical cancer(CC) patients, with a specific focus on the association between the co-expression levels of 4-1BB with PD-1 and clinical features, prognosis as well as immunotherapy response. The goal is to offer valuable insights into cervical cancer immunotherapy.

Methods: In this study, 50 treatment-naive patients diagnosed with CC were enrolled. Flow cytometry was used to detect PD-1/4-1BB, PD-1/ICOS and PD-1/CD28 co-expression on CD8+ T cells. Subsequent analysis aimed to investigate the differential co-expression between peripheral blood and cancer tissue, and also the correlation between co-expression and clinical features in these patients. Gene Expression Omnibus (GEO) datasets, The Cancer Genome Atlas (TCGA) cohort, The IMvigor210 cohort, The BMS038cohort and Immunophenoscores were utilized to investigate the correlation between PD-1/4-1BB and the immune microenvironment, prognosis, immunotherapy, and drug sensitivity in cervical cancer.

Results: The co-expression levels of PD-1/4-1BB, PD-1/ICOS, and PD-1/CD28 on CD8+ tumor-infiltrating lymphocytes (TILs) were significantly higher in cervical cancer patients compared to those in peripheral blood. Clinical feature analysis reveals that on CD8+ TILs, the co-expression of PD-1/4-1BB is more closely correlated with clinical characteristics compared to PD-1/ICOS, PD-1/CD28, PD-1, and 4-1BB. Pseudo-time analysis and cell communication profiling reveal close associations between the subgroups harboring 4-1BB and PD-1. The prognosis, tumor mutation burden, immune landscape, and immunotherapy response exhibit statistically significant variations between the high and low co-expression groups of PD-1/4-1BB. The high co-expression group of PD-1/4-1BB is more likely to benefit from immunotherapy.

Conclusion: PD-1/4-1BB, PD-1/ICOS, and PD-1/CD28 exhibit elevated co-expression on CD8+TILs of cervical cancer, while demonstrating lower expression in circulating T cells. The co-expression patterns of PD-1/4-1BB significantly contributed to the prediction of immune cell infiltration characteristics, prognosis, and tailored immunotherapy tactics. PD-1/4-1BB exhibits potential as a target for combination immunotherapy in cervical cancer.

Keywords: 4-1BB; PD-1; cervical cancer; co-expression; immunotherapy efficacy; prognosis.

Figure 1

PD-1/4-1BB, PD-1/ICOS and PD-1/CD28 co-expression on CD8+ TILs in cervical cancer patients was higher than in PBMCs (A) Gating strategy to identify PD-1/4-1BB, PD-1/ICOS and PD-1/CD28 double positive CD8+ T cells in PBMC and TILs using flow cytometry. Comparison of PD-1/4-1BB (B), PD-1/ICOS (C), and PD-1/CD28 (D) co-expression levels in TILs and PBMCs. Comparison of PD-1/4-1BB, PD-1/ICOS, and PD-1/CD28 co-expression levels on CD8+ PBMC (E) and CD8+ TILs (F). (G) 4-1BB expression according to differential PD-1 expression levels. The percentage of 4-1BB+ cells was compared among PD-1^positive and PD-1^neg subpopulations of total CD8+TILs. TILs, tumor infiltrating lymphocyte; PBMCs, peripheral blood mononuclear cells. ***p<0.001.

Figure 2

Establishment of the Single-Cell Landscape in Cervical Cancer. (A) UMAP view of total cells obtained from 1 CC sample in the GSE168652 dataset, color-coded by assigned cell type. (B) U-MAP analysis to recluster CD8+ T cells into six subpopulations based on subtype-specific gene markers. (C) Marker gene expression for each cell type, where dot size and color represent the percentage of the marker gene. (D) Violin plots showed elevated expression of PD-1 in the CD8_1 cluster and 4-1BB in the CD8_6 cluster.

Figure 3

Pseudo-time analysis reveal close associations between the subgroups harboring 4-1BB and PD-1. (A) Boxplot showing the comparison of CytoTRACE score between different CD8 subsets. Differentiation trajectory of CD8+T cells, with each color coded for pseudo-time (B) and clusters (C). (D) Fluctuations in PD-1 and 4-1BB gene expression during cell differentiation. (E) The differentially expressed genes (rows) along the pseudo-time (columns) were clustered hierarchically into six profiles. Color key differentially coding from blue to red indicated the relative expression levels from low to high. (F) Variations in the expression levels of PD-1 and 4-1BB across distinct cellular differentiation fates.

Figure 4

Intercellular ligand-receptor prediction among CD8+T cells and immune cells revealed by CellChat. (A) An overview of cell-cell interactions. (B) For the relative importance of each cell group based on the computed network centrality measures of signaling networks. Influencer represents a kind of cell that can control information flow within a signaling network, and a higher value indicates greater control on the information flow. The meaning of importance is the magnitude of the possibility of four roles (sender, receiver, mediator, and influencer) that the cell types play. The darker the color, the greater the role cells play. (C) Bubble plots of ligand-receptor pairs. Dot color reflects communication probabilities, and dot size represents computed p-values. Empty space means the communication probability is zero. p-values are computed from a two-sided permutation test. (D) Inferred incoming and outgoing communication patterns of CD8+T cells. The CD8_1 cluster predominantly serves as signal sender, while the CD8_6 cluster functions primarily as signal receiver. (E) The interplay between the CD8_1 cluster and CD8_6 cluster encompasses cellular signaling pathways and ligand-receptor interactions. (F) The expression of indicated transcriptional factors showed with heatmap. Combining single-cell datasets with TCGA, we performed overall survival (OS) analysis for the PD-1/4-1BB high and low co-expression groups, incorporating data from GSE168652 (G) and GSE171894 (H).

Figure 5

Co-expression of PD-1/4-1BB is closely linked with the immune microenvironment in cervical cancer. (A) Cytotoxic responses are heightened in the PD-1/4-1BB high co-expression group, featuring increased expression of (IFNG, GZMA, PRF1, GZMB). (B) Comparison of tumor mutation burden between high and low expression groups of PD-1/4-1BB. (C) Comparison of immune scores, StromaScore, ImmuneScore, and ESTIMATEScore, between the high and low expression groups. (D) Heatmap of the two groups based on ssGSEA scores for different immune regulatory factors. (E) PD-1/4-1BB high co-expression group exhibits elevated expression of both inhibitory and stimulatory immune checkpoints. (F, G) Waterfall plot of tumor somatic mutation in the high and low expression groups. ***p<0.001.

Figure 6

Prognostic and drug sensitivity analysis. (A) DSS and (B) PFI of patients with low or high expression groups. (C) Volcano plot of the distributions of all differentially expressed genes. (D) GO analysis; (E) KEGG analysis. (F) Stratification of PD-1/4-1BB predicts drug therapeutic benefits in CC. Proportion of normalized IC50 value of the 89 drugs between the low and high expression groups. (G) Comparison of IC50 values for six commonly used drugs in cervical cancer treatment, 5-Fluorouracil, Bleomycin, Etoposide, Gemcitabine, Mitomycin C, and Paclitaxel, between the high and low expression groups.

Figure 7

Immunotherapeutic response analysis. (A–D) The differences of IPS between high and low expression groups stratified by both CTLA4 and PD-1. (E) Proportion of patients with different treatment outcomes in high and low expression groups. The proportion of CR/PR patients in high expression group was significantly higher than that in low expression group in IMvigor210 cohort (p < 0.01). (F) The difference of the co-expression levels of PD-1/4-1BB between treatment outcome groups (p<0.01). The statistical difference above was compared by the Wilcoxon test. (G) The proportion of CR/PR patients in high expression group was significantly higher than that in low expression group in BMS038 cohort (p < 0.05). (H) In BMS038 cohort, The difference of the co-expression levels of PD-1/4-1BB between treatment outcome groups (p<0.01). IPS, Immunophenoscores; R=CR/PR, complete response/partial response; NR=SD/PD, stable disease/progressive disease. **p < 0.01.