Blocking CXCR4+ CD4+ T cells reprograms Treg-mediated immunosuppression via modulating the Rho-GTPase/NF-κB signaling axis

Abstract

Background: While clinical trials have shown that CXCR4 antagonists can enhance the efficacy of cancer immunotherapy, the molecular mechanisms by which CXCR4 modulates the tumor microenvironment remain poorly understood. We recently identified CXCR4 as a regulator of exhausted CD8⁺ T cell phenotypes in cancer. Here, we investigate its role in orchestrating regulatory T (T_reg) cell-mediated immunosuppression within tumors.

Methods: We conducted meta-analyses of single-cell RNA-seq datasets from pan-cancer tissues to characterize CXCR4 expression patterns in CD4⁺ T cells. Using CXCR4 antagonists and conditional knockout mice (Cxcr4^flox/flox, Lck^Cre), we inhibited T_reg phenotypes in vivo. Through single-cell transcriptomics and single-cell ATAC-seq of the cervical cancer mouse model, phosphoproteomics, and ChIP-seq analyses, we elucidated how CXCR4 blockade in CD4⁺ T cells suppresses activated T_reg phenotypes by modulating the Rho-GTPase/NF-κB signaling axis. We further integrated RNA-seq data, clinical trial datasets (NCT02826486 and NCT04516616), and human organoid models to validate the therapeutic potential of CXCR4 inhibition in enhancing antitumor immunotherapy.

Results: Single-cell transcriptomics of CD4⁺ T cells across multiple cancers revealed CXCR4 expression was associated with T_reg cell developmental trajectories. Pharmacological and genetic inhibition of CXCR4 inhibited T_reg phenotypes in cervical cancer and breast cancer. Mechanistically, phosphoproteomics and ChIP-seq analyses unveiled that blocking CXCR4⁺ CD4⁺ T cells reduced activated T_reg phenotypes by modulating the Rho-GTPase/NF-κB signaling axis. Single-cell transcriptomic and multi-omic analyses demonstrated that blocking CXCR4⁺ CD4⁺ T cells promoted immunotherapy via reprogramming T_reg-mediated immunosuppression. Furthermore, clinical trial data and human cervical cancer organoids confirmed that blocking CXCR4 enhances antitumor immunotherapy by reducing T_reg phenotypes.

Conclusions: Our study highlights the crucial role of CXCR4 in deriving T_reg-mediated immunosuppression via regulating the Rho-GTPase/NF-κB signaling axis, informing the potential of combining CXCR4 blockades with T cell-targeted immunotherapies.

Keywords: CXCR4; Immunosuppression; NF-κB2; Regulatory T (Treg) cells; Rho GTPase pathway.

Fig. 1

Single-cell transcriptomic analysis identified that CXCR4 expression was associated with T_reg cell developmental trajectories. A UMAP plot of CD4⁺ T cells (n = 45,363) in lung adenocarcinoma. B Correlation analysis of gene expression (CXCR4 and CTLA4) and the proportion of CD4-CTLA4-T_reg cells. C UMAP plot of CD4⁺ T cells (n = 11,166) in non-small cell lung cancer. D Correlation analysis of gene expression (CXCR4 and FOXP3) and the proportion of CD4-CTLA4-Treg cells. E UMAP plot of CD4⁺ T cells (n = 63,965) in multi-cancer. F Correlation analysis of CXCR4 expression and the proportion of CD4-LEF1-T_reg and CD4-CTLA4-T_reg cells. G Developmental trajectories of T_reg cells in breast cancer from GSE156728 via diffusion map. Expression levels of marker genes and CXCR4 were shown. H Developmental trajectories of T_reg cells in pancreatic cancer from GSE156728 via diffusion map analysis. Expression levels of marker genes and CXCR4 were shown

Fig. 2

Blocking CXCR4 in CD4⁺ T cells reduces the activated T_reg phenotypes in vivo. A Gross anatomy of subcutaneous tumors (U14) in mice treated with CXCR4 antagonists. B Flow cytometry and statistical analysis of tumor models (4T1 and ID8) under different treatments; p values were calculated using the unpaired, parametric t-test. C UMAP plot of CD4⁺ T meta-cell clusters of scRNA-seq data from pan-cancer analysis. Exhausted clusters are indicated. T_n, naïve T cells; T_m, memory T cells; T_em, effector memory T cells; T_emra, terminally differentiated effector memory or effector; Tfh, follicular helper T cell; Th1, T helper 1; ISG, interferon-stimulated genes. D Violin plots showing the expression levels of canonical marker genes across CD4⁺ T cell clusters. E Schematic diagram illustrating the experimental design for CXCR4 antagonist treatment in U14 tumor-bearing immunocompetent syngeneic mice. Flow cytometry analysis (F) and corresponding quantification (G) of CD4⁺ T cell subsets in tumors across different treatment groups. H Schematic diagram and experimental design for tumor-bearing (U14) Cxcr4-cKO and control mice. Flow cytometry analysis (I) and quantification (J) of CD4.⁺ T cell subsets in spleens from tumor-bearing control and cKO mice. p values were determined by unpaired, parametric t-test. **p < 0.01; ***p < 0.001

Fig. 3

Blocking CXCR4⁺ CD4⁺ T cells enhances anti-PD-1 immunotherapy efficacy by reprogramming T_reg-mediated immunosuppressive TME. A Mean tumor volumes and gross anatomy of U14 tumors in primary recipients with various treatments (saline and IgG, CXCR4 antagonist, anti-PD-1, combination). p values were assessed using two-way ANOVA. B Tumor volume analysis of 4T1-tumor-bearing animals treated with indicated drugs; each line represents an individual mouse. After 40 days of combination treatment, mice showing complete response to combined treatment were rechallenged with 4T1 tumors and treated with CXCR4 agonist and control vehicle. p values were assessed using two-way ANOVA. C RNA heatmaps of marker genes in U14 tumors from different treatments. D UMAP plots of 11 cell clusters and treatment samples based on scRNA-seq data. E UMAP plots of five CD4⁺ T cell subtypes and their treatment samples based on scRNA-seq data. F Venny plot of DGEs in CD4⁺ T cells compared to controls. The bottom panel shows pathway enrichment analysis of the overlapped DEGs. G Pseudotime analysis of differentially expressed heatmaps for CD4⁺ T cells. H Violin plot of gene expression in DC and T_reg cells from different treatments based on scRNA-seq data. I Schematic illustrating the mechanism by which blocking CXCR4⁺ CD4⁺ T cells regulates T_reg-mediated immunosuppression. CXCR4 inhibition reduces the recruitment and suppressive function of T_reg cells, characterized by downregulation of key immunosuppressive cytokines (e.g., IL-10, TGF-β) and immune checkpoint molecules (e.g., CTLA-4, TIM-3). This leads to diminished T_reg–APC interactions and decreased expression of co-inhibitory ligands (e.g., PD-L1, CD80, CD86) on APCs. ***p < 0.001

Fig. 4

Phosphoproteome and ChIP-seq analyses reveal that targeting CXCR4⁺ CD4⁺ T cells reduces T_reg-associated suppressive genes via modulation of the Rho-GTPase/NF-κB signaling axis. A Experimental design for phosphoproteome analysis in human T_reg cells treated with CXCR4 antagonist. Human T_reg cells were purified by FACS and stimulated in vitro. B Heatmap displaying the differentially regulated phosphorylation sites in T_reg cells before and after CXCR4 antagonist treatment. Each row represents a phosphorylation site, and each column represents a biological replicate. C Pathway enrichment analysis based on proteins with different phosphorylation sites in T_reg cells pre- and post-CXCR4 antagonist treatment. D Overview of phosphorylation level changes associated with the Rho-GTPase/NF-κB signaling axis induction by CXCR4 antagonist treatment. Differentially regulated phosphorylation sites related Rho-GTPase/NF-κB signaling axis were indicated. Protein–protein interaction networks represent differentially regulated phosphorylation sites. E The ex vivo co-culture system using human cervical cancer organoids and PBMCs derived from two patients at different treatment timepoints. Patient-derived PBMCs were co-cultured with autologous organoids under various treatment conditions. F Violin plots showing quantification of apoptosis rates in cervical cancer organoids under different treatment conditions, as assessed by caspase-3 staining. Each dot represents an independent measurement from different patient-derived samples. p values were calculated by the Mann–Whitney test in GraphPad Prism. G ChIP-seq analysis of chromatin occupancy in T_reg-associated genes based on NFKB2 and RelB ChIP-seq data. H ChIP-PCR validation of NFKB2 and RelB binding to the promoters of selected T_reg signature genes in T_reg cells before and after CXCR4 antagonist exposure. Data are presented as mean ± SD. *p < 0.05; **p < 0.01; ns, no significance

Fig. 5

Single-cell multi-omic analyses reveal that blocking CXCR4⁺ CD4⁺ T cells epigenetically reprogram T_reg-associated suppressive genes. A tSNE plots of 14 cell clusters based on scRNA-seq and scATAC-seq data from U14 tumor-bearing Cxcr4^flox/floxLck^Cre and Cxcr4^flox/flox mice. B Chromatin accessibility analysis of marker genes in cell clusters. C Representative mIHC staining of spleens from tumor-bearing Cxcr4^flox/flox Lck^Cre and Cxcr4^flox/flox mice. D tSNE plots of four subtypes of CD4⁺ T cells based on scRNA-seq and scATAC-seq data from tumor-bearing Cxcr4^flox/flox Lck^Cre and Cxcr4^flox/flox mice. E tSNE and violin plots displaying CXCR4 expression levels in T_reg cells based on scRNA-seq analysis. p values were determined by the Wilcoxon rank-sum test. F Motif scores in T_reg cells from scATAC-seq analysis. G Chromatin accessibility analysis of marker genes in T_reg cells. H Pathway enrichment analysis based on the top 50 different peaks in T_reg cells between Cxcr4^flox/flox Lck^Cre and Cxcr4^flox/flox mice. I Chromatin accessibility analysis of marker genes in B cells, macrophages, and DCs. J Representative mIHC staining of tumors from tumor-bearing Cxcr4^flox/flox Lck^Cre and Cxcr4^flox/flox mice. ***, p < 0.001

Fig. 6

Clinical trial data and human organoids confirm that blocking CXCR4 enhances antitumor immunotherapy efficacy. A Schematic overview of clinical trial NCT02826486, including treatment timeline and peripheral blood sampling schedule. B Longitudinal analysis of lymphocyte counts, CXCR4⁺ CD4⁺ T cells, CD4⁺ CD25⁺ FOXP3⁺ regulatory T cells, and CD69⁺ CD4⁺ T cells following CXCR4 antagonist monotherapy and combination therapy. Gray lines denote geometric means; blue and red lines indicate the lower and upper bounds of the 95% confidence intervals, respectively. M, monotherapy; Pre, pre-treatment; Post, post-treatment; D, day. Statistical comparisons were performed using two-sided t-tests based on geometric means and derived standard deviations. Significance thresholds were adjusted using Bonferroni correction. C Schematic of RNA-seq workflow from trial NCT04516616. Cervical cancer patients receiving neoadjuvant chemotherapy plus anti-PD-1 therapy were stratified into pathological complete response (pCR) and non-pCR groups, and tumor samples were subjected to transcriptomic profiling. D Heatmap showing immune cell-type enrichment scores, as assessed by xCell algorithm analysis, comparing pCR and non-pCR groups. E Paired comparisons of gene expression and immune cell scores before and after immune checkpoint blockade (ICB) in pCR and non-pCR groups. p values calculated by paired t-test in GraphPad Prism. F Correlation analysis of gene expression profiles of pre- and post-ICB in non-pCR patients. G Schematic of experimental setup in which PBMCs from non-pCR patients were co-cultured with cervical cancer-derived organoids. H Representative images of cervical cancer organoid and PBMC co-culture system under different treatments. Tumor cell apoptosis was evaluated by caspase-3 staining. ***, p < 0.001

Fig. 7

Main content analysis and findings of this study. This study integrates multi-omic profiling and functional validation experiments to investigate how CXCR4⁺ CD4⁺ T cells contribute to tumor immunosuppression and how CXCR4 blockade can reprogram this process. Approaches included single-cell RNA-seq, single-cell multi-omics, bulk RNA-seq, phosphoproteomics, mouse tumor models, patient-derived organoid PBMC co-culture systems, and clinical cohort analyses. Mechanistically, CXCR4 antagonism disrupted the Rho GTPase/NF-κB signaling pathways, leading to reduced transcription of T_reg-associated suppressive molecules (e.g., CTLA-4, PD-1, TIM-3, and TNFRSF members). This downregulation impaired T_reg-mediated suppression, ultimately rebalancing the tumor immune microenvironment in favor of antitumor immunity and enhancing the efficacy of anti-PD-1 immunotherapy