TBX3 shapes an immunosuppressive microenvironment and induces immunotherapy resistance

Abstract

Background: Identifying biomarkers that predict immunotherapy efficacy and discovering new targets for combination therapies are critical elements for improving the prognosis of bladder cancer (BLCA) patients. Methods: Firstly, we explored the expression patterns of TBX3 in normal and pan-cancer tissues and the correlation between TBX3 and the immune microenvironment using data from multiple public databases. Then, we combined various techniques, including bulk RNA sequencing, single-cell RNA sequencing, high-throughput cytokine arrays, functional experiments, ProcartaPlex multiplex immunoassays and TissueFAXS panoramic tissue quantification assays, to demonstrate that TBX3 shapes an immunosuppressive tumor microenvironment (TME) in BLCA. Results: We identified TBX3 as a key factor associated with the immunosuppressive microenvironment in BLCA through a systematic multi-omics analysis. We found that TBX3 is primarily expressed in malignant cells, where TBX3^high tumor cells increase the secretion of TGFβ1, which promotes the infiltration of cancer-associated fibroblasts (CAFs), thereby forming an immunosuppressive microenvironment. We further demonstrated that TBX3 enhances TGFβ1 expression by binding to the TGFβ1 promoter, and blocking TGFβ1 counteracts the immunosuppressive effects of TBX3. Moreover, TBX3 reduced the cancer-killing efficiency of CD8⁺ T cells by decreasing the proportion of GZMB⁺ CD8⁺ T cells, and knocking down TBX3 combined with anti-PD-1 treatment increased CD8⁺ T cell infiltration and reduced CAFs in vivo. We also validated the inverse relationship between TBX3⁺ malignant cells and CD8⁺ T cells and the positive relationship with CAFs in tissue microarrays. Lastly, we found that TBX3 predicted immunotherapy efficacy in our real-world immunotherapy cohort and multiple public cohorts. Conclusion: In summary, TBX3 promotes BLCA progression and immunotherapy resistance by inducing an immunosuppressive microenvironment, and targeting TBX3 could enhance the efficacy of immunotherapy for BLCA.

Keywords: CD8+ T cells; TBX3; bladder cancer; fibroblasts; immunosuppressive microenvironment; immunotherapy.

Figure 1

TBX3 is highly expressed in bladder cancer, and it is mainly expressed in malignant epithelial cells and fibroblasts. (A) Expression level of TBX3 in Xiangya bladder cancer cohort. (B) Expression level of TBX3 in the TCGA bladder cancer cohort. (C) The representative images of immunohistochemical staining showed the expression of TBX3 protein in bladder cancer and adjacent tissues. The yellow and brown nuclei represented positive. Scale bars, 100μm and 50μm. Expression levels of TBX3 were scored as four grades (-, +, + +, +++) by multiplying the percentage of positive cells and immunostaining intensity. (D)The positively stained nuclei (%) were analyzed by χ2 test. (E-F) The clustering of different cell types in Xiangya single cell cohort and PRJNA662018 cohort of bladder cancer tissue and the expression level of TBX3 in each cell type. (G)The heatmap shows the mRNA expression of TBX3 at the single-cell level across 33 tumor tissues. *P < 0.05, **P < 0.01, ***P < 0.001.

Figure 2

TBX3 correlated with a non-inflammatory tumor microenvironment in BLCA. (A) TBX3 expression and cancer immunity cycles in BLCA. The colors represent seven different steps. (B) Relationship between TBX3 and tumor-infiltrating immune cells (TIICs) by applying ssGSEA method within the TCGA-BLCA cohort. (C) Validation the interaction between TBX3 and TIICs in the Xiangya cohort. Taking the median expression level of tbx3 as the cut-off value, Xiangya cohort was divided into two groups. (D) Effector genes expression of CD8⁺ T cells, dendritic cells, NK cells, macrophages, and Th1 cells in high-TBX3 and low-TBX3 groups in TCGA-BLCA. (E) Correlation between TBX3 expression and T cell-inflamed scores in TCGA-BLCA. (F) Correlation between TBX3 expression and T cell-inflamed related genes (bottom left), and immune checkpoint genes (upper right) in the Xiangya cohort. *P < 0.05, **P < 0.01, ***P < 0.001.

Figure 3

TBX3 inhibited CD8⁺ T cell infiltration by promoting cancer associated fibroblasts. (A) Flow chart of in vivo studies, created using the BioRender.com website. (B) Macroscopic view of subcutaneous tumor in C57BL/6 mouse. (C) Subcutaneous tumor weight of two group. (D) The volume curve of subcutaneous tumor in two groups of C57BL/ 6 mice. (E) Survival rates of both groups of mice. (F) Single-cell RNA sequencing of subcutaneous tumors (3 vs 3) in both groups of mice was clustered into seven major cell types. (G) The heat map shows the difference in the proportion of immune cell clusters between the two groups. (H) Flow cytometric analysis of CD8⁺ T cell, CD4⁺ T in subcutaneous tumors. (I) Flow cytometric analysis GZMB⁺ CD8⁺ T cell in subcutaneous tumors. (J) IF staining images of CD8⁺ T cell in subcutaneous tumors. (K-L) Flow cytometric analysis and IF staining images of CAFs in subcutaneous tumors. Mean ± SD, *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.

Figure 4

TBX3 promotes the expression of TGFβ1 in bladder cancer cells. (A, B) GO and KEGG enrichment analysis of differential genes in patients with high and low expression of TBX3 in TCGA bladder cancer cohort. (C) KEGG enrichment analysis of bladder cancer cells Tccsup-vector vs Tccsup-oeTBX3 differential gene. (D) KEGG enrichment analysis of bladder cancer cells T24-shNC vs T24-shTBX3 differential gene. (E) Differential gene GSEA enrichment analysis of cancer cells in subcutaneous tumors of MB49-vector and MB49-oeTBX3 groups. (F) Cytokine microarray analysis of overexpressed Tccsup-oeTBX3 and Tccsup-vector differences in expression of cytokines. (G) The levels of TGFβ1 secreted in bladder cancer cell lines were analyzed by Elisa. (H) The levels of TGFβ1 secreted in bladder cancer cell lines were analyzed by WB. (I) Luciferase reporter gene system analysis of TGFβ1 promoter activity in Tccsup cells co-transfected with PGL3-TBX3. (J) Transcription factor TBX3 binds to the motifs of downstream target genes. (K) Prediction of possible promoter binding sites of TBX3 and TGFβ1 by JASPAR database (top 3 scores). (L) qPCR-ChIP experiments confirmed that TBX3 directly binds to promoter TGFβ1. (M) Construct a mutation vector for binding site-1 luciferase reporter. (N) The activities of serially mutated tgfb1 promoter reporter vectors in the HEK293T cells co-transfected with pCMV-TBX3. Mean ± SD, p < 0.0001, p <0.0001.

Figure 5

High expression of TBX3 in tumor cells shapes the non-inflammatory tumor microenvironment in a TGFβ1-dependent manner. (A) Co-culture workflow flowchart, created using the BioRender.com website. (B) Flow cytometry was used to analyze the proportion of IFN-r⁺ CD8⁺ T cells. (C) Flow cytometry was used to analyze the proportion of GZMB⁺CD8⁺ T cells. (D) Western blot analysis was performed on the labeled proteins of cancer-associated fibroblasts in the above four groups. (E) CAFs scores of cancer-associated fibroblasts in subcutaneous tumors in vector and oeTBX3 groups were compared. (F) The dot plot shows the average expression level of CAFs marker genes in vector and oeTBX3 groups. (G-J) Comparison of CAFs-related biological processes between vector and oeTBX3 groups.

Figure 6

Knocking down TBX3 inhibits tumor development and enhances anti-PD-1 therapeutic effect. (A) Construction of mouse subcutaneous tumor model and flow chart of immunotherapy, created using the BioRender.com website. (B) Four groups of tumor growth volume maps. (C) Gross view of subcutaneous tumors in four group. (D) Changes of tumor weight of mouse in each group. (E) Survival curves of four groups of mice. (F) The proportion of CD8⁺ T in tumor tissue was analyzed by flow cytometry. (G) The proportion of GZMB⁺ CD8⁺ T cells in tumor tissue was analyzed by flow cytometry. (H) The proportion of CAFs in tumor tissue was analyzed by flow cytometry. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.

Figure 7

TBX3⁺ tumor cells were negatively correlated with CD8⁺ T-cell infiltration and differentiation in human BLCA. (A-B) Representative multicolor staining of inflamed (A) and non-inflamed (B) phenotypes of patients with BLCA in the Xiangya BLCA tissue microarray (TMA): TBX3(pink), CK19 (azure), CD8 (green), a-SMA (yellow), and DAPI (blue). (C) Representative flow cytometry-like plots of TBX3⁺ CD8⁺(left), TBX3+a-SMA+ (middle) and TBX3⁺ CK19⁺cells (right) in TMA, respectively. (D) The histograms of different TBX3⁺ CD8⁺, TBX3⁺ a-SMA⁺, and TBX3⁺ CK19⁺ percent cells among the whole TMA. (E-F) Gradient analysis for multidimensional distances (0-25µm, 25-50µm, 50-100µm, 100-150µm) showed the spatial distribution of TBX3^- tumor cells (E) and TBX3⁺ tumor cells (F) between CD8+cells and a-SMA+cells. ***P < 0.001.

Figure 8

Relationship between TBX3 expression and immune checkpoint blockade (ICB) response. (A) Representative immunohistochemical (IHC) image of high TBX3 expression patient. Scale bar, 100 μm. (B) Representative CT image for patient with progressive disease after anti-PD-1 treatment. (C) Representative IHC image of low TBX3 expression patient. Scale bar, 100 μm. (D) Representative CT image for patient with complete response after anti-PD-1 treatment. (E) Relative percentage of patients with clinical response to immunotherapy between different TBX3 expression groups in Xiangya immune cohort. Red, immunotherapy response group; Blue, non-response group. (F) Disease-free survival (DFS) of patients with different TBX3 IHC scores in Xiangya immune cohort. (G) Expression of TBX3 on desert, excluded, and inflamed immune phenotypes in IMvigor210 cohort. (H-I) Expression of TBX3 on bladder cancer with different PD-L1 expression on immune cells (H) and tumor cells (I) in IMvigor210 cohort. (J) Correlation between TBX3 expression values and immunotherapy response in the desert phenotype of IMvigor210 cohort. Different color represents different response type. CR: complete response; PR: partial response; SD: stable disease; PD: progressive disease. (K-M) Relative percentage of patients with clinical response to immunotherapy between different TBX3 expression groups in GSE135222, GSE173839, PMID26359337 cohorts. Red, immunotherapy response group; Blue, non-response group. ns, not statistically significant. *p < 0.05; **p < 0.01; ***p < 0.001.