ATM Mutations Benefit Bladder Cancer Patients Treated With Immune Checkpoint Inhibitors by Acting on the Tumor Immune Microenvironment

Abstract

Immune checkpoint inhibitors (ICIs) have shown promising results in bladder cancer (BC). However, only some patients respond to ICIs. DNA repair defects (DDR) play an important role in the therapeutic response of bladder cancer. Therefore, we aimed to elucidate the association between ICIs in bladder cancer and ataxia telangiectasia mutated (ATM), a core component of the DNA repair system. From a collected immunotherapy cohort (n = 210) and The Cancer Genome Atlas (TCGA)-Bladder cancer cohort, which were both retrieved from publicly available resources, we performed a series of analyses to evaluate the prognostic value and potential mechanism of ATM in bladder cancer immunotherapy. We found that ATM-mutant (ATM-MT) bladder cancer patients derived greater benefit from ICIs [overall survival (OS), hazard ratio (HR) = 0.28, [95% confidence interval (CI), 0.16 to 0.51], P = 0.007] and showed a higher mutation load (P < 0.05) and immunogenicity (P < 0.05) than ATM-wild-type (ATM-WT) patients. The immune inflammatory response to antigenic stimulation, the regulation of the IFN pathway and the macrophage activation pathway were significantly enriched in the ATM-MT group (NES > 1, P < 0.05), while insulin-like growth factor receptor signaling pathways and vasculogenesis were significantly downregulated (NES < -1, P < 0.05). ATM mutation significantly upregulated the number of DNA damage repair pathway gene mutations (P < 0.05). ATM mutations resulted in increased bladder cancer sensitivity to 29 drugs (P < 0.05), including cisplatin and BMS-536924, an IGF-1R inhibitor. Our results demonstrate the importance of ATM as a prognostic signature in bladder cancer and reveal that ATM may impact the effects of ICIs by acting on the tumor immune microenvironment.

Keywords: ATM; DDR; bladder cancer; immune checkpoint inhibitors; tumor microenvironment.

FIGURE 1

Association between the ATM status and clinical outcomes in bladder cancer. (A) Kaplan–Meier curves comparing the overall survival (OS) of patients with ATM-MT with that of patients with ATM-WT in the immunotherapy cohort [n = 210, hazard ratio (HR) = 0.28, [95% confidence interval (CI), 0.16 to 0.51], P = 0.007]. ATM-MT, ATM mutant type; ATM-WT, ATM wild type; BLCA, bladder cancer. (B) Kaplan–Meier curves comparing the OS of patients with ATM-MT with that of patients with ATM-WT in the TCGA-BLCA cohort (n = 410, HR = 0.76, [95% CI, 0.49 to 1.16], P = 0.251). (C) Kaplan–Meier curves comparing the progression-free survival (PFS) of patients with ATM-MT with that of patients with ATM-WT in the TCGA-BLCA cohort (n = 411, HR = 0.65, [95% CI, 0.43 to 1.00], P = 0.09). (D) Kaplan–Meier curves comparing the disease-free survival (DFS) of patients with ATM-MT with that of patients with ATM-WT in the TCGA-BLCA cohort (n = 188, HR = 0.64, [95% CI, 0.26 to 1.57], P = 0.401).

FIGURE 2

Relationship between the ATM status and clinical phenotype. (A) Top 20 significantly mutated genes in the immunotherapy cohort of bladder cancer patients. Samples were ordered based on the somatic non-synonymous mutation burden, and genes were ranked by mutation frequencies (left panel). The ATM status, sample type (metastasis/primary), sex, drug type, OS, OS time (months), TMB score and MSI score are annotated in order in the top panel. (B) Top 20 significantly mutated genes in the TCGA-BLCA cohort. ATM status, age, sex, stage, ethnicity, OS, OS time (months), TMB and neoantigen load are annotated in order in the top panel. (C) The somatic mutations in ATM were evenly distributed in both the immunotherapy cohort and the TCGA-BLCA cohort. (D) Maftools was used to visualize the copy number alteration (CNV) analysis based on GISTIC2.0 of the TCGA-BLCA cohort.

FIGURE 3

ATM mutation was associated with enhanced tumor immunogenicity and activated antitumor immunity. (A) Heatmap showing average changes in the expression levels of immune-related gene between the ATM-MT and ATM-WT patients in the TCGA-BLCA cohort. The genes corresponding to the same lymphocyte or function are identified by the same color on the left side of the squares, and each square with an exact number represents the logFC of a gene, filled with different back colors, i.e., from red to gray. The logFC values marked in black font indicate that the absolute value of logFC is ≥1 with statistical significance (P < 0.05), while the logFC values in white font are non-significant (P > 0.05). Using Mann–Whitney U test tested the differences in tumor-infiltrating lymphocytes and immune-related gene expression between ATM-MT and ATM-WT samples. (B) Heatmap showing average changes in the expression levels of immune-cell-related gene between the ATM-MT and ATM-WT patients in the TCGA-BLCA cohort. The genes corresponding to the different cell types are identified by the different colors on the left side of the squares, and each square with an exact number represents the logFC of a gene, filled with different back colors, i.e., from red to blue. The logFC values marked in black font indicate that the absolute value of logFC is ≥1 with statistical significance (P < 0.05). Using Mann–Whitney U test tested the differences in immune-cell-related gene expression between ATM-MT and ATM-WT samples. (C) Comparison of the tumor mutational burden between the ATM-MT and ATM-WT tumors in the immunotherapy cohort. (D) Comparison of the tumor mutational burden between the ATM-MT and ATM-WT tumors in the TCGA-BLCA cohort. (E) Comparison of the neoantigen load between the ATM-MT and ATM-WT tumors in the TCGA-BLCA cohort. (F) Comparison of the tumor mutational burden between the ATM-MT and ATM-WT tumors in the GDSC-Bladder cancer cell line dataset. (G) CIBERSORT analyses quantifying the proportion of 22 immune cells in the ATM-MT and ATM-WT tumors in the TCGA-BLCA cohort. (C–G), *p < 0.05; ****p < 0.0001.

FIGURE 4

GSEA of the functional enrichment pathways of the ATM-MT and ATM-WT groups in the TCGA-BLCA cohort and GDSC-Bladder cancer cell line dataset. (A) The inflammatory response in the antigen pathway was significantly enriched in the ATM-MT group. (B) Interferon-gamma production and regulation pathways were significantly enriched in the ATM-MT group. (C) Macrophage activation involved in the immune response pathway was significantly enriched in the ATM-MT group. (D) The insulin-like growth factor receptor signaling pathway was significantly downregulated in the ATM-MT group. (E) The insulin-like growth factor receptor signaling pathway was significantly downregulated in the ATM-MT group. (F) Angiogenesis-related pathways were significantly downregulated in the ATM-MT group.

FIGURE 5

Comparison of drug sensitivity between ATM-MT and ATM-WT in the GDSC-Bladder cancer cell line cohort. We observed that the sensitivity to 29 drugs were increased in the ATM-MT group (P < 0.05), including Alectinib, AZD7762, Bleomycin, Carmustine, CCT_018159, Cisplatin, CX_5461, Cytarabine, Genentech_Cpd_10, YK_4_279, GW_2580, IOX2, KIN001_270, KU_55933, Lenalidomide, Motesanib, OSI_930, Pevonedistat, PFI_3, RO_3306, rTRAIL, Selisistat, Sepantronium_bromide, SN_38, STF_62247, Talazoparlb, WHI_P97, XMD14_99, and XMD15_27. IC50s reported in GDSC were log-transformed. Using the Mann–Whitney U test tested the differences in the ln(IC50) values of different drugs between the ATM-MT and ATM-WT cell lines. *P < 0.05; **P < 0.01.

FIGURE 6

Association between ATM mutation and mutated DNA damage repair pathway gene number variation. Mutated DNA damage repair pathway gene number variation increased in the ATM-MT group in the immunotherapy, TCGA-BLCA and GDSC-Bladder cancer cell line cohorts. The x axis indicates eight DDR-related pathways. BER, base excision repair; HR, homologous recombination; MMR, mismatch repair; SSB, single-stranded DNA binding; DSB, double-strand break repair; NER, nucleotide excision repair; NHEJ, non-homologous end-joining; FA, Fanconi anemia. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001.

FIGURE 7

Mechanism of ATM mutation enhancing the efficacy of immune checkpoint inhibitors in bladder cancer based on a hypothetical model.