APOBEC family reshapes the immune microenvironment and therapy sensitivity in clear cell renal cell carcinoma

Abstract

Emerging evidence suggests that the APOBEC family is implicated in multiple cancers and might be utilized as a new target for cancer detection and treatment. However, the dysregulation and clinical implication of the APOBEC family in clear cell renal cell cancer (ccRCC) remain elusive. TCGA multiomics data facilitated a comprehensive exploration of the APOBEC family across cancers, including ccRCC. Remodeling analysis classified ccRCC patients into two distinct subgroups: APOBEC family pattern cancer subtype 1 (APCS1) and subtype 2 (APCS2). The study investigated differences in clinical parameters, tumor immune microenvironment, therapeutic responsiveness, and genomic mutation landscapes between these subtypes. An APOBEC family-related risk model was developed and validated for predicting ccRCC patient prognosis, demonstrating good sensitivity and specificity. Finally, the overview of APOBEC3B function was investigated in multiple cancers and verified in clinical samples. APCS1 and APCS2 demonstrated considerably distinct clinical features and biological processes in ccRCC. APCS1, an aggressive subtype, has advanced clinical stage and a poor prognosis. APCS1 exhibited an oncogenic and metabolically active phenotype. APCS1 also exhibited a greater tumor mutation load and immunocompromised condition, resulting in immunological dysfunction and immune checkpoint treatment resistance. The genomic copy number variation of APCS1, including arm gain and loss, was much more than that of APCS2, which may help explain the tired immune system. Furthermore, the two subtypes have distinct drug sensitivity patterns in clinical specimens and matching cell lines. Finally, we developed a predictive risk model based on subtype biomarkers that performed well for ccRCC patients and validated the clinical impact of APOBEC3B. Aberrant APOBEC family expression patterns might modify the tumor immune microenvironment by increasing the genome mutation frequency, thus inducing an immune-exhausted phenotype. APOBEC family-based molecular subtypes could strengthen the understanding of ccRCC characterization and guide clinical treatment. Targeting APOBEC3B may be regarded as a new therapeutic target for ccRCC.

Keywords: APOBEC family; Clear cell renal cell carcinoma; Machine learning; Molecular subtypes; Multiomics; Tumor immunity.

Fig. 1

Dysregulation and genome alteration of the APOBEC family in cancers A Impact of the APOBEC family on cancer patient survival. B Difference expression level of the APOBEC family between tumor and normal tissues. C Impact of Correlation of DNA methylation and APOBEC expression level. D–E The CNV frequency and correlation of CNV with gene expression of the APOBEC family. F The genome locations of the APOBEC family on human chromosomes

Fig. 2

Screening of two subtypes of the APOBEC family in ccRCC A Consensus matrix based on APOBEC family member of TCGA-KIRC when k reached 2. B Principal component plot based on APOBEC family members. C KM curve of OS and PFI. D The expression profiles of the APOBEC family member between APCS1 and APCS2

Fig. 3

Functional enrichment analysis of DEGs from APCS1 and APCS2 A Volcano plot indicating all DEGs between APCS1 and APCS2. B–D Biological process, GSEA and GSVA analysis of DEGs. E Different transcription factor regulon activity between APCS1 and APCS2

Fig. 4

Investigation of the immune landscape A Correlation of the APOBEC family and immune cell infiltration. B–C Heatmap of immune-related signatures and immune cell infiltration between APCS1 and APCS2

Fig. 5

Immune signatures and ICI response difference A–B GSVA analysis indicating the different tumor-related immune signature enrichment scores between APCS1 and APCS2. C ICI response difference between APCS1 and APCS2

Fig. 6

Immune score and inner heterogeneity between subgroups A–C ESTIMATE scores, immune checkpoint signature expression level and TMB between APCS1 and APCS2. D Different level of immune antigens, cells and antitumor related pathway between APCS1 and APCS2

Fig. 7

Profiles of mutation landscape between APCS1 and APCS2 A Waterfall plot showing the top frequent mutation signatures between APCS1 and APCS2. B Potential druggable gene categories for APCS1 and APCS2 based on maftools. C The fraction of pathways of oncogenic pathways in different subgroups

Fig. 8

Copy number variations between APCS1 and APCS2 subtypes A–B GISTIC software was used to visualize the CNVs. C CNVs region for APCS1 and APCS2. D APCS1 and APCS2 subtypes copy number variations compared

Fig. 9

Drug sensitivity analysis of subtypes A Estimated IC50 of the clinically used chemotherapy targets between APCS1. B–C Estimated IC50 of the potential agents for APCS1 and APCS2

Fig. 10

Establishment and verification of the subtype biomarker-based risk model A Volcano plot indicating the univariate Cox coefficient of filtered biomarkers. B Random survival forest analysis identifying the most important signatures. C Optimal combination of signatures. D Risk score distribution in TCGA- and JAPAN-KIRC cohorts. E–F Survival analysis of OS and PFI in the TCGA-ccRCC cohort. G Time-dependent ROC curves for different risk groups in the training and test datasets

Fig. 11

The role of APOBEC3B in ccRCC A Number of tree indicating the importance proportion of APOBEC family members. B Protein levels of APOBEC3B in CPTAC-KIRC, classified by tumor grade and histological pathological stage. C–E The association between APOBEC3B and the immune signature, MMR, and DNA methyltransferases at the pancancer level. (DNMT1: red, DNMT2: blue, DNMT3A: green, and DNMT3B: purple). ∗ P < 0:05, ∗  ∗ P < 0:01, and ∗  ∗  ∗ P < 0:001. F The correlation between APOBEC3B expression and clinicopathologic features. G The relationship between APOBEC3B and clinical characteristics in ccRCC