Integrative genomic study of Chinese clear cell renal cell carcinoma reveals features associated with thrombus

Abstract

Clear cell renal cell carcinoma (ccRCC) is a heterogeneous disease with features that vary by ethnicity. A systematic characterization of the genomic landscape of Chinese ccRCC is lacking, and features of ccRCC associated with tumor thrombus (ccRCC-TT) remain poorly understood. Here, we applied whole-exome sequencing on 110 normal-tumor pairs and 42 normal-tumor-thrombus triples, and transcriptome sequencing on 61 tumor-normal pairs and 30 primary-thrombus pairs from 152 Chinese patients with ccRCC. Our analysis reveals that a mutational signature associated with aristolochic acid (AA) exposure is widespread in Chinese ccRCC. Tumors from patients with ccRCC-TT show a higher mutational burden and genomic instability; in addition, mutations in BAP1 and SETD2 are highly enriched in patients with ccRCC-TT. Moreover, patients with/without TT show distinct molecular characteristics. We reported the integrative genomic sequencing of Chinese ccRCC and identified the features associated with tumor thrombus, which may facilitate ccRCC diagnosis, prognosis and treatment.

Fig. 1. Landscape of mutations and copy number alterations of Chinese ccRCC.

a Alteration landscape of 152 Chinese ccRCC primary tumors. Top histogram, the number of silent and non-slient mutations in each sample. Upper heat map, gender, thrombus, and tumor stage information. Middle heat map, distribution of ccRCC-associated cancer genes and top 15 genes across samples, with genes ranked by mutation frequency. Bottom heat map, copy number gains (red) and losses (blue), with potential driver genes encompassed by the cytobands shown on the left. SMG genes are marked by a red asterisk. b The nonsynonymous mutational burdens of the Chinese and TCGA cohorts were compared. The box plot displays the first and third quartiles (top and bottom of the boxes), the median (band inside the boxes), and the lowest and highest point within 1.5 times the interquartile range of the lower and higher quartile (whiskers). Wilcoxon rank-sum test, *p < 0.05, ** p < 0.01, ***p < 0.001. c The mutation frequencies of genes in the Chinese cohort and TCGA cohort. The gene list is derived from a. Genes with significantly different mutation rates between two cohorts are marked by black asterisks. Fisher’s exact test, *p < 0.05, **p < 0.01, ***p < 0.001. The source data underlying Fig.  1a–c are provided as a Source Data file.

Fig. 2. Inferred mutational signatures and their contributions in Chinese ccRCC patients.

a ‘Lego’ plots display the frequency of 96 subtypes of base substitutions in Chinese ccRCC (left) and TCGA ccRCC (right). b Identifying the number of processes operating in a set of 152 ccRCC samples based on the reproducibility of their signatures and average Frobenius reconstruction error. c Three mutational signatures deciphered from the base substitutions identified in 152 ccRCC genomes. d The mutational burden was associated with the AA signature, and patients in the AA signature group had a heavier mutation load. The box plot displays the first and third quartiles (top and bottom of the boxes), the median (band inside the boxes), and the lowest and highest point within 1.5 times the interquartile range of the lower and higher quartile (whiskers). Wilcoxon rank-sum test, *p < 0.05, **p < 0.01, ***p < 0.001. e The somatic mutation load was positively associated with the contribution of the AA signature (Pearson’s correlation coefficient, two-tailed t test). f Contributions of each mutational signature per sample. The upper heat map shows sample gender information, tumor stage, mSigAct results, and the mutation landscape of 11 genes. The source data underlying Figs. 2c–f are provided as a Source Data file.

Fig. 3. Diverse mutation patterns between Chinese ccRCC and ccRCC-TT cohorts.

a The mutation load was compared between Chinese ccRCC and ccRCC-TT cohorts. The box plot displays the first and third quartiles (top and bottom of the boxes), the median (band inside the boxes), and the lowest and highest point within 1.5 times the interquartile range of the lower and higher quartile (whiskers). Wilcoxon rank-sum test, *p < 0.05, **p < 0.01, ***p < 0.001. b Nine significantly different mutated genes between the ccRCC and ccRCC-TT cohorts. Fisher’s exact test, *p < 0.05, **p < 0.01, ***p < 0.001. c Distribution of BAP1 and SETD2 mutations in ccRCC samples, ccRCC samples in the late stage and ccRCC-TT samples. BAP1 and SETD2 mutations were enriched in patients with TT. d Somatic mutations in signaling pathways across three cohorts. Non-silent mutations and indels were counted. The table shows the fraction of samples with alterations in each of the selected signaling pathways. In the pathway chart, the edges show pairwise molecular interactions, whereas boxes outlined in red denote alterations leading to pathway activation, whereas boxes outlined in blue indicate inactivation. The source data underlying Fig.  3a–c are provided as a Source Data file.

Fig. 4. Genomic differences between primary tumors and thrombi.

a Anatomical diagram of patients with thrombus. b The top bar plots display the numbers of shared or specific non-silent mutations between primary tumors and thrombi from 42 ccRCC-TT patients. The inset plot shows the distribution of heterogeneity between the primary tumors and thrombi. The bottom bar plots show the proportions of shared or specific mutations. c The contributions of six substitution patterns in primary tumors and thrombi. d The top 15 mutated genes in primary tumors and thrombi. The box plot displays the first and third quartiles (top and bottom of the boxes), the median (band inside the boxes), and the lowest and highest point within 1.5 times the interquartile range of the lower and higher quartile (whiskers). e Volcano plot of differentially expressed genes between primary tumors and thrombi with a threshold fold-change of 2 and p < 0.01. f The GSEA results revealed that the set of genes expressed in the thrombi were enriched in the cell cycle pathway, immunological pathway, and change of extracellular matrix and structure pathway. The source data underlying Fig.  4b–d are provided as a Source Data file.

Fig. 5. Gene expression subtypes.

a Tumors were separated into four clusters by unsupervised analyses based on differentially expressed mRNA patterns (showing 540 representative genes). Top to bottom: AA signature, tumor stage information; normalized abundance heatmap for 98 mRNAs; profile of silhouette width calculated from the consensus membership heatmap, Wcm; covariates for recurrent copy number alteration regions, and mutations in BAP1, CSMD3, SETD2, MTOR, PTEN, TP53, ARID1A, PIK3CA, and PBRM1. Some important genes related to the cell cycle, angiogenesis, and the EMT are listed on the left. b Overall, the scores of gene sets associated with EMT, angiogenesis, and cell cycle process in patients with thrombus were increased compared with patients with no thrombus. The phenomenon of immune cell infiltration was complex; CAF infiltration was more obvious in patients with thrombus, but T-cell and B-cell infiltration was more common in non-thrombus patients. Each dot presents one sample. The box plot displays the first and third quartiles (top and bottom of the boxes), the median (band inside the boxes), and the lowest and highest point within 1.5 times the interquartile range of the lower and higher quartile (whiskers). Wilcoxon rank-sum test, *p < 0.05, **p < 0.01, ***p < 0.001. The source data underlying Fig.  5a are provided as a Source Data file.