Characterization of the genomic landscape and actionable mutations in Chinese breast cancers by clinical sequencing

Abstract

The remarkable advances in next-generation sequencing technology have enabled the wide usage of sequencing as a clinical tool. To promote the advance of precision oncology for breast cancer in China, here we report a large-scale prospective clinical sequencing program using the Fudan-BC panel, and comprehensively analyze the clinical and genomic characteristics of Chinese breast cancer. The mutational landscape of 1,134 breast cancers reveals that the most significant differences between Chinese and Western patients occurred in the hormone receptor positive, human epidermal growth factor receptor 2 negative breast cancer subtype. Mutations in p53 and Hippo signaling pathways are more prevalent, and 2 mutually exclusive and 9 co-occurring patterns exist among 9 oncogenic pathways in our cohort. Further preclinical investigation partially suggests that NF2 loss-of-function mutations can be sensitive to a Hippo-targeted strategy. We establish a public database (Fudan Portal) and a precision medicine knowledge base for data exchange and interpretation. Collectively, our study presents a leading approach to Chinese precision oncology treatment and reveals potentially actionable mutations in breast cancer.

Fig. 1. Schematic of the study and sample distributions.

a Schematic of the study. Patients in cohort 3 were referred to genome-guided clinical trials when meeting criteria. b Purpose of the three different investigated cohorts. c Treatment information. ^†Clinical trial. ^‡Fudan Umbrella Trial. dd dose densed, EC epirubicin + cyclophosphamide, P paclitaxel, T docetaxel, NE vinorelbine + epirubicin, PE paclitaxel + epirubicin, CbP carboplatin, AI aromatase inhibitor. d Clinical features of our prospective cohort compared with those found in previous sequencing studies of breast cancers (MSKCC and TCGA). Source data for d are provided as a source data file.

Fig. 2. Genomic landscape and characteristics of prospectively sequenced Chinese breast cancer.

a Sequencing data of 1134 Chinese breast cancer samples classified by the molecular subtype and mutation profile and annotated with the variation type and mutation frequency. The mutation counts in each sample and each gene are provided above and on the right side, respectively. b Hotspot mutations (frequency higher than 2%) in Chinese breast cancer. c Copy number variations (CNVs) of 1114 Chinese breast cancer samples classified by the molecular subtype in our cohort. d Distribution of variant allele fractions (VAFs) in the recurrently mutated genes. e Recurrent genomic mutations (left and right) and their association with different molecular subtypes (middle). The asterisks indicate a statistically significant association with the subtype (FDR < 0.25).

Fig. 3. Population-specific genomic mutations in Chinese breast cancer compared with the MSKCC data.

a Scatter plots of the prevalence of mutated genes in primary breast cancer samples from the FUSCC (x-axis) and MSKCC (y-axis) datasets. b Scatter plots of the prevalence of mutated genes in primary breast cancer samples with the HR+/HER2− (top, left), HR+/HER2+ (top, right), HR−/HER2+ (bottom, left) and triple-negative (bottom, right) subtypes from the FUSCC (x-axis) and the MSKCC (y-axis) datasets. c Scatter plots of the prevalence of mutated genes in advanced breast cancer samples from the FUSCC (x-axis) and the MSKCC dataset (y-axis) datasets.

Fig. 4. Characteristics of mutations in oncogenic signaling pathways in prospectively sequenced Chinese breast cancer.

a Landscape of pathway mutations in 1134 Chinese breast cancer samples classified by the molecular subtype and annotated with the variation type. The mutation counts in each sample and each pathway are provided above and on the right side, respectively. b Comparison of mutations in oncogenic signaling pathways in primary samples between our cohort and the MSKCC dataset. The middle circular spot corresponds to a value for odds ratio and the lines represent 95% confidence intervals. A red or blue horizontal line represents the significant or non-significant result of the comparison of mutation frequencies between our cohort and the MSKCC’s cohort in each signaling pathway, respectively. The red line indicates a higher mutation frequency in the corresponding pathway favors in the MSKCC’s cohort, while the blue line indicates a higher mutation frequency in the corresponding pathway favors in our cohort. A total of 1025 primary breast cancer samples from FUSCC are compared with 869 primary breast cancer samples from MSKCC by different mutation status in oncogenic signaling pathways using Fisher’s exact test, adjusted by false discovery rate (FDR). The asterisks indicate FDR < 0.05. c Significant enrichment of pathway mutations in different molecular subtypes of breast cancer. d Significant mutual exclusivity (blue) and co-occurrence (red) of gene mutations in pathways in Chinese breast cancer (all samples, right; luminal B/HER2−, left). Spectrum bar: log10 (odds ratio (OR)); the color intensity represents the scale of the value. e Circus plot displaying the co-occurrent patterns among the oncogenic signaling pathways in our cohort. The line thickness corresponds to the number of mutations in two co-occurrent pathways. The significant co-occurrent patterns of mutations in the Hippo pathways are illustrated. Source data for b are provided as a source data file.

Fig. 5. Actionable and oncogenic alterations revealed by clinical sequencing.

a Fractions of alterations annotated based on their clinical actionability according to PGI in different molecular subtypes of breast cancer. b Distribution of breast cancer samples assigned with the level of the most significant alteration. c Fractions of samples with multiple oncogenic alterations annotated in different molecular subtypes of breast cancer. d Actionable alterations annotated in different molecular subtypes of breast cancer.

Fig. 6. NF2 mutations promote sensitivity to a YAP inhibitor.

a NF2 missense mutations discovered in Chinese breast cancer samples. The mutations reported in the COSMIC dataset were annotated next to the residue site. NF2 L75I, G240W, P257T/Q, and Q324K were identified as recurrent spots in Chinese breast cancer. b Western blot showing the activation of Hippo and cell-cycle pathway effectors in NF2-wild type and NF2-mutated breast cancer cells. c Hs578T cells stably expressing Flag-NF2-WT and Flag-NF2-G240W/Q324K were treated with an increasing dose of Verteporfin. d Hs578T cells stably expressing Flag-NF2-WT and Flag-NF2-G240W/Q324K were treated with 1 μM Verteporfin for 0, 6, 12, and 24 h. All western blots experiments are repeated three times. e–h Relative percentage of cell viability (%) of NF2-WT and NF2-mutated (G240W (e); Q324K (f); L75I (g); P257Q/T (h)) Hs578T cells treated with Verteporfin. The half-maximal inhibitory concentration (IC₅₀) values were calculated based on the day 3 data of various doses of drug treatment. The assays were performed with five replicates in three independent experiments; representative results are shown. These data represent the mean values, with error bars indicating the SEM (*P < 0.05). Source data for b–h are provided as a source data file.