Alterations in homologous recombination repair genes in prostate cancer brain metastases

Abstract

Improved survival rates for prostate cancer through more effective therapies have also led to an increase in the diagnosis of metastases to infrequent locations such as the brain. Here we investigate the repertoire of somatic genetic alterations present in brain metastases from 51 patients with prostate cancer brain metastases (PCBM). We highlight the clonal evolution occurring in PCBM and demonstrate an increased mutational burden, concomitant with an enrichment of the homologous recombination deficiency mutational signature in PCBM compared to non-brain metastases. Focusing on known pathogenic alterations within homologous recombination repair genes, we find 10 patients (19.6%) fulfilling the inclusion criteria used in the PROfound clinical trial, which assessed the efficacy of PARP inhibitors (PARPi) in homologous recombination deficient prostate cancer. Eight (15.7%) patients show biallelic loss of one of the 15 genes included in the trial, while 5 patients (9.8%) harbor pathogenic alterations in BRCA1/2 specifically. Uncovering these molecular features of PCBM may have therapeutic implications, suggesting the need of clinical trial enrollment of PCBM patients when evaluating potential benefit from PARPi.

Fig. 1. Summary of histologic and genetic alterations in prostate cancer brain metastases.

a Histology and immunohistochemistry results, b Summary of non-synonymous mutations/Mb. c Relative contribution of mutational signatures from COSMIC. d Summary of genes showing recurrent mutation or fusion (altered in >5 patients) from WES and targeted RNA-seq data. CNA are shown in large blocks, coding alterations in small blocks. Alteration types are colored according to the legend. Samples are grouped by patient and site (primary or metastasis). Source data are provided in Supplementary Tables 4 and 5.

Fig. 2. HRR genes alterations (PROfound genes) are highly represented and the HRD signature is enriched in PCBM.

a Comparison of HRD signature scores, estimated by SigMA for TCGA (n = 495), PCBM primary (n = 63), CRPC500 non-brain metastatic (n = 411), and PCBM metastatic (n = 105) samples (Wilcoxon test, two sided). Horizontal lines in boxplots show median, hinges show interquartile range, whiskers show 1.5 x interquartile range, points beyond 1.5 x IQR past hinge are shown. b Summary of mutations and copy number alterations affecting HR genes included in the PROfound clinical trial. Alterations (coding and copy number) are colored according to the legend. Ploidy annotation from FACETS and samples passing the HRD Signature ‘strict’ cutoff from SigMA are indicated. Source data are provided as a Source Data file.

Fig. 3. Clonal evolution in PCBM Clonal evolution in four patients from the PCBM cohort.

Patients P4 (a) and P14 (b) had dural metastases, and patients P5 (c) and P8 (d) had parenchymal metastases. Trace plots show cancer cell fraction (CCF) for each mutational cluster in each patient sample (P = primary, M = metastasis). Ribbons show 95% confidence interval, center of bands show mean cluster CCF estimate. Phylogenetic trees show best solution for evolutionary relationship between clones with different clusters of mutations where each node (numbered) is a cluster of mutations. Numbers on each branch show the number of mutations distinguishing a clone from the previous (all genes). Potential driver genes mutated in the distinction between a clone and the previous are indicated in colors corresponding to the branch. Solid branches show clusters of mutations which become clonal in metastatic samples. e Enrichment for canonical pathways associated with genes in mutational clusters, that expand from subclonal in primary samples to clonal in at least one metastatic sample from all 20 patients with primary and metastatic samples calculated using Ingenuity Pathway Analysis. f As for e but showing enrichment for gene networks associated with metastatic-clonal genes. g As for e but with metastatic-clonal genes in dural and parenchymal metastases analyzed separately. h As for g but examining gene sets defined by their upstream regulator. Enrichment was calculated using two-sided Fisher’s exact test. Plotted P-values are unadjusted. Source data are provided as a Source Data file.