Genomic landscape of platinum resistant and sensitive testicular cancers

Abstract

While most testicular germ cell tumours (TGCTs) exhibit exquisite sensitivity to platinum chemotherapy, ~10% are platinum resistant. To gain insight into the underlying mechanisms, we undertake whole exome sequencing and copy number analysis in 40 tumours from 26 cases with platinum-resistant TGCT, and combine this with published genomic data on an additional 624 TGCTs. We integrate analyses for driver mutations, mutational burden, global, arm-level and focal copy number (CN) events, and SNV and CN signatures. Albeit preliminary and observational in nature, these analyses provide support for a possible mechanistic link between early driver mutations in RAS and KIT and the widespread copy number events by which TGCT is characterised.

Fig. 1. Study overview.

We combined data from six tumour sequencing studies of TGCT, performing small variant, copy number and tumour evolution analyses to define the molecular hallmarks of TGCT and their response to platinum-based therapies. RES: resistant, SEN: sensitive, UNS: unselected, PRI: primary, MET: metastasis, WES: whole exome sequencing, PNL: gene panel sequencing, MIP: molecular inversion probe, ARY: genotyping array, TMB: tumour mutation burden.

Fig. 2. Tumour mutation burden in TGCT.

Box plot showing that tumour mutation burden (TMB) is higher in platinum-resistant tumours vs. platinum-sensitive/unselected tumours (a, b) and metastatic tumours vs. primary tumours (c, d). Analyses were performed on 269 tumours with exome sequencing data (from the ICR1, ICR2, DFCI, and TCGA sample series) after excluding those samples with missing histology and/or missing tumour purity estimate (n = 20), and one sample with an excessively high TMB (DFCI_C13_T1) (n = 269, a, c). Analyses were also restricted to a subset of tumours comprising either only primary tumours (n = 239, b) or tumours from cases with platinum resistant disease (n = 64, d). Boxes show the median ± 25–75th percentiles, whiskers show 1.5× interquartile range below and above the 25th and 75th percentiles, respectively. p values are derived via two-sided multiple linear regression models adjusting for sample stage, histology, platinum response, dataset, and tumour purity.

Fig. 3. Mutational landscape of cancer driver genes in TGCT.

a Oncoplot showing the cancer driver genes most frequently mutated across 631 index TGCTs from the full sample series (ICR1, ICR2, DFCI, TCGA, MSK and FDM). Only TGCTs carrying a mutation in one of the displayed genes are shown (n = 207). b Oncogenic mutation frequencies in individual genes/gene sets across index TGCTs with the relevant available data, comparing: platinum-sensitive/unselected (n = 284, orange) vs. platinum-resistant (n = 244, red) tumours; primary (n = 435, dark blue) vs. metastatic (n = 196, light blue) tumours; seminomas (n = 204, yellow) vs. nonseminomas (n = 417, purple); and testicular (n = 584, black) vs. extragonadal (n = 47, green) primary sample site. Asterisks denote p values derived from two-sided multivariable logistic regression, adjusting for platinum response, histology, stage, capture type, sample material and primary sample site.*p < 0.05, **p < 0.01, ***p < 0.001. Exact p values are provided in Supplementary Table 4. Putative TGCT genes: CBL, RAC1, PIK3CA, KMT2C, CREBBP, BCOR and CTNNB1. c WNT/CTNNB1 pathway alterations are overrepresented in platinum-resistant metastatic TGCTs (n = 22) with the majority consistent with pathway activation. Pathway components that negatively regulate CTNNB1 are shown in green.

Fig. 4. Autosomal arm level aneuploidy frequencies.

The frequency of whole arm autosomal aneuploidy events in 183 TGCTs with available copy number data and tumour purity ≥ 0.4 from series ICR1, ICR2, DFCI and TCGA, comparing platinum-sensitive/unselected (orange, n = 134) vs. platinum-resistant (red, n = 29) primary TGCT, and comparing primary (dark blue, n = 29) vs metastatic (light blue, n = 20) platinum resistant TGCT (only a single representative metastatic tumour from each patient with platinum resistant TGCT was included-see Supplementary Data 7). Asterisks denote p values derived from two-sided multivariable logistic regression: *p < 0.05, **p < 0.01, Exact p values are provided in Supplementary Data 8.

Fig. 5. Absolute copy number of driver genes.

Plot showing the absolute copy number in 183 TGCTs with available copy number data and tumour purity ≥0.4 (ICR1, ICR2, DFCI and TCGA sample series) for fourteen genes including the putative/known TGCT driver genes (KIT, KRAS, NRAS, RAC1, CBL, PIK3CA, CREBBP, KMT2C, CTNNB1), TP53 and PTEN, and three genes previously implicated as focally amplified in TGCT, namely FSIP2, MDM2 and MYCN^,,. Genes are ordered vertically from most to least frequently demonstrating copy gain.

Fig. 6. COSMIC mutational signatures.

COSMIC SNV mutation signatures identified in TGCT, organized by histology and platinum response. Analyses were performed on 247 primary tumours with exome sequencing data (ICR1, ICR2, DFCI and TCGA series) after excluding nine primary tumours with missing histology. Signatures were derived from somatic SNVs from primary tumours pooled together within each group as indicated. COSMIC signatures 1, 3 and 19 are displayed separately, whilst other represents the contribution from the remaining signatures combined.

Fig. 7. TGCT copy number signatures.

Five copy number signatures (CN-Sig-1, CN-Sig-3, CN-Sig-4, CN-Sig-5, CN-Sig-6) identified using TCGA SNP array data (n = 105). a Defining features of the CN signatures, showing each feature (SegSize, osCN, ChangePoint, BP10, CopyNumber and BPChr) split into 36 constituent components, as defined in ref. . The mean value for each component is shown on the x axis, with component weights shown on the y axis. Features are defined as follows: SegSize, segment size (Mb); ocCN, region length with neighbouring oscillating copy number segments (Mb); ChangePoint, difference in copy number between neighbouring segments; BP10, number of break points (10 Mb⁻¹), CopyNumber, absolute copy number of segment; BPChr, breakpoints per chromosome arm. b CN signature exposure correlations with clinical and molecular features. Two-sided Pearson’s correlations for association of q < 0.05 are illustrated. Age, aneuploidy and chromosome 12p copies: analysed across histologies. RAS pathway-related features: analysed in seminomas only. c Box plots showing CN signature exposure comparisons for KIT (n = 12) and RAS (n = 7, NRAS/KRAS) mutated seminomas, wildtype seminomas (n = 18) and nonseminomas (n = 66). Boxes show the median ± 25–75th percentiles, whiskers show 1.5× interquartile range below and above the 25th and 75th percentiles, respectively. p values are derived from two-sided Wilcoxon Rank Sum tests adjusted for multiple comparisons via the Benjamini-Hochberg method.