Characterization of Genetic Heterogeneity in Recurrent Metastases of Renal Cell Carcinoma

Abstract

Metastatic renal cell carcinoma (RCC) exhibits poor prognosis. Better knowledge of distant metastases is crucial to foster personalized treatment strategies. Here, we aimed to investigate the genetic landscape of metastases, including synchronous and/or recurrent metastases to elucidate potential drug target genes and clinically relevant mutations in a real-world setting of patients. We assessed 81 metastases from 56 RCC patients, including synchronous and/or recurrent metastases of 19 patients. Samples were analysed through next-generation sequencing with a high coverage (~1000× mean coverage). We therefore established a novel sequencing panel comprising 32 genes with impact on RCC development. We observed a high frequency of mutations in known RCC driver genes (e.g., >40% carriers of VHL and PBRM1 mutations) in metastases irrespective of the metastatic site. The somatic mutational composition was significantly associated with cancer-specific survival (p(logrank) = 0.03). Moreover, we identified in 34 patients at least one drug target gene as well as clinically relevant mutations listed in the VICC Meta-Knowledgebase in 7%. In addition to significantly higher mutational burden in recurrent metastases compared to earlier ones, synchronous and/or recurrent metastases of individual patients, even after a time-period >2 yrs, shared a high proportion of somatic events. Our data demonstrate the importance of somatic profiling in metastases for precision medicine in RCC.

Keywords: metastasis; next-generation sequencing; personalized therapy; pharmacogenomics; renal cell carcinoma.

Figure 1

Somatic variants in metastasis samples of primary RCC analysed through NGS panel approach. Two metastasis samples were excluded from variant analysis due to a hypermutated genetic landscape, resulting in a final cohort of 79 samples from 55 patients. Frequency distribution, including information about mutation types in selected panel genes and patient information (cases with multiple metastases, therapy, patient’s age, sex, subtype of primary tumour, BMI (kg/m²), site of metastases, and survival), is shown.

Figure 2

Somatic variants of RCC metastases in different metastatic sites. (A): Treatment timeline indicating course of therapy and patient’s survival. Time of surgical resection of metastasis is marked by asterisks. (B): Heatmap showing low (yellow) and high (red) mutational burden for each gene and organ for metastasis (n = 68) of cases without prior systemic therapy. Mean mutational load is displayed. (C): Association of the site of metastases with patient’s survival (p(logrank) = 0.00369) (upper panel). Time of occurrence of metastases in different organs after surgery of primary tumours is shown in the lower panel. (D): Kaplan–Meier plot showing the association of the somatic mutational composition and cancer-specific survival in our cohort (p(logrank) = 0.0334). Metastases (n = 32) harbouring multiple somatic drivers and VHL wildtype alleles were compared to PBRM1, SETD2, and VHL monodrivers (n = 47). (E): Mapping of somatic mutational events in our cohort to drug target information (TARGET drug recommendation, https://software.broadinstitute.org/cancer/cga/target, accessed on 15 March 2021) and to data on clinical significance using the VICC Meta-Knowledgebase (MetaKB). Pie plots indicate number of cases with recommendations in either of the databases.

Figure 3

Somatic variants in recurrent metastases over time. (A): Mutational burden in synchronously resected and/or recurrent metastases. Cases are grouped according to timespan between metastases resections (0 months: synchronous resection of metastases; <6 months, 6 months–2 yrs, >2 yrs: timespan between resection of different metastases from the same patient). Mutational burden in recurrent metastases increased significantly over time (analysed by Wilcoxon signed-rank test; p(Wilcoxon) = 0.023) between the first metastases and the later ones. (B): Pie plot indicates the number of shared variants in metastases resected from the same patient.

Figure 4

Somatic mutations in synchronously resected and/or metachronous metastases of individual patients ((A): case 004, (B): case 001, (C): case 010, and (D): case 050). Functional annotation of somatic variants using SIFT and PolyPhen, as well as COSMIC, MetaKB, and TARGET annotation is displayed. Phylogenetic trees of each case were constructed using MesKit [25]. Branches are coloured according to the distribution of mutations in different metastases. Lengths of the branches are proportional to the number of mutations. Support values of internal nodes are annotated within trees.