Substantial inter-individual and limited intra-individual genomic diversity among tumors from men with metastatic prostate cancer

Authors

Akash Kumar

Ilsa Coleman

Colm Morrissey

Xiaotun Zhang

Lawrence D. True

Roman Gulati

Ruth Etzioni

Hamid Bolouri

Bruce Montgomery

Thomas White

Jared M. Lucas

Lisha G. Brown

Ruth F. Dumpit

Navonil DeSarkar

Celestia Higano

Evan Y. Yu

Roger Coleman

Nikolaus Schultz

Min Fang

Paul H. Lange

Jay Shendure

Robert L. Vessella

Peter S. Nelson

Doi

10.1038/nm.4053

PMID: 26928463 · DOI: 10.1038/nm.4053 · Journal: Nature Medicine (2016)

TL;DR

Kumar et al. performed whole-exome sequencing, array CGH, and transcript profiling on 176 tumors from 63 men with metastatic castration-resistant prostate cancer (mCRPC) collected at rapid autopsy (prad_fhcrc). They found substantial inter-individual genomic diversity but limited intra-individual diversity — multiple metastases within the same patient were highly concordant for major oncogenic drivers (AR, TP53, RB1, ERG), CNA burden, mutation count, and AR/cell-cycle activity signatures. They also showed that elevated Fanconi anemia (FA) gene expression tracks with high cell-cycle progression, and that men with somatic FA-pathway or ATM aberrations had significantly longer responses to carboplatin chemotherapy.

Cohort & data

  • 63 men with metastatic CRPC enrolled in the University of Washington Prostate Cancer Donor Program; tissue collected at rapid autopsy (within 6 hours of death).
  • 176 primary or metastatic tumors total: WES on 141 tumors from 56 men; array CGH on 149 tumors from 60 men; expression microarray on 171 tumors from 63 men.
  • Histology: 156 adenocarcinomas; 20 small-cell neuroendocrine tumors (from 2 men).
  • Cancer type: prostate adenocarcinoma (PRAD).
  • Dataset: prad_fhcrc (cBioPortal prad_fhcrc); molecular profiling deposited at GEO accession GSE74685.
  • All men received androgen deprivation therapy (ADT); most also received an additional AR pathway-targeted agent and at least one systemic chemotherapy, most commonly docetaxel.
  • Assays: whole-exome-seq (Nimblegen V2/V3 capture, Illumina HiSeq 2000, 50/100 bp PE; bwa v0.7.1, GATK indel realignment, MuTect calling, MutSigCV), array-cgh-agilent-1m (Agilent 2x400K SurePrint G3 CGH; CBS segmentation; GISTIC 2.0), microarray-gene-expression (Agilent 44K), fish for TMPRSS2-ERG, and immunohistochemistry for AR, GR, ER alpha/beta, chromogranin A, and Ki-67.

Key findings

  • Recurrent driver landscape. Aberrations recurred in AR, ERG, TP53, RB1, SPOP, CHD1, ZBTB16/PLZF, plus 8q gain (including MYC) and 8p loss. Recurrent mutations were also identified in FOXA2, enriched in tumors with neuroendocrine features.
  • Hypermutated subset. Five men had hypermutated genomes with complex structural aberrations in MSH2 and MSH6; all metastases and matched primaries from these men were hypermutated, suggesting MMR deficiency arose early.
  • Mutation/CNA burden. Excluding hypermutators, tumors averaged 44 protein-coding mutations and a CNA burden of 38% of the genome; tumors from the same man had similar burdens.
  • AR aberrations. AR amplification or mutation occurred in 63% of men (vs very rare in untreated primary PC; Fisher’s exact P < 0.0001). 88% of patients had tumors with robust AR transcriptional activity despite prior AR-suppressive therapy.
  • Intra-individual concordance. AR activity and AR expression were highly concordant within individuals; AR genomic status was concordant in 32 of 41 informative men (8 all-negative, 24 all gain/mutation, 9 discordant). One man showed convergent evolution with separate AR-mutation and AR-amplification metastases.
  • TMPRSS2-ERG concordance. FISH on 53 tumors from 13 men showed 100% intra-individual concordance for TMPRSS2-ERG status; array-CGH-based assessment across 34 men showed 32 of 34 (94%) fully concordant.
  • Diversity index. Median proportion of CNA aberrations differing between tumors from the same individual was 6.7% vs 22.1% between different individuals (median inter-vs-intra difference 15.7%, 95% CI 15.5–15.9%, P < 0.001, two-sided 2-sample Wilcoxon rank sum).
  • Clustering. Unsupervised CNA clustering grouped tumors by patient with near 100% accuracy; expression clustering also grouped by patient except for bone metastases.
  • AR vs proliferation (inverse). AR activity was inversely correlated with cell cycle progression (CCP) score (r = -0.33, P < 0.001) and with E2F1 expression (P < 0.001); replicated in an independent mCRPC dataset. CCP score correlated with Ki-67 IHC (r = 0.48, P < 0.005).
  • AR-overexpression growth experiment. LNCaP-AR (overexpressing AR) proliferated more than LNCaP-WT at castrate ligand levels (10^-11 M R1881, P = 0.002), but at non-castrate levels (10^-9 M) AR overexpression suppressed growth; AR re-introduction into AR-null PC3 cells inhibited growth at 10^-9 M R1881 (P = 0.02).
  • FA pathway and cell cycle. High CCP tumors showed elevated expression of FA complex genes (FANCA, FANCI, FANCD2, BRCA1, BRCA2). RB1 loss associated with elevated CCP (P < 0.01), with E2F1 expression, and with a 15-gene FA pathway signature (r = 0.78, P < 0.001).
  • Functional FA validation. siRNA knockdown of individual FA genes reduced proliferation in LNCaP, 22Rv1, VCaP, and PC3 cells, and increased gamma-H2AX (DNA double-strand-break marker) following carboplatin exposure in LNCaP cells.
  • Carboplatin response biomarker. Of 20 men in the cohort treated with carboplatin, those with a somatic DNA-repair pathway aberration (homozygous deleterious event in any FA-pathway gene or heterozygous inactivating event in ATM) had significantly longer time on carboplatin treatment (Kaplan-Meier log-rank P = 0.02).

Genes & alterations

  • AR — amplification or mutation in 63% of men; 88% of patients had AR-active tumors. Outliers had AR activity without AR expression, with NR3C1 (GR) or ESR1 potentially substituting; PGR also discussed as a candidate alternative activator.
  • ERG / TMPRSS2 — TMPRSS2-ERG fusion status 100% concordant across metastases by FISH (53 tumors, 13 men); 94% concordant by array CGH (32/34 men).
  • TP53, SPOP, CHD1, ZBTB16 — recurrent mCRPC driver mutations consistent with prior reports.
  • MYC — recurrent 8q copy gain.
  • FOXA2 — recurrent mutations enriched in neuroendocrine-feature tumors.
  • MSH2, MSH6 — complex structural aberrations underlying hypermutation in 5 men; MMR deficiency apparent in matched primaries.
  • RB1 — loss associated with elevated CCP, elevated E2F1, and a 15-gene FA signature.
  • E2F1 — high expression correlates with CCP (r = 0.8, P < 0.001) and FA pathway activity (r = 0.78, P < 0.001).
  • FA pathway — FANCA, FANCB, FANCC, FANCD2, FANCE, FANCF, FANCG, FANCI, FANCL, FANCM, BRIP1, PALB2, SLX4, BRCA1, BRCA2 — homozygous deleterious events used as a DNA-repair-defect classifier for the carboplatin analysis; siRNA knockdown reduced proliferation in PC cell lines.
  • ATM — heterozygous inactivating events (or germline mutation) included with FA-gene events in the DNA-repair-defect group with longer carboplatin response.
  • PTEN — homozygous CN loss assessed for site-distribution analysis (no metastatic-site association detected).
  • CSMD1, NCOR2 — only two genes (out of 51 with metastasis-private NSVs in patient 99-091) altered at >5% frequency in prad_tcga or prad_su2c_2015, arguing most metastasis-private mutations are not drivers.

Clinical implications

  • Single-biopsy adequacy. Because major oncogenic drivers and AR/cell-cycle activity are largely shared across metastases within a patient, biopsy of a single metastatic site is a reasonable basis for clinical decision-making in mCRPC. The authors caution that this conclusion does not extend to primary tumors (more diverse) and is tempered by the small number of men in the cohort treated with potent AR-pathway antagonists (abiraterone or enzalutamide, n = 3) that could induce divergent resistance mechanisms (AR amplification, AR ligand-binding-domain mutation, AR splice variants).
  • DNA-repair defect as a carboplatin biomarker. Men with somatic homozygous loss of any FA-pathway gene or hetero-/homozygous loss/mutation of ATM had significantly longer time on carboplatin (P = 0.02), supporting platinum-based therapy in DNA-repair-deficient mCRPC and dovetailing with the contemporaneous olaparib PARP-inhibitor responses in the same biomarker subset.
  • Targeting the FA pathway. FA-gene over-expression in highly proliferative tumors (linked to RB1 loss / E2F1) plus the increased gamma-H2AX after FA knockdown + carboplatin nominate FA components as candidate therapeutic targets in proliferative mCRPC.
  • AR-low phenotypes. Tumors with AR transcriptional output but absent AR protein, accompanied by NR3C1 (GR) or ESR1 expression, suggest that GR or ER alpha may sustain AR-pathway output and could be exploited for combination targeting.

Limitations & open questions

  • Only 3 of 63 men received abiraterone or enzalutamide; inter-metastasis heterogeneity arising from these potent AR antagonists is therefore not assessed.
  • Sampling was at a single timepoint (rapid autopsy); the analysis cannot directly resolve polyclonal composition within metastases or temporal clonal dynamics.
  • Bone metastases clustered separately on expression profiling, which the authors attribute to microenvironment, lower tumor-cell content, or processing artifacts — potentially limiting comparability of bone vs visceral expression measurements.
  • Conclusions on intra-individual concordance assume that low-frequency private mutations are non-driver; some functionally important rare events may be missed.
  • The FA-pathway / carboplatin response analysis is based on 20 carboplatin-treated men and uses time-on-drug as a surrogate for response duration; prospective validation is needed before clinical biomarker use.
  • The CCP-FA-RB1-E2F1 axis is correlative in patient samples; mechanistic causality between RB1 loss, E2F1 activation, and FA pathway upregulation is not directly tested in vivo.

Citations from this paper used in the wiki

  • “We resected 176 primary or metastatic tumors from 63 men with mCPRC at the time of rapid autopsy and determined mutations by WES (n = 141 tumors from 56 men), copy number aberrations by array CGH (n = 149 tumors from 60 men), and transcript expression by microarray hybridization (n = 171 tumors from 63 men).” (Results, p.3)
  • “Amplification or mutation of the AR gene occurred in 63% of men … 88% of patients had tumors with robust AR activity, despite extensive prior treatment to suppress AR function.” (Results, p.4)
  • “The median inter- versus intra-individual tumor difference was 15.7% (95% CI 15.5-15.9%; P < 0.001; 2-sided 2-sample Wilcoxon rank sum test).” (Results, p.5)
  • “AR activity was inversely associated with CCP activity (r = -0.33; P < 0.001) … AR activity was also inversely associated with E2F1, a key regulator of cell proliferation (P < 0.001).” (Results, p.6)
  • “Of the men comprising this rapid autopsy cohort, 20 received carboplatin chemotherapy during their clinical course. Those with a somatic aberration in a DNA repair pathway gene … had significantly greater durations of response to carboplatin chemotherapy as determined by time on drug treatment (P = 0.02).” (Results, p.7)
  • “Patients with a DNA repair defect were defined as those with any combination of the following criteria: homozygous or heterozygous loss in ATM by copy number loss and/or mutation, homozygous loss of one of the following 15 Fanconi anemia associated genes (FANCA, FANCB, FANCC, FANCD2, FANCE, FANCF, FANCG, FANCI, BRIP1, FANCL, FANCM, PALB2, SLX4, BRCA1) or germline mutation in BRCA2 and/or ATM.” (Methods, p.11)
  • “The data are available for visualization and analysis in the cBioPortal for Cancer Genomics at: http://www.cbioportal.org/study.do?cancer_study_id=prad_fhcrc” (Methods, p.11)

This page was processed by crosslinker on 2026-05-14.