Genetic Determinants of Cisplatin Resistance in Patients With Advanced Germ Cell Tumors

Authors

Aditya Bagrodia

Byron H. Lee

William Lee

Eugene K. Cha

John P. Sfakianos

Gopa Iyer

Eugene J. Pietzak

Sizhi Paul Gao

Emily C. Zabor

Irina Ostrovnaya

Samuel D. Kaffenberger

Aijazuddin Syed

Maria E. Arcila

Raju S. Chaganti

Ritika Kundra

Jana Eng

Joseph Hreiki

Vladimir Vacic

Kanika Arora

Dayna M. Oschwald

Michael F. Berger

Dean F. Bajorin

Manjit S. Bains

Nikolaus Schultz

Victor E. Reuter

Joel Sheinfeld

George J. Bosl

Hikmat A. Al-Ahmadie

David B. Solit

Darren R. Feldman

Doi

10.1200/JCO.2016.68.7798

PMID: 27646943 · DOI: 10.1200/JCO.2016.68.7798 · Journal: Journal of Clinical Oncology (2016)

TL;DR

Bagrodia et al. performed whole-exome sequencing on a discovery cohort of 19 advanced germ cell tumors (GCTs) and validated findings with MSK-IMPACT targeted sequencing on a prospective cohort of 161 additional patients (combined N=180), enriched for the cisplatin-resistant phenotype. TP53 alterations were found exclusively in cisplatin-resistant tumors (17 of 104 [16.3%] vs 0 of 76; P<.001), and combined TP53/MDM2 pathway alterations were significantly more frequent in resistant disease (24.0% vs 2.6%; P<.001). TP53/MDM2 alterations independently predicted shorter progression-free survival (hazard ratio 1.83; 95% CI, 1.12 to 2.98; P=.016) after adjusting for the IGCCCG risk model. Strikingly, 72% of primary mediastinal nonseminomas harbored TP53 alterations, providing the first genetic basis for the dismal prognosis of this subgroup. Actionable alterations were detected in 55% of cisplatin-resistant GCTs, including novel functionally validated RAC1 hotspot mutations.

Cohort & data

  • N=180 men with advanced germ cell tumor receiving first-line cisplatin-based chemotherapy at Memorial Sloan Kettering Cancer Center.
  • Discovery cohort: 19 tumors profiled by whole-exome sequencing (10 cisplatin resistant, 9 cisplatin sensitive); mean coverage 116× (range 93–134×); 90% of target bases covered at >30×.
  • Validation cohort: 161 prospective patients profiled by whole-exome-seq and the msk-impact-panel (IMPACT410)-based targeted exon-capture assay (>300 cancer-related genes; 500–1,000× depth).
  • Cancer types: GCT — 70% NSGCT (n=126), 30% SEM (n=54). Primary site: 87.2% testis (TT), 12.2% mediastinum, 0.6% retroperitoneum. 49 resistant samples were MGCT containing teratoma.
  • Cisplatin sensitivity: 76 sensitive vs 104 resistant. Resistance defined by incomplete response to first-line therapy, nonteratomatous progression, or viable nonteratomatous GCT at postchemotherapy surgery.
  • IGCCCG risk distribution (combined cohort): good 51.1%, intermediate 15.6%, poor 32.8%. Poor-risk patients were heavily enriched among resistant cases (49% vs 10.5%).
  • Dataset: gct_msk_2016 — all mutational and clinical data publicly available on cBioPortal.
  • First-line regimens: BEP (bleomycin+etoposide+cisplatin) 37.2%, EP 42.2%, TIP/VIP 20%, PVB 0.6%.

Key findings

  • Mutation burden differs by sensitivity: Median total mutations 46 vs 21 (P=.05) and nonsynonymous mutations 36 vs 14 (P=.04) in resistant vs sensitive discovery tumors. The mean MSK-IMPACT mutation rate of 0.9/Mb is very low compared with other adult solid tumors.
  • 12p gain was present in 74% of discovery tumors, consistent with the well-characterized GCT cytogenetic signature.
  • TP53 alterations exclusive to resistance: 17 of 104 (16.3%) cisplatin-resistant tumors harbored TP53 mutations or deletions vs 0 of 76 sensitive tumors (P<.001). Mutations included previously reported recurrent missense (e.g. V173M, R248Q/W) and truncating events, plus one homozygous deletion.
  • MDM2 amplifications: 8 of 104 (7.6%) resistant vs 2 of 76 (2.6%) sensitive (P=.195); mutually exclusive with TP53 alteration. 5 of 7 (71.4%) MDM2-amplified tumors were cisplatin resistant.
  • Combined TP53/MDM2: 25 of 104 (24.0%) resistant vs 2 of 76 (2.6%) sensitive (P<.001).
  • Mediastinal primary enrichment: TP53 alterations were 13 of 22 (59.1%) in mediastinal vs 4 of 158 (2.5%) in testicular primaries (P<.001). Among primary mediastinal nonseminomas, the rate was 13 of 18 (72.2%); none of the 4 primary mediastinal seminomas harbored TP53 mutations.
  • Independent prognostic value: In multivariable Cox regression including IGCCCG risk, TP53/MDM2 alteration independently predicted shorter PFS (HR 1.83; 95% CI 1.12–2.98; P=.016).
  • MYCN amplifications in 5 patients, all cisplatin resistant; 4 of 5 had wild-type TP53/MDM2.
  • RAC1 mutations in 9 patients (5% incidence — the highest reported across cancer types per TCGA at the time): G12V (n=3), G12R (n=2), P34R, Q61R, Q61K (n=2). The mutants were functionally validated in HEK293 cells, showing increased phospho-PAK1 and phospho-MEK1/2 — the first demonstration that these RAC1 alleles activate downstream Rho-family signaling in GCT.
  • KIT mutations in 19 of 180 patients (mostly exon 17 hotspots associated with imatinib resistance), enriched in SEM (29.6% vs 4% in nonseminoma; P<.001).
  • KRAS mutations in 22 of 180 patients (15 at G12); enriched in seminomas overall (20% vs 8.7%; P=.045) but in nonseminomas 8 of 11 KRAS mutations occurred in cisplatin-resistant tumors.
  • PI3K pathway alterations in 24 of 180 (13.3%) — including 4 PIK3CA E542K mutations, 5 loss-of-function PTEN mutations, AKT1 amplification, MTOR, TSC1, and TSC2 alterations.
  • Other notable mutations: APC (n=3), FAT1 (n=4), AXIN1 (n=2), EP300/SETD2/PTPRD (n=3 each), BRCA2 deletions, three BRAF mutations (D594N, D594G, G466E), one GNAQ Q209P, four KDR amplifications, one MET amplification.

Genes & alterations

  • TP53 — recurrent missense (V173M, R248Q/W) and truncating mutations; exclusive to cisplatin-resistant tumors. Strongest single-gene biomarker of resistance in this study.
  • MDM2 — focal amplifications, mutually exclusive with TP53 alteration; 71% of MDM2-amplified tumors are cisplatin resistant. Therapeutic target via Nutlin-3 / nutlin-3 and other MDM2 inhibitors.
  • MYCN — amplification in 5 patients (all cisplatin resistant); transcriptionally targets both TP53 and MDM2; predicts MDM2-inhibitor sensitivity by analogy to neuroblastoma.
  • RAC1 — novel hotspot mutations at codons 12, 34, 61 (G12V/R, P34R, Q61R/K); functionally validated to activate PAK1 and MEK1/2 phosphorylation; 5% incidence — the highest reported across cancer types in TCGA at publication.
  • KIT — 20 hotspot mutations in 19 patients, mostly exon 17; enriched in SEM; imatinib-resistance pattern distinct from GIST.
  • KRAS — 23 hotspot mutations in 22 patients; G12 dominant; enriched in seminomas overall but in nonseminomas associated with cisplatin resistance.
  • NRAS — 4 mutations; 3 in cisplatin-resistant tumors.
  • BRAF — 3 hotspot mutations (D594N, D594G, G466E); rate within cohort 1.7%.
  • CBL — 7 alterations including hotspot W408R (RING finger), three X410 splice sites, one homozygous deletion.
  • GNAQ — Q209P (uveal-melanoma hotspot) in one patient.
  • PIK3CA — four E542K mutations.
  • PTEN — five loss-of-function plus four missense VUS.
  • AKT1, MTOR, TSC1, TSC2 — additional PI3K/mTOR pathway events.
  • APC, AXIN1, FAT1 — Wnt-pathway alterations.
  • EP300, SETD2, PTPRD, BRCA2 — additional tumor-suppressor / chromatin / DNA-repair events flagged as actionable.
  • MAX, MCL1, CCND3, KDR, MET — additional alterations contributing to the actionable list.

Clinical implications

  • Genomic profiling for risk stratification: TP53/MDM2 alteration status improves prognostication beyond the IGCCCG model and the authors recommend prospective genomic profiling of advanced GCT, particularly intermediate- and poor-risk patients, to facilitate clinical trials of novel strategies.
  • MDM2 inhibitor rationale: MDM2 amplification (and MYCN amplification in TP53 wild-type tumors) is the most common actionable alteration. The authors highlight 7 MDM2 inhibitors in clinical trials; in vitro, nutlin-3 showed antiproliferative and apoptotic synergy with cisplatin in TP53 wild-type, cisplatin-resistant GCT cell lines.
  • Targeted therapy candidates (Appendix Table A2 of the paper) for cisplatin-resistant patients include: imatinib/sunitinib for KIT mutations; MEK inhibitors (trametinib, selumetinib, binimetinib) for KRAS/NRAS/GNAQ alterations; sorafenib for BRAF D594; PI3K and mTOR inhibitors for PI3K-pathway alterations; PARP inhibitors for BRCA2 deletion; CDK4/6 inhibitors for CCND3 amplification; and RAC1 inhibitors (NSC23766) for novel RAC1 mutations.
  • Mediastinal nonseminoma biology: The 72% TP53-alteration rate in primary mediastinal nonseminoma provides the first molecular explanation for its uniquely poor prognosis within IGCCCG and supports prioritizing this subgroup for genomic profiling and trial enrollment.

Limitations & open questions

  • A large subset of cisplatin-resistant patients have no identified genetic basis for resistance by exome/targeted DNA sequencing; the authors propose follow-up whole-genome, transcriptome, and epigenome studies.
  • 24 (23.1%) cisplatin-resistant patients died of disease — lower than expected, attributed to short follow-up after first-line progression and high response to salvage high-dose chemotherapy with stem-cell transplantation; recurrent mutations were too sparse to associate with specific salvage regimens.
  • 49 resistant samples were mixed GCTs containing teratoma; microdissection was not performed, so a small amount of teratoma DNA may have been sequenced. TP53/MDM2 alterations in teratoma-free tumors argue against a meaningful confounding effect.
  • Each individual actionable alteration is rare, posing challenges for clinical trial enrollment and targeted drug development in this disease.
  • Pure teratoma and pure malignant transformation tumors were excluded by design, so findings do not extend to those histologies.
  • Functional validation was limited to RAC1; the causal role of other recurrent alterations in driving cisplatin resistance remains to be established.

Citations from this paper used in the wiki

  • “TP53 alterations were present exclusively in cisplatin-resistant tumors and were particularly prevalent among primary mediastinal nonseminomas (72%).” (Abstract)
  • “Despite this association, TP53 and MDM2 alterations predicted adverse prognosis independent of the IGCCCG model.” (Abstract; multivariable HR 1.83; 95% CI 1.12–2.98; P=.016)
  • “Actionable alterations, including novel RAC1 mutations, were detected in 55% of cisplatin-resistant GCTs.” (Abstract)
  • “Nine mutations were identified in RAC1 (G12V, n=3; G12R, n=2; P34R, Q61R, and Q61K, n=2)… Expression of the RAC1 mutations identified in GCT results in increased phosphorylation of PAK1 and ERK.” (Fig 3 / Results)
  • “5% incidence of RAC1 mutations in this cohort makes GCT the cancer type with the highest prevalence of RAC1 mutations reported to date.” (Discussion)
  • “All mutational and clinical data are available on the cBio Cancer Genomics Portal.” (Results — corresponds to study gct_msk_2016)

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