Genomic analyses of gynaecologic carcinosarcomas reveal frequent mutations in chromatin remodelling genes

Authors

Siân Jones

Nicolas Stransky

Christine L. McCord

Ethan Cerami

James Lagowski

Devon Kelly

Samuel V. Angiuoli

Mark Sausen

Lisa Kann

Manish Shukla

Rosemary Makar

Laura D. Wood

Luis A. Diaz Jr

Christoph Lengauer

Victor E. Velculescu

Doi

10.1038/ncomms6006

PMID: 25233892 · DOI: 10.1038/ncomms6006 · Journal: Nature Communications (2014)

TL;DR

Jones et al. performed whole-exome sequencing on 22 gynaecologic carcinosarcomas (malignant mixed Müllerian tumours: 17 uterine, 5 ovarian) matched to normal tissue. Microsatellite-stable tumours carried an average of 43 non-synonymous mutations each; four mismatch-repair–deficient tumours harboured 904–5,913 somatic alterations. Beyond expected hits in TP53 (67%), KRAS (27%), PIK3CA (41%), and PTEN (41%), they discovered that nearly two-thirds of cases harbour mutations in chromatin-remodelling genes — most prominently ARID1A (32%), ARID1B (14%), KMT2C/MLL3 (27%), SPOP (14%), and BAZ1A (18%) — and that more than three quarters of cases carry alterations in genes with potential clinical actionability (PI3K pathway, homologous recombination, mismatch repair).

Cohort & data

  • 22 carcinosarcoma patients (17 uterine, 5 ovarian) with matched tumour/normal exomes; one tumour (MM02) was a recurrence at the primary site, others were primary. Samples from OHSU Knight BioLibrary; 11 patients received adjuvant chemotherapy or radiation after surgery.
  • Cancer types: UCS (uterine carcinosarcoma, primary) and MXOV (mixed ovarian carcinoma; five ovarian cases).
  • Dataset: ucs_jhu_2014 — exome data deposited at the European Genome-phenome Archive under accession EGAS00001000941.
  • Assay: whole-exome-seq using Agilent SureSelect 51 Mb (v4) capture + Illumina HiSeq 2000 paired-end 100 bp, average 191× coverage with ≥90% of bases covered ≥10×; aligned to hg19 with Eland/CASAVA 1.7.
  • 10 fresh-frozen tumours; 12 formalin-fixed paraffin-embedded tumours; ~85% mean neoplastic cellularity after macrodissection.

Key findings

  • Total burden: 777 somatic mutations in 702 genes across 18 microsatellite-stable tumours; 18,371 alterations across 4 mismatch-repair–deficient tumours (range 904–5,913 in the hypermutators).
  • Mismatch-repair deficiency: biallelic truncating mutations in MSH6 (MM12T) plus single nonsense/frameshift hits in MSH6 (MM04T, MM18T) and MLH1 (MM20T) — driving the mutator phenotype. The MLH1-mutant case showed ~50% small indels (vs ~10% in MSS); MSH6-mutant cases showed elevated C:G→A:T transversions (30%) suggestive of unrepaired oxidative damage.
  • MSS substitution spectrum was dominated by C:G→T:A transitions at CpG sites (deamination of 5-methylcytosine).
  • TP53 mutated in 67% (16/22) — a higher fraction than previously appreciated for carcinosarcoma. Seven cases had TP53 mutation allele frequencies >80% (clonal, homozygous after normal-tissue adjustment).
  • PI3K pathway activated in >50% of cases: PIK3CA 41% (9/22), PTEN 41% (9/22), PIK3R1 14% (3/22). PIK3CA and PIK3R1 mutations were mutually exclusive; PTEN loss co-occurred with other pathway hits. Unlike prior reports, PI3K pathway mutations were not confined to uterine cases — they also occurred in ovarian cases.
  • Chromatin-remodelling genes are the most prominently mutated class: ARID1A 32% (7/22), ARID1B 14%, KMT2C/MLL3 27% (6/22) plus one KMT2D/MLL2 splice-site, SPOP 14% (3/22), BAZ1A 18% (4/22). Three of the ARID1A-mutated tumours also carried ARID1B frameshift/nonsense alterations. Approximately two-thirds of cases harbour mutations in genes involved in chromatin remodelling or modification.
  • Site-specific signature: among MSS tumours, ARID1A mutations were unique to ovarian-origin carcinosarcomas — 4 of 5 ovarian tumours, none of the MSS uterine tumours.
  • Other recurrently mutated genes: KRAS hotspots in 27% (5/22; 5 cases for actionability counting); CTNNB1 hotspot mutations in two MSS cases; FBXW7 in 23% (5/22, including the canonical R505C hotspot in MM19, three WD-repeat missense, one frameshift); PPP2R1A hotspot mutation M180R in one uterine carcinosarcoma; ERBB3 extracellular-domain hotspot V104M in one tumour.
  • Clonality: TP53/known cancer-gene mutations cluster at ~50% (heterozygous) or ~100% (homozygous) allele frequencies after purity adjustment — consistent with monoclonal origin of carcinosarcomas (supporting the conversion theory of histogenesis).
  • DNA-repair: somatic frameshift in BRCA1 (1 case), nonsense mutations in BRCA2 (3 cases), splice-site in FANCM (2 cases). One germline BRCA2 nonsense (K3326X) identified in patient MM08 with family history of cervical and lung cancer.

Genes & alterations

  • TP53 — 16/22 (67%); seven cases clonal homozygous (>80% VAF).
  • KRAS — 5/22 hotspot mutations (27%); flagged as actionable via MEK/BRAF inhibitors.
  • PIK3CA — 9/22 (41%), hotspot activating; mutually exclusive with PIK3R1.
  • PIK3R1 — 3/22 (14%) in-frame indels and frameshift; first implication in carcinosarcoma.
  • PTEN — 9/22 (41%) missense/nonsense/frameshift; co-occurs with other PI3K hits.
  • ARID1A — 7/22 (32%) truncating; ovarian-restricted among MSS cases (4/5 ovarian, 0/uterine MSS).
  • ARID1B — 3/22 (14%) frameshift/nonsense; co-occurs with ARID1A in 3 cases.
  • KMT2C (MLL3) — 6/22 (27%) mostly missense, with truncations.
  • KMT2D (MLL2) — 1 splice-site alteration (co-occurring with an MLL3 mutation in the same tumour).
  • SPOP — 3/22 (14%) missense in/near the MATH domain (E46K, E78K, M117V).
  • BAZ1A — 4/22 (18%); first carcinosarcoma report.
  • CTNNB1 — 2 MSS cases with hotspot mutations.
  • FBXW7 — 5/22 (23%) including R505C WD-repeat hotspot in MM19.
  • PPP2R1A — 1 uterine case, M180R HEAT-domain hotspot.
  • ERBB3 — 1 case, V104M extracellular-domain hotspot.
  • BRCA1 — 1 frameshift.
  • BRCA2 — 3 nonsense (somatic) + 1 germline (K3326X in patient MM08).
  • FANCM — 2 splice-site/frameshift.
  • MLH1 — 1 frameshift (MM20T, hypermutator).
  • MSH6 — 3 cases (MM04T, MM12T biallelic, MM18T) nonsense/frameshift — drivers of the mutator phenotype.

Clinical implications

  • More than three quarters of cases carry an alteration in a gene with potential clinical actionability — a striking yield given that no targeted therapies are currently approved for carcinosarcoma and prognosis remains poor.
  • PI3K pathway hits (PIK3CA, PIK3R1, PTEN) in >50% of cases support trials of PI3K/mTOR/AKT/MEK inhibitors — already in Phase I/II testing in endometrial cancer.
  • KRAS hotspot mutations (5 cases) flag candidates for MEK/BRAF inhibition.
  • ERBB3 V104M (1 case) is targetable with ERBB-family inhibitor antibodies in Phase II development.
  • FBXW7 loss-of-function (5 cases) may sensitize tumours to HDAC inhibitors (preclinical evidence).
  • Homologous-recombination repair defects — somatic BRCA1/BRCA2 (4 cases) and germline BRCA2 (1 case) — suggest sensitivity to platinum agents and PARP inhibitors; FANCM defects flag mitomycin sensitivity.
  • The four mismatch-repair–deficient hypermutators (MLH1, MSH6) are candidates for anti–PD-1 immunotherapy (MK-3475/pembrolizumab being evaluated in MSI-unstable tumours at time of publication).
  • Mutational profiling has potential diagnostic utility: detection of carcinosarcoma-defining mutations in cell-free DNA or liquid Pap smear specimens could enable early detection.

Limitations & open questions

  • Modest cohort (n=22) with an imbalance between uterine (17) and ovarian (5) primaries, limiting power to fully resolve site-specific differences.
  • No transcriptomic, methylation, or copy-number data presented — only exomes; functional consequence of many chromatin-remodelling missense alterations is inferred, not measured.
  • The site-specific finding that ARID1A mutations are restricted to ovarian-origin MSS carcinosarcomas is provocative but rests on 4/5 ovarian vs 0/MSS-uterine — open to replication in larger series.
  • Clinical actionability claims are inferred from mutation status; the paper does not test response to any of the suggested targeted agents in this cohort.
  • Whether the observed chromatin-remodelling dysregulation drives the biphasic carcinoma+sarcoma histology (the “conversion” hypothesis) or merely co-occurs with it remains untested mechanistically.
  • The germline BRCA2 K3326X carrier was identified opportunistically; the paper does not pursue a systematic germline analysis.

Citations from this paper used in the wiki

  • “On average, we identify 43 mutations per tumour, excluding four cases with a mutator phenotype that harboured inactivating mutations in mismatch repair genes.” (Abstract)
  • “In addition to mutations in TP53 and KRAS, we identify genetic alterations in chromatin remodelling genes, ARID1A and ARID1B, in histone methyltransferase MLL3, in histone deacetylase modifier SPOP and in chromatin assembly factor BAZ1A, in nearly two thirds of cases.” (Abstract)
  • “Alterations in genes with potential clinical utility are observed in more than three quarters of the cases and included members of the PI3-kinase and homologous DNA repair pathways.” (Abstract)
  • “Mutations in TP53 were observed in 67% (16/22) of cases, a significantly higher fraction than previously appreciated.” (Results)
  • “The phosphoinositide 3-kinase (PI3-kinase) pathway was affected by activating mutations in PIK3CA (9/22 cases) and PTEN (9/22 cases). We identified mutations in PIK3R1, a member of the PI3-kinase pathway that had not been previously implicated in carcinosarcomas, in 3 of 22 cases.” (Results)
  • “Truncating mutations in ARID1A were detected in 32% (7/22) of tumours… Four of the five ovarian tumours had ARID1A mutations and these were not observed in the microsatellite-stable tumours of the uterus.” (Results)
  • “Alterations of the histone methyltransferase MLL3 were identified in 27% (6/22) of carcinosarcomas. One of these tumours also harboured a splice site alteration in the related gene, MLL2.” (Results)
  • “SPOP… was found to be altered in 3 cases (14%)… BAZ1A… was mutated in 4 of the 22 carcinosarcomas.” (Results)
  • “Mutations in the tumour suppressor gene, FBXW7, were identified in 23% (5/22) of cases… Tumour MM19 carried the previously described hotspot mutation, 505R>C.” (Results)
  • “A known hotspot missense mutation, 104V>M, in the extracellular domain of the kinase ERBB3 was observed in a single tumour… KRAS hotspot mutations were identified in five cases and tumours of these patients may respond to MEK/BRAF inhibitors.” (Results)
  • “Two somatic nonsense mutations in BRCA2, a frameshift in BRCA1 and a splice site alteration of FANCM were also identified providing further evidence for the role of homologous recombination repair defects in carcinosarcomas.” (Results)
  • “Carcinosarcoma exome sequence data has been deposited at the European Genome-phenome Archive… under the accession code EGAS00001000941.” (Additional information)

This page was processed by crosslinker on 2026-05-12.