Molecular analysis of aggressive renal cell carcinoma with unclassified histology reveals distinct subsets

Authors

Ying-Bei Chen

Jianing Xu

Anders Jacobsen Skanderup

Yiyu Dong

A. Rose Brannon

Lu Wang

Helen H. Won

Patricia I. Wang

Gouri J. Nanjangud

Achim A. Jungbluth

Wei Li

Virginia Ojeda

A. Ari Hakimi

Martin H. Voss

Nikolaus Schultz

Robert J. Motzer

Paul Russo

Emily H. Cheng

Filippo G. Giancotti

William Lee

Michael F. Berger

Satish K. Tickoo

Victor E. Reuter

James J. Hsieh

Doi

10.1038/ncomms13131

PMID: 27713405 · DOI: 10.1038/ncomms13131 · Journal: Nature Communications (2016)

TL;DR

Chen et al. perform the first in-depth molecular characterization of 62 high-grade primary renal cell carcinomas with unclassified histology (URCC) at MSKCC, integrating MSK-IMPACT targeted sequencing, RNA-seq, OncoScan SNP arrays, FISH and IHC. They identify 29 recurrently mutated genes and define four molecularly distinct subsets that together cover ~76% of the cohort: NF2 loss with dysregulated Hippo–YAP signalling (26%), hyperactive mTORC1 signalling driven by MTOR/TSC1/TSC2/PTEN mutations (21%), FH deficiency (6%) and ALK translocation (2%), plus a chromatin/DNA-damage regulator group (21%). The NF2 loss and FH-deficient subsets are associated with the worst clinical outcome, whereas mTORC1-hyperactive uRCC has a comparatively better course — providing biological rationale for subset-specific therapy in this otherwise standard-of-care-naive disease (PMID:27713405).

Cohort & data

  • 62 high-grade primary uRCC tumours from MSKCC, all re-reviewed by three genitourinary pathologists (Y.B.C., V.E.R., S.K.T.) per WHO and ISUP consensus criteria; MiTF family translocation tumours excluded by TFE3/TFEB IHC and FISH (PMID:27713405).
  • Disease: renal cell carcinoma, unclassified — a high-grade non-clear-cell RCC subgroup that comprises 4–5% of all RCC and lacks standard therapy.
  • Clinicopathology: 58% locally advanced (≥pT3) at nephrectomy, 32% with regional lymph-node involvement; 42% (n=26) developed metastases and 35% (n=22) died of RCC during follow-up.
  • Assays:
    • Targeted DNA sequencing on the 230-gene MSK-IMPACT platform (avg coverage 348× tumour / 280× normal); matched normal in 61/62 cases. Variant calling with MuTect (SNVs) and GATK Somatic Indel Detector; reads aligned with BWA-MEM to hg19.
    • Genome-wide copy-number / LOH via Affymetrix OncoScan FFPE SNP array on 15/16 NF2-loss cases.
    • RNA-seq on 7 uRCC (4 NF2-loss, 3 NF2-WT) on Illumina HiSeq 2500, mapped with STAR; GSEA used to evaluate YAP/TAZ transcriptional signatures.
    • FISH: custom three-probe NF2/22q11/Cen10 assay; ALK break-apart probes (Abbott) for fusion confirmation.
    • Immunohistochemistry: NF2, YAP/TAZ, p-YAP, p-S6, p-4EBP1, FH, 2SC (succination), H3K36me3, INI1.
    • In vitro validation in 293T (MTOR mutant constructs + HA-S6K) and the NF2-loss nccRCC cell lines ACHN and LB996-RCC (shYAP1 knockdown, soft-agar colony formation, cell-cycle FACS).

Key findings

  • Mutation landscape: 29 recurrently mutated genes with average 2.6 (range 0–8) protein-coding somatic mutations per tumour. Most frequent: NF2 18% (11/62), SETD2 18%, BAP1 13%, KMT2C 10%, MTOR 8%; only one VHL mutation (T08), in stark contrast to ~75% in ccRCC (PMID:27713405).
  • NF2 loss subset (n=16, 26%): Defined by NF2 mutation and/or 22q12 loss. 22q hemizygous loss in 14 cases by IMPACT + FISH; copy-neutral LOH of 22q in T22 and T64. 13 of 16 (81%) show concurrent NF2 mutation + 22q LOH (biallelic inactivation), a feature not previously reported in RCC. NF2 IHC is significantly lower in this subset (Mann–Whitney P<0.001).
  • Hippo–YAP dysregulation: NF2-loss tumours show significantly stronger nuclear YAP/TAZ and lower p-YAP signal vs other uRCC (P<0.001). GSEA on RNA-seq confirms enrichment of an established YAP/TAZ transcriptional signature in NF2-loss tumours.
  • Functional YAP dependency: shRNA knockdown of YAP1 in the NF2-loss nccRCC lines ACHN and LB996-RCC reduces S- and G2/M-phase cells (Student’s t-test, P<0.001) and decreases soft-agar colony formation.
  • SETD2 enrichment in NF2-loss subset: SETD2 mutation rate 44% in NF2-loss vs 9% in remaining uRCC (Fisher’s exact P=0.004); all 7 NF2-loss/SETD2-mutated tumours show complete loss of the H3K36me3 mark, vs 1/55 retention in the remaining cohort. Concurrent 1p and/or 3p loss in >50% of NF2-loss cases — without VHL mutation, distinguishing them from ccRCC.
  • mTORC1-hyperactive subset (n=13, 21%): Mutually exclusive mutations in MTOR (5), TSC1 (4), TSC2 (3) or PTEN (4) in 16 cases; 13/16 confirmed hyperactive by p-S6 and p-4EBP1 IHC. MTOR L2427R recurred 3× and is functionally activating in 293T co-transfection assays (elevated p-T389-S6K and p-T37/46-4EBP1); V2475M behaves like wild-type and is likely a passenger. All 7 TSC1/TSC2-mutated tumours have maximal p-4EBP1 (H-score=300); only 2/4 PTEN-mutated tumours show similar staining.
  • NF2 loss vs mTORC1 are mutually exclusive subsets; NF2-loss uRCC, unlike NF2-deficient mesothelioma/meningioma, does not show mTORC1 hyperactivation.
  • FH-deficient subset (n=4, 6%): All 4 tumours are FH IHC-negative / 2SC-positive. Three are confirmed HLRCC cases with germline FH mutations; T41 carries a somatic homozygous deletion of FH. T71 (FH G401V missense) is FH+/2SC-negative and reclassified as a passenger.
  • TPM3–ALK fusion (n=1, 2%): T12 carries a TPM3–ALK fusion detected by IMPACT and confirmed by ALK break-apart FISH — the second adult RCC case with this fusion reported.
  • Chromatin/DNA-damage regulator group (n=13, 21%): Eight cases with chromatin-modulator mutations (SETD2, BAP1, KMT2A/C/D, PBRM1) without other driver; 5 with DNA-damage-response mutations (TP53, CHEK2, BRCA2).
  • Other (n=15, 24%): No recurrent feature. T62 has a MET H1094Y mutation; T69 has a BRAF Y472C mutation. Three SMARCB1-mutated tumours (T23, T38, T41) retain INI1 expression and were assigned to mTORC1, NF2-loss or FH subsets respectively, distinguishing them from renal medullary carcinoma.
  • Clinical outcome by subset (log-rank): NF2-loss and FH-deficient uRCC have the worst progression-free and cancer-specific survival; mTORC1-hyperactive and unspecified uRCC fare best; chromatin/DNA-damage group is intermediate. Single-gene SETD2 or BAP1 mutation status alone does not stratify outcome in this cohort.

Genes & alterations

  • NF2 — recurrent truncating + missense mutations (18%); biallelic inactivation by mutation + 22q12 LOH defines the dominant uRCC subset. Drives Hippo–YAP dysregulation; shYAP1 reverses proliferation phenotype in NF2-loss cells.
  • SETD2 — 18% overall, but 44% within the NF2-loss subset (Fisher P=0.004). Loss of H3K36me3 IHC mark co-occurs with mutation; suggests synthetic-lethal opportunity with WEE1 inhibition.
  • BAP1 — 13%, no significant outcome stratification on its own.
  • KMT2C, KMT2D, KMT2A — chromatin modulators recurrently mutated (combined 16% across KMT2 family).
  • MTOR — 8% missense; recurrent L2427R (×3) functionally activating; V2475M is a passenger; I1973F previously known activating.
  • TSC1, TSC2, PTEN — together with MTOR define the mTORC1-hyperactive subset; mutually exclusive across the 16-case set.
  • FH — defines FH-deficient subset; somatic homozygous deletion (T41) is a novel non-germline mechanism for FH-deficient RCC.
  • ALK and TPM3 — TPM3–ALK fusion in T12 marks an emerging RCC entity.
  • VHL — only 1/62 mutation (T08); the absence of VHL alteration despite frequent 3p loss distinguishes uRCC from ccRCC.
  • YAP1 and WWTR1 (TAZ) — not mutated, but their nuclear accumulation marks the NF2-loss subset; shYAP1 knockdown is functionally validating.
  • ATRX (7%), DNMT3A (5%), SMARCB1 (5%), KLF6 (5%), NOTCH2 (5%), TP53 (5%), PBRM1 (3%), CHEK2 (3%), BRCA2 — additional recurrent mutations.
  • MET H1094Y and BRAF Y472C — pathogenic mutations in single uRCC cases (T62, T69) suggesting overlap with pRCC and providing a candidate therapeutic target.

Clinical implications

  • Diagnosis: uRCC should be subdivided into molecularly defined entities. NF2 IHC + 22q FISH, FH/2SC IHC, and ALK FISH can prospectively partition >50% of uRCC into actionable groups; biallelic NF2 inactivation in particular is proposed as a defining marker of an aggressive uRCC entity.
  • Prognosis: NF2-loss and FH-deficient subsets confer significantly worse cancer-specific and progression-free survival than mTORC1-hyperactive or unspecified uRCC (log-rank, P<0.05–0.001); NF2 status should be considered in risk stratification.
  • Treatment hypotheses generated by the authors:
    • mTORC1-hyperactive uRCC carries alterations similar to those seen in ccRCC patients with long-term benefit from mTOR inhibitors, suggesting MTOR/mTORC1 inhibition is a readily available targeted therapy for this subset.
    • NF2-loss uRCC is candidate for YAP-pathway inhibitors (e.g., verteporfin) and, by virtue of H3K36me3 loss in concurrent SETD2-mutant cases, for synthetic-lethal WEE1 inhibition.
    • FH-deficient cases warrant genetic counselling (HLRCC).
    • The TPM3–ALK case raises the possibility of ALK inhibition in molecularly selected uRCC.
    • The MET H1094Y case provides a potential MET-directed therapeutic option.
  • Pathology workflow: FH and 2SC IHC alone are insufficient to distinguish HLRCC from sporadic FH-deficient RCC; genetic counselling should accompany pathologic suspicion.

Limitations & open questions

  • Small sample size (n=62) limits statistical power for subset-specific outcome analysis and for assessing rare events such as ALK translocation (n=1).
  • Single-institution MSKCC cohort with predominantly retrospective FFPE material; reproducibility in independent cohorts is needed.
  • Authors note that the histological boundary between NF2-loss uRCC and emerging entities such as type 2 pRCC or collecting-duct RCC is unresolved; small numbers of NF2-mutated pRCC and collecting-duct RCC have been reported elsewhere.
  • Whether somatic FH-deficient RCC behaves clinically like germline HLRCC remains unclear; current IHC criteria do not reliably distinguish the two.
  • The 230-gene MSK-IMPACT panel does not capture the whole exome, so private drivers in the 24% “other” group cannot be ruled out by this study.
  • Functional dependency on YAP was demonstrated only in two NF2-loss nccRCC lines (ACHN, LB996-RCC); in vivo or PDX validation is not presented.
  • Whether the NF2/SETD2 co-occurrence reflects a shared clonal origin or sequential selection is not addressed.

Citations from this paper used in the wiki

  • “We identify recurrent somatic mutations in 29 genes, including NF2 (18%), SETD2 (18%), BAP1 (13%), KMT2C (10%) and MTOR (8%).” (Abstract)
  • “Integrated analysis reveals a subset of 26% uRCC characterized by NF2 loss, dysregulated Hippo–YAP pathway and worse survival, whereas 21% uRCC with mutations of MTOR, TSC1, TSC2 or PTEN and hyperactive mTORC1 signalling are associated with better clinical outcome.” (Abstract)
  • “FH deficiency (6%), chromatin/DNA damage regulator mutations (21%) and ALK translocation (2%) distinguish additional cases.” (Abstract)
  • “13 of the 16 uRCC tumours with MTOR, TSC1, TSC2 or PTEN mutations exhibited hyperactive mTORC1 signals.” (Results, mTORC1 subset)
  • “the occurrence of SETD2 (3p21) mutation was significantly higher in the NF2 loss than in the remaining uRCC tumours (44% versus 9%, Fisher’s exact test, P=0.004).” (Results, NF2 loss)
  • “MTOR L2427R (recurred three times in our uRCC cohort) … L2427R exhibited higher activity, whereas V2475M showed baseline mTORC1 kinase activity comparable to the wild-type MTOR.” (Results, mTORC1)
  • “Knockdown of YAP in ACHN or LB996-RCC cells resulted in a decrease of proliferating cells (S and G2/M phases) … as well as a reduced colony formation in soft agar.” (Results, NF2 loss)
  • “NF2 loss and FH-deficient uRCC appeared to have worse clinical outcome than mTORC1 hyperactive and thus far unspecified uRCC.” (Results, outcomes)
  • “one uRCC (T12) carried a TPM3–ALK fusion, which was further confirmed by FISH analysis … the second RCC case with this specific fusion reported in adults.” (Results & Discussion)

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