Comprehensive genomic profiles of small cell lung cancer

Authors

Julie George

Jing Shan Lim

Se Jin Jang

Yupeng Cui

Janine Altmüller

Christian Becker

Martin Peifer

Jens Sage

Roman K. Thomas

Doi

10.1038/nature14664

PMID: 26168399 · DOI: 10.1038/nature14664 · Journal: Nature (2015)

TL;DR

George et al. performed whole-genome sequencing on 110 small cell lung cancer (SCLC) tumours — at the time, the first comprehensive somatic genome analysis of this disease — and showed that bi-allelic inactivation of TP53 and RB1 is obligatory in SCLC (often through complex genomic rearrangements). They discovered recurrent somatic rearrangements of TP73 producing N-terminally truncated, dominant-negative oncogenic isoforms (TP73Δex2/3), identified inactivating mutations in NOTCH-family genes in 25% of cases, and validated NOTCH as a tumour suppressor and master regulator of neuroendocrine differentiation in a Trp53;Rb1;Rbl2 triple-knockout mouse model. They also found rare but potentially actionable kinase mutations (BRAF, KIT, PIK3CA).

Cohort & data

  • 152 fresh-frozen SCLC clinical specimens collected under IRB approval; 110 (primary lung n=148, metastatic n=4) passed QC for whole-genome sequencing of tumour + matched normal (median tumour content 84%). Most cases were treatment-naive; only 5 were obtained at relapse.
  • Cancer type: SCLC (Small Cell Lung Cancer; OncoTree code).
  • Dataset: sclc_ucologne_2015 — cBioPortal study ID for this cohort.
  • Transcriptome sequencing (rna-seq) on 81 specimens (71 of the 110 WGS cases + 10 additional).
  • Affymetrix SNP 6.0 arrays (affymetrix-snp6) on 142 specimens (103 of the WGS cases + 39 additional).
  • Independent validation cohort (n=112): exome sequencing of 28 fresh-frozen tumours + 9 SCLC cell lines re-analysed from prior work, plus custom-panel targeted-dna-seq on 8 fresh-frozen + 67 FFPE samples (22-gene custom Agilent SureDesign panel, ≥200× coverage).
  • Mouse models: 8 SCLC tumours from Trp53;Rb1 double-knockout and Trp53;Rb1;Rbl2 triple-knockout (TKO) mice analysed by whole-exome-seq (n=6) or whole-genome-seq (n=2).
  • Assays: WGS via Illumina HiSeq 2000 (TruSeq DNA PCR-free libraries, paired-end 2×100 bp, ≥30× tumour and normal). Pathology review by ≥2 independent expert pathologists with immunohistochemistry for chromogranin A, synaptophysin, CD56, Ki67. Reference genome NCBI37/hg19 (mouse: NCBI37/mm9), BWA v0.6.1-r104 alignment.
  • Median follow-up 69 months; 31% of patients alive at last follow-up; smoking status known for 88% (median 45 pack-years among smokers).

Key findings

  • Mutation burden: SCLC genomes show 8.62 nonsynonymous mutations per Mb, with 28% C:G>A:T transversions on average — a heavy-smoking signature. Mouse SCLC tumours showed far lower burden (~28.5 protein-altering mutations per sample).
  • Universal TP53/RB1 loss: In the 108 tumours without chromothripsis, TP53 had bi-allelic loss in 100% and RB1 in 93% of cases. Inactivating events spanned mutations, translocations, homozygous deletions, hemizygous losses, copy-neutral LOH, and LOH at higher ploidy. Many RB1 mutations occurred at exon–intron junctions and caused protein-damaging splice events (confirmed by RNA-seq). Authors conclude TP53 and RB1 follow the classical “two-hit” Knudson tumour-suppressor pattern in SCLC.
  • Chromothripsis as alternative RB1-loss mechanism: 2 tumours had wild-type RB1 but exhibited massive rearrangements between chromosomes 3 and 11 that retained CCND1, producing significant cyclin D1 overexpression (confirmed by IHC). These cases had fewer Ki67-positive cells. Authors propose that since cyclin D1 negatively regulates Rb family proteins, chromothripsis-driven CCND1 overexpression can substitute for genomic RB1 loss.
  • Significantly mutated genes (q < 0.05): TP53, RB1, KIAA1211 (canonical: CRACDL), COL22A1, RGS7, and FPR1. RGS7 and FPR1 are both involved in G-protein-coupled receptor signalling.
  • Mutation clustering (P < 0.05): Hotspot/clustering analysis flagged the histone acetyltransferases CREBBP and EP300 (with recurrent inactivating translocations), centrosome genes ASPM, ALMS1, PDE4DIP, the RNA-regulating gene XRN1, the tetraspanin gene PTGFRN, and TP73.
  • Damaging mutations (P < 0.01): TP53, RB1, CREBBP, COL22A1, plus FMN2 and NOTCH1. NOTCH-family genes (NOTCH1, NOTCH2, NOTCH3, NOTCH4) were recurrently inactivated in 25% of human SCLC; Notch3 was also mutated in a Trp53−/−,Rb1−/−,Rbl2−/− mouse tumour.
  • Mouse-validated accessory tumour suppressors: PTEN (previously known), and RBL1 and RBL2 — both showed inactivating translocations and mutations in human tumours; Trp53/Rb1/Rbl2 triple-knockout mice develop SCLC with shorter latency than Trp53/Rb1 double-knockouts, confirming RBL2 as an accessory tumour suppressor.
  • Mutual exclusivity: Mutations in CREBBP, EP300, TP73, RBL1, RBL2, and NOTCH-family genes were largely mutually exclusive, suggesting overlapping pro-tumorigenic functions.
  • Rare potentially actionable kinase mutations: Four tumours carried oncogenic driver mutations from other cancers: BRAF, KIT, and PIK3CA — see Extended Data Fig. 3e.
  • Copy-number landscape: Focal 3p losses at 3p14.3–3p14.2 (FHIT) and 3p12.3–3p12.2 (ROBO1); FHIT expression was reduced in cases with focal deletions. Homozygous CDKN2A losses. Recurrent amplifications of MYC-family genes (MYC, MYCN, MYCL — paper uses MYCL1 nomenclature), FGFR1, and IRS2. Focal IRS2 amplifications occurred in 2% of cases.
  • TP73 oncogenic rearrangements: Genomic breakpoints clustered precisely in the TP73 locus in 7% of cases (n=8), with recurrent breaks in introns 1, 2, and 3 producing intragenic fusions that exclude exon 2 (p73Δex2) or exons 2+3 (p73Δex2/3). One tumour had genomic exclusion of exon 10. p73Δex2/3 was found only in cases with genomic rearrangements, confirmed not to be a natural splice variant. Overall TP73 was somatically altered (mutation or rearrangement) in 13% of cases. These N-terminally truncated isoforms have dominant-negative activity on wild-type p73 and p53 and are confirmed oncogenes in vivo.
  • Subclonality: SCLC showed threefold lower subclonal diversity than lung adenocarcinoma (P = 0.00023), and heterogeneity did not correlate with clinical stage (in contrast to adenocarcinomas).
  • Transcriptional subtypes & NOTCH: Unsupervised clustering of RNA-seq (n=69) revealed two major groups. The majority (77%, n=53/69) had high neuroendocrine markers CHGA (chromogranin A) and GRP (gastrin-releasing peptide), high DLK1 (a non-canonical Notch inhibitor), and high ASCL1 (a neuroendocrine lineage oncogene whose expression is repressed by active Notch). The remaining 23% (n=16/69) expressed SYP and NCAM1. Major SCLC driver mutations (TP53, RB1, CREBBP) did not differ between transcriptional subtypes.
  • NOTCH activation suppresses SCLC in vivo: Crossing Rosa26-LSL-N2ICD or Rosa26-LSL-N1ICD into Trp53;Rb1;Rbl2 TKO mice significantly reduced tumour number (P < 0.001) after Ad-Cre induction and extended survival. The recombination efficiency of an innocuous reporter (Rosa26mT/mG) was much higher than that of the N2ICD allele, indicating strong negative selection against active Notch during SCLC development. Ectopic N1ICD also inhibited growth and cell-cycle progression in mouse and human SCLC cell lines, and upregulated HES1, Hey1, Hey2 while abrogating neuroendocrine marker expression.
  • Other recurrent rearrangements: Clustered breakpoints affected TTC28 on chromosome 22 (inactivating translocations), and CDKAL1 on chromosome 6. Breakpoints also clustered downstream of the L1HS retrotransposon, supporting a role for this element in cancer.

Genes & alterations

  • TP53 — bi-allelic inactivation in 100% of non-chromothriptic tumours; missense mutations cluster in the DNA-binding domain.
  • RB1 — bi-allelic loss in 93% of non-chromothriptic tumours; many alterations at exon–intron junctions cause protein-damaging splice events; frequent complex translocations.
  • CCND1 — retained and overexpressed in 2 chromothriptic tumours with wild-type RB1; proposed alternative mechanism of Rb-pathway inactivation.
  • TP73 — recurrent intragenic genomic rearrangements (breakpoints in introns 1–3) producing dominant-negative N-terminally truncated isoforms p73Δex2 and p73Δex2/3 in 7% of cases; total TP73 alteration rate 13%.
  • NOTCH1, NOTCH2, NOTCH3, NOTCH4 — inactivating mutations (often in the extracellular domain) in 25% of human SCLC; mouse models confirm Notch activation suppresses SCLC initiation and prolongs survival.
  • CREBBP and EP300 — clustered mutations and recurrent inactivating translocations.
  • RBL1, RBL2 — inactivating translocations and mutations; RBL2 validated as accessory tumour suppressor in TKO mice (shorter SCLC latency).
  • PTEN — confirmed accessory tumour suppressor (previously known role in murine SCLC).
  • BRAF, KIT, PIK3CA — rare oncogenic kinase mutations with potential targeted-therapy implications (4 tumours total).
  • FGFR1, MYC, MYCN, MYCL, IRS2 — recurrent focal amplifications.
  • CDKN2A, FHIT, ROBO1 — recurrent focal deletions on 3p (FHIT, ROBO1) and CDKN2A homozygous loss.
  • ASCL1, DLK1, CHGA, GRP, SYP, NCAM1 — neuroendocrine and Notch-axis expression markers defining the two SCLC transcriptional subtypes.
  • COL22A1, RGS7, FPR1, CRACDL (paper nomenclature: KIAA1211) — significantly mutated genes of unclear function.
  • ASPM, ALMS1, PDE4DIP — clustered mutations in centrosome-associated genes.
  • XRN1, PTGFRN, FMN2 — clustered/damaging mutations; PTGFRN also mutated in murine SCLC.
  • TTC28, CDKAL1 — clustered chromosomal breakpoints (chromosomes 22 and 6 respectively).
  • HES1 — upregulated upon Notch activation in mouse models, consistent with canonical Notch target activation suppressing neuroendocrine differentiation.

Clinical implications

  • Universal TP53/RB1 loss as defining feature: Authors argue complete genomic loss of both TP53 and RB1 function is obligatory in SCLC pathogenesis. Tumours with wild-type RB1 should be examined for chromothripsis-driven CCND1 overexpression as an alternative Rb-pathway inactivation mechanism.
  • Rare actionable kinase mutations: Mutations in BRAF, KIT, and PIK3CA — found in only 4 tumours total — suggest that genotyping SCLC patients may identify rare individuals who could benefit from targeted kinase inhibitor therapy.
  • TP73Δex2/3 as a potential therapeutic target: Given the frequent occurrence of oncogenic TP73 rearrangements (13% of cases), and prior reports that therapeutic options exist to restrict p73-dependent tumour growth in vivo (including in Trp53-deficient tumours), these approaches may be promising in SCLC.
  • NOTCH pathway as a candidate therapeutic axis: Since NOTCH activation suppresses SCLC initiation and extends survival in TKO mice, and abrogates neuroendocrine marker expression, Notch-pathway agonism (or de-repression) is suggested as a therapeutic strategy — though the paper does not name specific clinical agents.
  • Prognostic null result: Mutations in CREBBP, EP300, TP73, RBL1, RBL2, and NOTCH-family genes were NOT significantly associated with total mutation number, overall survival, or other clinical parameters (Extended Data Fig. 4). Global mutational signatures also did not correlate with these mutations.

Limitations & open questions

  • Functional uncertainty for novel hits: The authors explicitly note that the biological role of most newly discovered genes (KIAA1211/CRACDL, COL22A1, ASPM, PDE4DIP, PTGFRN) “is much less clear” and that functional experiments will be required.
  • Stage enrichment: The 110-tumour cohort was enriched for earlier-stage SCLC (most obtained by surgical resection), since SCLC is rarely surgically managed. This may limit generalizability to the typical advanced-stage clinical SCLC population.
  • Few relapsed cases: Only 5/110 tumours were obtained at relapse, so the genomic landscape of chemo-resistant/recurrent SCLC is not addressed here.
  • Sample exclusions: 42/152 cases were excluded for insufficient DNA quality/quantity, potentially introducing selection bias.
  • No survival association for most novel drivers: Although TP73 and NOTCH alterations are mechanistically validated, the lack of overall-survival association leaves their prognostic and predictive value open.
  • Chromothripsis sample size: Only 2 tumours with wild-type RB1 and chromothripsis-driven CCND1 retention were identified — a striking but very rare alternative mechanism that needs validation in larger cohorts.
  • Notch therapeutic translation: While mouse models clearly support Notch as a tumour suppressor in SCLC, clinically tractable Notch-activating agents and patient-selection biomarkers remain to be defined.

Citations from this paper used in the wiki

  • “We have sequenced the genomes of 110 small cell lung cancers (SCLC) … In nearly all the tumours analysed we found bi-allelic inactivation of TP53 and RB1, sometimes by complex genomic rearrangements.” (Abstract)
  • “Two tumours with wild-type RB1 had evidence of chromothripsis leading to overexpression of cyclin D1 (encoded by the CCND1 gene), revealing an alternative mechanism of Rb1 deregulation.” (Abstract)
  • “We discovered somatic genomic rearrangements of TP73 that create an oncogenic version of this gene, TP73Δex2/3.” (Abstract)
  • “Finally, we observed inactivating mutations in NOTCH family genes in 25% of human SCLC. Accordingly, activation of Notch signalling in a pre-clinical SCLC mouse model strikingly reduced the number of tumours and extended the survival of the mutant mice.” (Abstract)
  • “SCLC genomes exhibited extremely high mutation rates of 8.62 nonsynonomous mutations per million base pairs (Mb). C:G>A:T transversions were found in 28% of all mutations on average, a pattern indicative of heavy smoking.” (Results, Recurrent somatic alterations)
  • “Among the significantly mutated genes … were TP53 and RB1, KIAA1211 and COL22A1, as well as RGS7 and FPR1, both of which are involved in G-protein-coupled receptor signalling.” (Results, p. 3)
  • “In the 108 tumours without chromothripsis, TP53 and RB1 had bi-allelic losses in 100% and 93% of the cases, respectively.” (Results, Universal inactivation of TP53 and RB1)
  • “By contrast, genomic breakpoints affecting chromosome 1 clustered precisely in the TP73 locus in 7% of the cases (n = 8). … the majority of breaks caused intragenic fusions and, thus, exclusion of either exon 2, or exons 2 and 3.” (Results, Oncogenic genomic events affecting TP73)
  • “Altogether, TP73 was somatically altered by mutations and genomic rearrangements in 13% of the cases.” (Results, p. 6)
  • “Overall, the NOTCH family was affected by genomic alterations in 25% of human SCLC.” (Results, Tumour suppressive roles of Notch in SCLC)
  • “We crossed Rosa26Lox-stop-Lox-Notch2ICD (LSL-N2ICD) mice that conditionally express an activated form of Notch2 … to TKO mice and found a significant reduction in the number of tumours that arose in the presence of N2ICD (P < 0.001).” (Results, p. 6)
  • “A comparison to lung adenocarcinoma indicated a threefold lower subclonal diversity in SCLC (P = 0.00023).” (Results, p. 2–3)

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