Distinct clinical outcomes and biological features of specific KRAS mutants in human pancreatic cancer

Authors

Caitlin A. McIntyre

Adrien Grimont

Jiwoon Park

Henry Walch

Nikolaus Schultz

Lukas E. Dow

William R. Jarnagin

Rohit Chandwani

Doi

10.1016/j.ccell.2024.08.002

PMID: 39214094 · DOI: 10.1016/j.ccell.2024.08.002 · Journal: Cancer Cell (2024)

TL;DR

In a consecutive MSK cohort of 1,360 resected PDAC patients, KRASG12R is enriched in stage I disease, associated with increased node-negativity, decreased distant recurrence, and improved overall survival relative to KRASG12D. Spatial profiling (NanoString CosMx SMI, n=20) and bulk RNA-seq (n=100 tumors) reveal enhanced oncogenic KRAS signaling and EMT in KRASG12D versus increased TNF/NF-κB signaling in KRASG12R tumors. Isogenic Krasmut;Trp53KO mouse organoids recapitulate the reduced migratory potential and improved orthotopic survival phenotype, validating the clinical findings.

Cohort & data

  • Primary cohort: 1,360 consecutive PAAD (PDAC) patients undergoing surgical resection at Memorial Sloan Kettering Cancer Center (MSKCC) between April 2004 and April 2019; 397 (29%) stage I, 963 (71%) stage II–III (AJCC 8th edition) PMID:39214094. Median follow-up of survivors: 84.9 months (95% CI 76.5–94.7 months); median OS 30.7 months (95% CI 28.0–32.8).
  • Sequenced subset: 397 patients with targeted tumor sequencing on MSK-IMPACT 351-/410-/468-gene panels (IMPACT341, IMPACT410, IMPACT468); 103 stage I, 294 stage II–III; tumor DNA from FFPE resection specimens matched to germline blood controls PMID:39214094.
  • Transcriptomic sub-cohort: Bulk RNA-seq on 100 upfront resected tumors (63 KRASG12D, 37 KRASG12R) from the Canadian COMPASS trial (NCT02750657) plus resected patients from the same consortium PMID:39214094.
  • Spatial profiling: High-plex in situ RNA + protein profiling via CosMx SMI on tissue microarrays of 20 patients (7 KRASG12R PDAC, 7 KRASG12D PDAC, 6 normal pancreas); ~400 0.5×0.5 mm fields-of-view interrogated PMID:39214094.
  • Mouse models: Isogenic ex vivo-transformed pancreatic organoids from LSL-KrasG12D and LSL-KrasG12R mice, both engineered with CRISPR Trp53 loss; orthotopic transplantation into immunocompetent C57BL/6 mice PMID:39214094.
  • Genomic data: Deposited at cBioPortal as study pancreas_msk_2024. RNA-seq deposited at the European Genome-Phenome Archive (EGAD0001004548, EGAD00001009409); mouse organoid RNA-seq at SRA PRJNA578549 PMID:39214094.
  • External validation: Pooled 950-patient multi-dataset cohort comprising the TCGA-Firehose Legacy (paad_tcga, n=148), ICGC-PACA-AU + ICGC-PACA-CA (paad_icgc, n=375 + n=232), and the Sausen resectable cohort (n=100); survival analyses used the 855 patients with matched genomic and survival data PMID:39214094.

Key findings

  • KRAS altered in 361/397 (90%) of sequenced patients; TP53 in 71%, CDKN2A in 24%, SMAD4 in 17%; no significant difference in driver-gene frequencies between early- and late-stage disease after multiple-testing correction PMID:39214094.
  • KRAS allele distribution in the sequenced cohort: G12D 36.5% (n=145), G12V 32.5% (n=129), G12R 13.9% (n=55), other KRAS 8.1% (n=32), WT 9.1% (n=36) PMID:39214094.
  • KRASG12R enriched in stage I vs stage II–III PDAC (23% vs 11%; p=0.022); 43.6% of KRASG12R patients were stage I vs 23.6% for KRASG12D (p=0.022) PMID:39214094.
  • Stage migration for KRASG12R driven by node-negativity, not primary tumor size: KRASG12R node-negative in 47% vs KRASG12D 26% (p=0.019); T-stage was not different between alleles PMID:39214094.
  • KRASG12R patients more likely to have a family history of pancreatic cancer (16.4% vs 4.2% for KRASG12D, p=0.003) and tended to be non-smokers (58.2% vs 43.1%, not significant) PMID:39214094.
  • KRASG12R tumors carried more CDKN2A mutations than KRASG12D (40% vs 22.1%, p=0.046); two-or-more tumor-suppressor co-mutations across TP53/SMAD4/CDKN2A more common in KRASG12R (51% vs 30%, p=0.029); TP53 and SMAD4 individual frequencies not allele-dependent PMID:39214094.
  • Time-to-recurrence: distant recurrence less frequent in KRASG12R (2-year CIF 24.2% vs 43.2% for KRASG12D; p=0.017); local recurrence elevated in KRASG12R (2-year CIF 20.6% vs 10.6%; p=0.045) PMID:39214094.
  • Overall survival (Kaplan-Meier): other KRAS 47 months, KRASWT 46, KRASG12R 39, KRASG12V 39, KRASG12D 30 (p=0.026); TP53 mutation associated with worse OS (33.7 vs 43.6 months, p=0.014); SMAD4 and CDKN2A status alone not associated with OS PMID:39214094.
  • External validation (n=855 pooled): KRASG12D median OS 15.6 months vs KRASG12R 20.8, KRASG12V 22.4, KRASWT 27 months (p=0.006); multivariate stratified by data source: KRASG12R HR 0.77 (0.60–0.99), p=0.038 and KRASG12V HR 0.71 (0.58–0.88), p=0.02 independently predictive of better OS; absence of tumor-suppressor co-mutation HR 0.74 (0.57–0.95), p=0.02 PMID:39214094.
  • CosMx SMI spatial profiling and bulk RNA-seq: KRASG12D enriched for EMT (vimentin, fibronectin), mTORC1, E2F/G2M Hallmarks; KRASG12R enriched for cytokeratin, NF-κB (p65), CD31, and TNFα/NF-κB signaling. KRASG12R tumors negatively enriched for canonical mutant-KRAS PDAC signatures PMID:39214094.
  • Mouse organoid orthotopic transplantation: KrasG12R;p53KO median OS 90 days vs KrasG12D;p53KO 44 days (p<0.001); KrasG12R organoids show reduced migration without growth/proliferation differences and depleted MYC activation PMID:39214094.

Genes & alterations

  • KRAS — PDAC allele-specific biology; G12R is prognostically favorable (early-stage, node-negative, longer OS, less distant recurrence, increased local recurrence) while G12D drives canonical oncogenic KRAS/EMT transcriptional programs. G12V confers improved OS without distinct clinicopathologic features. Allele frequencies: G12D 36.5%, G12V 32.5%, G12R 13.9% in the 397-patient sequenced cohort PMID:39214094.
  • TP53 — mutated in 71% of sequenced PDAC; frequency not different by KRAS allele; TP53 mutation associated with shorter OS (33.7 vs 43.6 months, p=0.014) PMID:39214094.
  • CDKN2A — mutated in 24% overall; enriched in KRASG12R tumors (40% vs 22.1% for KRASG12D, p=0.046); copy-number deletion frequency on the IMPACT panel was low and not incorporated into allele comparisons PMID:39214094.
  • SMAD4 — mutated in 17% of sequenced PDAC; frequency not different by KRAS allele PMID:39214094.
  • BRAF — putative oncogenic mutations enriched in early-stage vs late-stage disease (3.9% vs 0.7%, p=0.042) but not significant after multiple-comparison correction (q>0.95) PMID:39214094.
  • TGFBR2 — putative oncogenic mutations enriched in early-stage vs late-stage disease (8.7% vs 2.4%, p=0.008) but not significant after multiple-comparison correction (q>0.95) PMID:39214094.
  • MYC — KrasG12R organoids show loss of Myc activation relative to KrasG12D organoids; MYC is highlighted as a key downstream effector of KRAS in PDAC progression PMID:39214094.

Clinical implications

  • Routine KRAS allele-specific mutation profiling is warranted in resected PAAD: KRASG12R identifies a subset with better intrinsic prognosis marked by reduced nodal spread and distant recurrence, suggesting greater net benefit from local control via surgery and reduced need for intensive neoadjuvant chemotherapy PMID:39214094.
  • KRASG12R tumors may respond more favorably to neoadjuvant fluorouracil-based chemotherapy (most frequently FOLFIRINOX: fluorouracil + irinotecan + oxaliplatin + leucovorin, or gemcitabine + nab-paclitaxel): enrichment of KRASG12R among NAC-followed-by-resection patients suggests higher chemosensitivity rather than borderline-resectable selection (32.7% vs 17.9%; p=0.094) PMID:39214094.
  • Allele-specific transcriptional programs (NF-κB in G12R vs KRAS/EMT in G12D) suggest divergent actionable vulnerabilities across KRAS-mutant PDAC, rather than a uniform pan-KRAS therapeutic strategy; novel KRAS inhibitors in preclinical development should be tested in allele-specific contexts PMID:39214094.
  • Tumor-suppressor co-mutation burden (number of co-mutated TP53/SMAD4/CDKN2A) is an independent prognostic factor in external multi-cohort validation; absence of tumor-suppressor co-mutations: HR 0.72 (0.56–0.92), p=0.009 for improved OS PMID:39214094.

Limitations & open questions

  • Sequenced cohort (n=397) is a subset of the full 1,360 resected patients; MSK-IMPACT panel-based sequencing incompletely captures CDKN2A/B deletions (acknowledged by authors) PMID:39214094.
  • Spatial profiling (n=14 tumors after exclusion of one IL6-mutant outlier) and bulk RNA-seq (n=100) sample sizes are modest; allele-specific transcriptional differences are internally consistent but require larger external validation PMID:39214094.
  • Mechanistic basis for reduced metastatic propensity of KRASG12R beyond the NF-κB/EMT axis and reportedly deficient PI3K/macropinocytosis engagement remains unresolved PMID:39214094.
  • Cohort restricted to surgically-resectable stage I–III disease; findings may not generalize to unresectable stage III or metastatic stage IV PDAC, which were not studied PMID:39214094.
  • How KRAS allele status modifies biological responses to conventional chemotherapy or novel targeted therapy was not addressed PMID:39214094.
  • KRASG12V shows improved OS similar to KRASG12R but without distinctive clinicopathologic or transcriptional features; the underlying mechanism is an open question PMID:39214094.

Citations from this paper used in the wiki

  • “KRASG12R mutations are enriched in early-stage (stage I) disease, owing not to smaller tumor size but increased node-negativity.” — Abstract/Summary PMID:39214094.
  • “KRASG12R tumors are associated with decreased distant recurrence and improved survival as compared to KRASG12D.” — Abstract/Summary PMID:39214094.
  • “Overall survival in the individual datasets was quite comparable (median OS: ICGC AU 19.0 months; ICGC-CA 19.3; Sausen 18.26; TCGA 19.55; p>0.95)” — Results, external validation PMID:39214094.
  • “Orthotopic transplantation of KrasMUT; p53KO organoids into the pancreas of immunocompetent B6 mice showed prolonged survival for KrasG12R (median OS 90 days versus 44 days for KrasG12D; p<0.001)” — Results, organoid section PMID:39214094.

This page was processed by crosslinker on 2026-05-04.