Molecular Characterization of Acquired Resistance to KRASG12C-EGFR Inhibition in Colorectal Cancer

Authors

Rona Yaeger

Riccardo Mezzadra

Jenna Sinopoli

Yu Bian

Michelangelo Marasco

Esther Kaplun

Yijuan Gao

Huiyong Zhao

Andraud De Cruz Paulo

Yingjie Zhu

Amedea Chirvassu Perez

Kalyana Chadala Vada

Edisson Tsa

Sudhir Chowdhry

Sydney Bowker

Qiang Chang

Besnike Qeriqi

Britta Weigelt

Gouri J. Nanjangud

Michael F. Berger

Hierak Der-Torossian

Kenna Anderes

Nicholas D. Socci

Jinru Shia

Gregory J. Riely

Yonina R. Murciano-Goroff

Bob T. Li

James G. Christensen

Jorge S. Reis-Filho

David B. Solit

Elisa de Stanchina

Scott W. Lowe

Neal Rosen

Sandra Misale

Doi

10.1158/2159-8290.CD-22-0405

PMID: 36355783 · DOI: 10.1158/2159-8290.CD-22-0405 · Journal: Cancer Discovery (2023)

TL;DR

This study characterizes the genetic mechanisms of acquired resistance to combined KRASG12C and EGFR inhibition in KRASG12C-mutant colorectal cancer. Using cell lines, a patient-derived xenograft model, and serial ctDNA from 12 patients treated with sotorasib or adagrasib plus anti-EGFR antibodies (cetuximab or panitumumab), the authors identify KRASG12C amplification as a recurrent resistance mechanism that drives oncogene-induced senescence upon drug withdrawal, creating a therapeutic vulnerability to mTOR-targeting senolytic agents such as AZD8055.

Cohort & data

  • 2 KRASG12C-mutant colorectal cancer cell lines (C106, RW7213) grown to resistance in sotorasib plus cetuximab.
  • 1 KRASG12C-mutant colorectal cancer patient-derived xenograft model (CLR113) treated with sotorasib plus cetuximab until acquired resistance (~10 months).
  • 12 patients with KRASG12C-mutant colorectal cancer treated with combination KRASG12C + EGFR inhibition: adagrasib plus cetuximab (n = 8) or sotorasib plus panitumumab (n = 4).
  • Serial ctDNA collected approximately every 6 weeks in 4 patients for longitudinal resistance dynamics.
  • Sequencing platforms: MSK-IMPACT (cell lines, PDX, tissue), Guardant360 CDx (patients 1–5, 11), ctDx FIRST (patients 6–10), MSK-ACCESS (patient 12).
  • Dataset deposited at cBioPortal: coadread_mskresistance_2022.

Key findings

  • Resistance to combined KRASG12C + EGFR inhibition is heterogeneous, with multiple low-frequency alterations co-occurring in individual patients. Categories include: RAS mutations (58.3% of patients), MAPK pathway alterations (58.3%), RTK activation (75%), PI3K-mTOR pathway alterations (25%), nuclear function alterations (41.7%), and others (16.7%).
  • In cell line models, C106-resistant cells acquired clonal NRASG12D mutation and subclonal APCQ879* nonsense mutation; RW7213-resistant cells acquired high-level KRASG12C amplification (>20 copies by FISH).
  • In the CLR113 PDX model, resistance was associated with KRASG12C amplification (CCF 100%), BRAFK601E (CCF 13%), and RAF1S259F (CCF 10%).
  • Emergent resistance alterations in patient ctDNA included KRAS switch mutations (G12A/D/F/L/R/S/V), KRAS secondary mutations (H95L/Q/R, Y96D/H/N), NRAS mutations (Q61K/R), BRAF mutations and fusions, MAP2K1 mutations, MET amplification/fusion, RET fusion, EGFR mutations, and MYC amplification.
  • All acquired resistance alterations were detected at low VAF (one tenth to one hundredth of baseline alteration frequency), with no clonal sweep observed. Resistance alterations appeared and disappeared during treatment, but KRASG12C amplification was the only alteration that increased steadily in proportion to tumor markers.
  • Upon drug withdrawal, KRASG12C-amplified resistant cells (RW7213-R) entered oncogene-induced senescence with increased beta-galactosidase activity, decreased Ki-67 staining (from 60% to 1%), elevated p16 expression, and SASP cytokine secretion.
  • Drug rechallenge after withdrawal failed to suppress mTOR signaling or restore apoptotic potential (cleaved PARP), indicating intermittent therapy cannot reestablish drug sensitivity.
  • In patient 12, tissue biopsy at progression (8 days after drug stop) showed acquired KRAS amplification (KRAS/Cen12 ratio 13.2 vs. 1.8 pretreatment), increased p-ERK (3+, >90% cells), elevated p-S6 (2+, 70% cells vs. absent pretreatment), and increased p16 (2+, 65% cells vs. 5% pretreatment), confirming senescence markers in the clinical setting.
  • ctDNA monitoring in 2 patients after drug withdrawal showed approximately 2-fold reduction in KRAS amplification signal, while other alterations remained unchanged, indicating fitness cost of KRASG12C amplification without drug pressure.
  • The mTOR inhibitor AZD8055 selectively inhibited S6K and S6 phosphorylation in drug-withdrawn resistant cells and showed enhanced cytotoxicity in the senescent (drug-off) condition compared with continuous drug exposure (P ≤ 0.0001). The BCL2 inhibitor navitoclax did not show selective activity.

Genes & alterations

  • KRAS: KRASG12C is the primary driver mutation; KRASG12C amplification (>20 copies) is a recurrent resistance mechanism. Secondary KRAS mutations at G12, H95, Y96, and R68 positions emerge at resistance.
  • NRAS: NRASG12D acquired clonally in C106-resistant cells; NRASQ61K/R detected in patient ctDNA at resistance.
  • BRAF: BRAFK601E in PDX; BRAFV600E and BRAF fusions (SEC23A-BRAF) in patients.
  • EGFR: Target of combination therapy (cetuximab/panitumumab); EGFR mutations detected at resistance in some patients.
  • MAP2K1: MEK1 mutations including R47_E62delinsQ detected at resistance.
  • RAF1: RAF1S259F detected in resistant PDX (CCF 10%).
  • MET: MET amplification and fusion detected at resistance in patient ctDNA.
  • RET: CCDC6-RET fusion detected in patient ctDNA at resistance.
  • MYC: MYC amplification detected at resistance.
  • APC: Subclonal APCQ879* in C106-resistant cells; various APC alterations in patients.
  • TP53: Truncal TP53 mutations (R273C, R248W, R248Q) tracked as baseline alterations.
  • PIK3CA: PIK3CA mutations (E545K/Q, H1047R, P104L, E542K) detected in patient ctDNA.
  • ERBB2: ERBB2E363V detected in patient ctDNA.
  • ERBB3: ERBB3V104M in C106-resistant cells.
  • ERBB4: ERBB4K901N detected at resistance.
  • FGFR2: FGFR2 mutations (I654V, D725V) detected at resistance.
  • NF1: NF1 mutations (T676P, R2637*) detected in patient ctDNA.
  • PTEN: PTENR142W detected at resistance.
  • RB1: RB1 splice-site and missense mutations detected at resistance.

Clinical implications

  • Combined KRASG12C + EGFR inhibition in colorectal cancer produces responses (27% for sotorasib + panitumumab; ~40% for adagrasib + cetuximab) but resistance develops through highly heterogeneous, subclonal mechanisms that predominantly reactivate ERK signaling.
  • KRASG12C amplification is a recurrent, clinically relevant resistance mechanism that correlates with tumor markers and response, unlike the many low-frequency point mutations that appear and disappear.
  • Intermittent therapy (drug holiday followed by rechallenge) does not reestablish sensitivity in KRASG12C-amplified resistant cells because the senescent state induced by drug withdrawal elevates mTOR signaling and protects from apoptosis.
  • Drug withdrawal combined with mTOR-targeting senolytic therapy (e.g., AZD8055) represents a potential strategy to overcome acquired resistance in KRASG12C-amplified tumors.
  • Senescence-associated uPAR expression upon drug withdrawal could provide a target for CAR-T cell therapy; SASP-mediated immune recruitment could be exploited with immune checkpoint inhibitors.

Limitations & open questions

  • The senescence phenotype was characterized in detail only in the RW7213 cell line with KRASG12C amplification; the C106 line with NRASG12D did not exhibit senescence, limiting generalizability to non-amplification resistance mechanisms.
  • The authors were unable to generate KRASG12C-overexpressing cell lines directly, suggesting that KRASG12C-amplified cells require long-term drug exposure adaptation mechanisms that are not fully understood.
  • ctDNA analysis may overestimate heterogeneity relative to single-biopsy specimens, as it exposes multi-lesion tumor heterogeneity.
  • The clinical validation of the senolytic approach (drug withdrawal + mTOR inhibition) is preclinical only; no patient data support this strategy.
  • Longer-term drug-off kinetics and the mechanistic basis of mTOR activation in the senescent state require further investigation.
  • Whether the low-frequency resistance alterations that appear and disappear are drug-induced mutations or pre-existing subclones remains unclear.
  • The study does not evaluate newer KRASG12C inhibitors (e.g., divarasib) or tri-therapy combinations.

Citations from this paper used in the wiki

  • “all acquired alterations were identified at low VAF, at one tenth or one hundredth of the frequency of alterations identified at baseline” (Results, Resistance Dynamics section).
  • “KRASG12C amplification as a recurrent resistance mechanism that tracks with tumor markers and response and engages oncogene-induced senescence when the drug is removed” (Discussion).
  • “the mTOR inhibitor AZD8055, which has been previously proposed as a senolytic agent, was able to inhibit S6K and S6 phosphorylation selectively in RW7213 resistant cells in which drug was removed” (Results, Exploiting the New Steady State section).
  • “RW7213 cells off drug are more sensitive to AZD8055 than cells maintained with continuous drug exposure” (Results).
  • “Intermittent therapy… the senescent phenotype that has high mTOR pathway activation prevents a strategy of intermittent therapy from overcoming resistance” (Discussion).
  • “KRASG12C amplification… after drug stop, tumor tissue has elevated MAPK and mTOR pathway signaling” with p16 increase from 5% to 65% of cells (Results, patient 12 biopsy data).

This page was processed by crosslinker on 2026-05-06.