The genetic landscape of clinical resistance to RAF inhibition in metastatic melanoma

Authors

Eliezer M. Van Allen

Nikhil Wagle

Antje Sucker

Daniel Treacy

Cory Johannessen

Eva M. Goetz

Chelsea S. Place

Amaro Taylor-Weiner

Steven Whittaker

Gregory Kryukov

Eran Hodis

Mara Rosenberg

Aaron McKenna

Kristian Cibulskis

Deborah Farlow

Lisa Zimmer

Uwe Hillen

Ralf Gutzmer

Simone M. Goldinger

Selma Ugurel

Helen J. Gogas

Friederike Egberts

Carola Berking

Uwe Trefzer

Carmen Loquai

Benjamin Weide

Jessica C. Hassel

Stacey B. Gabriel

Scott L. Carter

Gad Getz

Levi A. Garraway

Dirk Schadendorf

Doi

10.1158/2159-8290.CD-13-0617

PMID: 24265153 · DOI: 10.1158/2159-8290.CD-13-0617 · Journal: Cancer Discovery (2014)

TL;DR

Van Allen, Wagle, and colleagues performed whole-exome sequencing on FFPE tumors from 45 patients with BRAF V600 metastatic cutaneous melanoma treated with single-agent vemurafenib or dabrafenib (the DeCOG cohort, skcm_broad_brafresist_2012). Genetic alterations in known or putative RAF inhibitor resistance genes were observed in 23 of 45 patients (51%). Beyond previously characterized NRAS, MAP2K1, BRAF amplification, and NF1 events, they discovered a “long tail” of new MAPK-pathway alterations — most notably MAP2K2 (MEK2) mutations and MITF amplification — that experimentally confer resistance to RAF/MEK inhibition. Three tumors carried multiple co-occurring resistance mechanisms within the same biopsy.

Cohort & data

  • 45 patients with BRAF V600 metastatic cutaneous melanoma treated with single-agent vemurafenib or dabrafenib from the Dermatologic Cooperative Oncology Group of Germany (DeCOG); cBioPortal dataset skcm_broad_brafresist_2012.
  • 14/45 (31%) had early resistance (progression in <12 weeks); 31/45 (68.9%) developed acquired resistance.
  • FFPE-derived tumor DNA underwent whole-exome sequencing on Illumina HiSeq, with mean target coverage of 200X (tumor) and 92X (germline).
  • Somatic SNVs called with MuTect, indels with Indelocator, annotation with Oncotator; alignments against hg19 via the Broad Picard/Firehose pipeline.
  • For 32 patients matched pre-treatment and post-resistance tumors were available; the rest contributed either pre-treatment (rapid progressors) or resistant biopsies (clinical responders).
  • Functional validation used A375, SKMEL19, UACC62, A2058, and WM266.4 BRAF V600E melanoma cell lines treated with dabrafenib, trametinib, selumetinib, PLX4720 (RAF tool compound), GDC-0941 (PI3K inhibitor), and VRT11E (ERK tool compound).

Key findings

  • BRAF V600 mutations were detected in 100% of pre-treatment biopsies; 44/45 missense at codon V600, plus one in-frame deletion (Val600_Lys601delinsGlu in Patient 11) predicted to phenocopy V600E.
  • Resistance alterations were enriched in MAPK-pathway genes (NRAS, BRAF amplification, MAP2K1, MAP2K2, MITF, NF1), accounting for 44.4% (20/45) of the cohort.
  • Recurrent classes by frequency:
    • NRAS somatic mutations: 17.8% (seven Q61, one T58); acquired NRAS mutations occurred exclusively in patients on therapy >12 weeks (P = 0.04).
    • MAP2K1 mutations: 15.6%; not all preclude clinical benefit.
    • BRAF amplification: 8.9%.
  • MEK2 (MAP2K2) is a new resistance gene. Four MAP2K2 mutations (V35M, L46F, C125S, N126D) identified in resistant tumors. When introduced into A375 cells, all four conferred resistance to both dabrafenib (RAF) and trametinib (MEK), with MEK2 C125S — homolog of the previously known MEK1 C121S — yielding >100-fold change in GI50; all four mutants remained sensitive to ERK inhibition (VRT11E) and sustained MEK/ERK phosphorylation under RAF-inhibitor treatment.
  • MAP2K1 mutations span pre-treatment and resistant tumors. Five MEK1 mutations identified in resistant or rapidly-progressing samples (V60E, G128V, V154I, P124S, P124L) and two (G276W, F53Y) in pre-treatment tumors of patients with clinical benefit. All tested MEK1 mutants, when doxycycline-induced in A375, produced 10–80-fold resistance to dabrafenib and 3–20-fold resistance to trametinib versus wild-type MEK1, but no effect on ERK-inhibitor (VRT11E) sensitivity.
  • MITF amplification is a transcriptional resistance mechanism. Focal relapse-specific MITF amplification was identified in one patient lacking other known resistance lesions. Overexpression of wild-type MITF (but not the DNA-binding-impaired R217Δ mutant) in WM266.4, SKMEL19, and UACC62 cells produced 30–80-fold increases in PLX4720 GI50 and cross-resistance to RAF, MEK, and ERK inhibitors.
  • Intra-tumor heterogeneity of resistance mechanisms. Three patients had multiple resistance alterations within the same post-progression biopsy:
    • Patient 41: NRAS Q61R + NRAS T58I (on mutually exclusive reads) plus a MAP2K1 mutation.
    • Patient 08: acquired NRAS Q61K + focal BRAF amplification.
    • Patient 02: MAP2K2 mutation + BRAF amplification.
  • PI3K-pathway alterations are common but contextual. PTEN missense mutations (9/11 in the phosphatase domain), a PTEN deletion (Patient 36), a PIK3CA H1047R acquired in a resistant tumor (Patient 48), and PIK3R1 variants were observed. PI3K alterations did not always preclude initial response (Patient 1: pre-treatment PTEN H93D, 66-week partial response). In vitro, combined PLX4720 + GDC-0941 markedly suppressed proliferation and induced apoptosis only in BRAF V600E / PTEN-null A2058 cells, not in A375 (BRAF V600E / PTEN WT).
  • RAC1 P29S marks early resistance. Pre-treatment RAC1 P29S in 3/14 early-resistance patients (Patients 26, 34, 46) and 0 patients with sustained response (P = 0.026).
  • HOXD8 nonsense mutation in one early-resistance tumor, consistent with prior RNAi-screen evidence that HOXD8 suppression promotes RAF-inhibitor resistance.

Genes & alterations

  • BRAF — V600 missense (44/45 pre-treatment), one Val600_Lys601delinsGlu (Patient 11); focal amplification in 8.9% of resistant tumors (e.g., Patients 02, 08).
  • NRAS — Q61R/K/H, T58I, Q60H mutations in 17.8% of resistant tumors; acquired NRAS restricted to >12-week responders (P = 0.04).
  • MAP2K1 (MEK1) — V60E, G128V, V154I, P124S, P124L (resistance/early-progression) plus G276W, F53Y (pre-treatment with clinical benefit). All experimentally validated mutants conferred 10–80-fold dabrafenib and 3–20-fold trametinib resistance in A375.
  • MAP2K2 (MEK2) — V35M, L46F, C125S, N126D; novel RAF-inhibitor resistance gene reported by this paper. C125S confers >100-fold cross-resistance to RAF and MEK inhibition.
  • MITF — relapse-associated focal amplification (Patient 58 background); overexpression confers pan-MAPK-inhibitor cross-resistance (30–80-fold) in WM266.4/SKMEL19/UACC62.
  • NF1 — included among MAPK-pathway resistance gene set (44.4% aggregate).
  • PTEN — missense mutations (11 patients; 9 in phosphatase domain), nonsense R233*, frameshift Y86fs, and a copy-number deletion (Patient 36); combination of RAF + PI3K inhibition synergistic in PTEN-null A2058 cells.
  • PIK3CA — acquired H1047R in Patient 48’s resistant tumor (oncogenic catalytic-subunit hotspot).
  • PIK3R1 — variants of uncertain significance in resistant tumors.
  • RAC1 — recurrent P29S in 3/14 pre-treatment tumors of early-resistance patients (P = 0.026); known gain-of-function melanoma oncogene.
  • HOXD8 — nonsense mutation in one early-resistance tumor; candidate transcriptional-effector resistance gene.

Clinical implications

  • Cross-resistance between RAF and MEK inhibitors (e.g., MAP2K1, MAP2K2, MITF lesions) implies combined RAF/MEK regimens may have limited durability in this subset; ERK inhibitors remain experimentally effective against MEK1/MEK2-mutant cells in vitro, motivating clinical evaluation of ERK-pathway inhibitors.
  • Patients with acquired NRAS mutations or BRAF amplification may respond to subsequent MEK-inhibitor regimens, though existing data suggest reduced benefit after RAF-inhibitor progression.
  • MITF-driven and other transcriptional-effector resistance mechanisms (potentially HOXD8) may confer pan-MAPK-pathway resistance and require non-MAPK therapeutic approaches.
  • Combined MAPK + PI3K inhibition merits clinical evaluation in BRAF-mutant melanomas with concurrent PI3K-pathway lesions, particularly PTEN loss; in PTEN-WT contexts the combination’s value is unclear.
  • Pre-treatment RAC1 P29S genotyping may identify patients at higher risk of early progression on single-agent RAF inhibitors.
  • Demonstrates feasibility of FFPE-based whole-exome sequencing for systematic resistance profiling on multicenter clinical cohorts.

Limitations & open questions

  • Whole-exome sequencing cannot detect non-genetic resistance mechanisms (BRAF alternative splicing, COT/MAP3K8 upregulation, RTK/ligand overexpression, HGF secretion), so true intra-tumor resistance heterogeneity and overall resistance prevalence are underestimated.
  • 13 patients lacked complete pre-treatment/post-resistance/germline trios; absent pre-treatment samples (e.g., Patients 02, 08, 36) preclude formal assignment of some alterations as acquired vs. pre-existing.
  • Single-focus biopsies cannot resolve heterogeneity from pre-existing tumor variability.
  • Several candidate resistance alterations (RAC1, HOXD8, many PTEN and PIK3R1 variants) require larger cohort validation and additional functional experiments.
  • The cellular contexts in which PI3K-pathway activation drives proliferative versus survival-only resistance remain unresolved; combination MAPK+PI3K therapy may benefit only a subset.
  • Cohort size (45 patients) is too small to reach statistical significance for individual rare-tail resistance genes.

Citations from this paper used in the wiki

  • “Genetic alterations in known or putative RAF inhibitor resistance genes were observed in 23 of 45 patients (51%).” (Abstract)
  • “Resistance alterations predominantly involved the MAPK pathway or downstream effectors (NRAS, BRAF, MAP2K1, MAP2K2, MITF, NF1), representing 44.4% (20/45) of the patient cohort.” (Results)
  • “One of these mutations (MEK2C125S) is homologous to a previously described MEK1C121S mutation that confers cross-resistance to RAF and MEK inhibitors in vitro.” (MEK2 section)
  • “MEK2C125S conferred profound resistance to both RAF and MEK inhibition, with fold change in GI50 greater than 100.” (Results)
  • “In both cell lines, MITF overexpression conferred a 30–80 fold increase in the PLX4720 GI50 values relative to control (LacZ) gene expression.” (MITF section)
  • “Among the 14 patients exhibiting early disease progression in the setting of RAF inhibition (disease progression within 12 weeks of initial RAF inhibitor therapy), three pre-treatment tumor biopsies harbored RAC1P29S mutations… (P = 0.026).” (Early-resistance section)
  • “As expected, acquired NRAS mutations occurred exclusively in patients on therapy for more than 12 weeks (P = 0.04).” (Results)

This page was processed by crosslinker on 2026-05-09.