An Acquired HER2T798I Gatekeeper Mutation Induces Resistance to Neratinib in a Patient with HER2 Mutant-Driven Breast Cancer

Abstract

We report a HER2T798I gatekeeper mutation in a patient with HER2L869R-mutant breast cancer with acquired resistance to neratinib. Laboratory studies suggested that HER2L869R is a neratinib-sensitive, gain-of-function mutation that upon dimerization with mutant HER3E928G, also present in the breast cancer, amplifies HER2 signaling. The patient was treated with neratinib and exhibited a sustained partial response. Upon clinical progression, HER2T798I was detected in plasma tumor cell-free DNA. Structural modeling of this acquired mutation suggested that the increased bulk of isoleucine in HER2T798I reduces neratinib binding. Neratinib blocked HER2-mediated signaling and growth in cells expressing HER2L869R but not HER2L869R/T798I In contrast, afatinib and the osimertinib metabolite AZ5104 strongly suppressed HER2L869R/T798I-induced signaling and cell growth. Acquisition of HER2T798I upon development of resistance to neratinib in a breast cancer with an initial activating HER2 mutation suggests HER2L869R is a driver mutation. HER2T798I-mediated neratinib resistance may be overcome by other irreversible HER2 inhibitors like afatinib.Significance: We found an acquired HER2 gatekeeper mutation in a patient with HER2-mutant breast cancer upon clinical progression on neratinib. We speculate that HER2T798I may arise as a secondary mutation following response to effective HER2 tyrosine kinase inhibitors (TKI) in other cancers with HER2-activating mutations. This resistance may be overcome by other irreversible HER2 TKIs, such as afatinib. Cancer Discov; 7(6); 575-85. ©2017 AACR.This article is highlighted in the In This Issue feature, p. 539.

Trial registration: ClinicalTrials.gov NCT01953926.

Conflict of interest statement

Disclosure of potential conflicts of interest: R.N. and R.L. are employees of Guardant Health. J.H. and V.M. are employees of Foundation Medicine. R.E.C. and A.S.L. are employees of Puma Biotechnology. D.C. is an employee of AstraZeneca. Other authors declare no conflicts of interest.

©2017 American Association for Cancer Research.

Figures

Fig. 1. HER2L869R exhibits a gain-of-function phenotype that is blocked by neratinib

(A) The amino acid sequences of human BRAF, ERBB2, and EGFR were aligned using Clustal Omega. BRAF V600, ERBB2 L869, and EGFR L861 residues are highlighted in yellow. (B) The structure of HER2L869R was modeled. The mutation from Leucine (cyan) to Arginine (highlighted in blue) permits favorable charge interaction (dashed yellow lines) with Asp769. (C) Residue pair energies involving residue 869 reveal the addition of a strong attractive (negative) interaction at Asp769 in the HER2L869R model. (D) MCF10A cells stably expressing HER2WT or HER2L869R were treated with vehicle (DMSO), 0.01–1.0 μM neratinib, or 1 μM lapatinib for 4 h in serum-free media. Cell lysates were probed with the indicated antibodies. Scans are all from the same gel/film; the vertical black line indicates an irrelevant lane that was removed from the figure for clarity. (E) Stably transduced MCF10A cells were seeded in 96-well plates in MCF10A starvation media (1% charcoal-stripped serum, no EGF). After 7 days, nuclei were stained with Hoechst and scored using the ImageXpress system. Data points represent the average ± standard deviation (SD) of 4 replicate wells (****, p<0.0001, ANOVA followed by Tukey’s multiple comparisons test). (F) Stably transduced MCF10A cells were plated in 3D Matrigel in presence of the indicated drugs (100 nM). Colonies were grown in media containing 5% charcoal-stripped serum without EGF and insulin. After ~2 weeks, colonies were stained with MTT and counted using the Gelcount system. Data represent the average ± SD of 3 replicates. Representative fields (10X objective) of wells are shown below (****p<0.0001, ANOVA followed by Tukey’s multiple comparisons test). (G) Chest wall skin metastases of patient with invasive lobular breast cancer harboring HER2L869R at baseline and 20 days after starting treatment with neratinib.

Fig. 2. HER2T798I induces acquired resistance to neratinib

(A) The patient’s plasma was drawn at the time of clinical trial screening (SCR) and the indicated cycles of neratinib therapy (1 cycle = 28 days). Plasma cfDNA was subjected to ERBB2 targeted capture and sequenced. (B) The amino acid sequences of human KIT, ERBB2, and EGFR were aligned using Clustal Omega. KIT T670, ERBB2 T798, and EGFR T790 gatekeeper residues are highlighted in yellow. (C) The structure of HER2WT and a model of HER2T798I are shown with neratinib bound to the kinase pocket. The threonine (WT) or isoleucine (mutant) residue at position 798 is shown in pink. (D) HEK293 cells were transiently transfected with V5-tagged HER2WT, HER2L869R, HER2T798I or HER2L869R/T798I and treated with the indicated concentrations of neratinib for 4 h in serum-free media. Cell lysates were subjected to immunoblot analyses with the indicated antibodies. The bar graph represents quantification of immunoblot bands using ImageJ software. (E) MCF10A cells stably expressing V5- tagged HER2WT, HER2L869R, HER2T798I or HER2L869R/T798I were cultured in MCF10A growth factor-depleted media. Cell lysates were subjected to immunoblot analyses with the indicated antibodies. The bar graph represents quantification of immunoblot bands using ImageJ software. (F) MCF10A cells from (E) were treated ± 123 nM neratinib in growth factor-depleted media for 6 days. Nuclei were stained with Hoechst and scored using the ImageXpress system. Data represent the average ± SD of 4 replicate wells (*p<0.05, ****p<0.0001, ANOVA followed by Tukey’s multiple comparisons test). (G) MCF10A cells stably expressing HER2L869R (blue) or HER2L869R/T798I (red) were treated with increasing concentrations of neratinib for 6 days. Nuclei were stained with Hoechst and scored using the ImageXpress system. Data represent the average ± SD of 4 replicate wells. IC50 values were calculated using GraphPad Prism.

Fig. 3. Afatinib and AZ5104 block HER2L869R/T798I signaling

(A–D) Computational modeling of the HER2 kinase domain in complex with (A) neratinib, (B) afatinib, (C) osimertinib, and (D) AZ5104 was performed. The N-terminal lobe and part of the C-terminal lobe of the tyrosine kinase domain (TKD) is shown in ribbon style. Each inhibitor is represented as sticks bound in the substrate-binding pocket. The T798I mutation is shown as red spheres deep in the pocket. The L869R mutation is shown as blue and green spheres on the far side of the alpha-C helix. (E) NR6 cells stably expressing V5-tagged HER2WT, HER2L869R, HER2T798I or HER2L869R/T798I (LR/TI) were treated with the indicated drugs at 100 nM for 4 h in serum-free media. Cell lysates were subjected to immunoblot analyses with the indicated antibodies. Scans are all from the same gel/film; the vertical black line indicates an irrelevant lane that was removed from the figure for clarity. The bar graph represents quantification of immunoblot bands using ImageJ software. (F) Stably transduced MCF10A cells were treated with the indicated drugs at 100 nM for 4 h in EGF- and serum-free media. Cell lysates were subjected to immunoblot analyses with the indicated antibodies as described in Methods.

Fig. 4. Afatinib and AZ5104 block HER2L869R/T798I-induced growth

(A) MCF10A cells stably expressing HER2L869R and HER2L869R/T798I were treated with increasing concentrations of neratinib, afatinib, osimertinib, or AZ5104. After 5 days, nuclei were stained with Hoechst and scored using the ImageXpress system. Data represent the average ± SD of 4 replicate wells. The fold change in IC50 values of MCF10A L869R/T798I cells relative to L869R cells is shown. (B) Stably transduced MCF10A cells were plated in 3D Matrigel in presence of the indicated drugs (100 nM). After 9 days, colonies were stained with MTT and counted using the Gelcount system. Data represent the average ± SD of 3 replicates. Representative fields (10X objective) of wells treated with vehicle (DMSO), 100 nM neratinib, and 100 nM afatinib are shown (****p<0.0001, ANOVA followed by Tukey’s multiple comparisons test). (C) BT474GFP (control) and BT474HER2-T798M were treated with the indicated drugs for 4 h in serum-free media. Cells lysates were tested in immunoblot anlayses using the indicated antibodies.