Efficacy and Determinants of Response to HER Kinase Inhibition in HER2-Mutant Metastatic Breast Cancer

Abstract

HER2 mutations define a subset of metastatic breast cancers with a unique mechanism of oncogenic addiction to HER2 signaling. We explored activity of the irreversible pan-HER kinase inhibitor neratinib, alone or with fulvestrant, in 81 patients with HER2-mutant metastatic breast cancer. Overall response rate was similar with or without estrogen receptor (ER) blockade. By comparison, progression-free survival and duration of response appeared longer in ER+ patients receiving combination therapy, although the study was not designed for direct comparison. Preexistent concurrent activating HER2 or HER3 alterations were associated with poor treatment outcome. Similarly, acquisition of multiple HER2-activating events, as well as gatekeeper alterations, were observed at disease progression in a high proportion of patients deriving clinical benefit from neratinib. Collectively, these data define HER2 mutations as a therapeutic target in breast cancer and suggest that coexistence of additional HER signaling alterations may promote both de novo and acquired resistance to neratinib. SIGNIFICANCE: HER2 mutations define a targetable breast cancer subset, although sensitivity to irreversible HER kinase inhibition appears to be modified by the presence of concurrent activating genomic events in the pathway. These findings have implications for potential future combinatorial approaches and broader therapeutic development for this genomically defined subset of breast cancer.This article is highlighted in the In This Issue feature, p. 161.

©2019 American Association for Cancer Research.

Figures

Figure 1.. Response in the monotherapy and combination therapy cohorts.

(A) Distribution of HER2 mutations observed in 34 monotherapy cohort patients (top) and 47 combination therapy cohort patients (bottom) positioned by their amino acid across the respective ERBB2 protein domains. Each unique mutation is represented by a circle and colored by their best overall response as indicated in the legend. (B) Treatment response and outcome for 34 monotherapy cohort patients (left) and 47 combination therapy cohort patients (right). Top graph represents percent best change of target lesion from baseline according to the appropriate response criteria (RECIST [version 1.1] or PET) with each bar colored by the respective HER2 allele as indicated in the legend. Bottom graph represents PFS with arrows indicating patients with ongoing treatment. CDK, cyclin-dependent kinase; HR, hormone receptor; PET, positron-emission tomography; PFS, progression-free survival; RECIST, Response Evaluation Criteria in Solid Tumors.

Figure 2.. Clonality and co-mutation of ERBB2.

(A) Plot of the ERBB2 clonality of 44 evaluable patients represented by cancer cell fractions with 95% confidence intervals and colored by additional ERBB2/ERBB3 activating events as indicated by the legend. (B) Bar plot showing the overall percent of ERBB2-mutant cases and the number of cases with multiple ERBB2 mutations (in dark blue) in the top mutated tumor types. (C) Allele-specific copy-number plot showing copy-neutral loss of heterozygosity (CN-LOH) at the ERBB2 locus (left) and plot of the expected (dotted line) and observed allele frequencies with 95% binomial confidence intervals of the mutations to infer the phase (in cis) of the ERBB2 mutations (right). (D) Proportion of all phaseable ERBB2 mutations across the broader prospective sequencing cohort occurring in cis versus in trans. Het, heterozygosity; OR, odds ratio.

Figure 3.. ERBB2 and ERBB3 co-mutation.

(A) OncoPrint of 47 evaluable patients grouped by clinical benefit (left, no clinical benefit, n=28; right, clinical benefit, n=19). Top bar chart represents the TMB shown in mutations per megabase (mut/Mb). MSI, allele domain and therapy type as indicated in the legend. Comprehensive oncoPrint showing alterations and clonality of ERBB2 and other co-alterations in genes associated with RTK/RAS/RAF and other pathways. (B) Heatmap of co-alteration patterns in the MAPK pathway with significant associations highlighted and represented by the number of cases observed across the broader prospective sequencing cohort. (C) Condensed oncoPrint showing ERBB2 missense and in-frame insertion or deletion (indel) mutations grouped by their respective protein domain and their co-occurrence patterns with ERBB3 and other MAPK alterations. *Significant nominal Fisher’s P-value; **Significant two-sided Fisher’s P-value. EC, extracellular; KD, kinase domain; MAPK, mitogen-activated protein kinase; MSI, microsatellite instability; TMB, tumor mutational burden.

Figure 4.. Mutant ERBB2 evolution on therapy.

(A) Bar plot of nine patients with paired pre- and post-treatment tissue samples showing the proportion of alterations that were shared or exclusive. (B) Three-dimensional modeling structure showing two mutations (gatekeeper T798I, L785F) conferring steric hindrance to neratinib binding. (C) Overall ERBB2 evolution in eight patients who acquired additional ERBB2 alterations in either the tissue and/or cell-free DNA. Each circle represents an ERBB2 mutation, colored by their respective allele/domain. (D) Conceptual schematic showing the impact of multiple activating events in ERBB2/ERBB3 and potential mechanisms of de novo and acquired resistance to pharmacologic inhibition to neratinib over time. Tx, treatment.