To the Editor

We investigated the incidence of ERBB2 mutations in a large panel of tumors analyzed by next-generation sequencing (NGS) and reported the clinical impact of ERBB2 targeting in this setting as well as potential mechanisms of secondary resistance.

We analyzed the AACR Project Genomics Evidence Neoplasia Information Exchange (GENIE) [1] and the Bergonie Institute Profiling study (ClinicalTrials.gov Identifier: NCT02534649) [2] databases (see Additional file 1: Supplementary Methods).

Seventeen thousand eight hundred seventy-eight patients were included in the study. Figure 1a describes the distribution of tumor types. Five hundred seventy-one ERBB2 mutations involving all the domains of the receptor (Fig. 1b) were found in 510 patients (2.85%): 472 missense mutations, 66 in frame mutations, 14 fusions, 11 frame shift mutations, 5 non-sense, and 3 splice mutations. Figure 1c describes the incidence of mutations according to tumor types. 49.4% (n = 282) of the mutations identified were described as oncogenic according to COSMIC and were more frequently detected in the bladder (9.4%), small bowel (7.1%), ampullar (6.5%), cervical cancer (4.1%), and nerve sheath tumor (2.9%), respectively (Fig. 1d). The most common mutations are represented in Table 1.

Fig. 1
figure 1

Landscape of ERBB2 mutations in cancer. a Proportion of histological subtypes included in the study. b ERBB2 hotspot mutations. c Proportion of patients with ERBB2 mutations according to the histological subtype. d Proportion of patients with activating ERBB2 mutations according to the histological subtype

Full size image
Table 1 Most common ERBB2 mutations in cancer patients
Full size table

The three most frequent alterations co-occurring with ERBB2 mutations were the following: TP53 mutation (n = 557, 59.5%), CDK12 amplification (n = 216, 24%), and PI3KCA mutation (n = 200, 21.4%).

Among the 39 tumors bearing both ERBB2 amplification and mutation, 16 (32%) were breast ductal carcinoma, 8 (16%) lung adenocarcinoma, and 6 (12%) bladder urothelial carcinoma.

Four patients with ERBB2 mutation-bearing tumor were treated with dual ERBB2 blockade with trastuzumab + lapatinib and experienced clinical benefit (Additional file 1: Table S2 and Figure S1). Despite tumor shrinkage at first tumor assessment (− 27%), patient 4 presented with secondary resistance with the occurrence of unique brain metastasis requiring neurosurgery. The molecular analysis of the resected brain metastasis revealed the presence of a L869R ERBB2 mutation, recently described as a mutation of resistance to lapatinib, whereas the two activating ERBB2 mutations present in the primary tumor were not identified [3].

The recent results from the SUMMIT study have shown that clinical benefit from ERBB2 tyrosine kinase inhibitor (neratinib) may be dependent on the type of ERBB2 alteration and type of tumor. For instance, no clinical activity was observed in bladder and colorectal cancers [4]. Of note, the patient with colorectal cancer included in our study had significant tumor shrinkage with a combination of anti-HER monoclonal antibody and of tyrosine kinase inhibitor. Further studies are needed to investigate if genomic determinants of response may differ when a monoclonal antibody is added to a tyrosine kinase inhibitor in ERBB2-mutated patients.

Interestingly, PI3KCA mutations represent one of the most frequent co-alterations identified in ERBB2-mutated tumors. The PI3K/Akt/mTOR pathway has been shown to play a potential important role in resistance to ERBB2 targeting therapy in ERBB2-overexpressing breast cancer. Moreover, preclinical studies indicated that inhibitors of this pathway can act synergistically with trastuzumab in resistant models [5]. This finding should be taken in consideration in ongoing and future trials investigating ERBB2-targeted therapy in ERBB2-mutated tumors.

Unfortunately, targeted therapies suffer from a major limitation, that is, the duration of any observed clinical benefit is invariably limited in length, owing to the relatively rapid acquisition of drug resistance. We report here for the first time a case of a ERBB2 activation loop mutation in a patient with non-amplified ERBB2 mutant colorectal cancer with acquired resistance to trastuzumab combined with lapatinib. The L869R mutation is located within the activation loop of the kinase domain and associated with gain of function activity. This mutation, which was identified in the secondary progressing brain lesion was not present in the primary tumor, was shown to confer resistance to lapatinib in vitro but to be sensitive to second-generation ERBB2/EGFR inhibitor such as neratinib in the clinical setting [6]. Of note, we cannot exclude the possibility that the L869R mutant cells might be already present in a minority of clones of the primary tumor (given tumor heterogeneity and limits in the sensitivity of NGS technologies) and selected under pressure of trastuzumab and lapatinib combination treatment.

The significant mutation rate of ERBB2 in several tumor types and the promising preliminary activity of dual ERBB2 targeting reported here deserved further clinical investigation aiming to demonstrate that ERBB2-mutational driven therapy can improve patient care irrespective of histology.