Large-scale DNA sequencing of tumor genomes has enabled the identification and targeting of oncogenic drivers in individual tumors, helping to fulfill the promise of precision medicine. Currently, there are more than 30 promising genomic drug targets in many cancer types and the list continues to grow (Cell 2017;168(4):584-599). It is anticipated that genomic profiling will soon become routine in clinical cancer care, expanding the number of clinically validated predictive biomarkers and enabling drug discovery efforts targeted to particular mutations.
From these recent genomic sequencing efforts, scientists discovered a new way by which tumors can activate the HER2 (ERBB2) oncogene: by DNA sequence mutation rather than amplification. Three decades ago, HER2 gene amplification was linked to approximately 20 percent of breast cancers (Science 1987;235(4785):177-182), resulting in the clinical development and FDA approval of a handful of HER2 inhibitors.
Trastuzumab, the first approved HER2 inhibitor, has cured thousands of patients with early-stage breast cancer and significantly prolonged the lifespan of patients with HER2-amplified metastatic breast cancer (N Engl J Med 2005;353(16):1659-1672, N Engl J Med 2015;372(8):724-734). It is now appreciated that the HER2 gene is mutated (point mutations or small in-frame insertions/deletions) in approximately 3 percent of all cancers, including breast, lung, and colorectal cancers (Clin Cancer Res 2006;12(1):57-61, Oncologist 2015;20(1):7-12). Many of these are “hotspot” mutations that are recurrently identified in tumor samples. The majority of the recurrent HER2 mutations activate HER2 kinase activity and HER2-mediated signaling (Cancer Discov 2013;3(2):224-237). HER2 is a member of the ErbB receptor tyrosine kinase family, which also includes EGFR, HER3, and HER4. HER2 relays signal transduction through oncogenic pathways such as the phosphoinositide 3-kinase (PI3K)/AKT/mTOR and RAS/RAF/MEK/ERK pathways, resulting in cell proliferation and cell survival.
In breast cancer, HER2 mutations are predominantly found in HER2-nonamplified (so-called HER2-negative) tumors. Therefore, most patients with HER2-mutant breast cancer are not standardly treated with anti-HER2 therapies. However, this is beginning to change, as there are now several clinical trials investigating HER2 inhibitors specifically in patients with HER2-mutant cancers. Preclinical studies have shown that irreversible EGFR/HER2 tyrosine kinase inhibitors (TKIs) such as afatinib and neratinib block mutant HER2 function (Cancer Discov 2013;3(2):224-237). In addition, recent clinical studies suggest that some HER2-mutant, nonamplified breast cancers respond to neratinib, suggesting that HER2 mutations are indeed drivers of tumor progression (Can Res 2017; doi:10.1158/1538-7445.SABCS16-PD2-08, Clin Cancer Res 2017;23(19):5687-5695).
Characterization of Novel HER2 Variant
We recently reported the case of a 54-year-old woman with estrogen receptor-positive (ER+), HER2-nonamplified metastatic lobular breast cancer progressing on endocrine therapy and chemotherapy (Cancer Discov 2017;7(6):575-585).
Next-generation sequencing identified the novel HER2L869R kinase domain mutation in a skin metastasis. We searched the literature and tumor sequencing databases (cBioPortal, Project GENIE, COSMIC, Foundation Medicine, and Guardant Health) for this mutation and found the HER2L869R/Q mutation in 21 additional tumors, 16 of which were breast cancers. The frequency with which HER2L869R was detected suggests that it is selected for during oncogenesis and not merely a “passenger” mutation. This mutation is homologous to EGFRL861R/Q, a known activating mutation in lung cancer, and BRAFV600E, a well-characterized oncogenic driver mutation in melanoma, suggesting that the HER2L869R mutation probably leads to activation of the HER2 protein.
To characterize the novel HER2 mutation, we performed structural modeling, biochemical studies, and cell proliferation/oncogenesis assays in noncancerous human mammary cells (MCF10A cells) that we transduced with the mutated HER2 gene. We found that HER2L869R expression increased HER2-mediated signal transduction and growth of colonies in 3-dimensional matrix (an in vitro assay mimicking tumor growth) compared to cells expressing wild-type (WT) HER2. Importantly, this gain-of-function phenotype was blocked completely by neratinib.
Based on these laboratory studies, the patient was enrolled in the SUMMIT clinical trial (NCT01953926) and exhibited a rapid and sustained partial response to neratinib (and later, the combination of neratinib and the ER antagonist fulvestrant). More recently, in a similar clinical trial, another patient harboring HER2L869R-mutant breast cancer was treated with neratinib and exhibited stable disease that lasted more than 7 months (Clin Cancer Res 2017;23(19):5687-5695). Together, these data suggest that the HER2L869R mutation is an oncogenic driver in breast cancer and can be blocked by neratinib.
While initial responses to targeted therapies are often impressive, unfortunately, advanced cancers are rarely cured by targeted therapies alone.
Acquired resistance almost always develops, and this patient was no exception. After 16 months of treatment with neratinib, the patient progressed with new skin metastases. Using next-generation DNA sequencing circulating tumor DNA (ctDNA) from the patient's blood, we detected a secondary HER2 mutation, the HER2T798I mutation, in addition to the HER2L869R mutation. Deep sequencing of the patient's ctDNA at the time of study enrollment or during the first 9 months of neratinib therapy failed to detect the HER2T798I mutation, confirming it was acquired during therapy.
HER2T798I is homologous to other well-characterized TKI-resistant “gatekeeper” mutations, most notably EGFRT790M, the most common cause of acquired resistance to first-generation EGFR TKIs in lung cancer (PLoS Med 2005;2(3):e73). This led us to hypothesize that, like EGFRT790M, HER2T798I would sterically block drug binding.
We performed computational modeling of the structure of HER2T798I bound to neratinib and found that the increased bulk of the isoleucine amino acid relative to the threonine would indeed result in steric hindrance, blocking the access of neratinib to the ATP-binding pocket of HER2. In contrast to cells expressing HER2WT or L869R, neratinib failed to block HER2 phosphorylation, HER2-mediated signaling, or HER2-induced cell growth in cells expressing HER2T798I, confirming that this mutation mediates resistance to neratinib.
Overcoming Neratinib Resistance
We next asked whether other EGFR/HER2 TKIs could block HER2T798I. We profiled the covalent EGFR/HER2 inhibitor afatinib, the mutant EGFR-selective inhibitor osimertinib (both FDA-approved to treat EGFR-mutant lung cancers), and the osimertinib metabolite AZ5104.
Afatinib and AZ5104, but not neratinib or osimertinib, blocked HER2 phosphorylation, HER2-induced signaling, and cell growth of cells expressing HER2T798I. Since afatinib is smaller than neratinib, it does not extend as deeply into the ATP binding pocket of HER2, and therefore is not expected to be significantly affected by the T798I mutation. AZ5104 is not being developed independently of osimertinib. Therefore, we propose that afatinib is worthy of clinical investigation in neratinib-resistant breast cancers harboring the HER2 gatekeeper mutation.
Following our report, HER2T798I was detected in the plasma of another patient with HER2-mutant (non-amplified) breast cancer with acquired resistance to neratinib (Clin Cancer Res 2017;23(19):5687-5695). We anticipate that as more patients with acquired resistance to neratinib are profiled, this mutation will be observed more frequently.
Recently, neratinib was FDA-approved to treat HER2-amplified breast cancer patients. We found that expression of the HER2 gatekeeper mutation rendered HER2 insensitive to neratinib in HER2-amplified breast cancer cells. To determine if HER2T798I promotes acquired resistance in HER2-amplified breast cancers, these patients' tumors should be profiled at progression.
Interestingly, the HER2T798I mutation was only detected in the patient's ctDNA, and not from a biopsy of the skin metastasis at the time of neratinib progression, suggesting that HER2T798I was acquired in only a subset of the metastases. It is probable that multiple different acquired mutations were driving drug resistance in different metastatic sites. This finding supports the idea that ctDNA sequencing is more comprehensive than tissue sequencing; it can offer a snapshot of mutations in all tumor subclones and metastatic sites at a given time. I can envision a future in which ctDNA profiling is used during treatment with targeted therapies to detect acquired resistance prior to clinical progression and the mutations that cause resistance. These acquired mutations, in turn, can then be targeted with therapy to overcome initial resistance.
Precision Medicine Era
In conclusion, our study showed that HER2L869R is a driver mutation in breast cancer that can be initially blocked by neratinib treatment. However, acquisition of the HER2T798I gatekeeper mutation promotes acquired resistance to neratinib by sterically blocking inhibitor binding. The acquisition of a second-site mutation in HER2 that prevents inhibitor binding strongly suggests that the tumor is highly dependent on mutant HER2 activity. Afatinib overcomes HER2T798I-induced resistance, probably because it is smaller than neratinib. Our results suggest that HER2-mutant patients with acquired resistance to neratinib should undergo molecular profiling at the time of progression. Patients harboring HER2T798I may then respond to other irreversible HER2 TKIs such as afatinib.
Future drug discovery efforts should focus on agents that selectively block mutant HER2, including the gatekeeper mutation (analogous to the lung cancer drug osimertinib, which selectively targets mutant EGFR), thereby sparing the considerably toxicity associated with inhibiting WT HER2 and EGFR (Expert Opin Drug Saf 2017;16(10):1111-1119).
We are now entering an era in which precision medicine can be used to target not only specific genes, but also specific single-nucleotide mutations within that gene. Precision medicine is getting more precise, ultimately leading to safer and more effective treatments for cancer patients.
ARIELLA B. HANKER, PHD, is a research instructor, Department of Medicine, Division of Hematology/Oncology in the Vanderbilt-Ingram Cancer Center at Vanderbilt University Medical Center, Nashville, Tenn.
Have Stories Delivered Directly to You
Sign up for our free e-newsletter and receive email notifications when each issue becomes available and when other new content is posted online. New to Oncology Times? Visit the publication's website at www.oncology-times.com where you can get free access to archived issues, subscribe to the print magazine, and create your own personal online article collections.