Drug Resistance in HER2-Positive Breast Cancer Brain Metastases: Blame the Barrier or the Brain?

Abstract

The brain is the most common site of first metastasis for patients with HER2-positive breast cancer treated with HER2-targeting drugs. However, the development of effective therapies for breast cancer brain metastases (BCBM) is limited by an incomplete understanding of the mechanisms governing drug sensitivity in the central nervous system. Pharmacodynamic data from patients and in vivo models suggest that inadequate drug penetration across the "blood-tumor" barrier is not the whole story. Using HER2-positive BCBMs as a case study, we highlight recent data from orthotopic brain metastasis models that implicate brain-specific drug resistance mechanisms in BCBMs and suggest a translational research paradigm to guide drug development for treatment of BCBMs. Clin Cancer Res; 24(8); 1795-804. ©2018 AACR.

Figure 1. Brain-specific drug resistance mechanisms offer novel therapeutic targets for HER2-positive breast cancer brain metastases

(Left panel) The blood-tumor-barrier associated with BCBMs has increased permeability mediated by desmin-positive pericytes and altered tight junctions. Although this allows trastuzumab to penetrate the brain parenchyma, inhibition of HER2 alone (Right panel, dashed arrow) is counteracted by brain-specific resistance mechanisms (see text for details). (Right panel) Brain-specific drug resistance mechanisms in BCBMs include loss of PTEN expression and activation of the PI3K-AKT-mTOR pathway as well as activation of parallel signaling pathways via neuregulin-HER3 axis. Drug inhibition of these targets (colored bar-headed line) shows promising efficacy in pre-clinical models.

Figure 2. HER2-positive BCBM (DFBM-355) has differential drug sensitivity when implanted in the brain compared to mammary gland of SCID mouse

A) Kaplain-Meier survival curves of DF-BM355/-bearing mice show improved survival by treatment with dual PI3K + mTOR inhibition [BKM120 (30 mg/kg, PO) + RAD001 (7.5 mg/kg, QD)] but not with single agent PI3K, HER2 (lapatinib, 100 mg/kg PO) or mTOR drug inhibition (n=5–18). Adapted from Ni et al, Nat Med 2016. B) When implanted in the mammary gland (MG), DFBM-355 tumor growth inhibition by PI3K+mTOR (BKM+RAD) is equivalent to PI3K+HER2 (BKM+LAP) or HER2+mTOR (LAP+RAD) inhibition. Data shown as mean ± SEM, (n=6–8). Difference tested by two-way ANOVA, followed by Dunnett’s multiple comparison test ** P < 0.01. C) AKT/mTOR gene signature expression is reduced in DFBM-355 implanted in mammary gland compared to the brain. Data are represented as mean ± SD (n = 3–5 per group; difference tested by t-test. P = 0.0004).

Figure 3. DFCI Breast Cancer Brain Metastasis translational workflow

This sample workflow adopted at our institution, Dana Farber Cancer Institute (DFCI) allows for rapid and multi-level analysis of genomic, protein and functional changes (translational research) specific to breast cancer brain metastases. Data from descriptive and functional studies is integrated to identify candidate targets for drug testing in clinical trials (patient care). Images reproduced from Ni et al, Nat Med 2016 and NCI Visuals Online of ‘Treatment Resistant Breast Cancer Cells’ https://visualsonline.cancer.gov/details.cfm?imageid=10574

Figure 4. Model clinical trial scheme testing an investigational drug in patients with HER2-positive BCBMs

A clinical trial testing a rationally chosen drug in patients with HER2-positive BCBMs should consist of two cohorts: Cohort A, an open-label, single-arm, two-stage, phase II cohort: the ‘efficacy cohort’; and, Cohort B, a pre-surgical window cohort: the ‘biomarker evaluation / discovery’ cohort (see text for details). Patients are treated with a trastuzmab (T) backbone in combination with drug X. For Cohort A, the primary endpoint would be CNS objective response rate as measured by RANO-BM criteria. For Cohort B, the primary endpoint would be inhibition of drug-specific pharmacodynamic markers in resected brain tumor tissue. Secondary clinical endpoints will include other pharmacodynamic biomarkers e.g. change in cfDNA with treatment, in addition to clinical outcomes.