A HER2-heterogeneous breast cancer model points to combination targets that may reduce T-DXd resistance
Background
Intratumoral heterogeneity in human epidermal growth factor receptor 2 (HER2) is a factor contributing to resistance to HER2-targeted treatment in HER2-positive breast cancer. Cells with high and low HER2 expression coexist within the same tumor, but preclinical models that reproduce this variation have been limited. This study analyzed resistance to antibody–drug conjugates (ADCs) in human breast cancer models composed of two cell populations and identified candidate targets for combination therapy that may reduce resistance.
Key findings
The researchers isolated cell populations with high (HER2hi) and low (HER2lo) HER2 expression from the 21NT, 21PT, and HCC1954 breast cancer cell lines. Within each cell line, the two populations originated from the same tumor. Fluorescence in situ hybridization (FISH) confirmed ERBB2 gene amplification in HER2hi cells and no amplification in HER2lo cells. HER2lo is a designation for cells in these research models and should not be equated directly with the clinical HER2-low classification.
Experiments using these models and deoxyribonucleic acid (DNA) barcodes to trace cell lineages showed cooperation between the two subpopulations. Both populations tended to grow faster when cultured together than when cultured alone. HER2lo cells were less sensitive to HER2-targeted ADCs, including trastuzumab deruxtecan (T-DXd), but showed sensitivity to HER2 kinase inhibitors. However, drug responses varied by model, and an increase in the HER2lo fraction after T-DXd or neratinib treatment was also reported in patient-derived organoids.
HER2 heterogeneity also affected treatment outcomes in mouse xenografts. After T-DXd treatment, the residual mixed tumors consisted mainly of HER2lo cells and tended to recur earlier after treatment cessation than tumors composed uniformly of HER2hi cells. All mixed tumors recurred within 92 days, whereas 2 of 5 mice bearing uniform tumors remained recurrence-free for more than 140 days after treatment stopped. In the mixed tumors that recurred, HER2hi cells again became dominant, showing that the cellular composition of residual tumors differed from that seen during regrowth.
The researchers conducted a CRISPR gene-knockout screen in heterogeneous co-cultures and selected ABCC1 and USP9X as candidate targets for combination therapy. This does not mean that the genes themselves increased sensitivity: genetic deletion or pharmacological inhibition of these targets increased HER2lo cell sensitivity to T-DXd. ABCC1 inhibition enhanced the growth-inhibitory effect of DXd, the cytotoxic drug carried by T-DXd. USP9X inhibition increased HER2 ubiquitination and trafficking to lysosomes during T-DXd treatment; the researchers interpreted this as a mechanism that promotes drug release from the ADC. In mice, combining each inhibitor with T-DXd enhanced tumor suppression and prolonged the recurrence-free period compared with T-DXd alone.
Significance and outlook
This study explains the functional significance of HER2 heterogeneity and proposes a treatment strategy targeting vulnerabilities in resistant cells. However, the key therapeutic effects were demonstrated in preclinical cell and mouse models. Cell-line models do not fully reproduce the complexity and microenvironment of human tumors, and the findings do not establish improved patient survival or the safety of these combinations. USP9X inhibitor development and evaluation of toxicity to normal tissues remain outstanding, so the potential for clinical application must be assessed in further studies.
Source: Full text of the Cancer Discovery paper
Intratumor heterogeneity for HER2 in HER2-positive breast cancer is a driver of resistance to HER2-targeted therapies. The advancement of treatments for HER2 heterogeneous tumors has been hindered by the lack of preclinical models that accurately mimic the human disease. Here we describe human HER2 heterogeneous breast cancer models composed of ERBB2 amplified (HER2hi) and non-amplified (HER2lo) cell populations derived from the same tumor. Utilizing these models, together with cellular barcoding, we demonstrate subclonal cooperation between HER2hi and HER2lo subpopulations. Furthermore, HER2lo cells drive resistance to HER2-targeting antibody-drug conjugates (ADC) like T-DXd but are sensitive to HER2 kinase inhibitors. CRISPR screens in heterogeneous co-cultures identified sensitizers of HER2lo cells to T-DXd including ABCC1 and USP9X. USP9X inhibition enhances the lysosomal targeting of HER2, thereby potentiating ADC payload release and reducing tumor recurrence after T-DXd treatment. Our results elucidate the functional relevance of HER2 heterogeneity and propose improved therapies for these tumors.
These findings provide evidence that evaluations of HER2-targeted drugs should consider not only the response of the tumor as a whole but also the surviving cell populations and regrowth after treatment stops. In follow-up drug development, mixed HER2hi/HER2lo models could be used to compare T-DXd alone with combinations involving ABCC1 or USP9X inhibition, while tracking tumor size, residual cell composition, and recurrence. The two targets address distinct processes—DXd sensitivity and HER2 trafficking to lysosomes, respectively—providing scope to evaluate the combination candidates separately. Before these findings can inform treatment choices for individual patients, inhibitor selectivity, toxicity to normal tissues, and efficacy in patients require further evaluation.