Search

Background Metastatic breast cancer (MBC) is incurable, with a 5-year survival rate of 28%. In the USA, more than 42,000 patients die from MBC every year. The most common type of breast cancer is estrogen receptor-positive (ER+), and more patients die from ER+ breast cancer than from any other subtype. ER+ tumors can be successfully treated with hormone therapy, but many tumors acquire endocrine resistance, at which point treatment options are limited. There is an urgent need for model systems that better represent human ER+ MBC in vivo, where tumors can metastasize. Patient-derived xenografts (PDX) made from MBC spontaneously metastasize, but the immunodeficient host is a caveat, given the known role of the immune system in tumor progression and response to therapy. Thus, we attempted to develop an immune-humanized PDX model of ER+ MBC. Methods NSG-SGM3 mice were immune-humanized with CD34+ hematopoietic stem cells, followed by engraftment of human ER+ endocrine resistant MBC tumor fragments. Strategies for exogenous estrogen supplementation were compared, and immune-humanization in blood, bone marrow, spleen, and tumors was assessed by flow cytometry and tissue immunostaining. Characterization of the new model includes assessment of the human tumor microenvironment performed by immunostaining. Results We describe the development of an immune-humanized PDX model of estrogen-independent endocrine resistant ER+ MBC. Importantly, our model harbors a naturally occurring ESR1 mutation, and immune-humanization recapitulates the lymphocyte-excluded and myeloid-rich tumor microenvironment of human ER+ breast tumors. Conclusion This model sets the stage for development of other clinically relevant models of human breast cancer and should allow future studies on mechanisms of endocrine resistance and tumor-immune interactions in an immune-humanized in vivo setting.

Model systems that recapitulate the complexity of human tumors and the reality of variable treatment responses are urgently needed to better understand cancer biology and to develop more effective cancer therapies. Here we report development and characterization of a large bank of patient-derived xenografts (PDX) and matched organoid cultures from tumors that represent some of the greatest unmet needs in breast cancer research and treatment. These include endocrine-resistant, treatment-refractory, and metastatic breast cancers and, in some cases, multiple tumor collections from the same patients. The models can be grown long-term with high fidelity to the original tumors. We show that development of matched PDX and PDX-derived organoid (PDxO) models facilitates high-throughput drug screening that is feasible and cost-effective, while also allowing in vivo validation of results. Our data reveal consistency between drug screening results in organoids and drug responses in breast cancer PDX. Moreover, we demonstrate the feasibility of using these patient-derived models for precision oncology in real time with patient care, using a case of a triple negative breast cancer with early metastatic recurrence as an example. Our results uncovered an FDA-approved drug with high efficacy against the models. Treatment with the PDxO-directed therapy resulted in a complete response for the patient and a progression-free survival period more than three times longer than her previous therapies. This work provides valuable new methods and resources for functional precision medicine and drug development for human breast cancer.Graphical Abstract

Background Targeted therapies for triple-negative breast cancer (TNBC) are limited; however, the epidermal growth factor receptor (EGFR) represents a potential target, as the majority of TNBC express EGFR. The purpose of these studies was to evaluate the effectiveness of two EGFR-targeted antibody-drug conjugates (ADC: ABT-414; ABBV-321) in combination with navitoclax, an antagonist of the anti-apoptotic BCL-2 and BCL-XL proteins, in order to assess the translational relevance of these combinations for TNBC. Methods The pre-clinical efficacy of combined treatments was evaluated in multiple patient-derived xenograft (PDX) models of TNBC. Microscopy-based dynamic BH3 profiling (DBP) was used to assess mitochondrial apoptotic signaling induced by navitoclax and/or ADC treatments, and the expression of EGFR and BCL-2/XL was analyzed in 46 triple-negative patient tumors. Results Treatment with navitoclax plus ABT-414 caused a significant reduction in tumor growth in five of seven PDXs and significant tumor regression in the highest EGFR-expressing PDX. Navitoclax plus ABBV-321, an EGFR-targeted ADC that displays more effective wild-type EGFR-targeting, elicited more significant tumor growth inhibition and regressions in the two highest EGFR-expressing models evaluated. The level of mitochondrial apoptotic signaling induced by single or combined drug treatments, as measured by DBP, correlated with the treatment responses observed in vivo. Lastly, the majority of triple-negative patient tumors were found to express EGFR and co-express BCL-XL and/or BCL-2. Conclusions The dramatic tumor regressions achieved using combined agents in pre-clinical TNBC models underscore the abilities of BCL-2/XL antagonists to enhance the effectiveness of EGFR-targeted ADCs and highlight the clinical potential for usage of such targeted ADCs to alleviate toxicities associated with combinations of BCL-2/XL inhibitors and systemic chemotherapies.

Aim Breast cancer is the second leading cause of cancer death in women, yet only a few antigens are available for directing immune therapies to a patient’s breast tumor. In this study we establish a new system for identifying tumor-specific antigens presented by the class I HLA of breast tumors. Our approach relies on generating a breast cancer HLA peptide-ligand reference database coupled with isolation of peptides from the HLA of patient-derived breast cancer xenografts (PDX). This HLA PDX approach can rapidly identify targets for existing and emerging cancer immune therapies. Methods A tumor-specific HLA reference peptide ligand library was constructed using 3 breast epithelial tumor cell lines and 1 healthy breast epithelial line. Cells were transfected to secrete HLA-A∗02:01, secreted HLA was affinity purified, HLA-bound peptides were eluted, separated by HPLC, and HLA-A∗02:01 breast cancer peptides were analyzed by mass spectroscopy. To validate the peptides as relevant targets for human breast tumors in vivo, HLA was extracted from PDX-derived tumors and the presence of specific peptides were confirmed in these primary tumors. Results In 20 PDX tumor lines, 7 of the tumors were A∗02:01, half of which expressed detectable A∗02:01. When evaluated for candidate peptide:HLA expression, the HCI-015 triple negative PDX xenograft expressed high levels of the Macrophage Inhibitory Factor 19–27 (MIF) peptide, a peptide that is expressed solely in breast cancer cell lines and primary tumors but not healthy tissues. We are currently expanding HCI-015 PDX for additional HLA extractions. Conclusions These data suggest that PDX tumors can be used as surrogates to human tumors for the direct identification of tumor-specific antigens. The use of a peptide library generated from several breast cancer cell lines allows peptides to be identified with increased sensitivity. Furthermore, expression of the candidate MIF 19-27 peptide on HCI-015 provides compelling evidence that certain tumor-antigens are conserved among cell lines, tumors, and in vivo models of breast cancer. Such antigens provide a unique therapeutic window that can be exploited for targeted immunotherapies.