Single cell heterogeneity and evolution of breast cancer bone metastasis and organoids reveals therapeutic targets for precision medicine

Bone metastasis (BoM) is a major cause of morbidity and mortality from breast cancer.1Harbeck N. Penault-Llorca F. Cortes J. et al.Breast cancer.Nat Rev Dis Primers. 2019; 5: 66Crossref PubMed Scopus (1358) Google Scholar BoM treatment is hindered by tumor evolution and heterogeneity,2Aftimos P. Oliveira M. Irrthum A. et al.Genomic and transcriptomic analyses of breast cancer primaries and matched metastases in AURORA, the breast international group (BIG) molecular screening initiative.Cancer Discov. 2021; 11: 2796-2811Crossref PubMed Scopus (67) Google Scholar with limited clinically relevant models to test and prioritize treatments.3Weilbaecher K.N. Guise T.A. McCauley L.K. Cancer to bone: a fatal attraction.Nat Rev Cancer. 2011; 11: 411-425Crossref PubMed Scopus (953) Google Scholar Furthermore, heterogeneity between unique BoMs within a patient may drive treatment resistance.4Wu J.M. Fackler M.J. Halushka M.K. et al.Heterogeneity of breast cancer metastases: comparison of therapeutic target expression and promoter methylation between primary tumors and their multifocal metastases.Clin Cancer Res. 2008; 14: 1938-1946Crossref PubMed Scopus (178) Google Scholar To address these challenges, we investigated BoM evolution and heterogeneity in a case of primary estrogen receptor-positive/progesterone receptor-negative/human epidermal growth factor receptor 2-negative (ER+/PR−/HER2−) invasive lobular breast carcinoma (ILC), which during adjuvant letrozole treatment progressed to the left pelvis (BoML) and right tibia (BoMR) (Supplementary Figure S1A, available at https://doi.org/10.1016/j.annonc.2022.06.005). The primary tumor, bilateral BoM, and BoM patient-derived organoids (PDOs) were subjected to bulk DNA/RNA and single-cell RNA sequencing (scRNAseq). Histopathology demonstrated evolution of disease from ER+ primary ILC to ER− BoM with mixed lobular/ductal carcinoma features (Figure 1A). DNAseq revealed 286 somatic substitutions gained in BoM compared with the primary tumor, with 169 shared between BoMs, 19 unique to BoML, and 98 unique to BoMR (Figure 1B, Supplementary Table S1, available at https://doi.org/10.1016/j.annonc.2022.06.005). Clinically actionable mutations included PI3K-E545K (NM_006218) and BRCA1-D1834H (NM_007300) (Supplementary Figure S1B and C, available at https://doi.org/10.1016/j.annonc.2022.06.005) with BRCA1-D1834H gained in the BoM. These mutations were readily detected in circulating free DNA (Supplementary Figure S1D, available at https://doi.org/10.1016/j.annonc.2022.06.005). RNAseq revealed shared and unique transcriptomic evolution in bilateral BoM, including loss of luminal and gain of basal/HER2 features (Supplementary Figure S1E, available at https://doi.org/10.1016/j.annonc.2022.06.005), up-regulation/reduction of expression of actionable genes frequently altered in BoMs5Priedigkeit N. Watters R.J. Lucas P.C. et al.Exome-capture RNA sequencing of decade-old breast cancers and matched decalcified bone metastases.JCI Insight. 2017; 2e95703Crossref PubMed Scopus (81) Google Scholar including EPHA3, ROS1, and PTPRD (Figure 1C), and enhanced cancer hallmark pathways including PI3K-AKT-mTOR, angiogenesis, epithelial to mesenchymal transition (EMT), and androgen response (Figure 1D). scRNAseq of BoMs identified six cell populations (Figure 1E), with BoML showing more stroma cells and BoMR containing more epithelial cells (Supplementary Figure S1F, available at https://doi.org/10.1016/j.annonc.2022.06.005). Six epithelial subpopulations were identified (Figure 1F), with BoML having more EMT and BoMR containing more luminal/EGFR and proliferative cells (Supplementary Figure S1G, available at https://doi.org/10.1016/j.annonc.2022.06.005), consistent with RNAseq data (Figure 1D). EMT cells showed the highest hypoxia signature, and their feature genes were mostly regulated by PRRX/TWIST (Supplementary Figure S1H and I, available at https://doi.org/10.1016/j.annonc.2022.06.005). Nine fibroblast clusters featuring diverse functions were identified (Supplementary Figure S1J, available at https://doi.org/10.1016/j.annonc.2022.06.005). Consistent with bulk RNAseq showing higher hypoxia/angiogenesis signatures (Figure 1D), BoML had more mCAF/vCAF compared with BoMR (Supplementary Figure S1K, available at https://doi.org/10.1016/j.annonc.2022.06.005). BoMs showed comparable immune cell types (Supplementary Figure S1L and M, available at https://doi.org/10.1016/j.annonc.2022.06.005), but BoML contained more naive T/B cells whereas BoMR had more regulatory T cells (Tregs)/macrophages. Expression of immune checkpoints and ligands were detected in CD8+ T/natural killer/Treg cells (Supplementary Figure S1N, available at https://doi.org/10.1016/j.annonc.2022.06.005). In summary, multiomics profiling uncovered potential treatments, highlighting PI3K and BRCA1 for both BoMs, EMT, angiogenesis and programmed cell death protein 1/programmed death-ligand 1 for BoML, and androgen response and proliferation for BoMR. PDOs were derived from BoML and BoMR. scRNAseq demonstrated that PDOs retained the subclonal heterogeneity of BoM epithelial cells (Figure 1G) and showed highly correlated gene expression with BoM (Supplementary Figure S1O, available at https://doi.org/10.1016/j.annonc.2022.06.005). PDOs preserved >90% of somatic mutations and COSMIC mutational signatures of BoM (Supplementary Figure S1P and Q, Supplementary Table S2, available at https://doi.org/10.1016/j.annonc.2022.06.005). Consistent with the PI3K-E545K and BRCA1-D1834H mutations, both PDOs showed robust in vitro growth inhibition in response to alpelisib and talazoparib (Figure 1H and I) and sensitivity to alpelisib with PDO left as a murine xenograft (Figure 1J). In conclusion, bulk and single cell profiling of bilateral BoMs, primary tumor, and PDO revealed intratumor heterogeneity and evolution, and therapeutic opportunities for precision medicine. SO and AVL are Hillman Fellows. We would like to thank UPMC Genome Center, University of Pittsburgh HSCRF Genomics Research Core, and the University of Pittsburgh Center for Research Computing for supporting this study. This work was supported by the Breast Cancer Research Foundation (no grant number, to AVL and SO); Susan G. Komen Scholar awards [grant numbers SAC110021 to AVL, SAC160073 to SO]; the Metastatic Breast Cancer Network Foundation (no grant number, to SO); the National Cancer Institute [grant number R01 CA252378 to SO/AVL]; Magee-Women’s Research Institute and Foundation, Nicole Meloche Foundation, and the Shear Family Foundation (no grant numbers). The China Scholarship Council and Tsinghua University provided financial support for FC. This project used the Pitt Biospecimen Core/UPMC Hillman Cancer Center Tissue and Research Pathology Services supported in part by National Institutes of Health grant award [grant number P30CA047904]. Funding was in part provided by the Institute for Precision Medicine of the University of Pittsburgh and UPMC (no grant number).