Transcriptional Hallmarks of Drug Tolerance in Hormone-Dependent Cancers

Drug tolerance is a key step allowing the development of drug resistance in hormone-dependent breast cancer. While transcriptional heterogeneity has been correlated with tolerance, the transcriptional programs specifically driving it remain poorly understood. To address this, we profile and computationally mine 300’000 single-cell transcriptomes following drug adaptation in models of breast and prostate cancer, spanning seven different treatments (endocrine therapies and CDK-inhibitors), and validate these via integration with novel and published profiles of lesions treated with neo-adjuvant endocrine therapies. We identify a comprehensive spectrum of transcriptional programs and upstream regulators, correlated to different stages of drug tolerance, and conserved across first- and second-line therapies, and therapeutic approaches. Pre-existing programs cooperate with adaptive ones, with the former being more conserved across therapies. These newly identified programs suggest fine-tuning of existing cell states, rather than cell-type or lineage switches, as a main driver of tolerance and resistance. Notably, JAK-STAT signalling emerges as one of the critical regulators of these programs in both the early and late stages of drug tolerance and serves as a predictor of progression in primary hormone-dependent tumours treated with endocrine therapies.