EPIPROPER

RATIONALE
Acquired resistance to therapy is common across cancer types and leads to disease relapse. Accumulating evidence suggests that non-genetic mechanisms strongly contribute to therapy failure. We hypothesize that tumors contain epigenetically-defined, drug-tolerant persister cells (PCs) that are predisposed to survive treatment and fuel disease recurrence. Using triple negative breast cancer (TNBC) as a paradigm, we propose:
(i) to identify prospective PCs in treatment-naïve tumors, characterize their epigenetic state with a focus on histone and DNA modifications, and uncover molecular changes that drive adaptive behaviors upon treatment;
(ii) to leverage this knowledge to define predictive biomarkers of patient response to treatment;
(iii) to uncover vulnerabilities of PCs that can be exploited to inhibit disease recurrence.

APPROACHES
(i) We will combine analysis of clinical samples and pre-clinical models to define robust epigenetic signatures of PCs at multiple stages during therapy, with a focus on early responses. As a first strategy, we will employ single-nuclei-based methods to profile selected histone modifications in PCs directly in TNBC patients, comparing residual tumors following neo-adjuvant chemotherapy with matched treatment-naive tumors. In parallel, we will use next-generation molecular barcoding and a suite of orthogonal -omics methods to comprehensively characterize histone and DNA modifications of prospective PCs in xenograft models. This approach relies on CRISPR-based molecular barcode readers that enables the identification and isolation of prospective PCs a posteriori from untreated tumors, as well as monitoring and perturbation of PC fates over time.
ii) We will test the predictive value of defined PC epigenetic signatures using multiple biobanks available in our organizations, focusing on treatment-naïve samples with associated information on response to treatment and disease progression. We will also use EPINUC, a non-invasive method for sensitive and inexpensive detection of single-molecule epigenetic modifications from plasma samples. The availability of independent patient cohorts will aid the identification of reliable biomarkers of drug tolerance.
iii) Using genetic and pharmacological perturbation in preclinical models, we will identify actionable epigenetic targets in PCs, which can be exploited to prevent or to target residual disease.

EXPECTED RESULTS AND IMPACT
By dissecting epigenetic intratumor heterogeneity in TNBC prior to and during therapy, we will identify heritable traits that predispose subsets of tumor cells to survive treatment, as well as adaptive changes that fuel disease relapse. Our ability to identify and isolate prospective PCs from treatment-naive tumors will allow us to define, for the first time, predictive biomarkers that can guide clinicians in the management of patients, with major implications for personalized disease monitoring.