Cell-free circulating chromatin profiling for epigenomic characterization of mechanisms of response and resistance to sacituzumab govitecan in breast cancer.

Abstract

1081 Background: The efficacy of sacituzumab govitecan (SG), a TROP2-directed antibody-drug conjugate (ADC), in hormone receptor-positive/HER2-negative (HR+/HER2-) metastatic breast cancer (mBC) has been demonstrated, yet biomarkers predicting response and resistance remain an unmet clinical need. We applied a novel multimodal epigenomic liquid biopsy assay to characterize tumor-specific transcriptional activation of relevant genes of interest and resistance mechanisms in the phase 2 SACI-IO HR+ trial (NCT04448886). Methods: Baseline plasma samples were collected from patients (pts) with HR+/HER2- mBC enrolled in SACI-IO HR+, which compared SG alone to SG combined with pembrolizumab (SG-pembro). Genome-wide signals from promoters, enhancers and DNA methylation were profiled from 1 mL of plasma from 95 pts, of which 80 met the assay quality control thresholds and ctDNA metrics required for downstream analysis (ctDNA ≥ 0.5%, N SG = 42, N SG-pembro = 38). We used epigenomic and RNA-seq datasets from 23 breast cancer cell lines to train a model to predict TROP2 expression (r = 0.66, P < 0.01) and tested for association with progression free survival (PFS). We used Gene Set Variation Analysis to score samples for HALLMARK gene set activities using gene-proximal epigenomic signals and tested for independent association of those activities with PFS via CoxPH models, with baseline ctDNA fraction included as a known prognostic covariate. Statistical significance was determined based on improved model fit compared to ctDNA alone. Results: Compared to healthy donors, plasma from trial pts was enriched for breast cancer specific signatures such as estrogen response and hedgehog signaling (FDR < 0.05), highlighting the ability of the liquid biopsy platform to extract tumor specific signal. Baseline ctDNA fraction was prognostic in both treatment arms (Hazard Ratio [HR] SG = 0.38, P < 0.01; HR SG-pembro = 0.28, P < 0.01). Conversely, predicted TROP2 expression was not associated with PFS in either treatment arm. In the SG arm, higher activity of pathways such as epithelial to mesenchymal transition and Wnt signaling were associated with shorter PFS (FDR < 0.1), highlighting potential mechanisms of resistance. Gene signatures related to cell cycle such as mitotic spindle and E2F targets were associated with longer PFS (FDR < 0.1). The above pathway associations with PFS were not statistically significant in the SG-pembro arm. Conclusions: This study demonstrates the feasibility of a multimodal epigenomic liquid biopsy platform for non-invasive characterization of therapeutic response and resistance to SG with or without pembrolizumab in HR+/HER2- mBC. By providing real-time insight into transcriptional regulation, this approach may improve patient stratification and guide ADC treatment strategies. Clinical trial information: NCT04448886 .