Aberrant cytoplasmic localization of MLH1 characterizes a sub-clonal breast cancer cell population that seeds recurrence

Estrogen receptor positive (ER+) breast cancer is one of the most common causes of cancer-related death in women. Mortality is largely driven by recurrence of treatment-resistant disease after many years of apparent response, making the molecular events that cause recurrence a critical area of investigation. Loss of expression of MLH1, a tumor suppressor best studied in its role in mismatch repair, induces resistance of ER+ breast cancer cells to standard estrogen-targeting therapies. It does so by delinking cell cycle progression from estrogen regulation, a role distinct from its function in mismatch repair. MLH1 loss, as currently clinically diagnosed by detecting genomic instability or by immunohistochemistry for absence of protein, occurs in 12-15% of all cancers. Here, we demonstrate that sub-clonal, patient-derived mutations in MLH1, which neither impact protein abundance nor contribute sufficiently to genomic instability to be detected diagnostically, seed endocrine treatment resistance by enabling estrogen-independent growth in vitro, ex vivo in patient-derived organoids (p=0.005) and in vivo (p=0.0001). The mechanism underlying this endocrine treatment resistance is aberrant localization of MLH1 to the cytoplasm in vitro and in vivo (p=0.04), which precludes cell cycle arrest in response to endocrine therapy while simultaneously rendering cells acutely dependent on CDK4/6 activity. Consequently, administration of CDK4/6 inhibitors causes extreme regression in cells with cytoplasmic MLH1 compared to control cell populations with nuclear localization of MLH1 in vitro (p=0.00000009), ex vivo (p=0.01) and in vivo (p=0.01). As aberrant cytoplasmic localization occurs in an additional ~12% of ER+ breast cancer patients, it constitutes a new, major contributor to MLH1 dysregulation. The potential applicability of cytoplasmic MLH1 as a predictor of responsiveness to existing targeted therapies in a hard-to-treat breast cancer subtype posits an update of current clinical diagnostic criteria and therapeutic strategies. This is particularly important in the adjuvant setting where identification of biomarkers predicting responsiveness to CDK4/6 inhibitors remains an urgent, unmet clinical need. ### Competing Interest Statement The authors have declared no competing interest.