Combined anti-CSF-1R and anti-TIM-3 overcome macrophage-mediated mechanisms of PARP inhibitor (PARPi) resistance in BRCA1-associated triple negative breast cancer

Background: Poly (ADP-Ribose) polymerase inhibitors (PARPi) have improved outcomes for BRCA -associated triple negative breast cancer (TNBC); however, resistance develops, resulting in lack of durable responses. Our prior work demonstrated that PARPi activates cGAS/STING signaling, driving CD8+ T-cell recruitment, essential for tumor clearance. These studies led to clinical trials testing PARPi plus immune checkpoint blockade (ICB). However, clinical trials demonstrated no benefit compared to PARPi monotherapy, indicating T-cells in the tumor microenvironment (TME) are inhibited. We previously showed that PARPi induce suppressive tumor-associated macrophages (TAMs) which contribute to PARPi resistance. Removing TAMs with anti-CSF-1R therapy significantly enhanced overall survival (OS) when combined with PARPi and is now being tested in clinical trials (NCT03604692). Here, we test if ICB can enhance the CSF-1R +PARPi combination in both treatment naïve and PARPi-resistant tumors. Methods: Mice bearing naïve or PARPi-resistant BRCA1-deficient TNBC ( K14-Cre;Brca1 f/f ;Trp53 f/f ) tumors were treated with PARPi ± CSF-1R inhibition (CSF-1Ri) ± ICB and followed for tumor size and OS. Flow cytometry was employed to define immune mechanisms of response. Results: In PARPi-naïve tumors, PARPi + CSF-1Ri significantly increased OS compared to PARPi monotherapy. The combination of PARPi + CSF-1Ri + anti-PD-1 led to a modest improvement in efficacy compared to PARPi + CSF-1Ri treatment. In contrast, the addition of anti-TIM3 to PARPi + CSF-1Ri resulted in durable therapeutic benefit, with a significant increase in OS compared to PARPi + CSF-1Ri. The combination of PARPi + CSF-1Ri + anti-TIM-3 significantly increased CD8+ T-cell infiltration, granzyme B production, and reduction in Tregs indicating an increased anti-tumor immune response. Additionally, anti-TIM-3 induced a pro-inflammatory phenotype in TAMs. In PARPi-resistant tumors TAMs expressed higher levels of CSF-1R compared to PARPi-naïve tumors. Additionally, PARPi-resistant tumors had increased infiltration of Tregs, and higher expression of PD-1 on T-cells, strongly indicating an increased immunosuppressive TME. In PARPi-resistant tumors, anti-CSF-1R therapy restored PARPi efficacy and significantly increased OS. Furthermore, addition of anti-TIM-3 to the anti-CSF-1R + PARPi combination significantly increased OS in PARPi-resistant tumors, whereas anti-PD-1 did not enhance therapy efficacy. Conclusion: PARPi-resistant tumors have increased T-cell exhaustion and infiltration of immunosuppressive TAMs. Importantly, our data shows that targeting TAMs through the CSF-1R axes can overcome acquired PARPi resistance, which can be further enhanced with anti-TIM3 therapy, defining a novel strategy addressing a critical unmet medical need.