A potent TOP2 poison with robust preclinical efficacy and a favorable therapeutic window supporting clinical evaluation in pancreatic cancer

Background. Topoisomerase 2 alpha ( TOP2A) is overexpressed in pancreatic ductal adenocarcinoma (PDAC) and its expression strongly correlates with poor prognosis and increased metastatic potential. Several TOP2A targeting drugs, including doxorubicin (DOX) are available, but none represent effective PDAC treatment modalities. Annamycin (ANN), a potent, non-cardiotoxic analog of DOX, but in contrast to DOX exhibits high accumulation in the pancreas and displays robust antitumor activity in orthotopic human PDAC models. It is currently being evaluated as a liposome formulation (L-ANN) in Phase 3 clinical trials in AML patients. Objective. Our objective was to assess the preclinical anticancer efficacy of L-ANN using the syngeneic KPC cell line model for PDAC, and the potential impact of L-ANN on the tumor microenvironment. Methods. In vivo efficacy of L-ANN was assessed in a KPC syngeneic model of PDAC in fully immunocompetent mice (C57BL/6). Tumor progression was measured using bioluminescence imaging (BLI) and ultrasound imaging (USI). ANN levels in plasma and tumors were assessed using LC/MS/MS. Immunohistochemistry analysis of residual tumor tissue were quantified using Aperio Image Scope. Results. The KPC model is derived from genetically engineered mice, in general characterized by profoundly immunosuppressive tumor microenvironment. All mice were injected orthotopically with established KPC cells and showed consistent tumor growth with an average weight of 654.6± 83.8 mg at day 20. However, mice that received 3 weekly L-ANN injections at 4 mg/kg (IV, starting day 7) had significantly reduced tumors (103.5±34.1 mg), with 3 mice exhibiting complete regression. Tumor weight correlated with BLI and USI data and confirmed the dramatic inhibition of tumor growth in the L-ANN cohort. High levels of ANN were measured in residual tumor tissue 1h post injection (5.8±0.5 µg/kg) and were roughly 7-fold higher than the average plasma levels (0.83±0.1 µg/ml). Ex vivo analysis revealed increased levels of DNA damage in residual tumors when compared to vehicle-treated mice (pH2A gamma 35.3±5.4% vs. 5.8±0.3%, respectively). Immune profiling of the tumors revealed significant infiltration of CD8+ and CD4+ cells only in L-ANN treated mice, suggesting their contribution to the antitumor activity of ANN. T cell ablation experiments warrant verification of the role played by CD8+ T cells in the antitumor response. Conclusion. Our preclinical studies demonstrate robust antitumor activity of L-ANN in different PDAC models, including prior human PDAC models and now an implantable syngeneic immunocompetent KPC model, that correlates with high tumor uptake of L-ANN. Along with low toxicity prolife of L-ANN (e.g. lack of cardiotoxicity) established in ongoing clinical trials, these data substantiate the clinical evaluation of L-ANN in PDAC patients.