A Hyperactive Signalosome Results in High Sensitivity to HSP90 Inhibitors in AML

HSP90 is well established in supporting tumorigenesis by stabilizing oncogenic client proteins. Given this crucial role, a number of HSP90 inhibitors have been tested in various types of cancer, including leukemia. However, clinical trials thus far revealed only a subset of AML patients benefited from the treatment. Therefore, precision medicine approaches to define parameters that predict the patients' response to HSP90 inhibitors are needed to select patients who are most likely to benefit.

We have previously demonstrated that PU-H71, a novel purine scaffold HSP90 inhibitor with selectivity for a tumor-specific HSP90 and currently translating into Phase 2 clinical evaluation, is capable of ablating malignant blasts, progenitor and stem cells in AML patient samples using in vitro studies. We found that leukemia cell lines (n=18) and primary AML patient samples (n=26) with greater numbers of simultaneously activated signaling networks, including PI3K-AKT and JAK-STAT, were the most sensitive to HSP90 inhibition. Using different genetic models, our studies revealed that diverse oncogenic transformations that converge upon simultaneous hyperactivation of PI3K-AKT and JAK-STAT promote sensitivity to PU-H71.

To validate the efficacy of PU-H71 in vivo, we generated AML-GFP-luciferase xenograft models using cell lines with hyperactive signalosome. Xenotransplanted mice were treated with PU-H71 one week post-engraftment. In vivo imaging indicated that MOLM-13 xenografted leukemia was rapidly and significantly reduced by PU-H71 treatment. Six doses of PU-H71 produced robust anti-leukemic activity as indicated by in vivo imaging and flow cytometric analysis of post-treatment bone marrow (no disease detected). In addition, we generated 7 AML patient-derived xenografts (PDX) cohorts with samples that displayed varied levels of activation of PI3K-AKT and JAK-STAT signaling pathways. After initial validation that status of the PI3K-AKT and JAK-STAT signaling pathways were preserved in the PDX, we initiated treatment with PU-H71 and found that, as predicted, the AML-PDX with the most hyperactive signalosome were the most sensitive to in vivo treatment to PU-H71. Importantly, samples with hyperactive PI3K-AKT and JAK-STAT signaling also demonstrated a significant reduction in LSC using secondary transplants.

Taken together, we found that a hyperactive signalosome results in increased sensitivity to the HSP90 inhibitor PU-H71 in vitro and in vivo. Our study suggests that evaluation of PI3K-AKT and JAK-STAT signaling pathways may provide a means to select patients who are most likely to benefit from HSP90 inhibitory therapy.