Retinoic acid and proteotoxic stress induce myeloid leukemia progenitors cell death overcoming the protective effects of the bone marrow niche mesenchymal cells

Background: Acute myeloid leukemia (AML) patients bearing the ITD mutation in the tyrosine kinase receptor FLT3 (FLT3-ITD), have poor prognosis and high risk of relapse after initial remission. FLT3-ITD is retained in the endoplasmic reticulum (ER) and generates intrinsic proteotoxic stress. We recently set up a therapeutic strategy combining low doses of the differentiating agent retinoic acid (R), the ER stress inducer tunicamycin (T) and the oxidative stress inducer arsenic trioxide (A), that exerted strong cytotoxic activity on FLT3-ITD+ AML cell lines and primary blasts isolated from patients, but, importantly, not on normal hematopoietic progenitors treated ex vivo. Objectives: In order to facilitate possible translational applications, here we substituted T with the proteasome inhibitor bortezomib (B) to induce ER stress and alteration of proteostasis. We evaluated AML cells sensitivity to low doses of RBA combination in comparison with the effects of each drug alone. Since it is well established that the bone marrow niche favours AML cells survival in response to treatments, leading to therapy resistance, we analyzed the impact of bone marrow stromal cells (BMSCs) on AML responsiveness to the treatment, in vitro and in vivo. Last, we investigated the molecular pathways involved in BMSC-AML crosstalk and possible strategies to overcome niche-mediated AML protection. Methods: We treated FLT3-ITD+ AML cells with low doses of the aforementioned drugs, used alone or in combinations and we investigated the induction of ER and oxidative stress. We then performed the same experiments in an in vitro co-culture system of FLT3-ITD+ AML cells and BMSCs to assess the protective role of the niche on AML blasts. Eventually, we tested the combination of drugs in an orthotopic murine model of human AML. Results: FLT3-ITD+ cell lines and human primary blasts are highly sensitive to the combination RBA but not to the same doses of drugs used as single agents. The cytotoxic effect is mostly due to the generation of oxidative stress. Importantly, BMSCs co-cultured with AML are able to protect leukemic blasts from RBA toxicity by attenuating oxidative stress. However, the use of pharmacological doses of ascorbic acid (Vit C) as an adjuvant pro-oxidant agent tampers with BMSCs protective effects. Indeed, high doses of Vit C have been tested in clinical trials as adjuvant for cancer treatment, showing no general toxicity. Of note, the combination of RBA with ascorbic acid impacts on viability and proliferation of leukemic blasts but not of BMSCs. Intriguingly, BMSC-AML crosstalk is mostly contact-dependent and we observed for the first time that it involves BMSCs actin microfilaments dynamics, mechanotransduction, and the Hippo signalling pathway. Remarkably, this combined approach showed anti-leukemic effects also in vivo. Conclusions: This study demonstrates that FLT3-ITD+ AML cells are sensitive to RBA combination. Importantly, the bone marrow niche can protect leukemic blasts from oxidative stress, but addition of Vit C hampers its protective effect. These results highlight the importance of studying AML in the context of the bone marrow niche and open to possible applications in AML treatment.