A Niche Driven Mechanism Determines Response to ATRA and a Mutation-Independent Therapeutic Approach for MDS and AML

Myeloid cancers such as myelodysplastic syndromes (MDS) and acute myeloid leukemia (AML) remain resistant to standard of care (SOC) and targeted therapies. Here, we identify an oncogenic signal from the niche, constitutive activation of b-catenin/JAG1 signaling in osteoblastic lineage cells, as a mechanism determining response to all-trans-retinoic acid (ATRA), a regimen with disparate results in MDS and AML. B-catenin signaling in the osteoblastic lineage is activated following hypermethylation of its regulators in MDS patients; occurs in ~40% of MDS and AML patients and patients with del(5q)-associated myeloid malignancies; its levels increase with disease severity; and correlate with MDS to AML transformation and a worse overall survival.

By analyzing publicly available RNA-seq data and bone marrow biopsies collected from 14 ATRA-treated patients from four different study sites we found a strong association between inhibition of b-catenin activity in osteoblasts and/or Notch signaling in AML cells with ATRA, leading to AML improvement. An overall response rate of 100% was reached in patients with active baseline osteoblastic b-catenin, including 4 patients with a complete response (CR) and 1 patient with CR without absolute neutrophil counts of more than 10^{^}9^/L (CRn). Among patients without baseline active osteoblastic b-catenin, only 2 patients out of the 9 examined achieved a CR resulting in an overall response rate of 22%. Responsiveness was observed across patients with diverse cytogenetic and mutational profiles belonging to both intermediate and adverse risk groups. Mechanistically, ATRA-mediated inhibition of osteoblastic b-catenin activity suppressed cell growth and survival and promoted differentiation of MDS/AML cells solely from patients with active b-catenin signaling.

Improvement occurs despite a diverse ATRA dosing schedule in combination with other therapeutic agents (AZA, VEN, VPA). The beneficial effect in treatment outcome appears to be dependent on baseline osteoblastic b-catenin activation since, in contrast to its high association with treatment response (Fisher's exact test, p=0.021, n=14), there was no association between response and the degree of global DNA hypomethylation or histone acetylation among patients treated in combination with AZA/VPA, even though acetylation and methylation were decreased upon treatment. Moreover, prior AZA/VEN treatment of an AML patient resulted in relapse and CR was reached only upon the subsequent ATRA addition to AZA/VEN backbone. Supporting this conclusion, treatment with ATRA monotherapy of an MDS patient with high levels of osteoblasts with activated b-catenin improved disease status and was associated with complete inhibition of b-catenin activity in the patient's osteoblasts and a parallel decrease in Notch signaling in the patient's MDS cells. The patient remained transfusion independent for at least 9 years, the entire duration of follow-up. These results suggest that responsiveness to ATRA can be due to inhibition of activated b-catenin signaling in osteoblasts.

Indeed, testing this hypothesis in leukemic mice with constitutive active osteoblastic b-catenin, ATRA administration inhibited b-catenin activity in osteoblasts and improved disease with no evidence of relapse and a superior safety profile to SOC. Inactivation of ATRA receptor, RARA, specifically in osteoblasts of these mice, abrogated the protective effects of ATRA and its inhibitory effect on osteoblastic b-catenin activity demonstrating that ATRA acts through osteoblasts to improve disease progression and survival.

These results provide an explanation for the differential response to ATRA suggesting ATRA repurposing in osteoblastic b-catenin associated myeloid malignancies. A circulating skeletal cell population expressing activated b-catenin, reflecting b-catenin activation status in marrow-resident osteoblastic lineage cells, can serve as a mechanistic biomarker allowing patient stratification and monitoring of treatment response. They also highlight the therapeutic potential of targeting the niche that has the potential to evade relapse and overcome SOC toxicity.

Jurcic:Pfizer: Research Funding; Syros Pharmaceuticals: Consultancy, Membership on an entity's Board of Directors or advisory committees; Incyte: Consultancy; Gallop Oncology: Consultancy, Membership on an entity's Board of Directors or advisory committees; Sumitomo Pharma: Research Funding; Blueprint Medicines: Research Funding; Forma Therapeutics: Research Funding; BMS/Celgene: Research Funding; Rigel Pharmaceuticals: Consultancy. Berman:Novartis: Honoraria. Garcia-Manero:Amphivena: Research Funding; Helsinn: Research Funding; Novartis: Research Funding; AbbVie: Research Funding; Bristol Myers Squibb: Other: Personal fees, Research Funding; Astex: Research Funding; Onconova: Research Funding; H3 Biomedicine: Research Funding; Merck: Research Funding; Curis: Research Funding; Janssen: Research Funding; Genentech: Research Funding; Forty Seven: Research Funding; Aprea: Research Funding; Astex: Other: Personal fees; Helsinn: Other: Personal fees; Genentech: Other: Personal fees. Raza:TFC Therapeutics: Membership on an entity's Board of Directors or advisory committees, Patents & Royalties.