AML Differentially Alters Fate and Function of Mesenchymal Stem- and Osteolineage Cells in the Leukemic Niche

Acute Myeloid Leukemia (AML) alters the composition of the supportive bone marrow (BM) stroma, leading to a leukemia-supportive microenvironment that promotes extrinsic chemotherapy resistance. We previously reported functional and gene expression changes in patient derived mesenchymal stem cells (MSC) that included the increased secretion of chemoprotective cytokines, sustained in the absence of AML cells through serial passage ex vivo . Other groups demonstrated altered DNA methylation patterns in patient MSCs that correlated with increased extrinsic drug resistance in vitro . However, a more rigorous analysis of changes in stromal fate and function requires longitudinal- including pre diagnostic- observations, impossible to obtain from patients and further complicated by patient-patient heterogeneity as well as expansion culture induced in vitro artifacts. Here, we developed a protocol for flow-cytometric isolation and culture- free analysis of stringently purified endosteal MSCs (CD45^-, CD31^-, CD51⁺, SCA-1⁺) and osteoblastic progenitor cells (OPCs; CD45^-, CD31^-, CD51⁺, SCA-1^-). This provides a robust platform in our established AML NSG xenograft model (using Molm14, HL60 or U937 cell lines) for a direct analysis of two principle mesenchymal populations known to be involved in leukemia drug resistance. At a leukemic burden of 60-90% in the BM, we initially observed a significant decline in the proportion of OPC in xenograft-bearing mice compared to healthy controls. We hypothesized that this could be caused by a block in osteogenic differentiation in the MSCs, or an increase in turnover of the OPCs. Surprisingly, we found an increase in osteoblastic differentiation in MSCs derived from xenograft-bearing mice, evidenced by strong induction of Osterix, Runx2, and other regulators of osteogenesis. We also found a concomitant increase in Annexin V and phosphorylated P53 within the OPC population, indicating significant apoptotic turnover, whereas no apoptosis in MSCs was observed. It was recently demonstrated that AML blasts can induce an unfolded protein response (UPR) in bystander cells. Here, we found that while both MSCs and OPCs demonstrate increased expression of Bip, Chop, and Xbp1, OPCs appear to be more sensitive to AML-induced UPR stress, an observation entirely consistent with their increased secretory function. These studies also provide a potential mechanism to explain experimental and clinical observations of osteoblastic cell loss as an aspect of disease progression in the AML BM. The induction of UPR stress in the expanding MSCs in the AML niche along with prior evidence of persistent functional changes led us to determine if an epigenetic mechanism was involved. First, in a candidate screen we discovered that AML significantly dysregulates the expression of key regulators of DNA methylation patterning in MSCs. Notably, the de novo methyltransferases Dnmt3a and 3b, as well as Tet2, were significantly upregulated in MSCs from all three AML xenografts. This highly reproducible in vivo observation prompted us to perform whole genome bisulfite sequencing on MSCs derived from our Molm-14 xenograft model, revealing substantial changes between AML xenografts and control MSCs in the DNA methylation signatures of CpG islands within annotated promoter regions. Gene ontology analyses of these promoter regions identified significant AML induces changes in apoptotic, UPR stress, and differentiation programs, consistent with both our transcriptional profiling and the change in MSC/ OPC composition in the niche. In sum, this work provides a link to explain prior studies in patients and demonstrates the power of xenograft models to experimentally dissect the complex and dynamic relationship between leukemia and discrete stromal populations in the AML niche. We provide direct evidence, free of tissue culture bias, that AML effects changes in the BM niche that differentially affect two key cellular components responsible for extrinsic drug resistance and relapse in AML.