Acute Myeloid Leukemia Cells Acquire Chemo-Resistance By Inducing Osteoblast Differentiation in Mesenchymal Stem Cells through up-Regulation of RUNX2

The leukemia bone marrow micro-environment (BME) is comprised of the endosteal and vascular niches, provides vital support for cellular growth and conveys drug resistance to leukemia cells. Several reports suggest that mesenchymal stem/stromal cells (MSCs) present in the bone marrow niche induce cell survival and anti-apoptotic proteins in acute myeloid leukemia (AML) cells and protect them from chemotherapy. The mechanisms underlying BME-mediated chemo-resistance however have not been fully elucidated. Here, we hypothesize that AML cells induce functional changes and prime MSCs to protect leukemia cells from chemotherapy. To test our hypothesis, we have compared age matched (between 40-60 years) bone marrow derived MSCs from AML patients (AML-MSC, n=10) and normal (N-MSC, n=10) individuals and analyzed their proliferation, cell surface phenotype, multi-lineage differentiation and chemo-protection potential. AML-MSCs are phenotypically different, with their polygonal morphology and larger cell size compared to N-MSCs which are elongated and spindle shaped appearance. The average cell doubling time of AML-MSCs is 52±8hrs compared to 34±6hours for N-MSCs during their exponential growth phase (p<0.01). Cell surface phenotyping by flow cytometry revealed that most of the markers known to be expressed on N-MSCs including CD105, CD90, CD73, CD51, CD44, SUSD2, CD106, CD140b, CD140a, CD106 and CD271 were also expressed on AML-MSCs at similar levels. Interestingly, tissue non-specific alkaline phosphatase (TNAP, clone W8B2), a cell surface protein highly expressed in naïve-MSCs and osteoblast progenitors (Battula VL et al., Haematologica, 2009) was 10-14 fold higher in AML- as compared to N-MSCs. Since TNAP is also a osteoblast specific marker, we compared osteoblast differentiation potential of N- vs AML-MSCs. Surprisingly, a dramatic increase in alkaline phosphatase activity (by BCIP/NBT substrate) was observed in AML-MSCs even without induction of osteoblast differentiation. mRNA analysis by qRT-PCR revealed that osteoblast specific genes including osteopontin, TNAP, osteocalcin, and osterix were 5-10 fold up-regulated in AML-MSCs compared to N-MSCs before induction. In N-MSCs, the expression of these markers was induced only under osteoblast differentiation conditions. These data indicate that AML-MSCs are primed to differentiate into-osteoblasts. Adipocyte differentiation was assessed by Oil-Red O staining for lipid droplets and revealed a > 95% reduction (p<0.0001) in the number mature adipocytes in AML-MSCs compared to N-MSCs suggesting that AML-MSCs lack the ability to differentiate into adipocytes. To understand the mechanism inducing osteogenic specific differentiation of AML-MSCs, we performed mRNA expression analysis of genes that regulate this process. We found RUNX2, a transcription factor that induces osteogenic but inhibits adipogenic differentiation, was 4-5 fold increased in AML-MSCs compared to N-MSCs. To validate these observations, we co-cultured N-MSCs in the presence or absence of OCI-AML3 cells for 3-5 days and FACS sorted the MSCs for gene expression analysis. We observed a 3-4 fold up-regulation of TNAP protein expression by flow cytometry and 4-6 fold up-regulation of osteoblast specific markers including osteopontin, alkaline phosphatase and osterix in MSCs co-cultured with OCI-AML3 cells. In addition, RUNX2 was up-regulated in MSCs when co-cultured with OCI-AML3 cells. These data suggest that AML cells induce osteogenic differentiation in BM-MSCs by up-regulation of RUNX2. To identify the clinical significance of these observations, we examined the ability of AML- and N-MSCs to protect AML cells from chemotherapy. Co-culture of OCI-AML3 cells with either AML- or N-MSCs and treatment with Cytarabine revealed a 15±4.5% increase in the number of live leukemia cells when co-cultured with AML-MSCs compared to N-MSCs. These data indicate that AML-MSCs protect leukemia cells better from chemotherapy than normal MSCs. In conclusion, AML cells induce osteogenic differentiation in MSCs through up-regulation of the RUNX2 transcription factor. Increased chemo-protection of AML cells by AML-MSCs suggests a prominent role of these cells in AML relapse. Targeting RUNX2 and thereby inhibition of osteoblast differentiation of MSCs may provide enhanced treatment options for AML therapy.