Reverse Phase Protein Analysis of Mesenchymal Stem Cells from AML Patients and Healthy Donors Reveals Distinct Patterns of Protein Expression Reflecting Differences in Senescence, Differentiation and Survival Signaling

The bone marrow microenvironment (BME) critically supports hematopoietic stem cells and protects leukemia cells from chemotherapy, immune surveillance, and related stresses. A critical component of the BME is the mesenchymal stem cell (MSC). Dan Link’s group demonstrated that MSC are essential for human hematopoiesis, particularly as a source of SDF-1, which regulates homing, proliferation, and differentiation of HSC. Moreover, studies from our group and others have demonstrated that MSC protect leukemia cells from chemotherapy. At present, very little is known about MSC derived from AML patients, and an understanding of the proteomic makeup of these cells in the leukemia microenvironment could help to elucidate mechanisms involved in supporting their pro-tumor function. We used reverse phase protein array analysis (RPPA) to compare the expression of 151 proteins in MSC derived from AML BMs (N = 106) with those from healthy donors (N = 71). The expression of 45 of these proteins was deemed significantly different (p < 0.01) between the two sets. AML MSC expressed higher levels of p53 and p21 (CDKN1A), and the expression of the latter was correlated with other proteins within each MSC set. Using beta-galactosidase staining, AML MSC were found to undergo senescence more frequently than normal MSC. Elevated p21 in AML MSC is consistent with this finding. While 15 proteins were positively, and 20 proteins negatively, correlated with p21 expression in normal MSC, there were only three proteins positively, and nine negatively, correlated in AML-derived MSC. In normal MSC, SMAD1 (a key component in MSC growth and differentiation involving multiple receptors like TGF beta and BMP) expression and AKT signaling were low when p21 is expressed. However, in AML MSC this association was not seen, albeit a negative correlation with ITGAL was observed. SMAD1 expression was higher in normal MSC. In normal MSC, the expression of SMAD1 was negatively correlated with PPARG and NPM1, and was positively correlated with the expression of phosphorylated ELK. The opposite relationship was seen in AML MSC (i.e., PPARG and NPM1 exhibited positive correlation with SMAD1 and phosphorylated ELK was negatively correlated with the protein). While the significance of these relationships remains to be determined it is interesting to note that PPARG is a key regulator of adipocyte differentiation in MSC, so perhaps this alteration of SMAD/PPARG in AML MSC could impede their differentiation potential. In an accompanying abstract from our group, we report that AML MSC are primed toward osteoblastic differentiation and do not differentiate into adipocytes (Battula VL et al, ASH 2014). The RPPA data on PPARG is consistent with this finding. SMAD1 also positively regulates miR-21. Since p21 is a miR-21 target, it seems possible that the differences in expression could be attributed to SMAD1 and miR-21 signaling. We analyzed miR-21 expression in normal and AML-derived MSC (N = 10, each) using qRT-PCR and found a statistically significant (p =0.014) increase in its expression in normal MSC relative to their disease counterparts. When anti-miR-21 was transduced into healthy donor MSC, which caused a 3-fold increase in p21 (but no difference in cyclin D1 expression, another miR-21 target whose expression was also increased in AML MSC). AML MSC also exhibited higher protein expression of the B55 alpha subunit (PPP2R2A) of protein phosphatase 2A (PP2A). This expression contrasted interestingly with that of leukemia cells, since we have previously reported low PPP2R2A levels in AML blasts associated with shorter remission durations (Ruvolo et al Leukemia 2011). Furthermore, AKT phosphorylation was negatively correlated with PPP2R2A expression in AML blasts, and normal MSC, but there was no correlation between PPP2R2A and phosphorylated AKT in AML MSC. Also, expression of PPP2R2A was positively correlated with the expression of the survival protein NOL3 (ARC) which may provide new clues to possible survival mechanisms in AML MSC. In summary, these findings represent insights into the proteomic profiling of normal and AML MSC. Results suggest that senescence (via p21), differentiation potential (involving SMAD/PPARG pathway), and survival signaling (including PP2A/AKT) are altered in AML MSC. Studies are underway to determine how these variations in MSC properties impact the AML microenvironment.