A Specific Mesenchymal Stem and Progenitor Cell (MSPC) Subpopulation with a Multi-Potent Gene Signature Is Transcriptionally Altered in the Setting of Myelodysplastic Syndrome (MDS) in Primary Human Bone Marrow Aspirates

Introduction: Myelodysplastic syndromes (MDS) are genetically diverse and heterogeneous diseases characterized by dysplasia and cytopenias and a dismal prognosis of ˂ 50% overall survival at 5-years. In vitro and in vivo experimental data have shown that the bone marrow microenvironment (BMME) may play a role in disease progression and, importantly, in murine models, transplantation of MDS to a healthy BMME was shown to mitigate disease. Mesenchymal Stem and Progenitor Cells (MSPCs), as part of the BMME, give rise to multiple non-hematopoietic progenitor cells and provide essential support to hematopoietic stem cells. MSPCs have also been shown to be functionally altered in patients with myeloid neoplasias. Murine studies have demonstrated distinct roles for specific subsets of bone marrow mesenchymal cells in myeloid malignancies. We hypothesized that specific subsets of human bone marrow MSPCs will play differential roles in the pathogenesis of MDS. Using flow-cytometry, high-throughput sequencing and gene set enrichment analysis (GSEA), we characterized human MSPCs in the bone marrow aspirates from patients with MDS and normal healthy controls (NBM).

Methods: Mononuclear cells were isolated by iliac crest bone marrow aspiration from 10-healthy donors and 14-patients at the University of Rochester Medical Center. Within the non-hematopoietic (CD45^-, CD235^-), non-endothelial (CD31^-) bone marrow compartment, we enriched and isolated 3 distinct-subpopulations of MSPCs based on cell-surface expression of CD271/NGFR, CD106/VCAM-1 and CD146/MCAM. Populations were defined as follows: CD271⁺/CD146^- (CD271⁺), CD271⁺/CD146⁺/CD106⁺ (DPCD106⁺), and CD271⁺/CD146⁺/CD106^- (DPCD106^-). RNA-seq analysis was performed on each subpopulation to define transcriptional signatures (TS) and gene set enrichment patterns. Statistically significant differentially expressed genes (DEGs) were defined by fold-change ≥ ±1 and p-value ˂0.05.

Results: We first set out to define differences in the TS and interrogate the function of MSPC populations in NBM. Principle component analysis (PCA) demonstrated the highest variance between the DPCD106+ and the CD271+ populations. The number of DEGs were also highest between the DPCD106⁺ and CD271⁺ populations (n=3,619 genes). GSEA identified 745 and 336 gene sets with positive enrichment in the DPCD106+ and CD271+ group, respectively, and illustrated that the DPCD106⁺ population was significantly enriched in gene sets involved in early embryonic developmental and "stem-like" pathways whereas the CD271⁺ population was enriched in cell cycling, DNA and chromosomal organization.

In the setting of MDS, the mean relative frequency of MDS CD271⁺ nearly tripled (0.4230/0.1445; p-value 0.045). Compared to NBM, MDS DPCD106+ cells had the highest variance by PCA and the highest number of DEGs (n=560). GSEA identified 19-gene sets with significant enrichment in the MDS DPCD106+ group and, intriguingly, 12 (63%) were identical to gene sets enriched in the CD271+ group. Furthermore, of the 560 DEGs in the MDS DPCD106+ MSPCs, 300 were upregulated and, of those, 160 (53%) were identical to upregulated genes in the CD271+ NBM group including the acquisition of a proliferative signature. Altogether, this data suggests a switch in the TS of theDPCD106+ population in the setting of MDS. Importantly, this TS clustered MDS DPCD106+ from NBM, regardless of MDS risk category.

Conclusion: We successfully characterized 3 subtypes of MSPCs in NBM and MDS. In NBM, we demonstrate that cell surface expression of CD271, CD146 and CD106 defined the most stem-like TS within the non-hematopoietic, non-endothelial bone marrow compartment. In the setting of MDS, the increase in population frequency of the CD271⁺ cells and the concomitant transcriptomic aberrations observed in the MDS derived DPCD106⁺ population support the hypothesis that specific MSPC populations have differential roles in MDS pathogenesis. Further, we identify a TS that discriminates MDS derived MSPCs from NBM irrespective of MDS-risk category. This suggests that alterations within specific MSPC populations may represent a unifying pathway in disease pathogenesis despite heterogeneity and genetic drivers intrinsic to the MDS clone. Thus, targeting the BMME represents a potentially novel therapeutic strategy aimed at mitigating disease and restoring normal hematopoiesis in patients with MDS.