Population dynamics modeling reveals that myeloid bias involves both HSC differentiation and progenitor proliferation biases
Apeksha Singh, Dennis J. Chia, Dinesh S. Rao, Alexander Hoffmann
Abstract
ABSTRACT: Aging and chronic inflammation are associated with overabundant myeloid-primed multipotent progenitors (MPPs) among hematopoietic stem and progenitor cells (HSPCs). Although hematopoietic stem cell (HSC) differentiation bias has been considered a primary cause of myeloid bias, whether it is sufficient has not been quantitatively evaluated. Here, we analyzed bone marrow data from the IκB- (Nfkbia+/-Nfkbib-/-Nfkbie-/-) mouse model of inflammation with elevated NFκB activity, which reveals increased myeloid-biased MPPs. We interpreted these data with differential equation models of population dynamics to identify alterations of HSPC proliferation and differentiation rates. This analysis revealed that short-term HSC differentiation bias alone is likely insufficient to account for the increase in myeloid-biased MPPs. To explore additional mechanisms, we used single-cell RNA sequencing (scRNA-seq) measurements of IκB- and wild-type HSPCs to track the continuous differentiation trajectories from HSCs to erythrocyte/megakaryocyte, myeloid, and lymphoid primed progenitors. Fitting a partial differential equations model of population dynamics to these data revealed not only less lymphoid-fate specification among HSCs but also increased expansion of early myeloid-primed progenitors. Differentially expressed genes along the differentiation trajectories supported increased proliferation among these progenitors. These findings were conserved when wild-type HSPCs were transplanted into IκB- recipients, indicating that an inflamed bone marrow microenvironment is a sufficient driver. We then applied our analysis pipeline to scRNA-seq measurements of HSPCs isolated from aged mice and human patients with myeloid neoplasms. These analyses identified the same myeloid-primed progenitor expansion as in the IκB- models, suggesting that it is a common feature across different settings of myeloid bias.