Bone Marrow Stromal Cell Remodeling Is a Common Feature of Diverse Fibrotic Myeloproliferative Neoplasm Models

Myeloproliferative neoplasms (MPN) are blood cancers initiated by driver mutations that transform hematopoietic stem cells. MPN exhibit gross pathologic bone marrow (BM) stromal remodeling, including damaging myelofibrotic change that leads to dependence on extramedullary hematopoiesis and more severe clinical diseases. Therapies targeting fibrotic change would have broad appeal in the treatment of these diseases. We previously demonstrated a critical role for malignant myeloid cells in remodeling endosteal mesenchymal stromal cells (MSC) into myelofibrotic osteoblast-lineage cells (OBC) in a model of chronic myelogenous leukemia (CML) driven by BCR/ABL (Schepers et al., Cell Stem Cell, 2013). In a separate study in a fibrotic MPN model driven by Jak2^V617F, neuropathy and nestin-positive MSC cell death were found critical to disease progression but their involvement in myelofibrosis was not investigated (Arranz et al. Nature. 2014). Our goal is to characterize the type of BM stromal remodeling occurring in non-CML MPN models driven by various mutations and representing a spectrum of disease severity and fibrosis. This includes a minimally fibrotic transgenic Jak2^V617F alone model (Jak2^V617F model, Xing et al., Blood, 2008) and more advanced fibrotic models driven by MPL^W515L expression (MPL^W515L model, Pikman et al., PLoS Med, 2006) or combined transgenic Jak2^V617F expression with conditional deletion of the polycomb gene EZH2 (Jak2^V617F/EZH2^-/- model, Sashida et al., JEM, 2016). We found common blood and BM hematopoietic changes in all three models, including thrombocytosis and expansion of myeloid-biased multipotent progenitor BM cells and confirmed the degree of fibrosis using picrosirius red staining of bone sections. Both MPL^W515L and Jak2^V617F/EZH2^-/- heavily fibrotic models demonstrated inhibition of total endosteal MSC, OBC and endothelial cell (EC) numbers during disease development - in most cohorts a greater than 50% decrease in absolute stromal cell numbers was found. In addition, we observed that whole BM cells from Jak2^V617F/EZH2^-/-mice contained a significantly lower number of totalfibroblast colony forming cells (CFU-F). In co-culture experiments designed to measure direct MSC remodeling induced by malignant cells, both MPL^W515L and Jak2^V617F/EZH2^-/- BM cells inhibited healthy endosteal MSC colony formation over time. In contrast, we found no inhibition of stromal cell numbers or co-culture MSC growth in the minimal fibrotic Jak2^V617F model. In initial experiments measuring rare central marrow perivascular MSC, we found reduced LepR⁺ MSC (Ding et al., Nature, 2012) in both MPL^W515L and Jak2^V617F/EZH2^-/- long bone sections using immunofluorescence. Our results show that fibrotic development in non-CML MPN inhibits stromal cell numbers and function likely via direct effects of malignant hematopoietic cells. This is in contrast to fibrotic CML development where myelofibrotic endosteal stromal cells are expanded. This difference could be partly explained by the type and localization of fibrosis in these various models. The CML model has focal endosteal collagen-I fibrosis which is heavily reliant on osteoblast remodeling, while the MPL^W515L and Jak2^V617F/EZH2^-/- models have more diffuse reticulin central marrow fibrosis which may be produced through a process of stromal cell senescence or differentiation. Overall, this study underscores that a “one size fits all“ approach to understanding myelofibrosis is insufficient. To tease out these differences, we are examining qualitative and quantitative changes in additional central marrow MSC populations, including PDGFR⁺, Sca-1⁺ and Gli-1⁺ MSC, during MPN development as well as assaying the molecular mediators of stromal remodeling. Our long-term goal is to identify therapies that can restore a more normal BM stroma and support healthy hematopoiesis in MPN.