SITE-1 protease ablation in the osterix-lineage results in bone marrow dysplasia with neutrophilia coupled to osteopenia

Purpose: Site-1 protease (S1P) is a proprotein convertase known primarily for its roles in lipid homeostasis and the unfolded protein response. In previous studies, we demonstrated in mice the importance of S1P to overall skeletal development. A recent study in a human with defective S1P functions coupled to skeletal abnormalities validated the importance of S1P to skeletal development. S1P ablation in chondroprogenitors results in mice with no endochondral bone. S1P ablation in postnatal chondrocytes completely eliminates the primary postnatal growth plate by preventing chondrocyte maturation with a profound effect on trabecular bone growth. Our studies to analyze a direct role for S1P in bone development by ablating S1P in the Osterix (Osx)-lineage demonstrated a vital role in bone development. These mice have osteopenia with a significant reduction in bone formation and mineral apposition rates. Besides having delicate bones, these mice also develop scoliosis very early postnatally. These mice also demonstrated drastically reduced skeletal stem cells in the bone marrow. Furthermore, bone marrow stromal cells were unable to differentiate into osteoblasts in vitro. To investigate the mechanistic link between S1P-ablation, osteopenia and bone marrow anomalies in these mice, we investigated bone marrow-derived cells by single cell sequencing and proteomic analysis. Methods: To generate mice with S1P ablation in the Osx-lineage (Cko), S1Pf/f mice (mice homozygous for the floxed exon 2 of S1P) were bred with S1P+/f;Osx-Cre (Osx1-GFP::Cre) mice. Male and female mice showed identical phenotypes. For proteomic analysis of the bone marrow, non-RBC bone marrow cells harvested from postnatal 21-day old (P21) WT and Cko mice were lysed and analyzed by tryptic digestion and nano-LC/MS/MS technology after tandem mass tag labeling. For single cell RNA sequencing of GFP+ cells from the bone marrow of P21 Cko mice, single GFP+ cells were FACS-sorted and sequenced by RNA-Seq. For a global understanding of the bone marrow, non-RBC, single bone marrow cells were also analyzed by the 10X Genomics platform. Bone marrow cells were also analyzed by FACS using fluorescent-conjugated antibodies to neutrophil and hematopoietic stem cell (HSC) markers. Immunofluorescence for stefin proteins were performed on formalin-fixed, paraffin-embedded mouse hind limbs using a rabbit anti-human Stefin A antibody. Results: LC/MS/MS proteomic analysis of lysates from bone marrow cells harvested from WT and Cko mice showed that the stefin A (Stfa) family of proteins (Stfa1/Stfa2/Stfa3/Stfa2l1) was significantly upregulated in the Cko, when compared to WT. This data suggested a disruption of the bone marrow niche in the Cko. The bone marrow in the Cko is also marked by the presence of GFP+ cells, not seen in WT, heterozygote or control Osx-Cre mice. FACS-sorting of these GFP+ cells from the Cko followed by sequencing revealed an expression profile similar to neutrophils, a hematopoietic cell; furthermore, these GFP+ cells also showed expression of the stefin protein family. As the GFP is expressed as a Cre-GFP fusion protein from the Osx1 promoter, this data indicated that Osx-Cre is also active in hematopoietic cells, rather than only in osteoblast progenitors in the bone marrow. Sequencing of several thousands of WT and Cko bone marrow cells by the 10X Genomics platform revealed that the Cko bone marrow, while similar to WT, was also much altered. tSNE plots identified ten clusters for both WT and Cko with seven clusters exhibiting almost identical expression profiles, and three clusters deviating significantly from each other. One of these three clusters in the Cko showed an expression profile that suggested duplication of the neutrophil population that also expressed additional members of the Stfa gene family, not seen in WT neutrophils. The Cko also had an additional Ly6C-expressing cluster indicative of an increase in other hematopoietic cells such as monocytes or dendritic cells in the Cko bone marrow. The doubling of neutrophils in the Cko bone marrow was confirmed by FACS analysis for CD45 CD11b Ly6G triple-positive markers for neutrophils and was statistically significant. FACS analysis for HSCs showed that the increase in neutrophils is due to an increase in granulocyte-macrophage progenitors (GMP) in the Cko bone marrow. Immunofluorescence analysis for stefin showed very high expression for stefin in the Cko bone marrow in mice as young as P3. Conclusions: Our study shows that S1P ablation in the Osterix lineage results in the disruption of the bone marrow niche. This disruption is manifested as an increase in neutrophils indicating an inflammatory disposition in the Cko bone marrow. The Cko bone marrow also demonstrates an increase in Stfa proteins, which are known cathepsin inhibitors. While stefins are part of the normal expression profile in neutrophils, more members of this family are expressed in Cko neutrophils, with the doubling of the neutrophils contributing to stefin overexpression in the Cko bone marrow. The increase in inflammatory components such as neutrophils via a skewed increase in GMPs may induce an undesirable cross talk between the bone marrow niche and stem cell development pathways, which may be responsible for the decrease in skeletal progenitors in the Cko. How exactly stefin proteins contribute to this scenario needs to be investigated further, but our study indicates that S1P regulates both stefin production and the stem cell pool in the bone marrow and may be required to balance the production of skeletal and hematopoietic stem cell populations. Our study thus suggests that S1P is involved in the mechanistic crosstalk between the osteogenic and hematopoietic developmental programs.