Refining the identity of mesenchymal cell types associated with murine periosteal and endosteal bone.

Single-cell RNA-seq has led to novel designations for mesenchymal cells associated with bone as well as multiple designations for what appear to be the same cell type. The main goals of this study were to increase the amount of single-cell RNA sequence data for osteoblasts and osteocytes, to compare cells from the periosteum to those inside bone, and to clarify the major categories of cell types associated with murine bone. We created an atlas of murine bone-associated cells by harmonizing published datasets with in-house data from cells targeted by Osx1-Cre and Dmp1-Cre driver strains. Cells from periosteal bone were analyzed separately from those isolated from the endosteum and trabecular bone. Over 100,000 mesenchymal cells were mapped to reveal 11 major clusters designated fibro-1, fibro-2, chondrocytes, articular chondrocytes, tenocytes, adipo-Cxcl12 abundant reticular (CAR), osteo-CAR, preosteoblasts, osteoblasts, osteocytes, and osteo-X, the latter defined in part by periostin expression. Osteo-X, osteo-CAR, and preosteoblasts were closely associated with osteoblasts at the trabecular bone surface. Wnt16 was expressed in multiple cell types from the periosteum but not in cells from endocortical or cancellous bone. Fibro-2 cells, which express markers of stem cells, localized to the periosteum but not trabecular bone in adult mice. Suppressing bone remodeling eliminated osteoblasts and altered gene expression in preosteoblasts but did not change the abundance or location of osteo-X or osteo-CAR cells. These results provide a framework for identifying bone cell types in murine single-cell RNA-seq datasets and suggest that osteoblast progenitors reside near the surface of remodeling bone.

RANKL-mediated osteoclast formation is required for bone loss in a murine model of Staphylococcus aureus osteomyelitis.

Staphylococcus aureus osteomyelitis leads to extensive bone destruction. Osteoclasts are bone resorbing cells that are often increased in bone infected with S. aureus. The cytokine RANKL is essential for osteoclast formation under physiological conditions but in vitro evidence suggests that inflammatory cytokines may by-pass the requirement for RANKL. The goal of this study was to determine whether RANKL-dependent osteoclast formation is essential for the bone loss that occurs in a murine model of S. aureus osteomyelitis. To this end, humanized-RANKL mice were infected by direct inoculation of S. aureus into a unicortical defect in the femur. Mice were treated with vehicle or denosumab, a human monoclonal antibody that inhibits RANKL, both before and during a 14-day infection period. The severe cortical bone destruction caused by infection was completely prevented by denosumab administration even though the bacterial burden in the femur was not affected. Osteoclasts were abundant near the inoculation site in vehicle-treated mice but absent in denosumab-treated mice. In situ hybridization demonstrated that S. aureus infection potently stimulated RANKL expression in bone marrow stromal cells. The extensive reactive bone formation that occurs in this osteomyelitis model was also reduced by denosumab administration. Lastly, there was a notable lack of osteoblasts near the infection site suggesting that the normal coupling of bone formation to bone resorption was disrupted by S. aureus infection. These results demonstrate that RANKL-mediated osteoclast formation is required for the bone loss that occurs in S. aureus infection and suggest that disruption of the coupling of bone formation to bone resorption may also contribute to bone loss in this condition.

A framework for defining mesenchymal cell types associated with murine periosteal and endosteal bone.

Single-cell RNA sequencing has led to numerous novel designations for mesenchymal cell types associated with bone. Consequently, there are now multiple designations for what appear to be the same cell type. In addition, existing datasets contain relatively small numbers of mature osteoblasts and osteocytes and there has been no comparison of periosteal bone cells to those at the endosteum and trabecular bone. The main goals of this study were to increase the amount of single cell RNA sequence data for osteoblasts and osteocytes, to compare cells from the periosteum to those inside bone, and to clarify the major categories of cell types associated with murine bone. To do this, we created an atlas of murine bone-associated cells by harmonizing published datasets with in-house data from cells targeted by Osx1-Cre and Dmp1-Cre driver strains. Cells from periosteal bone were analyzed separately from those isolated from the endosteum and trabecular bone. Over 100,000 mesenchymal cells were mapped to reveal 11 major clusters designated fibro-1, fibro-2, chondrocytes, articular chondrocytes, tenocytes, adipo-CAR, osteo-CAR, pre-osteoblasts, osteoblasts, osteocytes, and osteo-X, the latter defined in part by Postn expression. Osteo-X, osteo-CAR, and pre-osteoblasts were closely associated with osteoblasts at the trabecular bone surface. Wnt16 was expressed in multiple cell types from the periosteum but not in any cells from endocortical or cancellous bone. Fibro-2 cells, which express markers of skeletal stem cells, localized to the periosteum but not trabecular bone in adult mice. Suppressing bone remodeling eliminated osteoblasts and altered gene expression in pre-osteoblasts but did not change the abundance or location of osteo-X or osteo-CAR cells. These results provide a framework for identifying bone cell types in murine single cell RNA sequencing datasets and suggest that osteoblast progenitors reside near the surface of remodeling bone.