Contextualizing Causal Genes from Bone Mineral Density GWAS Using Single-Cell Transcriptomics in Diversity Outbred Mice

Genome-wide association studies (GWASs) have advanced our understanding of the genetics of human bone mineral density (BMD), a clinical predictor of fracture risk and osteoporosis. Aside from the identification of causal genes, other difficult challenges to leveraging GWAS include characterizing the roles of predicted causal genes in disease and providing additional functional context, such as the cell type predictions or biological pathways in which causal genes operate. Using single cell transcriptomics (scRNA-seq) can augment the utility of BMD GWAS by linking variants and genes to a cell type context in which these causal genes may drive disease; however, few population-level single-cell transcriptomics data sets have been generated on bone. To address this challenge, we demonstrate the utility of bone marrow–derived stromal cells cultured under osteogenic conditions (BMSC-OBs) in large populations of Diversity Outbred (DO) mice. The BMSC-OB model can be used to generate cell type–specific transcriptomic profiles of mesenchymal lineage cells to inform human genomic studies. By enriching for mesenchymal lineage cells in vitro, coupled with pooling of multiple samples for scRNA-seq preparation and downstream genotype deconvolution, we demonstrate the scalability of the BMSC-OB model for population-level studies. Furthermore, we show that BMSC-OBs are diverse and consist of bone-relevant cells, such as osteoblasts, osteocyte-like cells, marrow adipogenic lineage precursors (MALPs), and cells with characteristics of mesenchymal progenitors. Through the use of scRNA-seq analytical tools, we confirm the biological identities of BMSC-OBs and show that their transcriptomic profiles are similar to cells isolated in vivo. To contextualize BMD GWAS-implicated causal genes and prioritize targets for subsequent investigations, we generated cell type-specific gene co-expression networks (GCNs). Using pseudotime trajectories inferred from the BMSC-OB scRNA-seq data, we identify networks enriched with genes that exhibit the most dynamic changes in expression across trajectories. We discover 21 network driver genes, which are also causal genes that have human BMD GWAS associations that colocalize with expression/splicing quantitative trait loci (eQTL/sQTL). These driver genes, including Fgfrl1 and Tpx2 (along with their associated networks), are predicted to be novel regulators of BMD via their roles in the differentiation of mesenchymal lineage cells. In this work, we showcase the power of single-cell transcriptomics from mouse bone-relevant cells and human BMD GWAS to prioritize genetic targets with potential causal roles in the development of osteoporosis.