Integrative Single-Cell Transcriptomic, Mendelian Randomization and In Silico Perturbation Analyses Prioritize MUC20 as a Candidate Gene Associated with Osteoporosis and Metabolic Dysfunction-Associated Steatotic Liver Disease in the Liver–Bone Axis

Metabolic dysfunction-associated steatotic liver disease (MASLD) and osteoporosis (OP) are epidemiologically linked, but shared cell-type-specific molecular features remain unclear. We integrated public single-cell/single-nucleus transcriptomic datasets for OP (GSE147287) and MASLD (GSE289173) with two-sample Mendelian randomization (MR), colocalization, network annotation, macrophage-focused in silico perturbation, and exploratory serum assessment. After quality control, 13,753 OP cells and 42,438 MASLD cells/nuclei were analyzed. Macrophages were consistently identified in both datasets and showed disease-associated expansion. Directionally concordant macrophage differentially expressed genes yielded 147 shared candidate genes, with enrichment mainly involving lipid/sterol metabolism, extracellular matrix and adhesion processes, immune presentation, lysosomal processing, and phagocytic pathways. MR prioritized MUC20 as the only candidate with concordant odds ratios greater than 1 for both OP (OR = 1.044, 95% CI 1.003–1.086) and MASLD (OR = 1.111, 95% CI 1.038–1.189). Colocalization supported shared genetic signals for MUC20 in OP (PP.H4 = 0.855) and MASLD (PP.H4 = 0.816). In silico perturbation suggested a limited but pathway-enriched predicted transcriptional footprint. Serum MUC20 was higher in patients with OP+MASLD than in healthy controls. This integrative analysis identified shared macrophage-associated transcriptional themes and prioritized MUC20 as a candidate molecule for future liver–bone crosstalk studies.