Roles of circulating proteins in osteoporosis pathogenesis: Insights from Mendelian randomization.

BACKGROUND

Osteoporosis (OP) poses a significant health burden. Circulating proteins represent promising drug targets. However, previous Mendelian randomization (MR) studies have limited evidence for their causal role due to single-source protein quantitative trait loci (pQTL) data and lack of systematic, multi-layered validation.

OBJECTIVE

To identify and prioritize circulating proteins causally implicated in OP using a proteome-wide MR framework integrating five pQTL datasets, Bayesian colocalization, transcriptome-wide analyses, and single-cell expression mapping.

METHODS

A two-stage (discovery and replication) proteome-wide MR analysis was performed using cis-pQTLs from five large-scale studies (n = 2968 unique proteins). Associations with four bone mineral density (BMD) traits (total body, femoral neck, lumbar spine, and forearm) were assessed and validated in three independent OP case-control datasets. Significant proteins underwent Bayesian colocalization (coloc.abf and coloc.susie), summary-data-based MR (SMR), and heterogeneity in dependent instruments (HEIDI) tests. A systematic evidence-tiered framework (Tiers 1-3) was applied. Single-cell RNA sequencing mapped cell-type-specific expression of prioritized genes.

RESULTS

Seven circulating proteins showed significant MR associations with BMD. LRP4 was identified as a Tier 1 (high-confidence) target, demonstrating robust colocalization (PPH4 = 0.915) and passing SMR/HEIDI criteria. TNFSF11 and IBSP were classified as Tier 2 (moderate-confidence) targets (TNFSF11: PPH4 = 0.635; IBSP showed significant MR associations but lacked colocalization evidence). Remaining proteins (GCKR,ANPEP,ANGPTL7,LYAR) were categorized as Tier 3 (low-confidence). Single-cell analysis revealed enriched expression of TNFSF11, LRP4, and IBSP in bone marrow mesenchymal stem cells.

CONCLUSION

This study provides a rigorous, multi-layered prioritization of circulating proteins causally implicated in OP. LRP4 emerges as the strongest candidate, with TNFSF11 moderately supported. These findings offer prioritized targets for future drug development and experimental validation.