Genetic Anchoring of Stromal Cellular Programs Relevant to Ligamentum Flavum Remodeling in Lumbar Spinal Stenosis

Background/Objectives: Lumbar spinal stenosis (LSS) has a substantial inherited component, but the cellular contexts in which genetic susceptibility acts remain incompletely resolved. We integrated large-scale human genetics with a donor-resolved spinal-ligament single-cell resource and an external ligamentum flavum (LF) expression dataset to localize inherited susceptibility and prioritize regulatory candidates. Methods: We reconstructed a nine-donor spinal-ligament single-cell resource and derived donor-aware Broad Stromal and degenerative-LSS-associated Stromal transcriptional programs. These programs were evaluated using stratified LD score regression (S-LDSC) against an LSS genome-wide association dataset of 40,303 cases and 741,468 controls, followed by gene-based association, fine-mapping, target-aware molecular colocalization, donor-robustness analyses, and external LF tissue directional sensitivity. Results: The Broad Stromal annotation was the strongest and only multiple-testing-significant positive annotation in the primary prespecified one-at-a-time S-LDSC analysis (coefficient = 6.48 × 10−8, p = 0.0031, BH-FDR = 0.0245), whereas the disease-associated Stromal program was not significantly enriched. Donor-aware Stromal pseudobulk analysis identified 106 differentially expressed genes at FDR < 0.05, with high leave-one-donor-out directional stability. Cross-layer integration prioritized six genes. CRLF1 showed the most complete candidate-specific regulatory convergence, including an eight-variant credible set and GTEx skeletal-muscle CRLF1 colocalization at rs2074177 (PP.H4 = 0.974), while MMP14 showed stronger disease-associated differential expression and a narrower two-variant credible set. Conclusions: Inherited LSS susceptibility preferentially localizes to a broad spinal-ligament Stromal cellular/regulatory context rather than establishing genetic mediation of the observed disease-associated transcriptional state. CRLF1 and MMP14 represent complementary candidates for tissue-specific functional validation.