Single-cell and spatial transcriptomics characterisation of RSPO2⁺ nucleus pulposus cells reveals a WNT/FN1-CD44...

Intervertebral disc degeneration (IVDD) is a common cause of chronic low back pain, imposing a significant economic and physiological burden on individuals and society worldwide. Although dysregulation of the WNT/β-catenin pathway is considered an important factor contributing to the dysfunction of nucleus pulposus (NP) cells and degradation of the extracellular matrix, the mechanisms by which specific subgroups of NP cells are activated and the maintenance of excessive activation of specific pathways remain unclear.

We combined single-cell RNA sequencing and spatial transcriptomics with human disc specimens, a rat tail-compression model using static 1.5 MPa loading, an independent rat needle-puncture time-course model, primary NP-cell experiments using continuous static hydrostatic pressure (0.2 MPa for 24 h), WNT-pathway perturbation, molecular docking and dynamics, and targeted siRNA. Computational analyses were interpreted at the cell/spot level, whereas animal-level inference used the stated biological replicates.

Single-cell transcriptomics identified R-spondin 2 (RSPO2) as a selective marker of a homeostatic NP-resident Cluster 1 compartment whose proportion contracted after mechanical injury, while RSPO2-related WNT/FN1-CD44 signalling scores increased across expanded degenerative effector populations. Human panels were used as representative cross-sectional comparisons without inferential between-grade testing. In NP cells, 0.2 MPa pressure and exogenous RSPO2 enhanced WNT/β-catenin, FN1-CD44, matrix-catabolic and apoptotic responses; IWR-1 and DKK1 attenuated these changes. Molecular docking, 100-ns dynamics and a forward co-immunoprecipitation were consistent with an RSPO2-LGR4 association. RSPO2 worsened degeneration in the 1.5 MPa rat compression model, whereas combined IWR-1 treatment was protective. Spatial transcriptomics provided descriptive maps across the needle-puncture time course, and machine-learning models showed internal spot-level discrimination of the prespecified RSPO2-associated labels.

RSPO2 is a candidate regulator, rather than only a marker, of an IVDD-associated NP state. The perturbation data support an RSPO2-associated WNT amplification program coupled to FN1-CD44/MMP3 activation, but the proposed ordering remains provisional pending independent-siRNA and rescue validation.

This study identifies an RSPO2-associated NP state and a candidate WNT/FN1-CD44 programme linked to IVDD. The findings provide a testable basis for future WNT-directed intervention studies while recognising the limits of acute rat models and cross-species data.