Integration of single-cell and bulk RNA sequencing reveals divergent T-cell immune circuits and a shared osteo-immune...

Spinal tuberculosis (STB) and Brucella spondylitis (BS) are the two most common infectious diseases of the spine. Despite overlapping clinical features, their immunopathological mechanisms remain poorly defined.

We performed integrated transcriptomic profiling of infected nucleus pulposus tissues from STB and BS patients, combining single-cell RNA sequencing (STB n=3; BS n=3) with bulk RNA sequencing (STB n=7; BS n=5). Single-cell analysis resolved immune heterogeneity and T-cell subset architecture, whereas bulk RNA-seq defined global transcriptional signatures. Cross-platform integration identified concordant differentially expressed genes and shared pathways. AUCell was applied to quantify T-cell functional programmes. Pseudotime trajectory reconstruction, regulon inference, and cell-cell communication analyses were performed to delineate immune circuit differences.

A total of 67,274 high-quality cells were analysed, identifying nine major cell populations, with T cells as a dominant immune component. STB tissues were enriched in inflammatory cells, including monocytes/macrophages, neutrophils, and osteoclasts, whereas BS showed higher proportions of B and T cells. In STB, CD4 T cells were skewed toward IL26-CD4 T, Cytotoxic CD4 T, Proliferating CD4 T, and Treg subsets, while CD8 T cells were enriched in IFNG-AS1-CD8 T and exhaustion-like populations, reflecting sustained IFN-driven activation with proliferative and apoptotic stress. In contrast, BS exhibited expansion of GZMB-CTLs and IFNG-CTLs subsets, consistent with a focused cytotoxic programme with preserved memory features. Integrated single-cell and bulk analyses demonstrated convergence of overlapping genes on TCR, NF-κB, cytokine signalling, and osteoclast differentiation pathways. Functional scoring and pseudotime analysis indicated broader Th1/Th17 activation and progression toward terminal exhaustion in STB, whereas BS T cells maintained intermediate effector-memory states. Regulatory network analysis revealed a STAT1/IRF-dominated interferon axis in STB and an IRF4/EOMES/RFX3-centred effector-memory programme in BS. Cell-cell communication mapping showed an MHC-II-CD4 axis predominating in STB and an MHC-I-CD8 axis in BS. Despite divergent immune circuits, both conditions converged on osteoclast differentiation pathways, suggesting immune-bone metabolic coupling as a shared mechanism of vertebral destruction.

STB and BS exhibit pathogen-specific T-cell immune circuits that converge on a common osteo-immune pathway, providing mechanistic insight and potential targets for host-directed therapy.