Single-cell and spatial profiling highlights TB-induced myofibroblasts as drivers of lung pathology

Ian Mbano, Nuo Liu, Marc H. Wadsworth, Mark John. Chambers, Thabo Mpotje, Osaretin E. Asowata, Sarah K. Nyquist, Kievershen Nargan, Duran Ramsuran, Farina. Karim, Travis K. Hughes, Joshua D. Bromley, Robert Krause, Threnesan Naidoo, Liku B. Tezera, Michaela Reichmann, Sharie Keanne Ganchua, Henrik N. Kløverpris, Kaylesh J Dullabh, Rajhmun Madansein, Sergio Triana, Adrie J.c. Steyn, Bonnie Berger, Mohlopheni J. Marakalala, Sarah M. Fortune, JoAnne L. Flynn, Paul Elkington, Alex K. Shalek, Alasdair LeslieThe Journal of Experimental Medicine 2025 · DOI

TL;DR

Tuberculosis (TB) causes significant lung damage, but the specific cells driving this pathology are poorly understood. This study uses single-cell and spatial transcriptomics to reveal that a specific type of fibroblast, expressing a "myofibroblast-like" signature (MMP1+CXCL5+), plays a key role in TB-induced lung damage and fibrosis. This finding, along with the identification of its interaction with SPP1+ macrophages, opens new avenues for developing host-directed therapies.

What Problem Does This Paper Address?

Tuberculosis (TB) is a major global health threat, causing millions of deaths annually and leading to significant lung destruction and fibrosis. While the formation of granulomas is a hallmark of TB infection, the precise cellular mechanisms driving this lung damage and the subsequent development of post-TB lung disease (PTLD) remain unclear. Existing animal models often fail to fully recapitulate the complex pathology seen in human TB, particularly the extensive extracellular matrix (ECM) remodeling. This knowledge gap hinders the development of effective host-directed therapies (HDTs) that could complement traditional antibiotic treatments. The authors aim to dissect the cellular and molecular landscape of human TB lung tissue to identify key cell populations and interactions responsible for immunopathology and tissue remodeling.

Methods

The researchers employed a multi-pronged approach combining cutting-edge single-cell and spatial transcriptomics techniques:

  • Single-cell RNA sequencing (scRNA-seq): Lung tissue samples from TB-infected and TB-negative individuals were processed to isolate individual cells. These cells were then sequenced to capture their gene expression profiles, allowing for the identification and characterization of diverse cell types and states. The Seq-Well S³ platform was used for scRNA-seq.
  • Spatial Transcriptomics (10x Visium): This technique allows for the analysis of gene expression directly within the tissue context. By profiling RNA from specific locations on tissue sections, the researchers could map the spatial distribution of cell types and their gene expression patterns, providing crucial insights into cellular organization within TB granulomas. Formalin-fixed paraffin-embedded (FFPE) samples were used for this.
  • Bioinformatic Analyses:
    • Clustering and Subclustering: Unsupervised clustering algorithms were used to group cells based on their gene expression profiles, identifying distinct cell types and subtypes.
  • Differential Expression (DE) Analysis: Genes that were significantly up- or down-regulated between different cell types or conditions were identified.
  • Gene Set Enrichment Analysis (GSEA): This method was used to determine which biological pathways or gene sets were over-represented in specific cell populations or conditions.
  • Weighted Gene Co-expression Network Analysis (WGCNA): This approach was used to identify modules of co-expressed genes within fibroblast populations, revealing potential functional gene networks.
  • Cell-Cell Interaction Analysis (MultiNichenet and LIANA): These tools were used to infer potential communication pathways between different cell types by analyzing the expression of ligands and receptors.
  • Spatial Deconvolution (cell2location): This algorithm was used to estimate the abundance of different cell types within each spatial spot of the Visium data.
  • Validation Techniques:
    • Immunohistochemistry (IHC): Used to visualize the spatial localization of specific proteins within tissue sections, confirming findings from transcriptomic data.
  • Flow Cytometry: Used to quantify the frequency of specific cell populations in lung tissue.
  • Bulk RNA-seq Analysis: Used to validate findings in independent datasets, such as human lymph node granulomas.

Datasets

  • Human Lung Tissue: Samples were obtained from TB-diseased and TB-negative individuals undergoing lung resections.
  • Nonhuman Primate (NHP) Lung TB Granuloma Datasets: Previously published datasets from M.tb-infected NHPs were used for validation and comparison.
  • Human Lung Cell Atlas (HLCA): A large, integrated dataset of human lung cells was used as a reference for cell type annotation and label transfer.
  • Lymph Node (LN) TB Granuloma Bulk RNA-seq Data: An independent dataset of human LN TB granulomas was used for validation.
  • Tuberculin Skin Test (TST) Challenge Data: Bulk RNA-seq data from TST challenges in TB patients were used to assess in vivo responses.

Key Insights and Ideas

  • Fibroblasts as Key Drivers: The study strongly implicates fibroblasts, particularly a subset expressing MMP1 and CXCL5, as major contributors to the lung pathology in TB. This challenges previous assumptions that focused primarily on immune cells.
  • The "TB-Myofibroblast": The identification of a specific myofibroblast-like gene signature (MMP1+CXCL5+) in fibroblasts from TB-diseased lungs is a significant discovery. This specific cell type appears to be associated with more severe disease and higher bacterial burden.
  • Fibroblast-Macrophage Axis: The research highlights a critical cross-talk between these MMP1+CXCL5+ fibroblasts and SPP1+ macrophages within the granuloma cuff. This interaction is proposed to be a key mechanism driving granuloma progression and fibrosis.
  • Spatial Context Matters: The use of spatial transcriptomics was crucial in demonstrating that these key cell populations and their interactions are not randomly distributed but are specifically localized within the granuloma cuff, suggesting a coordinated pathological process.
  • Potential for Host-Directed Therapies: By identifying specific cell populations and their molecular signatures, the study provides promising targets for novel therapeutic interventions that could modulate the host's response to TB, potentially reducing lung damage and improving outcomes.

Key Results

  • 30 Distinct Cell Subsets Identified: scRNA-seq analysis of human TB lung tissue revealed a complex cellular landscape with 30 distinct immune, parenchymal, and stromal cell subsets.
  • Fibroblast Enrichment in TB Lungs: Fibroblasts were found to be enriched in TB-diseased lung tissue, while macrophages were depleted.
  • MMP1+CXCL5+ Fibroblast Association with Severe Disease: This specific fibroblast subset was associated with severe TB disease and higher bacterial burden in nonhuman primate granulomas.
  • SPP1+ Macrophages Co-localize with Myofibroblasts: IHC and spatial transcriptomics confirmed the co-localization of MMP1+CXCL5+ fibroblasts and SPP1+ macrophages within the granuloma cuff.
  • Fibroblasts Dominate Cell-Cell Communication: In TB-diseased lungs, fibroblasts were identified as both major senders and receivers of cellular signals, indicating a central role in tissue communication.
  • MMP1+CXCL5+ Fibroblasts are Prolific Signal Senders: Despite their relatively small proportion, these fibroblasts were among the most significant signal senders in TB-diseased lungs, contributing to pathways like ECM remodeling and inflammation.
  • TST Challenge Induces TB-Myofibroblast Signature: In vivo stimulation with mycobacterial antigens (tuberculin) in TB patients amplified the TB-myofibroblast signature, suggesting a direct link between M.tb exposure and fibroblast activation.

Limitations and Future Work

  • Modest Cohort Size: The authors acknowledge that their scRNA-seq cohort size is modest, which limits the power to detect subtle cell type proportion differences and introduces inter-patient variability. They addressed this by leveraging spatial transcriptomics and integrating with public datasets.
  • Limited Control Samples: The scarcity of fibroblast data from TB-negative controls may influence analyses comparing fibroblast populations.
  • Mechanistic Dissection Needed: While the study identifies key cell populations and interactions, further research is required to fully elucidate the mechanistic roles of the myofibroblast-like phenotype and its interaction with SPP1+ macrophages.
  • Future Directions: The authors suggest several avenues for future work, including:
  • Co-culture systems and conditioned media assays: To directly assess how fibroblasts influence macrophage behavior and vice versa.
  • Ex vivo stimulation experiments: To determine if TB-specific cues directly drive fibroblast and macrophage differentiation.
  • Genetic approaches: Such as targeted gene knockouts or CRISPR screens in fibroblasts to clarify causal relationships.
  • Studying earlier time points: To understand the origins of TB complications and disease progression.
  • Integrated spatial and temporal mapping: To create a comprehensive disease map that can guide optimal treatment strategies.

Summary updated

Abstract

Tuberculosis (TB) typically causes lung destruction and fibrosis, leading to ∼1.3 million deaths annually. The cellular drivers of human TB immunopathology remain poorly defined. We performed single-cell RNA sequencing and spatial transcriptomics on lung tissues from TB-infected and TB-negative individuals, identifying 30 distinct immune, parenchymal, and stromal cell subsets. Several were linked to TB pathology and corroborated through immunohistochemistry, flow cytometry, and independent human datasets. Fibroblasts were identified as major drivers in both active TB granuloma and TB-diseased lung tissue. In particular, the MMP1+CXCL5+ fibroblast subset, expressing a myofibroblast-like gene signature, was associated with severe disease and higher bacterial burden in nonhuman primate granulomas. Network analyses revealed cross talk between MMP1+CXCL5+ fibroblasts and SPP1+ macrophages within the granuloma cuff, which has been reported in other disease contexts, and may play an important role in TB immunopathology. Our findings highlight previously unappreciated cell populations and potential targets for novel TB therapies.
BibTeX
@article{mbano2025single,
  author = {Ian Mbano and Nuo Liu and Marc H. Wadsworth and Mark John. Chambers and Thabo Mpotje and Osaretin E. Asowata and Sarah K. Nyquist and Kievershen Nargan and Duran Ramsuran and Farina. Karim and Travis K. Hughes and Joshua D. Bromley and Robert Krause and Threnesan Naidoo and Liku B. Tezera and Michaela Reichmann and Sharie Keanne Ganchua and Henrik N. Kløverpris and Kaylesh J Dullabh and Rajhmun Madansein and Sergio Triana and Adrie J.C. Steyn and Bonnie Berger and Mohlopheni J. Marakalala and Sarah M. Fortune and JoAnne L. Flynn and Paul Elkington and Alex K. Shalek and Alasdair Leslie and Alasdair Leslie},
  title = {{Single-cell and spatial profiling highlights TB-induced myofibroblasts as drivers of lung pathology}},
  journal = {The Journal of Experimental Medicine},
  year = {2025},
  doi = {10.1084/jem.20251067},
}
Figure