Journal Club: Mapping Disease Niches - Spatial and Temporal Insights from Single-Cell & Spatial Transcriptomics

· 9 min read · Paper Feed

Journal Club: Mapping Disease Niches - Spatial and Temporal Insights from Single-Cell & Spatial Transcriptomics

Conceptual illustration: Mapping Disease Niches - Spatial and Temporal Insights from Single-Cell & Spatial Transcriptomics

Introduction

Single-cell RNA sequencing (scRNA-seq) has changed how we think about cellular heterogeneity in health and disease. But it has a blind spot: context. Cells don't operate in isolation. They live in three-dimensional tissue neighborhoods where position shapes function, local crosstalk drives pathology, and treatment reshapes the tissue around them. This month's journal club looks at three recent studies that tackle this gap by weaving temporal and spatial information into the definition of pathogenic cellular niches in chronic inflammatory and fibrotic diseases.

The common thread is a move beyond simply cataloging cell types toward defining disease-specific cellular states within their spatial and temporal tissue niches. Paper 1 (Francis et al.) uses longitudinal scRNA-seq to track the temporal evolution of cellular states in psoriatic skin resolving under IL-23 blockade. Paper 2 (Mayr et al.) uses spatial transcriptomics to map the spatial architecture of pathogenic niches in idiopathic pulmonary fibrosis (IPF). Together, they illustrate a workable approach: pair high-resolution cellular profiling with contextual information to break disease down into its functional units.

Paper 1: A Transient Inflammatory Fibroblast State in Psoriasis Resolution

Key Findings

  1. Early and substantial fibroblast response: IL-23 inhibitor therapy (guselkumab) produced the largest gene-expression shifts not in immune cells but in fibroblasts and myeloid populations as early as day 3 of treatment, which challenges the idea that keratinocytes are the primary early responders.
  2. A pathogenic fibroblast state: The study identified a transient fibroblast state found only in lesional skin and marked by co-expression of WNT5A and IL24. This state was enriched in psoriasis compared to healthy skin or atopic dermatitis.
  3. Fibroblast-to-keratinocyte signaling axis: Using in-silico ligand-receptor analysis and in-vitro validation, the authors showed that signals from these WNT5A+/IL24+ fibroblasts upregulate inflammatory genes (e.g., S100A7, DEFB4A) in keratinocytes, placing fibroblasts upstream of epidermal inflammation.
  4. Therapeutic targeting of the niche: The abundance of this pathogenic fibroblast state dropped after IL-23 blockade, and the finding was validated by deconvolution of bulk RNA-seq datasets and RNA in-situ hybridization. The state is therefore a candidate cellular target of effective biologic therapy.

Methodology & Study Design

  • Longitudinal single-cell profiling: The core approach was scRNA-seq (10x Genomics) on skin biopsies from psoriasis patients (n=8) at three time points: baseline (lesional and non-lesional), day 3, and day 14 after starting IL-23p19 inhibitor treatment. This produced 114,745 high-quality single-cell transcriptomes.
  • Multi-modal validation: Findings were supported by 1) in-silico deconvolution (CIBERSORTx) of four independent bulk RNA-seq psoriasis cohorts, 2) RNA in-situ hybridization (RNAscope) to spatially localize WNT5A and IL24 transcripts, and 3) in-vitro co-culture assays of primary fibroblasts and keratinocytes.
  • Analytical pipeline: Standard tools (Seurat) were used for clustering and differential expression, with NicheNet for modeling ligand-receptor interactions.

Significance

This work reframes the early mechanism of action for IL-23 blockade in psoriasis. It shifts attention from a purely immune-centric model to one in which stromal cells (fibroblasts) act as primary sensors and responders to therapy. By identifying a transient, disease-specific fibroblast state that drives epidermal inflammation and is rapidly cleared by treatment, the study underscores the dynamic nature of disease niches and points to fibroblast states as candidate biomarkers or therapeutic targets.

Paper 2: Spatial Architecture of Pathogenic Niches in Pulmonary Fibrosis

Key Findings

  1. Three distinct disease-associated niches: Spatial transcriptomics (Visium) of IPF and control lungs revealed three spatially segregated, pathological niches:
    • The fibrotic niche: Located around airways and composed of myofibroblasts and aberrant basaloid cells, representing the core effector site of extracellular matrix deposition.
    • The airway macrophage niche: Found in the airway lumen adjacent to the fibrotic niche and dominated by SPP1+ (osteopontin-positive) macrophages.
    • The immune niche: Characterized by organized lymphoid cell foci (T cells, B cells, plasma cells) within fibrotic tissue and surrounded by remodeled endothelial vessels.
  2. Spatial cohabitation defines function: The proximity of the SPP1+ macrophage niche to the fibrotic niche is consistent with a local, paracrine signaling axis that may drive myofibroblast activation. The immune niche's organization also suggests tertiary lymphoid organ (TLO)-like formation within the fibrotic milieu.
  3. Integration validates and refines cell states: By integrating spatial data with a large scRNA-seq atlas, the authors localized previously defined rare cell populations (such as aberrant basaloid cells) and confirmed their spatial restriction to specific pathological regions.

Methodology & Study Design

  • Spatial transcriptomics as the anchor: The study used the 10x Visium platform to profile gene expression in situ from 16 tissue sections taken from 8 IPF patients and 5 controls.
  • Integration with scRNA-seq: A key strength was integration with a publicly available IPF scRNA-seq atlas (~300,000 cells). This allowed cell-type deconvolution within each Visium spot and mapping of cell states to tissue locations.
  • Spatial analysis suite: The team used spatial differential expression analysis, neighborhood analysis, and cell-cell communication inference (CellChat) to define niches and interactions.

Significance

This paper provides a high-resolution spatial atlas of the IPF lung ecosystem. It moves IPF biology from a list of dysregulated cell types toward a map of functionally integrated, disease-driving neighborhoods. The identification of the SPP1+ macrophage niche next to the fibrotic front offers a spatially informed therapeutic target. Defining these niches also provides a blueprint for engineering more physiologically relevant in vitro models (e.g., organoids) that reproduce these cellular adjacencies.

Synthesis & Discussion

Read together, these two papers offer complementary lessons in deconstructing complex diseases. Francis et al. supply the temporal dimension—showing how a therapeutic perturbation dissolves a pathogenic cellular state over days. Mayr et al. supply the spatial dimension—showing how multiple pathogenic cell states coexist in architecturally defined neighborhoods that likely reinforce one another.

A pattern that stands out is the central role of specialized mesenchymal and myeloid cells as organizers of disease niches. In psoriasis it's the WNT5A+/IL24+ fibroblast; in IPF it's the myofibroblast and the SPP1+ macrophage. Both studies lean on computational integration (deconvolution, spatial mapping) to push from correlation toward causation, proposing testable signaling axes (fibroblast→keratinocyte; macrophage→myofibroblast).

These approaches also point to a shift from targeting a single cell type or cytokine to disrupting an entire pathogenic niche. The benefit of IL-23 blockade in psoriasis, as shown, may depend on its ability to collapse a multi-cellular unit. By the same logic, effective IPF therapy may need to target the fibrotic and macrophage niches simultaneously.

Questions worth chewing on:

  1. What upstream signals generate these pathogenic niches? In psoriasis, what initially produces the WNT5A+/IL24+ fibroblast?
  2. How plastic are these niches? If treatment stops, do the same cellular states re-emerge in the same spatial patterns?
  3. Can we identify master regulator genes for each niche that could be targeted for ablation?
  4. How generalizable is this niche paradigm across other chronic inflammatory diseases (e.g., rheumatoid arthritis, inflammatory bowel disease)?

For Your Lab Meeting

  1. Methodological trade-offs: Paper 1 traded spatial information for longitudinal single-cell depth, while Paper 2 traded longitudinal tracking for spatial context. For your disease of interest, which axis (temporal or spatial) is currently more limiting, and what experimental design would you prioritize?
  2. Causation vs. correlation: Both studies infer cellular crosstalk (NicheNet, CellChat) from transcriptional data. How confident should we be in these predictions? What orthogonal functional experiments (beyond the in-vitro assay in Paper 1) are needed to validate a proposed ligand-receptor axis as truly pathogenic?
  3. The fibroblast problem: Fibroblasts are notoriously heterogeneous and plastic. Does the WNT5A+/IL24+ state represent a distinct lineage, or a transient activation state that any dermal fibroblast can adopt? How would you design a fate-mapping experiment to test this?
  4. From niche to target: The SPP1+ macrophage in IPF is a spatially appealing target, but macrophages are multifunctional. How would you design a strategy to deplete or reprogram only this subset without compromising host defense or the homeostatic functions of other macrophages?
  5. Clinical translation: Both studies analyze human patient samples. What are the major challenges in moving from these high-resolution molecular maps to robust, simple biomarkers (e.g., a plasma protein) that report on niche activity in a clinical trial?

Key Terms

  • Spatial Transcriptomics: Technologies that capture gene expression while retaining the two-dimensional spatial coordinates of each measurement within a tissue section. Platforms like 10x Visium provide genome-wide expression for spots that contain multiple cells.
  • Deconvolution: A computational method for estimating the proportion of different cell types (or cell states) in a bulk tissue sample, or, as in Paper 2, within each spot of a spatial transcriptomics dataset. CIBERSORTx is a commonly used tool.
  • Pathogenic Niche: A localized, multicellular microenvironment within diseased tissue that sustains pathology through reciprocal signaling between resident cell states. It is more than a list of cells; it includes their spatial organization and functional interactions.
  • Aberrant Basaloid Cell: A unique epithelial cell state identified in IPF scRNA-seq studies, expressing both basal cell (KRT5, KRT17) and developmental markers. It is thought to be a maladaptive progenitor cell contributing to failed re-epithelialization and fibrosis.
  • Ligand-Receptor Analysis: A bioinformatics approach (using tools like NicheNet or CellChat) that infers potential cell-cell communication by pairing expressed ligands in one cell cluster with their cognate receptors in another, often combined with downstream target gene expression.

References

  1. L. Francis, D. McCluskey, C. Ganier, Treasa Jiang, X. Du-Harpur, J. Gabriel, et al. (2024). Single-cell analysis of psoriasis resolution demonstrates an inflammatory fibroblast state targeted by IL-23 blockade. Nature Communications. https://doi.org/10.1038/s41467-024-44994-w

  2. Christoph H. Mayr, Diana Santacruz, Sebastian Jarosch, M. Bleck, John Dalton, A. McNabola, et al. (2024). Spatial transcriptomic characterization of pathologic niches in IPF. Science Advances. https://doi.org/10.1126/sciadv.adl5473

Single-Cell Transcriptomics scRNA-seq Genomics