Spatial Transcriptomics for Fibrosis: Niche & Targets
Spatial transcriptomics (Visium, MERFISH, Xenium) maps fibrotic niches. Covers ligand-receptor analysis and target discovery in fibrosis models.
Introduction
Single-cell RNA-seq (scRNA-seq) has had a revolutionary impact on understanding fibrosis pathology. Myofibroblast subpopulations, macrophage polarization dynamics, and epithelial cell reprogramming — all invisible with bulk RNA-seq — have been uncovered at cellular resolution.
However, scRNA-seq has a fundamental limitation: tissue dissociation into single cells destroys the spatial information of where each cell was located.
Fibrosis is inherently a spatial disease. Within the fibrotic niche, activated fibroblasts, macrophages, epithelial cells, and vascular endothelial cells are arranged in specific spatial patterns, actively exchanging signals. Without understanding "who is next to whom," the full picture of fibrosis progression remains incomplete.
Spatial Transcriptomics (ST) is a family of technologies that measure gene expression with spatial coordinates on tissue sections, recovering the spatial context that scRNA-seq loses.
This article covers ST technology fundamentals, specific applications in fibrosis research, and practical use in preclinical drug discovery.
1. Limitations of scRNA-seq: Why Spatial Information Matters
What scRNA-seq Revealed
| Discovery | Example | Significance |
|---|---|---|
| Fibroblast subtypes | CTHRC1⁺ pathological fibroblast | Identification of subpopulations actively driving fibrosis |
| Macrophage phenotype continuum | SPP1⁺ → TREM2⁺ dynamic transition | Rejection of the M1/M2 dichotomy; discovery of pro-fibrotic phenotypes |
| Aberrant epithelial differentiation | KRT5⁺/KRT17⁺ aberrant basaloid cells | Elucidation of abnormal regeneration pathways in IPF |
What scRNA-seq Cannot Tell Us
- Spatial proximity: Where are CTHRC1⁺ fibroblasts positioned relative to other cells?
- Local signaling: Paracrine signal transmission is distance-dependent
- Tissue architecture relationships: Is the fibrotic focus peribronchial? Perivascular? Subpleural?
- Dissociation artifacts: Some cells are lost during tissue dissociation (large cells, fragile cells)
- Multicellular interactions: Simultaneous 3+ cell interaction patterns cannot be reconstructed
Key point: scRNA-seq tells us "who is there" (Who), but not "where they are" (Where) or "who they're talking to" (With whom). ST answers both of these questions.
For researchers tracking fibrosis & inflammation R&D
FDA approval alerts, trial readouts, preclinical model selection, and assay optimization — curated signal for bench-to-pipeline readers. 2 emails/month max.
2. Major Spatial Transcriptomics Technologies
Technology Classification
ST technologies fall into two major categories.
Sequencing-Based (Capture-Based)
mRNA is captured at specific locations on tissue sections and read by next-generation sequencing.
| Platform | Spatial Resolution | Genes Detected | Features |
|---|---|---|---|
| Visium (10x Genomics) | 55 µm (spot diameter) | Whole transcriptome | Most widely adopted. 1 spot = multiple cells. H&E overlay possible |
| Visium HD | 2 µm (bin) | HD WT Panel (probe-based / protein-coding genes only) / HD 3' Gene Expression (3' poly(A) capture / whole transcriptome) | Single-cell to subcellular resolution. Released 2024. Chemistry and coverage differ substantially across assays — see 10x official specs |
| Slide-seq V2 | 10 µm | Whole transcriptome | DNA bead-based. Cost-efficient but technically challenging |
| Stereo-seq (BGI) | 0.5 µm (theoretical) | Whole transcriptome | Nanoball technology. Ultra-high resolution but data processing-intensive |
Imaging-Based (In situ)
mRNA is directly visualized on tissue sections.
| Platform | Spatial Resolution | Genes Detected | Features |
|---|---|---|---|
| MERFISH (Vizgen) | Subcellular | 100–1,000 genes (panel) | High sensitivity and resolution. Requires target panel design |
| Xenium (10x Genomics) | Subcellular | 300–5,000 genes (panel) | Single-molecule detection in tissue. Automated workflow |
| CosMx SMI (NanoString / Bruker) | Subcellular | 1,000-gene panel / 6K Discovery (6,175 RNAs, ligand-receptor focus) / Human Whole Transcriptome Panel (18,000+ RNAs), select per study | FOV (field of view)-based high-throughput analysis. Gene count and cost vary substantially by assay choice |
| seqFISH+ | Subcellular | ~10,000 genes | Highest gene count but technically demanding |
Which Is Best for Fibrosis Research?
| Purpose | Recommended Technology | Rationale |
|---|---|---|
| Exploratory screening (broad overview first) | Visium / Visium HD | Whole transcriptome. Hypothesis-free discovery |
| High-resolution spatial distribution of known markers | MERFISH / Xenium | Single-cell/molecule-level resolution |
| Multi-sample comparison (drug efficacy) | Visium + deconvolution | Cost and throughput balance |
| Preclinical POC mechanism verification | Xenium (flexible panel design) | Focused measurement of target pathway genes |
3. Spatial Anatomy of the Fibrotic Niche
What Is the Fibrotic Niche?
The fibrotic niche is a microenvironment centered on activated fibroblasts, where specific immune cells, epithelial cells, and vascular cells spatially congregate. With the advent of ST, the composition and dynamics of this niche are being revealed at the molecular level.
Discoveries in Pulmonary Fibrosis (IPF)
Spatially Defined Cell Arrangement Patterns
ST analysis of IPF lung tissue has revealed the following spatial patterns:
- Fibrotic Front: The boundary region between normal alveoli and fibrotic foci. CTHRC1⁺ pathological fibroblasts aggregate and exchange TGF-β signals with adjacent KRT17⁺ aberrant basaloid cells
- Honeycomb cyst lining: MUC5B⁺ mucus-producing cells are aberrantly positioned. Surrounding SPP1⁺ macrophages promote fibrosis
- Peribronchial: Lymphocyte aggregation (tertiary lymphoid structures) forms, sustaining chronic inflammation
Spatial Ligand-Receptor Analysis
Analyzing ST data with CellChat or NicheNet enables estimation of ligand-receptor pairs between spatially proximate cells:
| Sender Cell | Ligand | Receiver Cell | Receptor | Signaling | Role in Fibrosis |
|---|---|---|---|---|---|
| Epithelial (basaloid) | TGF-β1 | Fibroblast | TGFBR1/2 | Smad2/3 | Myofibroblast differentiation |
| Macrophage (SPP1⁺) | SPP1 (OPN) | Fibroblast | CD44/Integrin | NF-κB/ERK | Proliferation & survival promotion |
| Fibroblast | PDGF-BB | Fibroblast | PDGFRα/β | PI3K/Akt | Autocrine proliferation loop |
| Vascular endothelium | ET-1 | Fibroblast | ETAR | Rho/ROCK | Enhanced ECM production |
Discoveries in Hepatic Fibrosis
From ST data of MASH/NASH models:
- Portal region: Periportal fibroblasts (HSCs) activate and interact closely with biliary epithelium
- Ductular Reaction: Proliferating biliary epithelial cells (CK19⁺) emit fibroblast-attracting signals
- Zone 3 → Zone 1 progression: Spatial tracking of fibrosis spreading from the centrilobular to portal regions
Discoveries in Renal Fibrosis
From ST analysis of UUO models and 5/6 nephrectomy models:
- Tubulo-interstitial interface: Injured proximal tubular epithelium (VCAM1⁺) directly activates adjacent fibroblasts
- Glomerular-Bowman's capsule: Accompanying glomerulosclerosis, Bowman's capsule epithelium undergoes aberrant differentiation (parietal epithelial cell activation)
4. Application to Drug Target Discovery
How ST Changes the Target Discovery Paradigm
Traditional bulk RNA-seq and scRNA-seq approaches selected target candidates from lists of differentially expressed genes (DEGs). ST adds spatial context, enabling more precise target selection.
Approach 1: Identifying Spatially Co-localized Ligand-Receptor Pairs
Procedure:
- Identify Fibrotic Niche regions in ST data (clustering)
- Estimate ligand-receptor pairs within the niche using CellChat/NicheNet
- Filter for pairs inactive in healthy tissue but activated only in the Fibrotic Niche
- → Prioritize as disease-specific signal target candidates
Advantage: Prioritizing signals spatially restricted to fibrotic foci rather than genes expressed throughout the tissue may help prioritize targets with more tissue-context specificity. Actual off-target risk still depends on target expression in other organs, delivery, and pathway redundancy — spatial restriction alone does not guarantee a safer profile
Approach 2: Cell Composition Changes via Spatial Deconvolution
Since Visium (55 µm spots) captures multiple cells per spot, deconvolution (RCTD, Cell2location, Tangram, etc.) estimates per-spot cell composition using scRNA-seq reference data.
Drug discovery application:
- Quantify spot-level cell composition changes: drug-treated vs. control
- Visualize "in which spatial regions did pathological fibroblast proportion decrease?"
- Detect localized drug effects invisible to conventional bulk assessment
Approach 3: Spatial Trajectory Analysis
Analyze spatial gene expression gradients along the fibrosis "progression front":
- Define a spatial axis: healthy region → transition zone → fibrotic focus
- Compute gene expression change patterns along this axis (Spatial Trajectory)
- Genes that sharply increase at the transition zone → fibrosis "trigger" candidates
5. Practical Guide to ST in Preclinical Research
When to Use ST
| Research Phase | ST Application | Recommended Platform |
|---|---|---|
| Target discovery | Whole-transcriptome comparison: healthy vs. disease model | Visium HD |
| Mechanism verification | Track spatial changes post-treatment with specific marker panels | Xenium / MERFISH |
| POC/efficacy assessment | Correlate fibrosis scores with ST data | Visium + PSR AI quantification |
| Biomarker discovery | Identify spatially restricted disease markers | Visium → Xenium (stepwise) |
Cost and Practicality
| Platform | Relative cost | Approximate per-sample range (example) | Sample throughput | Experimental duration |
|---|---|---|---|---|
| Visium | Medium | $1,500–2,500 (core facility / service quote examples) | 4 samples/slide | 2–3 days (library prep) |
| Visium HD | Medium–High | $2,000–3,500 (same as above) | 2 samples/slide | 2–3 days |
| Xenium | High | $1,000–2,000/sample (same as above) | High throughput | 1–2 days (run) |
| MERFISH | High | $2,000–4,000 (same as above) | 1–4 samples/run | 2–3 days |
[NOTE] The numbers above are illustrative ranges aggregated from publicly available core-facility and service-provider quotes. List prices, institutional rates, and analysis-included vs. analysis-excluded pricing are mixed in this space. Actual costs vary substantially with vendor, reagent grade, tissue type, and institutional throughput contracts. Use a direct quote from your vendor / service provider for any budgeting decision — the figures in this table are intended only to convey relative order-of-magnitude differences, not absolute values.
[TIP] Applying ST to every sample in a full preclinical study (N=8–10/group, 3 groups) becomes cost-prohibitive. A practical design is to run ST on representative samples only (2–3/group) and complement the remaining samples with histological evaluation (PSR + hydroxyproline).
Analysis Pipeline
[Data Acquisition]
Visium: Space Ranger → h5ad/AnnData format
Xenium: Xenium Ranger → Transcript table
[Preprocessing]
Quality filtering → Normalization → Dimensionality reduction (PCA/UMAP)
[Spatial Clustering]
BayesSpace / SpaGCN / STAGATE → Spatially-aware cluster identification
[Cell Type Annotation]
Deconvolution (Cell2location / Tangram) → Integration with scRNA-seq reference
[Ligand-Receptor Analysis]
CellChat / NicheNet / COMMOT → Spatial communication pattern estimation
[Visualization & Statistics]
Scanpy / Squidpy / STUtility → Spatial plot generation
6. Integration Strategy: scRNA-seq + ST
Why Integration Is Necessary
| ST Strength | ST Weakness | Solution |
|---|---|---|
| Spatial information preserved | Gene detection sensitivity lower than scRNA-seq (especially Visium) | Complement with scRNA-seq |
| Tissue architecture context | Sequencing-based: multiple cells per spot | Complement with deconvolution |
| In situ measurement (no dissociation) | High cost, limited sample numbers | Strengthen group comparisons with scRNA-seq |
Recommended Integration Workflow
- Phase 1: Build a cell atlas with scRNA-seq (define cell types and states)
- Phase 2: Spatial mapping with ST (Visium) + deconvolution (using Phase 1 reference)
- Phase 3: Validate key ligand-receptor pairs at single-cell resolution with imaging ST (Xenium)
This stepwise approach maximizes spatial insights while controlling costs.
Frequently Asked Questions (FAQ)
Does ST replace scRNA-seq?
At present, ST complements rather than replaces scRNA-seq. scRNA-seq excels in whole-transcriptome sensitivity at single-cell resolution, while ST provides spatial information. Integrating both yields the most powerful approach.
Can ST be applied to FFPE tissue?
All three major platforms (Visium CytAssist, Xenium, CosMx) offer FFPE-compatible assays. However, supported species (human, mouse, etc.), assay chemistry version, and required sample quality (DV200, block storage age, etc.) differ from assay to assay — whether a given sample is actually compatible should be confirmed against the official spec sheet and the latest application notes for the specific assay you plan to use. Retrospective analysis from existing FFPE archives is a major preclinical advantage, but older blocks may fall outside the RNA-quality envelope of current assays.
Can mouse model data translate to human tissue?
The broad cell type framework is conserved, but differences between mouse and human exist — particularly in macrophage subtypes and fibroblast heterogeneity. Using cross-species comparison tools (SAMap, etc.) to distinguish conserved versus species-specific pathways is essential.
How many samples should receive ST per study?
For discovery research, 2–3 samples/group is the practical minimum under cost constraints. For statistically robust group comparisons, 5–6 samples/group is ideal, but costs escalate rapidly. A common design limits ST to representative samples and validates with histological methods (PSR + hydroxyproline).
Summary
| Key Point | Details |
|---|---|
| scRNA-seq limitation | Loss of spatial information. The "structure" of the fibrotic niche is invisible |
| Value of ST | Reveals cell arrangements and interactions in the Fibrotic Niche with spatial coordinates |
| Key platforms | Discovery: Visium/HD; Validation: Xenium/MERFISH |
| Drug discovery application | Targeting spatially restricted signals, detecting localized drug effects, biomarker discovery |
| Practical approach | Run ST on representative samples only; complement with histology |
| Integration strategy | Stepwise: scRNA-seq → Visium deconvolution → Xenium validation |
References
Note: This article synthesizes publicly available findings in fibrosis scRNA-seq and Spatial Transcriptomics. The references below cover foundational scRNA-seq papers (identifying CTHRC1⁺ fibroblasts and KRT5⁺/KRT17⁺ aberrant basaloid cells) and the major analysis tools used for spatial cell–cell communication. Primary Spatial Transcriptomics atlases for IPF, liver, and kidney fibrosis are rapidly expanding; specific spatial findings discussed in the text should be cross-checked against the references below, official platform documentation, and the latest reviews.
- Adams TS, et al. Single-cell RNA-seq reveals ectopic and aberrant lung-resident cell populations in idiopathic pulmonary fibrosis. Sci Adv. 2020;6(28):eaba1983. PMID: 32832599
- Habermann AC, et al. Single-cell RNA sequencing reveals profibrotic roles of distinct epithelial and mesenchymal lineages in pulmonary fibrosis. Sci Adv. 2020;6(28):eaba1972. PMID: 32832598
- Jin S, et al. Inference and analysis of cell-cell communication using CellChat. Nat Commun. 2021;12(1):1088. PMID: 33597522
- Browaeys R, et al. NicheNet: modeling intercellular communication by linking ligands to target genes. Nat Methods. 2020;17(2):159–162. PMID: 31819264
- Cang Z, et al. Screening cell-cell communication in spatial transcriptomics via collective optimal transport (COMMOT). Nat Methods. 2023;20(2):218–228. PMID: 36690742
Platform official information:
- 10x Genomics Visium HD: https://www.10xgenomics.com/products/visium-hd-spatial-gene-expression
- 10x Genomics Xenium panels: https://www.10xgenomics.com/products/xenium-panels
- NanoString CosMx SMI: https://nanostring.com/products/cosmx-spatial-molecular-imager/ffpe-dataset/
Related Articles
- AI-Based Image Analysis for Improved Fibrosis Scoring — Digital pathology advances through image analysis
- Sirius Red Staining × AI Image Analysis — AI-powered PSR quantification (complementary assessment with ST)
- Hydroxyproline Assay: Principle, Protocol & Collagen Quantification — Biochemical collagen quantification
- Can We Stop IPF Fibrosis Progression? Next-Generation Antifibrotic Target Strategies — Drug discovery frontlines where ST validates target candidates
- TGF-β/Smad Signaling Pathway and Fibrosis — The central pathway of the fibrotic cascade