Full Citation
Wang S, Ma F, Bai Z, Zhang W, Song X, Zhang J, et al. Single-cell transcriptomic profiling identifies therapeutic subpopulations of adipose-derived mesenchymal stromal cells for human keloid management. Mol Cell Biochem. 2026;481:1369-1386.
Background and Question
Keloids arise from abnormal wound healing with excessive collagen deposition and invasive growth beyond the injury margin. Bulk adipose-derived stromal cells have been considered for scar therapy, but their heterogeneous composition makes mechanism, potency, and reproducibility uncertain.
Research question
Which keloid-associated molecular programs are linked to fibroblast and keratinocyte pathology, and can single-cell profiling identify ASC subpopulations with more rational therapeutic potential?
Methods and Evidence Chain
Integrated gene-expression profiling, functional annotation, protein-protein interaction mapping, and hub-gene identification.
Used scRNA-seq to resolve ASC subpopulations and nominate subsets with inhibitory effects on keloid development.
Connected keloid fibroblast and keratinocyte programs to hub genes, including NOG, IL6, APP, and NOTCH1.
Evaluated therapeutic efficacy of selected ASC subpopulations in a miniature pig hypertrophic-scar model.
Pathogenesis mapping
Integrated gene-expression profiling, functional annotation, protein-protein interaction mapping, and hub-gene identification.
Single-cell layer
Used scRNA-seq to resolve ASC subpopulations and nominate subsets with inhibitory effects on keloid development.
Disease logic
Connected keloid fibroblast and keratinocyte programs to hub genes, including NOG, IL6, APP, and NOTCH1.
In vivo check
Evaluated therapeutic efficacy of selected ASC subpopulations in a miniature pig hypertrophic-scar model.
Key Results
NOG and IL6 upregulation were strongly associated with keloid fibroblast formation.
APP and NOTCH1 were implicated in keloid keratinocyte development.
scRNA-seq identified ASC subpopulations with candidate inhibitory activity against scar development.
A large-animal hypertrophic-scar model was used to move beyond purely computational nomination.
Mechanism Interpretation
The study supports a two-compartment scar model: fibroblast-side cytokine and morphogen signals sustain ECM deposition, while keratinocyte-side APP/NOTCH1 programs may reinforce abnormal epidermal-mesenchymal communication. Therapeutic ASC subsets may act by secreting factors that rebalance inflammatory, matrix, and epithelial signaling.
Mechanism / workflow schematic
Mermaid source is included so the website can render the diagram in supported browsers.
flowchart TD A[Skin injury] --> B[Keloid keratinocyte program] A --> C[Keloid fibroblast program] B --> D[APP and NOTCH1 signaling] C --> E[NOG and IL6 hub signals] D --> F[Epithelial-mesenchymal crosstalk] E --> G[Collagen deposition and matrix persistence] H[Selected ASC subpopulation] --> I[Paracrine modulation] I --> F I --> G
Clinical and Translational Relevance
Clinical relevance
For keloid clinics, the practical message is that future cell-based therapy should define the active cell state and potency markers. It also reinforces IL6/NOG and NOTCH-linked crosstalk as targets to test alongside established intralesional treatment.
Translational value
The combination of single-cell resolution and a pig scar model is useful for selecting potency assays before human studies. The work could guide standardized ASC products or ASC-secretome products for refractory scars.
Limitations and Critique
Hypertrophic-scar models only partially recapitulate true human keloid invasion and recurrence.
ASC subpopulation isolation, expansion, and stability require GMP-compatible protocols.
Hub-gene associations need perturbation experiments to prove causality.
Pain, itch, recurrence after excision, and validated scar scales were not established as clinical outcomes.
Reviewer-style critique
The paper is valuable because it resists treating ASCs as a uniform product. The strongest next step is causal validation: knockdown, blockade, or rescue experiments for nominated hub pathways and direct comparison with TAC/5-FU or silicone/laser-based standards.
Practical Next Research Actions
Action 1
Build a local keloid scRNA-seq reference for fibroblast, keratinocyte, endothelial, immune, and nerve-associated cell states.
Action 2
Test ASC subsets or conditioned media on primary keloid fibroblast collagen contraction and ECM-production assays.
Action 3
Measure IL6, NOG, NOTCH1, COL1A1, COL3A1, POSTN, and ACTA2 as a compact potency panel.
Action 4
Design combination studies with intralesional TAC/5-FU to test whether cell-derived signals reduce recurrence biology.
Evidence-quality judgment
Moderate mechanistic evidence: strong discovery design and animal follow-up, but human therapeutic efficacy remains unproven.