This study applied bulk RNA sequencing to paired head and neck keloid tissue and matched adjacent normal skin tissue (MANST) from 14 patients to characterize how cell type composition contributes to differences in gene expression in keloid disease (KD). The cohort included 10 African American and 4 European American individuals. The analysis focused on cell type enrichment, tissue-specific associations between cell types, and differentially expressed genes after accounting for cell type composition.
Bulk RNA-Seq was performed on keloid and matched adjacent normal skin samples. Cell type enrichment for defined fibroblast subtypes and immune cell types was estimated by single-sample gene set enrichment analysis (ssGSEA). To test for differences across tissue types and for tissue-specific relationships, the authors used linear mixed-effects models for three primary analyses: differential cell type enrichment across tissue, tissue type–specific associations between fibroblast subtypes and immune cells, and identification of differentially expressed genes (DEGs) across tissue. An independent validation cohort of eight African American participants was used for replication; the abstract notes validation was conducted but does not report detailed validation statistics.
The analysis identified that three distinct fibroblast subtypes were differentially enriched between keloid tissue and MANST. The results underscore heterogeneity in fibroblast subtype composition in keloid lesions compared with adjacent normal skin, indicating that shifts in fibroblast subtype abundance contribute to the transcriptional profile of KD.
In addition to fibroblast subtypes, 14 immune cell types showed differential enrichment across tissue types. The detection of multiple immune cell enrichment differences highlights the role of immune cell composition in KD and suggests complex immune–stromal interactions in keloid tissue compared with adjacent normal skin.
Seventeen tissue type–specific associations between fibroblast subtypes and immune cells were identified. Of these, 14 associations demonstrated decreased correlation or association strength in keloid tissue relative to MANST. These findings indicate that many normal relationships between fibroblast subtypes and immune cells are altered in keloid lesions, which may reflect disease-specific microenvironment remodeling.
When gene expression differences were examined after adjusting for cell type enrichment, specific genes emerged with the largest adjusted fold-changes. The long noncoding RNA MIR31HG exhibited the largest positive fold-change in keloid tissue after adjustment, while NR4A2 showed the largest negative fold-change. These adjusted DEGs point to transcriptional changes that are not solely attributable to shifts in cell composition and may represent keloid-specific molecular signals.
The authors validated findings in an independent cohort of eight African American individuals. The abstract states validation was performed but does not provide detailed validation results, effect sizes, or statistical metrics in the summary. Therefore, specifics of replication strength and which associations were reproduced are not reported in the source abstract.
By estimating and adjusting for cell type enrichment in bulk tissue RNA-Seq, this study revealed that keloid transcriptional differences reflect both altered cell type composition and tissue-specific gene expression changes. The identification of differentially enriched fibroblast subtypes, multiple immune cell changes, altered fibroblast–immune associations, and adjusted DEGs such as MIR31HG and NR4A2 suggests that both cellular heterogeneity and intrinsic gene regulation contribute to KD pathophysiology. These results support the importance of accounting for cell type composition in bulk transcriptomic studies to uncover tissue-specific molecular mechanisms.
The authors declare no competing interests. Reported funders include Henry Ford Health, National Institutes of Health (K08GM128156), the American Academy of Facial Plastic and Reconstructive Surgery, and the National Institute of Arthritis and Musculoskeletal and Skin Diseases (R01AR083553). The preprint is posted on bioRxiv and made available under a CC-BY 4.0 International license.