iPSC-Derived Fibroblasts Adapt Tissue-Specific Gene Programs Through Co-Culture Signals

Fibroblasts are among the most functionally diverse cell types in the human body, yet fewer than 20% of fibroblast-enriched genes overlap between organs such as the heart, gut, skeletal muscle, and bladder. This extraordinary heterogeneity makes it challenging to develop accurate in vitro disease models, particularly for organs like the heart where primary fibroblast access is severely limited. iPSC-derived fibroblasts (iFBs) represent a potentially transformative alternative, but whether they can truly recapitulate tissue-specific biology — not just express generic fibroblast markers — had not been rigorously established.

This study from the Berlin Institute of Health tested whether co-culture with organ-specific cell types could drive iFBs toward tissue-appropriate transcriptional programs. iFBs were generated from iPSCs using a BMP4-based differentiation protocol with SB-431542 supplementation on days 6–10 to suppress myofibroblast transition. These cells were then co-cultured directly (on opposite sides of a porous transwell membrane) or indirectly (separated by a transwell insert, paracrine signalling only) with keratinocytes (ectoderm), iPSC-derived cardiomyocytes (mesoderm), normal human bronchial epithelial cells (endoderm/lung), and ASC-derived jejunal organoid cells (endoderm/gut) for 7 days. All experiments were performed in triplicate, with control iFBs maintained in the same media without co-culture partners.

Bulk RNA sequencing followed by principal component analysis (PCA) revealed clear transcriptional divergence between iFBs exposed to different co-culture environments. Pathway analysis using DESeq2 (adjusted p ≤ 0.05, |log2 fold change| ≥ 1) identified functional shifts consistent with tissue-specific biology: cardiac co-culture iFBs showed significant upregulation of TGF-β signalling pathways — a hallmark of cardiac fibroblast activation — while dermal co-culture iFBs exhibited enrichment of ECM remodelling gene sets consistent with dermal fibroblast function. Lung and intestinal co-culture conditions produced their own distinguishable transcriptional signatures, confirming that the adaptation was not a generic stress response but environment-specific.

Single-cell deconvolution was performed using the scSemiProfiler framework against the Tabula Sapiens reference atlas (>20,000 cells filtered to fibroblast subtypes, retaining subtypes with ≥1,000 cells and up to 2,000 cells each). This analysis revealed that while iFBs shifted their population-level transcriptional centre of gravity toward tissue-appropriate fibroblast subtypes, they retained mixed subpopulation compositions — meaning full tissue specification was not achieved. The deconvolution data also showed that indirect co-culture produced measurable but transient transcriptional changes: when the transwell insert was removed after 7 days and iFBs were cultured alone for a further 7 days, the tissue-specific signatures partially faded, suggesting that sustained direct cell-cell contact or continuous paracrine exposure is required to stabilise the phenotype.

Targeted RT-qPCR validation confirmed key differentially expressed genes identified by RNA-seq, including vimentin and PDGFRA upregulation in iFBs versus iPSCs as quality controls, and organ-specific markers in co-cultured iFBs. The authors acknowledge that demonstrating transcriptional changes is necessary but not sufficient — it remains to be shown whether these gene expression shifts translate into functional fibroblast behaviours such as altered ECM deposition, contractility, or paracrine immune modulation. Optimising co-culture duration, cell ratios, and media formulations will be critical next steps for leveraging iFBs in fibrosis research and personalised drug screening platforms.