Two adult fibroblast populations can look almost identical under the microscope yet carry strikingly different developmental signatures. This study asks whether HOX genes preserve a cell’s embryonic address long after development ends. Fibroblasts are the most abundant cell type in the human body and play crucial roles in inflammation and cancer progression.

They originate from the mesoderm or neural-crest-derived ectomesenchyme. Ectomesenchyme-derived fibroblasts contribute to facial formation and do not express HOX genes during development, but the expression and role of HOX genes in adult fibroblasts was not known.

HOX genes are a conserved family of transcription factors that regulate craniocaudal development, with expression tightly controlled during embryogenesis. Facial fibroblasts derived from neural-crest-originated ectomesenchyme are characteristically negative for HOX gene activity during development, while fibroblasts in the trunk and limbs show region-specific HOX activity before and after birth.

The stability of HOX gene activity in adult human fibroblasts of ectomesenchymal origin under different conditions remained poorly understood, and fibroblast heterogeneity was often overlooked in research design. The study therefore investigated whether the developmental HOX gene signature is preserved in adult fibroblasts under physiological and pathological conditions, including cancer-associated fibroblasts.

The main comparison used six facial dermal fibroblast samples and an age-matched set of six upper-forearm samples, with biopsy sites standardised for both locations. Samples from neuroectoderm-derived regions, including the face, oral cavity, and forebrain, were considered ectomesenchymal, while subgalear fibroblasts were selected as the best available control for intracranial cancer-associated fibroblasts.

The collection contained eighty-five samples from normal tissues and ninety-seven from pathological tissues. Transcriptome profiling used either microarrays or RNA sequencing. Raw data were processed with the oligo and limma packages in R and Bioconductor.

The data were background corrected using a normal–exponential model and quantile normalised. Batch effects were corrected with the sva package, and log two transformed normalised expression data were used for heatmap visualisation with ComplexHeatmap.

Figure one compares adult dermal fibroblasts from the face and forearm across cell size, growth, protein staining, and transcript expression. The violin plots show facial cells were somewhat larger, while the growth curves were essentially identical; the microscopy and boxplots then document origin-associated differences across markers including vimentin, nestin, alpha-SMA, CD271, S100A6, and MKI67.

This matters because the two populations look broadly similar in culture, yet retain molecular differences that support distinct developmental identities. Figure two compares gene-expression profiles across facial, forearm, and subgaleal fibroblasts.

In panel A, the heatmap shows distinct expression patterns between ectomesenchyme-originated facial cells and mesoderm-originated forearm cells, with subgaleal fibroblasts included for comparison; the authors report nine hundred fifty-nine differentially expressed genes. Panel B links these differences to developmental and morphogenetic Gene Ontology terms, including forelimb, cranial nerve, heart-valve, and mesenchymal development, supporting the idea that adult fibroblasts retain tissue-origin-associated programs.

RNA sequencing revealed significant transcription-profile differences between ectomesenchyme-originated fibroblasts from the face and mesoderm-originated fibroblasts from the upper forearm of adult donors. In total, nine hundred fifty-nine genes were differentially expressed.

Gene set enrichment analysis showed that Gene Ontology terms related to development and morphogenesis were enriched. Figure three compares homeobox-gene expression across facial, subgalear, and forearm fibroblasts. In the heatmap, facial samples show a distinct pattern from forearm samples, while subgalear fibroblasts display activity of HOXC5 and HOXC4.

The volcano plot identifies homeobox genes among the most upregulated genes in fibroblasts prepared from the forearm, supporting the authors’ use of HOX activity as a marker of fibroblast origin. Forearm fibroblasts showed significantly higher activity of HOXA9, HOXD9, HOXA10, HOXD10, HOXA11, HOXD11, HOXA13, and TBX5.

Every one of these reported differences had a false-discovery rate below zero point zero zero one. These genes are connected in the analysis to development of the upper and lower limbs, making the limb-associated expression pattern especially prominent in forearm fibroblasts.

This supports the authors’ use of gene-expression patterns to distinguish fibroblast origin. Subgalear fibroblasts from occipital to parietal regions exhibited activity of HOXC5 and HOXC4 genes. Their HOX activity differed strikingly from dermal fibroblasts isolated from the viscerocranium and forearm fibroblasts.

HOXC6, HOXC8, HOXD10, TBX4, and TBX5 were detected by immunocytochemistry in cultured fibroblasts from ectomesenchyme and mesoderm. No signal for these proteins was observed in the cell nucleus. The proteins appeared as cytoplasmic granules in all studied fibroblast types, with very low activity in facial cells and the strongest activity in trunk cells.

Figure six is a heatmap of HOX-gene expression across fibroblasts from different pathological tissues, with sample annotations for developmental location, cell description, and sex. The color scale represents log-two expression intensity, while the gene rows show that expression profiles vary with the cells’ mesodermal or ectomesenchymal origin.

This matters because it supports the authors’ use of HOX activity as a molecular trace of cellular origin, including in cancer-associated fibroblasts. Cancer-associated fibroblasts from tumours arising in ectomesenchyme-dependent areas were devoid of HOX gene expression, with scarce exceptions.

A similar trend appeared in mesenchymal cells from glioblastomas and secondary brain tumours. Fibroblasts from epileptogenic foci in the brain were also devoid of HOX gene activity. Figure seven is a heat map of RNA-sequencing HOX-gene activity across fibroblast samples from epileptogenic brain foci, glioblastomas, brain metastases, and subgaleal tissue.

The color scale represents expression z-scores, while the annotation bars identify each sample’s location, tumour type, and sex. The figure shows generally negligible HOX activity in the glioblastoma, metastasis, and epileptogenic-focus groups, although several glioblastoma-associated samples display variable HOX expression, highlighting differences among these brain-derived fibroblasts.

Adult facial dermal fibroblasts from ectomesenchyme were morphologically very similar to mesodermal forearm fibroblasts, yet they differed in homeobox-gene expression. Facial fibroblasts exhibited negligible HOX activity. HOX genes were expressed in all adult fibroblasts of mesodermal origin.

Glioblastoma cancer-associated fibroblast samples showed a variable number of HOX genes expressed at varying intensities. This contrasted with control samples from pharmacoresistant epilepsy foci, where HOX gene transcription was generally silent. The analysis used fibroblasts isolated from normal skin, tumours, and patients with systemic sclerosis, cultured from the second to fifth passage.

The authors describe the study as a pilot study with limitations. The study had a relatively large collection of solid tumours with uniform results, but the number of samples for each diagnosis was relatively small, limiting generalizability. Future studies should include comprehensive protein validation and single-cell sequencing.

The findings highlight the importance of considering embryonic origins of fibroblasts in biomedical research. The ectomesenchyme-based cancer microenvironment is often overlooked but is crucial for accurate experimental outcomes. Proper fibroblast controls are essential, and mixing ectomesenchyme and mesoderm fibroblasts is incorrect and can lead to misleading results, even with advanced omics methods.

The study finds that HOX gene patterns distinguish ectomesenchymal from mesodermal fibroblasts, even in pathological tissues. That makes embryonic origin an essential variable when comparing fibroblasts in cancer and regenerative-medicine research.