Orofacial clefts, which include cleft lip and cleft palate, are among the most common birth defects worldwide, affecting roughly one in 600 newborns. For decades, researchers have known that the gene IRF6 is a major player: mutations in it cause Van der Woude syndrome, the most frequent form of syndromic clefting, and common variants near it raise the risk of non-syndromic clefts. Yet the picture has always been incomplete, because IRF6 was thought to act almost exclusively in the oral epithelium, the thin layer of cells lining the mouth. A new study published in the Journal of Molecular Medicine overturns part of that assumption by showing that IRF6 is also active in cranial neural crest cells, the remarkable embryonic population that builds much of the face, and that its expression there is controlled by another famous craniofacial gene, SOX9.
The research team, led by Matthias Weider and Lina Gölz of University Hospital Erlangen together with colleagues in Bonn, began by making IRF6 visible in developing mouse embryos. Using transgenic animals in which all neural crest-derived tissue glows yellow thanks to a Wnt1::Cre-driven fluorescent reporter, the scientists stained embryonic sections at days 9.5, 10.5 and 11.5 of development. They found IRF6 protein in neural crest-derived cells of the pharyngeal arches, the embryonic structures that give rise to the jaw, palate and other facial components. Intriguingly, the fraction of neural crest cells carrying high levels of IRF6 shrank as development progressed: by embryonic day 11.5, only about 2.6 percent of these cells retained signal intensity comparable to that of epithelial cells, hinting at a narrow, precisely timed window of activity.
To confirm that this was not a quirk of the mouse or of the staining method, the team mined single-cell RNA sequencing data from embryonic craniofacial tissue of both mouse and human origin. The computational analysis revealed a subset of mesenchymal cells expressing IRF6 in both species. Crucially, many of these IRF6-positive cells also expressed SOX9, with 56 percent of IRF6-positive mesenchymal cells co-expressing the gene, and an even larger share, 86.8 percent, co-expressing TFAP2A, the transcription factor already known to regulate IRF6 in epithelium. The overlap placed IRF6 squarely inside the regulatory world of cranial neural crest cells, a transient, stem cell-like population whose migration, proliferation and differentiation must be choreographed with extraordinary precision for the face to form correctly.
SOX9 was a compelling suspect for several reasons. It is a bona fide cranial neural crest transcription factor, essential for assigning cells to the cartilage lineage, and it is the major gene mutated in Pierre Robin sequence, a craniofacial condition characterized by an underdeveloped lower jaw, a tongue that falls back into the throat, and frequently a cleft palate. Moreover, a single nucleotide variant called rs76145088, which genetic studies had previously linked to non-syndromic orofacial clefting, sits within a DNA sequence in the IRF6 enhancer MCS-9.7 that resembles a SOX9 binding site. That enhancer, located thousands of base pairs upstream of IRF6, was already famous: a different variant within it, rs642961, disrupts binding of TFAP2A and increases cleft risk, helping to establish MCS-9.7 as a hotspot of cleft-associated regulatory variation.
To test whether SOX9 really drives the enhancer, the researchers cloned MCS-9.7 into a luciferase reporter plasmid, a standard tool that glows when the enhancer is active. In neural crest-derived Neuro-2a cells, they first reproduced the known activation by TFAP2A, validating the system. They then added SOX9 and found that it, too, switched the enhancer on. The striking result concerned the rs76145088 variant: the ancestral allele, which is the risk allele for orofacial clefting, supported robust SOX9-dependent activation, while the derived allele, whose guanine base breaks the SOX9 consensus sequence, drove drastically weaker activation. Dose experiments sharpened the point, as 250 nanograms of SOX9 expression plasmid sufficed to activate the ancestral version of the enhancer, whereas three times that amount was needed for the derived version.
The team then moved from artificial reporters to endogenous biology. In O9-1 cells, a stable cell line derived from mouse cranial neural crest, they silenced Sox9 with short hairpin RNAs and watched the activity of the ancestral enhancer allele drop, confirming that normal cellular levels of Sox9 sustain MCS-9.7 activity. Electrophoretic mobility shift assays showed that purified SOX9 protein binds directly to the DNA sequence spanning rs76145088, and that it binds the ancestral allele more strongly than the derived one, exactly as the reporter experiments predicted. Chromatin immunoprecipitation in O9-1 cells, pulling down DNA bound by Sox9 inside living cells, enriched the region corresponding to MCS-9.7 relative to a negative control, albeit modestly, consistent with an enhancer that is active in only a small subset of cells.
The decisive experiment came next. Using CRISPR/Cas9 gene editing, the researchers knocked out Sox9 in O9-1 cells, generating two independent clones in which the protein was completely absent, one through a frameshift mutation and the other through a dramatic rearrangement including an inversion around the start codon. Quantitative PCR revealed that Irf6 transcript levels were virtually eliminated in both knockout clones. Other known upstream regulators of Irf6, including Tfap2a, Notch1, Foxe1, Trp63, the Pbx genes and several Wnt genes, showed only mild and inconsistent changes, arguing that SOX9 acts directly on the IRF6 enhancer rather than through intermediate regulators. Notably, the well-established IRF6 target gene Grhl3 was not expressed in these neural crest cells, suggesting that IRF6 may control a different set of downstream genes in cranial neural crest cells than it does in epithelium.
Perhaps the most thought-provoking twist is the direction of the risk effect. Usually, loss-of-function mutations in IRF6 cause clefting by reducing gene activity in epithelium. Here, the ancestral, risk-associated allele of rs76145088 is the one that binds SOX9 more strongly and drives higher enhancer activity. The authors therefore propose a gain-of-function mechanism: prolonged or excessive IRF6 expression in craniofacial mesenchyme could contribute to clefting, in the same way that enhanced binding of the transcription factor ETS1 to an SHH enhancer causes polydactyly. Because rs76145088 is common across human populations while clefts remain relatively rare, the finding fits the multifactorial threshold model of non-syndromic clefting, in which risk accumulates from many variants and environmental influences before a developmental threshold is crossed. The authors also note that the derived allele may be protective rather than the ancestral allele harmful, and that transcription factors like IRF6 must be kept within tight dosage windows, since both too little and too much activity can derail embryonic development.
The study does not close the case. Only a small fraction of cranial neural crest cells express IRF6, the enhancer carries both active and repressive histone marks suggestive of a poised or heterogeneous state, and previous mouse experiments deleting Irf6 in neural crest lineages produced bone mineralization defects but no palatal clefts. Still, by connecting a classic epithelial cleft gene to SOX9, the gene behind Pierre Robin sequence, the work draws two major cleft-associated pathways into a single regulatory axis and reveals that the developing face depends on IRF6 operating in more cellular neighborhoods than anyone had appreciated. Dysregulation of the SOX9-IRF6 axis in cranial neural crest cells, the authors conclude, may be a genuine contributor to the origins of orofacial clefting, and the heterogeneous mesenchymal cells that carry this signal are now an obvious target for the next generation of studies.
Subject of Research: Transcriptional regulation of the orofacial cleft risk gene IRF6 by SOX9 in cranial neural crest cells
Article Title: The orofacial cleft risk gene IRF6 is a target gene of SOX9 in cranial neural crest cells
Article References: Weider, M., Wagner, M. C., Schmid, T., Gehlen-Breitbach, S., Rodrian, G., Peschel, N., Schneider, H., Ludwig, K. U., Wegner, M., & Gölz, L. (2026). The orofacial cleft risk gene IRF6 is a target gene of SOX9 in cranial neural crest cells. Journal of Molecular Medicine, 104(1), Article 113. https://doi.org/10.1007/s00109-026-02720-7
Image Credits: AI Generated
DOI: 10.1007/s00109-026-02720-7
Keywords: orofacial cleft, IRF6, SOX9, cranial neural crest cells, cleft lip and palate, Pierre Robin sequence, MCS-9.7 enhancer, TFAP2A, CRISPR/Cas9, gene regulation, craniofacial development, single-cell RNA sequencing
Cite Scienmag News
APA MLA Chicago
Copy citation Download RIS
Tags: cleft lip and palatecranial neural crest cellscraniofacial developmentcraniofacial gene expression during embryogenesisCRISPR-Cas9embryonic neural crest cell differentiationGene regulationgenetic basis of Van der Woude syndromegenetic mutations linked to orofacial cleftsGenetic regulation of cleft lip and palateIRF6IRF6 gene role in craniofacial developmentlong-range gene regulation in cranioMCS-9.7 enhancermolecular mechanisms of craniofacial syndromesneural crest cell contribution to facial formationorofacial cleftPierre Robin sequenceSingle-Cell RNA SequencingSOX9SOX9 gene regulation in facial developmentTFAP2Atransgenic mouse models of facial development
