The research, led by Hanshu Zhang, Xueying Ma, and colleagues at Wuhan University and Hangzhou Normal University, tackles a long-standing puzzle in craniofacial biology. RUNX2 is the essential transcription factor driving osteoblast differentiation throughout the skeleton. Mice carrying only one functional copy of Runx2 develop cleidoccranial dysplasia, with failed ossification in both the posterior frontal (PF) suture, which normally fuses after birth, and the sagittal (SAG) suture, which must remain open for life. Yet earlier work had shown something strange: blocking a microRNA cluster called Mir338 rescued the ossification defect in the mesoderm-derived SAG suture but left the neural crest-derived PF suture untouched, even though Runx2 was equally compromised in both.

To understand this paradox, the team combined lineage tracing, bulk RNA and microRNA sequencing, and single-cell RNA sequencing of the two suture types in postnatal day 7 mice. Their comparative profiling revealed that the PF and SAG sutures harbor distinct miRNA-mRNA regulatory landscapes during osteogenic differentiation. Among the microRNAs downregulated during bone formation was miR-338-3p, known to target Runx2 itself. Genes upregulated in the PF suture included the neural crest marker Pax3, while the SAG suture specifically expressed Six2, a marker of coronal suture progenitors, confirming that lineage identity persists molecularly long after embryogenesis.

Genetic experiments confirmed the lineage-specific rescue in vivo. When the researchers crossed Runx2 haploinsufficient mice with Mir338 knockout mice, the double mutants showed a significant narrowing of the sagittal suture gap, indicating partial restoration of bone formation, while the posterior frontal defect persisted. Lineage tracing using Wnt1-Cre and Prrx1-Cre reporters showed that Mir338 ablation increased the number of mesoderm-derived Prrx1-positive osteoprogenitors in the SAG suture but failed to restore the depleted Wnt1-positive neural crest population in the PF suture. Proliferation, measured by Ki67 staining, recovered in both sutures, but alkaline phosphatase activity, a hallmark of osteogenic differentiation, was rescued only in the SAG suture.

Single-cell analysis then pointed to mitochondria as the decisive factor. The mature osteoblast cluster, marked by Bglap, Ibsp, and Dmp1, shrank in both sutures of Runx2 haploinsufficient mice, but Mir338 knockout restored this population only in the SAG suture. Differentially expressed genes in the rescued subcluster were strongly enriched for mitochondrial respiration components, including Cdk8 and electron transport chain genes such as mt-Co1, mt-Co2, and mt-Atp6. Mature osteoblasts are known to depend on fragmented, donut-shaped mitochondria and oxidative phosphorylation, making mitochondrial dynamics a plausible bottleneck for bone formation.

Imaging and functional assays bore this out. TOMM20 immunostaining revealed reduced mitochondrial content at the osteogenic fronts of Runx2 haploinsufficient mice, and stimulated emission depletion microscopy showed impaired mitochondrial fragmentation, with fewer and longer mitochondria in SAG-derived osteoblasts. Seahorse metabolic profiling demonstrated that basal respiration and ATP production dropped significantly in SAG suture-derived cells from Runx2 haploinsufficient mice, and both recovered with Mir338 ablation. In PF-derived cells, by contrast, respiration and ATP output were largely preserved despite the ossification defect, suggesting that neural crest-derived osteoblasts are not critically dependent on high respiratory output for their differentiation program.

The team traced the upstream wiring to a RUNX2-NDUFS5 axis. NDUFS5, a nuclear-encoded subunit of mitochondrial complex I, was downregulated in osteoblast lineages of Runx2 haploinsufficient mice, and integrated ChIP-seq and CUT&Tag data showed that RUNX2 directly binds the Ndufs5 promoter at an active enhancer marked by H3K27ac. Knocking down Ndufs5 in MC3T3-E1 pre-osteoblasts disrupted mitochondrial fragmentation and reduced mineralization. Most strikingly, transplanting isolated mitochondria from Mir338 knockout donor cells into Runx2-knockdown recipient cells restored mitochondrial DNA copy number, osteogenic marker expression, and even Runx2 levels. In an ex vivo cranial suture culture, mitochondrial transplantation narrowed the SAG suture osteogenic front by up to 306 micrometers while producing only a modest, non-significant effect on the PF suture, faithfully reproducing the lineage-specific rescue seen in living mice.

The molecular switch at the heart of this specificity turned out to be CDK8, a cyclin-dependent kinase dually regulated by RUNX2 and Mir338. RUNX2 directly activates Cdk8 transcription, while miR-338-3p directly binds and represses the Cdk8 3′ untranslated region, a interaction confirmed by dual-luciferase reporter assays. CDK8 in turn promotes mitochondrial fission by interacting with DRP1, the master regulator of mitochondrial division, as demonstrated by co-immunoprecipitation and colocalization on mitochondria. In the SAG suture, nearly all Prrx1-positive mesoderm cells expressed CDK8, and Mir338 ablation restored CDK8 levels in Runx2 haploinsufficient mice. Overexpressing Cdk8 in Runx2-knockdown cells partially rescued osteogenic markers, establishing CDK8 as the functional effector of the dosage-sensitive toggle between RUNX2 and Mir338.

The PF suture, however, plays by different rules. Neural crest-derived cells showed intrinsically low Cdk8 expression, and bisulfite sequencing revealed that CpG sites in the Cdk8 promoter are heavily hypermethylated in PF cells compared with SAG cells. This epigenetic silencing, likely a lasting memory of the neural crest’s developmental journey, prevents RUNX2 from effectively activating Cdk8 even though the transcription factor can physically reach the locus. Consequently, the entire RUNX2-Mir338-CDK8-mitochondrial circuit is functionally inert in the frontal suture, and no amount of microRNA inhibition or mitochondrial supplementation can rescue a defect whose bottleneck lies elsewhere.

Beyond solving a specific developmental puzzle, the work reframes how scientists think about congenital disease. The authors propose that a cell’s embryonic origin encodes a metabolic and epigenetic identity that determines which genetic lesions it can tolerate and which therapies will work. For disorders like cleidocranial dysplasia, this suggests future treatments must respect lineage-specific constraints: mesoderm-derived defects may respond to metabolic interventions such as mitochondrial transplantation or CDK8 modulation, while neural crest-derived defects may require epigenetic drugs to derepress silenced effectors first. As the cranial suture lineages and their regulatory networks are conserved between mice and humans, these findings offer a conceptual blueprint for precision craniofacial medicine, where therapy is matched not just to a mutation but to the developmental history of the tissue it must repair.

Subject of Research: Lineage-specific regulation of cranial suture osteogenesis by a Mir338-CDK8-mitochondrial axis in Runx2 haploinsufficiency

Article Title: A Mir338 -CDK8-mitochondrial axis governs lineage-specific osteogenesis and defines Runx2 pleiotropy in cranial sutures

Article References: Zhang, H., Ma, X., Peng, J., Shen, Q., Jin, R., Jia, Y., Zhou, Y., Chen, Z., & Liu, H. (2026). A Mir338-CDK8-mitochondrial axis governs lineage-specific osteogenesis and defines Runx2 pleiotropy in cranial sutures. Journal of Advanced Research. https://doi.org/10.1016/j.jare.2026.10.029

Image Credits: AI Generated

DOI: Not provided

Keywords: cranial sutures, RUNX2, Mir338, CDK8, mitochondrial dynamics, osteogenesis, neural crest, mesoderm, epigenetic methylation, cleidocranial dysplasia, single-cell RNA sequencing, mitochondrial transplantation