Skeletal stem and progenitor cells maintain cranial suture patency and prevent craniosynostosis

Abstract

Cranial sutures are major growth centers for the calvarial vault, and their premature fusion leads to a pathologic condition called craniosynostosis. This study investigates whether skeletal stem/progenitor cells are resident in the cranial sutures. Prospective isolation by FACS identifies this population with a significant difference in spatio-temporal representation between fusing versus patent sutures. Transcriptomic analysis highlights a distinct signature in cells derived from the physiological closing PF suture, and scRNA sequencing identifies transcriptional heterogeneity among sutures. Wnt-signaling activation increases skeletal stem/progenitor cells in sutures, whereas its inhibition decreases. Crossing Axin2^LacZ/+ mouse, endowing enhanced Wnt activation, to a Twist1^+/- mouse model of coronal craniosynostosis enriches skeletal stem/progenitor cells in sutures restoring patency. Co-transplantation of these cells with Wnt3a prevents resynostosis following suturectomy in Twist1^+/- mice. Our study reveals that decrease and/or imbalance of skeletal stem/progenitor cells representation within sutures may underlie craniosynostosis. These findings have translational implications toward therapeutic approaches for craniosynostosis.

Fig. 1. Temporal FACS profiling of CD51⁺;CD200⁺ cells in patent and fusing cranial sutures.

a Outline of calvarial suture harvesting and processing. b Time-course profiling of skeletal stem/progenitor cells isolated in vivo by flow cytometric analysis from freshly harvested posterior frontal (PF), sagittal (SAG), and coronal (COR) sutures. Tibial growth plate (GP) was used as reference. Values are given as a percentage of the P3 population, representing viable (propidium iodide-), non-hematopoietic (CD45−) cells isolated from the sutures. The fusing PF suture shows a dramatic decrease in the percentage of skeletal stem/progenitor cells between day pN7 and pN9, corresponding to the initiation of PF suture closure. Conversely, the patent SAG and COR sutures and GP maintain a sustained percentage of these cells over time. Data are represented as means ± SEM; PF vs SAG P = 0.002, PF vs COR P = 0.0147, PF vs GP P = 0.0179, *P ≤ 0.05 unpaired, two-tailed student t-test were performed. n = 60 animals/timepoint, experiments were repeated 6 independent times for pN3, 4 for pN17 and 2 independent times for pN5, 7, 9, 11, and 15. c Diagram illustrating the timing of the endochondral differentiation through which the PF suture closes. d Representative FACS plots of skeletal stem/progenitor (CD51⁺;CD200⁺) cells isolated in vivo from the PF, SAG, and COR sutures and GP at pN3, pN9, and pN17. CD51⁺;CD200⁺ cells decrease as a function of time in the physiologically fusing PF suture. Refer to Supplementary Fig. 1 for complete FACS gating and isolation strategy. Source data are provided as a Source data file.

Fig. 2. Transcriptomic profiling of CD51⁺;CD200⁺ cells from fusing and non-fusing sutures.

a Venn diagram of top differentially expressed genes (up- or downregulated by ≥3 fold) from bulk RNA-seq evaluation of skeletal stem/progenitor cells (CD51⁺;CD200⁺) isolated from different sutures. b Heatmap representation of top genes associated with syndromic craniosynostosis in CD51⁺;CD200⁺ cells isolated from PF sutures^–. Yellow: upregulation; blue: downregulation. c Schematic outlining single-cell RNA analysis. d UMAP embedding showing the distribution of transcriptional programs for single CD51⁺;CD200⁺ cells isolated from PF, SAG, and COR sutures partitioned into five clusters. Cells are colored either by cluster (i) or suture of origin (ii). e Heat map of top 10 differentially expressed genes characteristic to each of the five transcriptionally-defined clusters. Yellow: upregulation; purple: downregulation. f Pie chart showing the representation of CD51⁺;CD200⁺ cells from each suture within each cluster. Sequencing data were deposited in the GEO database as GSE69909. All data can be accessed from the Gene Expressions Omnibus (http://www.ncbi.nlm.nih.gov/geo/) using accession number GSE138882. PF posterior frontal, SAG sagittal, COR coronal.

Fig. 3. Imbalance of CD51⁺;CD200⁺ cells in syndromic and non-syndromic craniosynostosis.

a Representative Movat’s pentachrome staining showing synostosis of COR suture in day pN15 Twist1^+/− mouse, compared to wild-type suture. Scale bars: 100 μm. Magnification at ×20. n = 3 animals/group, experiments were repeated 3 independent times. b FACS profiling of CD51⁺;CD200⁺ cells isolated from explanted wild-type and Twist1^+/− COR sutures at day pN3 reveals a significant decrease of these cells in Twist1^+/− COR sutures in comparison to wild-type. Values are normalized as total CD51⁺;CD200⁺ cells per 500,000 events. Data are represented as means ± SEM; P = 0.0482, *P ≤ 0.05, unpaired, two-tailed student t-test were performed. n = 10 animals/timepoint, experiments were repeated 4 independent times. c Whole mount top view (left panel) and Movat’s pentachrome staining (right panel) of coronal sections of untreated and SB431542-treated SAG suture explants (day pN3) harvested after 8 days in culture. SB431542-treated SAG suture lacks the suture mesenchyme, with bony tissue (black arrow) replacing the suture mesenchyme and fused. Scale bars: 200 μm. Magnification at ×40. n = 5 animals/group, experiments were repeated 3 independent times. d–f PCR analysis of Smad6, Id2, and Bglap expression in SB431542-treated and untreated SAG sutures confirming the effectiveness of SB431542 treatment and inhibition of TGFβ signaling paralleled by activation of BMP signaling. Values represent mean ± SEM; Smad6; P = < 0.0001, Id2; P = 0.0001, Bglap; P = < 0.0001, ***P ≤ 0.001, ****P ≤ 0.0001, unpaired, two-tailed student t-test were performed. n = 5 animals/group experiments were repeated 3 independent times. g FACS profiling of CD51⁺;CD200⁺ cells isolated from day pN3 untreated and treated sutures harvested after 8 days in culture shows a decrease of these cells in SB431542-treated SAG suture explants. Values are normalized as a total CD51⁺;CD200⁺ cells per 500^,000 events. Values represent mean ± SEM; P = 0.0478, *P ≤ 0.05, unpaired, two-tailed student t-test were performed. n = 20 animals/group, experiments were repeated 3 independent times. SAG sagittal, COR coronal. Source data are provided as a Source data file.

Fig. 4. Modulation of cWnt signaling affects CD51⁺;CD200⁺ cell frequency in the suture mesenchyme.

a Schematic representation of the experimental procedures. TAM, Tamoxifen. b Movat’s pentachrome staining of coronal sections from CD-1 mouse PF sutures treated in vivo at day pN4 with Wnt3a (150 ng) and harvested at day pN8 (normal onset of fusion). Movat’s pentachrome staining shows a largely expanded suture mesenchyme in Wnt3a-treated sutures relative to PBS and untreated controls. Top: no treatment, middle: Wnt3a, bottom: PBS. Green bars mark the distance between the two osteogenic fronts. Scale bars: 100 μm. Magnification at ×20. n = 3 animals/group, experiments were repeated 3 independent times. c FACS profiling of freshly harvested Wnt3a-treated PF sutures at day pN8 reveals an increased representation of CD51⁺;CD200⁺ cells in these sutures relative to untreated and PBS controls. Values represent mean ± SEM; Wnt3a vs No Treatment, P = 0.0025, Wnt3a vs PBS, P = 0.046 *P ≤ 0.05, **P ≤ 0.01, unpaired, two-tailed student t-test were performed. n = 20 animals/group, experiments were performed 6 independent times. d FACS profile of freshly harvested SAG sutures at pN8 following in vivo treatment with cWnt signaling inhibitors sFrp-1 (2 μg) and Dkk-1 (2 μg) or PBS. Treatment with inhibitors decreased CD51⁺;CD200⁺ cells in comparison to PBS and untreated controls. Values represent mean ± SEM. n = 20 animals/group, experiments were performed two independent times. e Movat’s pentachrome staining of SAG sutures treated with cWnt signaling inhibitors sFrp-1 (2 μg) and Dkk-1 (2 μg) and relative controls at day pN17. Green bars mark the distance between the two osteogenic fronts. Scale bars: 100 μm. Magnification at ×20. Top: no treatment, middle: Dkk-1/sFrp-1, bottom: PBS. n = 3 animals/group, experiments were repeated 3 independent times. PF posterior frontal, SAG sagittal. Source data are provided as a Source data file.

Fig. 5. Sustained activation of cWnt signaling rescues COR suture craniosynostosis in Twist1^+/− mice.

a Pentachrome staining and histological analysis of pN15 COR sutures from wild-type (top panel), Twist1^+/− (middle panel) and Twist1^+/−:Axin2^LacZ/+ (bottom panel) mice. Similar to the wild-type suture, the Twist1^+/−:Axin2^LacZ/+ COR suture is patent with a clear undifferentiated tissue mesenchyme between the osteogenic fronts, thus demonstrating a rescue of the COR suture craniosynostosis. Conversely, the Twist1^+/− COR suture is fused, showing a bony bridge and lack of a suture mesenchyme. Scale bars, 100 μm. Magnification at ×20. n = 3 animals/group, experiments were repeated 3 independent times. b FACS analysis of CD51⁺;CD200⁺ cells isolated from COR sutures of wild-type, Axin2^LacZ/+, Twist1^+/−, and Twist1^+/−:Axin2^LacZ/+ mice indicates increased representation of CD51⁺;CD200⁺ cells in the Twist1^+/−:Axin2^LacZ/+ double transgenic. Values represent mean ± SEM; Axin2^LacZ/+ vs. Twist1^+/− P = 0.025, Twist1^+/− vs. Twist1^+/−:Axin2^LacZ/+, P = 0.0172; *P ≤ 0.05, unpaired, two-tailed student t-test were performed. n = 10 animals/group, experiments were repeated 3 independent times. c Screening for coronal suture fusion in Twist1^+/− versus Twist1^+/−:Axin2^LacZ/+ mice reveals a frequency of COR craniosynostosis in Twist1^+/−:Axin2^LacZ/+ mice restored to that of wild-type mice (i.e., patent COR suture phenotype), compared to the widespread fusion seen in Twist1^+/−. Screening was performed on mice ranging from day pN6 up to 6 months. n = 282 animals screened. d (top panel) X-gal staining identifies a of cWnt-activated cells within the mesenchyme of the patent COR suture in Twist1^+/−: Axin2^LacZ/+ double transgenic mice at day pN3, pN18, and 6 months postnatal. (bottom panel) Movat’s pentachrome staining performed on adjacent slide sections as for top panels. Scale bars: 100 μm. Magnification at ×20. n = 3 animals/group, experiments were repeated 3 independent times. pN post-natal. e Micro-CT analysis of wild-type, Twist1^+/−:Axin2^LacZ/+, Twist1^+/−, and Axin2^LacZ/+ skulls at 6 months postnatal reveals the presence of a wide patent COR suture in the Twist1^+/−:Axin2^LacZ/+ mouse, whereas a unilateral fused COR suture is observed in the Twist1^+/− mouse. n = 3 animals/group, experiments were repeated 3 independent times. COR coronal. Source data are provided as a Source data file.

Fig. 6. Transplantation of CD51⁺;CD200⁺ cells and Wtn3a prevents re-synostosis following surgical repair of craniosynostosis.

a Schematic representation of experimental procedure for COR suture ablation experiments. pN16 Twist1^+/− COR suture synostosis was confirmed by preoperative micro-CT scans. Fused COR sutures were excised using a 0.3 mm drill and replaced by wild-type CD51⁺;CD200⁺ cells (3 × 10⁵ cells) and recombinant Wnt3a protein (200 ng) loaded onto a collagen scaffold sponge, thus engineering a “suture mesenchyme”. An empty suturectomy was used as control. The extent of re-fusion was evaluated over a 14-week period by micro-CT and histological analysis. POD post-operative day, PO post-operative. b Time-course micro-CT analysis of treated and untreated COR suturectomies over a 14-week period. Untreated suturectomy re-fuses in contrast to the treated suturectomy which remains patent through 14-weeks. n = 2 animals/group, experiments were repeated 3 independent times. c Quantification of COR suturectomy re-fusion by micro-CT over a 14-week period measuring image intensity in a standardized ROI encompassing the suturectomy using the Adobe Photoshop measurement tool. Untreated suturectomies (blue) show increased bone surface measurements over the 14-week period compared to suturectomy treated with a combination of CD51⁺;CD200⁺ cells and Wnt3a protein (red) confirming the untreated suturectomy is refusing over-time. n = 2 animals/group, analysis was performed one time. d Movat’s pentachrome staining of sagittal sections of COR suturectomies at 14-weeks showing fusion of the untreated suturectomies compared to suturectomies treated with cells and Wnt3a protein. Magnification ×5, scale bars: 200 μm. n = 2 animals/group, experiments were repeated 3 independent times. e wild-type GFP⁺ CD51⁺;CD200⁺ cells (3 × 10⁵ cells) coached with recombinant Wnt3a protein (200 ng) loaded onto a collagen scaffold sponge were transplanted as above to monitor their presence within the ablated COR suture at 10-week post-operative. (top panel), Movat’s pentachrome staining of a COR suturectomies sagittal section. Magnification ×10, scale bar: 100 μm. (bottom panel), confocal micrograph of a sister slide of COR suturectomies sagittal section. Fluorescence indicates the presence of transplanted cells within the ablated suture. Magnification ×10, scale bar: 100 μm. f Magnification at ×20 of the boxed area in (e), suggests that transplanted CD51⁺;CD200⁺ cells participate in forming suture mesenchyme by recruiting neighboring cells Magnification ×20, scale bar: 50 μm. n = 2 animals/group, experiments were repeated 3 independent times. COR coronal. Source data are provided as a Source data file.