Graphical abstract

Open in a separate window

Abstract

Introduction

Cholesterol, which is abundant in cellular membranes, is responsible for their properties and plays a role in membrane trafficking. More than 20 enzymatic reactions are involved in cholesterol biosynthesis and degradation through negative and positive feedback mechanisms [1]. Mutations in the human lathosterol 5-desaturase (SC5D) gene, which catalyzes the conversion of lathosterol to 7-dehydrocholesterol in the second to last step of endogenous cholesterol synthesis, cause lathosterolosis, a disease characterized by craniofacial abnormalities, microcephaly, cataracts, and skeletal defects [1]. Although the metabolites and enzymes in cholesterol metabolic pathways play a crucial role in osteogenesis [2], it remains unclear how cholesterol metabolism aberrations contribute to craniofacial bone formation. Interestingly, Sc5d^-/- mice die at birth with various malformations, including cleft palate, micrognathia, and abnormal limb formation [3].

The bone extracellular matrix contains collagens, which provide a specific niche for the resident cells to function, and their production is regulated through a series of multiple regulatory pathways [4]. The primary cilium, a cholesterol-enriched antenna-like structure that senses extracellular cues on cell surfaces, recruits and retains receptors and membrane proteins [5] and serves as a signaling center for various signaling pathways, including hedgehog (HH) and WNT/β-catenin signaling. Defects in primary cilia lead to a group of disorders associated with ciliary abnormalities (a.k.a. ciliopathies) characterized by craniofacial anomalies such as micrognathia, craniosynostosis, hypertelorism, and cleft palate, as well as immature lungs and enlarged bladders [6], [7]. The primary cilium comprises the axoneme, motor dynein and kinesin, cargos, and the ciliary membrane, which is a cholesterol-enriched membrane distinct from the plasma membrane. Phenotypic similarities between ciliopathies and Sc5d deficiency in mice suggest that cholesterol metabolism plays a crucial role in bone development through the primary cilium. In this study, we show that cholesterol metabolism aberrations inhibit osteogenesis due to a failure in HH and WNT/β-catenin signaling associated with the primary cilia in Sc5d null mice.

Materials and methods

Results

Micrognathia in Sc5d null mice

Sc5d knockout (KO) mice exhibited craniofacial deformities, including cleft palate, open eyelids, short head, and micrognathia (extremely short bone in the anterior half of mandible) (Fig. 1A). In addition, Sc5d KO mice exhibited limb abnormalities (either syndactyly or polysyndactyly) in both forelimbs and hindlimbs, as well as defects in endochondral ossification in the radius/ulna and tibia/fibula (Fig. S1). To investigate bone formation in Sc5d KO mice, we performed skeletal staining and found that craniofacial bones, especially the mandible, in E18.5 Sc5d KO mice were smaller than the ones in wild-type (WT) control mice, with less mineralized frontal and parietal bones and no calcification in the intraparietal bone (Fig. 1B). The size and shape of the mandible were further evaluated with reconstructed 3D micro-computed tomography (microCT) analyses in Sc5d KO and WT control mice. We found that the mandible in Sc5d KO mice was shorter in length [distance between the most anterior point of the mandible (#1) and the posterior point of the condylar process (#4)], accompanied by the volume and surface area of the mandible, although the height [distance between the molar alveolus of dentary (#2) and the inferior point of the mandibular body (#3)] was not significantly altered in Sc5d KO mice compared with WT (Fig. 1C, D). Histological analyses showed that although osteoblast condensation at the osteogenic front occurred normally in the mandible of Sc5d KO mice, intramembranous ossification around the Meckel’s cartilage was compromised in Sc5d KO mice, resulting in a reduced calcified bone area in E18.5 Sc5d KO mice compared to WT control mice (Fig. 1E).

Open in a separate window

Fig. 1

Osteogenesis defects in Sc5d null mice. (A) Side view of the craniofacial region of E18.5 Sc5d null (right) and wild-type (WT) control (left) mice. (B) Skeletal staining with Alcian blue (cartilage) and Alizarin red S (bone) in E18.5 Sc5d null (right) and WT control (left) mice. Side view (top panels) and top view of the calvaria (bottom panels). (C) MicroCT images of the skull in E18.5 Sc5d null (right) and WT control (left) embryos. Side view (top panels), top view of the calvaria (middle panels), and mandible (bottom panels); Scale Bar: 1 mm 1. Most anterior point of the mandible; 2. molar alveolus of dentary; 3. inferior point of the mandibular body; 4. posterior point of the condylar process. (D) Measurement of height (2 – 3) and length (1 – 4) of the mandible in (C). WT (blue bars) and Sc5d null (orange bars). ***p < 0.001. (E) Hematoxylin and Eosin staining in Sc5d null (right) and WT control (left) mice at E13.5 and E18.5. The boxed areas are enlarged. Scale bars, 100 μm in top panels and 20 μm in bottom panels. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Unaltered cell survival in Sc5d KO mice

There are several possible causes for the micrognathia phenotype in Sc5d KO mice: reduced cell proliferation, increased cell death, dysregulated osteoblastogenesis, or a combination of these. To determine the cellular mechanism underlying the bone defects in Sc5d KO mice, we performed BrdU incorporation assays in Sc5d KO and WT control mice at E12.5, E13.5, and E14.5 and found that there was no defect in cell proliferation at the active site of intramembranous ossification of the mandible in Sc5d KO mice (Fig. 2A, B). We further confirmed that there were no cell proliferation defects in our immunohistochemical analysis for Ki67 in Sc5d KO osteoblasts labeled with RUNX2, an early osteoblast marker (Fig. S2A, B). Furthermore, cell proliferation defects in Sc5d KO osteoblasts were confirmed by cell proliferation assays using primary osteoblasts isolated from the calvaria of Sc5d KO and WT control mice (Fig. S3A). In addition, there was no change in cell death in the mandible of Sc5d KO mice, compared to that of WT control mice, at E12.5, E13.5, and E14.5 (Fig. 2C). Taken together, we conclude that the defects in osteogenesis in Sc5d KO mice were not caused by defects in osteoblast survival and proliferation.

Open in a separate window

Fig. 2

Suppression of osteoblast differentiation and collagen expression in Sc5d null mice. (A) BrdU incorporation assays at the intramembranous ossification region of the mandible in Sc5d null (bottom) and WT control (top) mice at E12.5, E13.5, and E14.5. Methylene blue was used for counterstaining. Scale bars, 50 μm. (B) Quantification of BrdU-positive cells per total number of cells (%) in B. WT (blue bars) and Sc5d null (orange bars). NS: not significant. (C) TUNEL staining (red) at the intramembranous ossification region of the mandible in Sc5d null (bottom) and WT control (top) mice at E12.5, E13.5, and E14.5. Scale bars, 50 μm. (D) Immunoblotting of COL1, SPP1, and GAPDH (internal control) in the mandible of Sc5d null (KO) and WT control mice at E14.5 and E18.5. (E) Immunohistochemical staining for RUNX2, COL1, and SPP1 at the intramembranous ossification region of the mandible in Sc5d null and WT control mice at E13.5 and E14.5. Methylene blue was used for counterstaining. Scale bars, 100 μm. (F) Relative mRNA expression of the indicated osteoblast differentiation markers in the mandible of Sc5d null (orange bars) and WT control mice (blue bars) at E13.5 and E14.5. ***p < 0.001. n = 6 per group. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Compromised primary cilia-mediated hedgehog and WNT/β-catenin signaling in Sc5d null osteoblasts

Our recent study shows that cholesterol metabolism plays a crucial role in the formation of primary cilia [8]. To investigate whether and how ciliogenesis was altered in Sc5d KO osteoblasts, we conducted immunocytochemical analyses for the primary cilia and found that their length in Sc5d KO osteoblasts was shorter than in WT control osteoblasts, and that the number of ciliated Sc5d KO osteoblasts was lower than that seen in WT control osteoblasts in cell culture and in the mandibles of E13.5 and E14.5 embryos (Fig. 3A, B; Fig. S4A–C). The primary cilia can regulate HH and WNT/β-catenin signaling [12], [13], [14]; therefore, we hypothesized that HH and WNT/β-catenin signaling was compromised due to a failure in primary cilium formation in Sc5d KO osteoblasts. First, to evaluate HH signaling activity, we employed Gli1-LacZ reporter mice and found that HH signaling activity was suppressed in osteoblasts in Sc5d^-/-;Gli1-LacZ KO mice compared to Sc5d^+/+;Gli1-LacZ WT control mice at E13.5 (Fig. 3C). Furthermore, we performed in situ hybridization and quantitative RT-PCR for genes regulated by HH signaling (Gli1, Gli3, and Ptch1) and found that HH signaling was downregulated in the developing mandible of Sc5d KO mice (Fig. 3D, E).

Open in a separate window

Fig. 3

Failure in ciliogenesis leading to the suppression of hedgehog signaling in Sc5d null osteoblasts. (A) Immunocytochemical analysis for acetylated tubulin (primary cilium: green) and γ-tubulin (basal body: red) in cultured osteoblasts isolated from E18.5 Sc5d null (right) and WT control (left) mice. DAPI was used for counterstaining (blue). Scale bar, 5 μm. (B) Quantification of the number of ciliated cells (left) and the length of the primary cilia (right) in cultured osteoblasts isolated from E18.5 Sc5d null and WT control mice using 6 randomly selected fields from A. ***p < 0.001. (C) β-galactosidase staining (blue) at the intramembranous ossification region of the mandible in Sc5d^-/-;Gli1-LacZ (right, mutant) and Sc5d^+/+;Gli1-LacZ (left, control) mice at E14.5. Nuclear Fast Red was used for counterstaining (red). MC, Meckel’s cartilage. Scale bars, 50 μm. (D) In situ hybridization for Gli1 and Ptch1 in coronal sections of the mandible of E14.5 wt (left) and Sc5d KO (right) mice. Hematoxylin was used for counterstaining (blue). Scale bars, 100 μm. (E) Relative mRNA expression for readout genes for the hedgehog signaling in the mandible of Sc5d null (orange bars) and WT control (blue bars) mice at E13.5 (left) and E14.5 (right). ***p < 0.001; *p < 0.05. (F) β-galactosidase staining (blue) at the intramembranous ossification region of the mandible in Sc5d^-/-;Topgal (right, mutant) and Sc5d^+/+;Topgal (left, control) mice at E14.5. Nuclear Fast Red was used for counterstaining (red). MC, Meckel’s cartilage. Scale bars, 50 μm. (G) In situ hybridization for Axin2 and Lef1 in coronal sections of the mandible from E14.5 wild-type (WT, left) and Sc5d KO (right) mice. Hematoxylin was used for counterstaining (red). Scale bars, 100 μm. (H) Relative mRNA expression for readout genes for the WNT/β-catenin signaling in the mandible of Sc5d null (orange bars) and WT control (blue bars) mice at E13.5 (left) and E14.5 (right). ***p < 0.001; *p < 0.05. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Next, we analyzed WNT/β-catenin signaling activity with Topgal reporter mice and found that signaling was suppressed in osteoblasts in Sc5d^-/-;Topgal mice, compared to Sc5d^+/+;Topgal control mice, at E13.5 (Fig. 3F). Furthermore, we confirmed that WNT/β-catenin signaling was downregulated in the developing mandible of Sc5d KO mice with in situ hybridization and quantitative RT-PCR for genes regulated by WNT/β-catenin signaling (Axin2 and Lef1) (Fig. 3G, H).

To confirm that expression of Col1a1 and Spp1 was directly regulated through WNT/β-catenin and HH signaling, we treated osteoblasts isolated from Sc5d KO and WT control mice with Lithium Chloride (LiCl) or WNT3A for inducing WNT/β-catenin signaling, and with Smoothened agonist (SAG) for inducing HH signaling, and found that expression of Col1a1 and Spp1 was induced in WT control osteoblasts, but not in Sc5d KO osteoblasts (Fig. 4A–C). In addition, we confirmed that the activation of either HH and WNT/β-catenin signaling pathway partially induced osteogenic differentiation in Sc5d KO osteoblasts (Fig. 4D). Our previous study showed that the Col1a1 promoter contains the GLI-binding motif (5ʹ-CACCACCCA-3ʹ: −68 bp to −60 bp) and LEF1-binding motifs (minimal core sites: 5ʹ-CTTTG-3ʹ; −29 bp to −25 bp) [8]. We confirmed that both HH and WNT/β-catenin signaling pathways could directly regulate Col1a1 expression (Fig. 4E, F). Similarly, we analyzed the promoter region of Spp1 and found that there were three GLI-binding motifs and three LEF1-binding motifs (Fig. S5). LEF1, but not GLI, bound to the Spp1 promoter region in chromatin immunoprecipitation (ChIP) assays (Fig. 4G, H), indicating that Spp1 expression is regulated through WNT/β-catenin signaling. Taken together, our results show that HH and WNT/β-catenin signaling can differentially regulate expression of COL1 and SPP1, leading to intramembranous ossification defects.

Open in a separate window

Fig. 4

Regulation of expression of Col1a1 and Spp1 by HH and WNT/β-catenin signaling. (A) Quantitative RT-PCR for the indicated genes in cultured Sc5d null (orange and red bars) and WT control (blue and green bars) osteoblasts treated with LiCl or vehicle. ***p < 0.001. (B) Quantitative RT-PCR for the indicated genes in cultured Sc5d null (orange and red bars) and WT control (blue and green bars) osteoblasts treated with WNT3A or vehicle. **p < 0.01; ***p < 0.001. (C) Quantitative RT-PCR for the indicated genes in cultured Sc5d null (orange and red bars) and WT control (blue and green bars) osteoblasts treated with SAG or vehicle. **p < 0.01; ***p < 0.001. (D) Alizarin Red staining at 14 days of osteogenic differentiation in cultured primary osteoblasts treated with vehicle, LiCl, WNT3A, or SAG. (E) Schematic diagram (top) of binding sites (BSs) for LEF1/active β-catenin (green) and GLI (orange) in the promoter region (5 kb upstream from transcription start site) of Col1a1. Conserved BSs among 3 species were selected for experimental validation. Chromatin immunoprecipitation (ChIP) assay using anti-GLI antibody on Col1a1 in cultured osteoblasts from Sc5d null (orange bars) and WT control (blue bars) mice. ns, not significant; ***p < 0.001. (F) ChIP assay using anti-LEF1 antibody on Col1a1 in cultured osteoblasts from Sc5d null (orange bars) and WT control (blue bars) mice. ns, not significant; ***p < 0.001. (G) Schematic diagram (top) of binding sites (BSs) for LEF1/active β-catenin (green) and GLI (orange) in the promoter region (5 kb upstream from transcription start site) of Spp1. Conserved BSs among 3 species were selected for experimental validation. ChIP assay using anti-GLI antibody on Spp1 in cultured osteoblasts from Sc5d null (orange bars) and WT control (blue bars) mice. ns, not significant. (H) ChIP assay using anti-LEF1 antibody on Spp1 in cultured osteoblasts from Sc5d null (orange bars) and WT control (blue bars) mice. ***p < 0.001. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

To further confirm the functional significance of proper cholesterol metabolism in the regulation of primary cilium-related HH and WNT/β-catenin signaling in osteoblast differentiation, we conducted rescue experiments with simvastatin, a drug that can normalize cholesterol metabolism. We found that gene and protein expression of Col1a1 and Spp1 was almost completely restored in osteoblasts treated with simvastatin (Fig. 5A, B). Furthermore, we confirmed that defects in osteogenic differentiation in Sc5d KO osteoblasts were partially rescued with simvastatin through primary cilia-mediated HH and WNT/β-catenin signaling pathways (Fig. 5C–F).

Open in a separate window

Fig. 5

Normalized osteogenic differentiation in Sc5d osteoblasts with simvastatin treatment. (A) Quantitative RT-PCR for the indicated genes in cultured Sc5d null (orange and red bars) and WT control (blue and green bars) osteoblasts treated with simvastatin or vehicle. ns, not significant; **p < 0.01; ***p < 0.001. (B) Immunoblotting of COL1, SPP1, and GAPDH (an internal control) in cultured primary osteoblasts treated with simvastatin or vehicle. (C) Alizarin Red staining at 14 and 28 days of osteogenic differentiation in cultured primary osteoblasts treated with simvastatin or vehicle. (D) Quantitative RT-PCR for Axin2 and Lef1 for WNT/β-catenin signaling and Gli1 and Ptch1 for HH signaling in cultured Sc5d null (orange and red bars) and WT control (blue and green bars) osteoblasts treated with simvastatin or vehicle. (E) Immunocytochemical analysis for acetylated tubulin (primary cilium, green) and γ-tubulin (basal body, red) in cultured Sc5d KO (right) and WT control (left) osteoblasts treated with simvastatin or vehicle. DAPI was used for counterstaining (blue). Scale bar, 5 μm. (F) Quantification of the number of ciliated cells (left) and length of the primary cilia (right) in cultured Sc5d KO (orange) and WT control osteoblasts. ***p < 0.001. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Discussion

Various molecules are modified with cholesterol, but the role of cholesterol modification remains largely unknown [15], [16], [17], [18], [19]. The accumulation of immature cholesterol intermediates caused by mutations in genes crucial for cholesterol synthesis may have a different compensational capability in function. Lathosterolosis is a very rare autosomal recessive congenital disorder caused by mutations in SC5D [1], [20]. To date, since only six patients have been reported, mutations in SC5D would lead to defects in early embryogenesis and miscarriages. Patients with lathosterolosis present a variety of clinical symptoms, including global developmental delay, cataracts, microcephaly and brain defects, as well as polydactyly and syndactyly, and craniofacial dysmorphology including micrognathia and a high arched palate [21], [22], [23]. In this study, we found that both HH and WNT/β-catenin signaling pathways are compromised in Sc5d KO mice. Our results show that the number of ciliated cells, as well as cilium length, is decreased in Sc5d KO osteoblasts, which is responsible for decreased HH and WNT/β-catenin signaling.

In the absence of HH ligands, Patched 1 (PTCH1) inhibits the translocation of Smoothened (SMO1) from the ciliary poach to the ciliary membrane. On the other hand, in the presence of HH ligands, PTCH1 undergoes endocytosis and ubiquitination, which induces translocation of SMO1 onto the primary cilia [24], [25]. Importantly, the interaction of cholesterol located on the ciliary membrane with SMO1 is crucial for transduction of HH signaling via inactivation of PTCH1 [26], [27], [28], [29], [30]. A recent study shows that cholesterol in the plasma membrane acts as a second messenger between PTCH1 and SMO1 [30].

Interestingly, DHCR7, which is a cholesterol biosynthesis enzyme, localizes near the base of the primary cilium and activates HH signaling, whereas stimulation of the HH signaling suppresses activity of DHCR7 and removes it from the primary cilium [31]. HH stimulation activates the cholesterol 7 alpha-hydroxylase (CYP7A1), which accumulates near the ciliary base and produces oxysterols that promote HH signaling [31]. Therefore, HH signaling might be affected through both deficient ciliogenesis and abnormal composition of the ciliary membrane. However, cholesterol amounts, as well as localization of cholesterol synthesis enzymes around the ciliary membrane, may directly affect the functions of primary cilia. Pharmacological cholesterol depletion in zebrafish embryos results in shorter and less frequent primary cilia, leading to ciliopathy-like phenotypes, such as left–right asymmetry and heart developmental defects [32]. In addition, depletion of TMEM135 (transmembrane protein 135) in Huh7 cells, a human hepatoma-derived cell line, causes lysosomal cholesterol accumulation, which results in shorter primary cilia by impaired trafficking of Rab8 to the centrioles [33]. Patients with Zellweger syndrome (ZS), an autosomal recessive peroxisome-deficient disorder, exhibit ciliopathy-like features. Fibroblasts from ZS patients and PEX1/14 double knockout cells contain less cholesterol in the ciliary membrane, resulting in downregulated HH signaling via delocalization of SMO1 on the primary cilium, although the length and number of the primary cilium remain intact [34]. These results suggest that not only cholesterol amounts in primary cilia but also intracellular cholesterol distribution plays a crucial role in ciliogenesis and HH signaling.

In contrast to HH signaling, the role of cholesterol metabolism in WNT signaling remains largely unknown. A previous study showed that binding of cholesterol-bound Dishevelled 2 (Dvl2) to Frizzled 7 (Fzd7) and subsequent activation of canonical Wnt/β-catenin signaling is greater than that of non-cholesterol-bound Dvl2 in Xenopus embryos [35]. Binding of cholesterol to Fzd5, a Wnt receptor, at the extracellular linker region enhances palmitoylation, which is crucial for maturation and trafficking of Fzd5 to the plasma membrane [36]. Exogenous cholesterol supplementation in Wnt-addicted pancreatic ductal adenocarcinoma cells enhances canonical WNT/β-catenin signaling via Fzd5, resulting in tumor growth [36]. Zebrafish with homozygous mutations in the 3-hydroxy-3-methylglutaryl-CoA synthase 1 gene (Hmgcs1) at the Vu57 allele, which encodes the first enzyme in the cholesterol synthesis pathway, exhibit severe craniofacial anomalies, including complete absence of Meckel’s, ceratohyal, and ceratobranchial cartilages due to suppression of Wnt/β-catenin signaling [37], [38]. Interestingly, activation of Wnt signaling can restore the craniofacial defects in Hmgcs1 mutant zebrafish [37], [38]. These results suggest that activation of WNT signaling is at least partially dependent on cholesterol metabolism.

Our previous study showed that mice with a deficiency in Dhcr7, another enzyme crucial for cholesterol synthesis that acts right after the SC5D enzymatic reaction, exhibit accelerated calvarial bone formation due to upregulated WNT/β-catenin signaling [8]. Interestingly, WNT/β-catenin signaling was altered in a different manner between Sc5d KO and Dhcr7 KO osteoblasts, while HH signaling was suppressed in both Sc5d KO and Dhcr7 KO osteoblasts due to defective ciliogenesis. There are several possibilities for the underlying cell and tissue-specific regulatory mechanisms: 1) difference in the toxicity of each cholesterol intermediate, and 2) difference in molecules (accessory proteins etc.) and complexes involved in the pathway. In this study, we found that there was no change in cell proliferation and survival so that accumulation of cholesterol intermediates might not be toxic together with the effects on HH and WNT/β-catenin signaling. Previous studies suggest that acylation, esterification, and conjugation of cholesterol and phospholipid on WNT3A, a canonical WNT ligand, play a crucial role in its stabilization and activation [39], [40], [41]. A recent study shows that 7-dehydrocholesterol (7-DHC) can partially compensate for the functions of mature cholesterol in Clathrin-mediated endocytosis [42]. Interestingly, stabilization of endocytic vesicles is affected by cholesterol’s synthetic status; vesicles containing 7-DHC are the most stable, followed by mature cholesterol, desmosterol, and lanosterol [43]. Therefore, aberrations in cholesterol synthesis may affect signaling activity through endocytosis.

Regarding the bone phenotype, previous studies show that HH signaling plays crucial roles in limb and digit development [44], [45], [46]. For example, Gli3 conditional KO (Shh-Cre;Gli3^Xt-J/Xt-J and Prx1-Cre;Gli3^Xt-J/Xt-J) as well as Gli3 null mice exhibit preaxial polydactyly [47], [48]. A hypomorphic mutation or spontaneous deletion of Ptch1 in Ptch1^dl/dl and Ptch1^mes/mes mice, respectively, results in preaxial polydactyly [49], [50]. Shh null mice lack all digits, tibiae, and fibulae [51], whereas mice with ectopic expression of Shh in epithelial cell lineages (K14-Shh) exhibit polysyndactyly, absence of calvarial bones, a hypoplastic mandible, and cleft palate due to expansion and increase of SHH signaling [52], suggesting that gradient SHH signaling plays a crucial role in the determination of digit number and identification. Moreover, mice without cholesterol modification on SHH (Sox2-Cre;Shh^N/+) exhibit preaxial polydactyly due to widespread SHH signaling, suggesting that proper cholesterol modification on SHH is essential for the determination of digit number and identity [53]. Mice with mutations in Ihh, an endochondral ossification morphogen (Ihh^E95K/E95K mice), exhibit brachydactyly, but not polydactyly [54]. Similar to mice with loss of HH signaling, mice with dysregulated WNT signaling display limb and digit phenotypes. For example, Dkk1 null mice exhibit polydactyly or syndactyly [55]; Lrp4 deficient mice exhibit oligodactyly or syndactyly [56]. The limb and digit phenotypes in Sc5d KO mice are different in severity and location from those in mice with loss of either HH or WNT signaling. Taken together, the phenotypic features in Sc5d KO mice are likely caused by a combination of dysregulated WNT and SHH signaling pathways due to defects in ciliogenesis. Indeed, loss of ciliary proteins results in defects in limb and digit development, as seen in Sc5d KO mice. For instance, mice deficient for ciliary proteins (e.g. Prx1-Cre;Ift88^F/- [57], Prx1-Cre;Kif3^F/- [57], and Ift80 null mice [58]) display preaxial polydactyly, dysregulated endochondral ossification, and shorter long bones. Tbx3 null mice exhibit agenesis of the ulna, fibula, and most of digits [59], and Prx1-Cre;Tbx3 conditional KO mice exhibit preaxial polydactyly and postaxial oligodactyly in the forelimbs [59].

There are some differences in endochondral and intramembranous ossification in Sc5d KO mice. Endochondral ossification forms the majority of trunk bones, whereas intramembranous ossification forms the majority of craniofacial bones. In the limbs and digits, cartilage formation is dysregulated, resulting in bent shorter bones, as well as polysyndactyly, in the middle phalanges or syndactyly. Interestingly, there is no obvious bone defect in the trunk, limbs, and digits in mice deficient for Dhcr7, which have loss of mature cholesterol as well as accumulation of cholesterol intermediates, 7-dehydrodesmosterol and 7-dehydrocholesterol. Taken together, the digit malformation and endochondral ossification defect in Sc5d KO mice are likely caused by abnormally accumulated cholesterol intermediates, lathosterol and dehydrolathosterol, but not by lack of mature cholesterol.

Conclusion

In this study, we demonstrated the role of cholesterol metabolism in the regulation of osteogenesis and ciliogenesis using Sc5d KO mice. Sc5d KO osteoblasts displayed fewer and shorter primary cilia compared to controls. The disruption of ciliogenesis led to the suppression of HH and WNT/β-catenin signaling, resulting in downregulation of the expression of Col1a1 and Spp1. The principles learned from this study may lead to the development of innovative therapeutic approaches for bone diseases related to cholesterol metabolic aberrations.

Compliance with ethics requirements.

This article does not include any studies with human subjects. All animal experiments were reviewed and approved by the Animal Welfare Committee (AWC) and the Institutional Animal Care and Use Committee of UTHealth (AWC 22–0087). All methods were performed in accordance with the relevant guidelines and regulations provided by ARRIVE (Animal Research: Reporting of In Vivo Experiments).

CRediT authorship contribution statement

Chihiro Iwaya: Investigation, Methodology, Visualization, Writing – review & editing. Akiko Suzuki: Conceptualization, Investigation, Methodology, Visualization, Writing – original draft, Writing – review & editing. Junichi Iwata: Funding acquisition, Supervision, Conceptualization, Investigation, Methodology, Writing – original draft, Writing – review & editing.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgements

Footnotes

References