Introduction

Cholesterol is a source of bile acids, steroid hormones, and oxysterols in the body and is a vital component of cellular membranes1. Cholesterol biosynthesis is regulated through multiple, highly coordinated steps, which involve over 30 reactions catalyzed by more than 15 enzymes2. Sterol-C5-desaturase (SC5D) converts cholesta-7,24-dien-3β-ol and lathosterol into 7-dehydrodesmosterol (7-DHD) and 7-dehydrocholesterol (7-DHC), whereas 7-dehydrocholesterol reductase (DHCR7) catalyzes 7-DHD and 7-DHC to desmosterol and cholesterol.

Recent studies indicate that dysregulation of cholesterol synthesis is associated with various bone disorders and diseases3. In humans, mutations in SC5D cause lathosterolosis, which is characterized by growth retardation and intellectual disability, short limbs, polydactyly/syndactyly, and craniofacial malformations including cleft palate, micrognathia, midfacial hypoplasia, and calvarial defects4. Mutations in DHCR7 cause Smith-Lemli-Opitz syndrome (SLOS), with cleft palate, postaxial polydactyly, 2–3 toe syndactyly, microcephaly, micrognathia, and intellectual disability as manifestations in humans5,6. Dysfunction of SC5D and DHCR7 leads to lathosterolosis and desmosterolosis, respectively7,8,9,10,11. Thus, these clinical manifestations and biochemical features suggest that the inborn errors associated with cholesterol biosynthesis are mainly due to the accumulation of cholesterol precursors, and not due to lower mature cholesterol levels.

As seen in humans, mice deficient for Sc5d and Dhcr7 display defects in bone formation1,3. Specifically, mice deficient for Sc5d (Sc5d−/− mice; hereafter Sc5d KO mice) exhibit cleft palate, micrognathia, agenesis of the lower incisors, calvaria hypomineralization (defects in intramembranous ossification), malformations in the long bones (defects in endochondral ossification), and syndactyly/polydactyly4,12, whereas mice deficient for Dhcr7 (Dhcr7−/− mice; hereafter Dhcr7 KO mice) exhibit accelerated calvarial bone formation and cleft palate, but only in 9% of the mutant mice13,14. Moreover, although increased levels of cholesterol precursors and lower levels of mature cholesterol in serum and tissues are commonly detected in newborn Sc5d KO and Dhcr7 KO mice4,15,16, these mouse models show different protein expression profiling in the brain17. In addition, Sc5d KO and Dhcr7 KO mice display craniofacial skeletal anomalies due to altered hedgehog and WNT/β-catenin signaling pathways at different extents and locations14,18. These results suggest that the phenotypic differences between Sc5d KO and Dhcr7 KO mice may be due to the accumulation of different cholesterol precursors.

Previous studies showed that cholesterol metabolism aberrations lead to craniofacial bone anomalies3; however, it remains unclear how an aberrant accumulation of cholesterol intermediates or loss of mature cholesterol specifically affects bone morphology. In this study, we investigated how the accumulated cholesterol precursors affect bone size in mouse models with a deficiency in cholesterol biosynthesis. Comparing the size of craniofacial bones across these mouse models allows us to identify specific areas that are affected during craniofacial development. To determine the contributions of ectopic accumulation of different cholesterol intermediates, as well as of the absence of mature cholesterol, to the pathogenesis of these diseases, we analyzed bone morphology in Sc5d KO and Dhcr7 KO mice with high-resolution μCT, and measured length and volume using 3D-reconstructed images.

Materials and methods

Animals

Sc5d+/- and Dhcr7+/− mice were a gift from Dr. Forbes D. Porter (The Eunice Kennedy Shriver National Institute of Child Health and Human Development, National Institutes of Health, Bethesda, Maryland, USA). Genotyping was performed using PCR primers as previously described4,14. All mice were bred under pathogen-free conditions, with free access to water and food and a 12-h light/dark cycle, in the UTHealth animal facility. 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).

μCT scanning and three-dimensional (3D) reconstruction

Pregnant mice were euthanized through carbon dioxide (CO2) inhalation, and embryos were euthanized through CO2 inhalation followed by decapitation, according to the American Veterinary Medical Association (AVMA) Guidelines for the Euthanasia of Animals. Embryos were collected at embryonic day (E) E18.5, fixed with 4% paraformaldehyde overnight, and stored in cold phosphate-buffered saline (PBS) until μCT scanning (n = 6 each genotype). The embryos were placed in a 12-mm diameter sample holder and stabilized with polypropylene straws during scanning. The μCT scans were performed at a 12-µm resolution using the SCANCO µCT-40 system (SCANCO medical USA Inc., USA; 55 kVp and 145 µA); 70% ethanol was used as the scan medium. The scanned images were analyzed using 3D-reconstructed μCT images generated with the Dragonfly software [Version 2021.1 for Windows. Object Research Systems (ORS) Inc., Montreal, Canada] with DICOM files. Craniofacial bones were isolated and labeled using the Dragonfly’s semiautomatic segmentation editor19,20. The square root of the sum of the squared distances from the centroid of the landmark configuration to each landmark was quantified as the deviation of the landmark configuration of each specimen. Basic morphometric data were analyzed with principal component analysis (PCA). The distribution of each landmark in mutants (Sc5d KO and Dhcr7 KO) and control littermates (Sc5d WT and Dhcr7 WT) along with each principal component (PC) axis was explained by PC1 and PC2. Landmarks were identified as previously described19,20 and are provided in Table 1. We used one of the landmarks as a fixed point in each bone to superimpose the images.

Table 1 Selected landmarks of craniofacial bones.
Full size table

Statistical analysis

All results obtained were analyzed with the Prism software (GraphPad Software, California, USA). The statistical significance for multiple pairs of groups was evaluated using a two-way or one-way ANOVA with Tukey’s multiple comparison test. An adjusted p < 0.05 was statistically significant. Data are represented as dots and mean ± standard deviation (SD) in the graphs and tables. A multiple variable PCA was conducted using a standardized model, and the PCs were selected based on the percent of total explained variance.

Results

Throughout this study, we analyzed littermate wild-type (WT) mice as controls in each Dhcr7 and Sc5d strain. In addition, we confirmed that there was no significant difference between WT mice of the Dhcr7 and Sc5d strains (Tables S1-S3).

Premaxilla

We first defined the anatomical landmarks on the premaxilla in the 3D μCT images (Fig. 1A and Table 1). Next, to identify differences between WT controls and Sc5d KO and Dhcr7 KO mice, we measured the length between landmarks, as well as total bone volume, using the 3D-reconstruction images. We found that there was no significant difference in volume in the premaxilla of Dhcr7 KO and WT controls, whereas Sc5d KO mice displayed smaller premaxilla compared with either Dhcr7 KO or WT control mice (Fig. 1B and Tables 2 and 3). In Dhcr7 KO mice, the bone outline was almost indistinguishable from that in WT controls, although the volume was smaller than in WT controls. The palatal process of the premaxilla [length between point 5 and 6, adjusted P value (PAdj) = 0.035] and the body (length between point 1 and 4, PAdj < 0.001) were statistically different between Dhcr7 KO and WT control mice; however, total length (length between point 1 and 6) and width (length between point 3 and 6) were the same as those in WT mice (Fig. 1C and Tables 2 and 3). On the other hand, in Sc5d KO mice, bone volume was significantly smaller than in WT control and Dhcr7 KO mice. Anterior–posterior lengths (between 1 and 6, and 5 and 6, respectively) and height (point 3 and 4) were significantly shorter than in WT and Dhcr7 KO mice. Interestingly, the most anterior part of the premaxilla (e.g. length between point 1 and 7) were indistinguishable in all genotypes (Fig. 1C, D and Table 2, and Tables S4-S6), indicating that these areas of the bone are not affected by cholesterol metabolism aberrations. Thus, in the premaxilla, Dhcr7 KO mice displayed no major defects, whereas Sc5d KO exhibited hypoplastic premaxilla with major changes in its posterior part.

Figure 1
figure 1

µCT analysis of the premaxilla. (A) 3D reconstruction of the premaxilla in E18.5 WT, Sc5d KO, and Dhcr7 KO mice. Definitions of landmarks: 1. most anterior superior point of the premaxilla; 2. most lateral point of the premaxillary-maxillary suture; 3. tip of the frontal process of the premaxilla; 4. most medial point of the premaxillary-maxillary suture; 5. most anterior point of the anterior palatine foramen; 6. most posterior point of the premaxilla; and 7. most posterior point of the incisive foramen. Scale bar: 1 mm. (B) Wiring trace of the premaxilla in E18.5 WT (blue), Dhcr7 KO (orange), and Sc5d KO (red) mice. Arrows indicate the missing portion in Sc5d KO mice. (C) Quantification of the size (length, width, height, and volume) of the maxilla from Dhcr7 WT (green bars), Dhcr7 KO (yellow bar), Sc5d WT (blue bars), and Sc5d KO (red bars) mice. *p < 0.05; **p < 0.01; ***p < 0.001; ns, not significant. N/A, not available. (D) Scatter plots of individual scores of PCA displaying the degree of morphological variances (length, width, height, and volume) of the premaxilla in Dhcr7 WT (green dots), Dhcr7 KO (yellow dots), Sc5d WT (blue dots), and Sc5d KO (red dots) mice, shown by PC1 and PC2. Distribution in mutants (Sc5d KO and Dhcr7 KO) and control littermates (Sc5d WT and Dhcr7 WT) along with 10 principal components (blue arrows) are shown.

Full size image
Table 2 Measurements of selected landmarks of craniofacial bones from E18.5 control, Dhcr7 KO, and Sc5d KO mice.
Full size table
Table 3 Comparison of the volume of craniofacial bones from E18.5 control, Dhcr7 KO, and Sc5d KO mice.
Full size table

Maxilla

We defined the anatomical landmarks on the maxilla in the 3D μCT images (Fig. 2A and Table 1). We then measured the length between each landmark, as well as total bone volume, using the 3D-reconstruction images, and found that there were no substantial differences in the main body between Dhcr7 KO and WT control mice (Fig. 2B). Interestingly, although there was no obvious cleft palate in Dhcr7 KO mice, the palatal process of the maxilla (represented as the length between point 7 and 8, PAdj < 0.001; and between 8 and 9, PAdj < 0.001) was smaller than that in WT controls (Fig. 2C and Table 2, and Table S4). In Sc5d KO mice, due to cleft palate, the right-left distance between paired maxillary bones (L6‒R6 and L9‒R9) was longer compared to Dhcr7 KO and WT control mice (Fig. 2C and Table 2, and Table S4). The distal part of the maxilla (length between point 1 and 3, PAdj = 0.324; and between 3 and 6, PAdj = 0.342) was not altered in Sc5d KO mice, whereas the palatal process of the maxilla (point 7 and 8) was missing due to cleft palate (Fig. 2C, D and Table 2, and Tables S4-S6). Interestingly, the posterior part of the maxilla was curved outward (represented by the width between left and right point 5 [L5‒R5], PAdj < 0.001) and was longer in Sc5d KO mice than in WT and Dhcr7 KO mice (Fig. 2C, D and Table 2, and Tables S4-S6). In addition, the frontal process of the maxilla (height between point 1 and 10) was significantly underdeveloped in Sc5d KO mice compared to Dhcr7 KO and WT control mice (Fig. 2B and C). Overall, the volume of the maxilla was reduced in both Dhcr7 KO and Sc5d KO mice, but more significantly in Sc5d KO mice (Fig. 2C, D and Table 3, and Tables S4-S6). Thus, Dhcr7 KO mice showed mild bony defects in the maxilla, whereas Sc5d KO mice exhibited major defects in the medial part due to cleft palate.

Figure 2
figure 2

µCT analysis of the maxilla. (A) 3D reconstruction of the maxilla in E18.5 WT, Dhcr7 KO, and Sc5d KO mice. Definitions of landmarks: 1. anterior point of the maxilla; 2. lateral inferior intersection of the frontal and zygomatic process of the maxilla; 3. tip of the zygomatic process of the maxilla; 4. anterior-medial point to the zygomatic process of the maxilla; 5. posterior point of the maxilla; 6. posterior-lateral point of the palatine process of the maxilla; 7. posterior-medial point of the palatine process of the maxilla; 8. most anterior-medial point of the palatine process of the maxilla; 9. anterior-lateral point of the palatine process of the maxilla; and 10. medial point of the premaxillary-maxillary suture. Scale bar: 1 mm. (B) Wiring trace of the maxilla in E18.5 WT (blue), Dhcr7 KO (orange), and Sc5d KO (red) mice. Arrows indicate the missing portion in Sc5d KO mice. (C) Quantification of the size (length, width, height, right-left distance, and volume) of the maxilla from Dhcr7 WT (green bars), Dhcr7 KO (yellow bar), Sc5d WT (blue bars), and Sc5d KO (red bars) mice. *p < 0.05; **p < 0.01; ***p < 0.001; ns, not significant. N/A, not available. (D) Scatter plots of individual scores of PCA displaying the degree of morphological variances (length, width, height, right-left distance, and volume) of the maxilla in Dhcr7 WT (green dots), Dhcr7 KO (yellow dots), Sc5d WT (blue dots), and Sc5d KO (red dots) mice, shown by PC1 and PC2. Distributions in mutant (Sc5d KO and Dhcr7 KO) and control littermate (Sc5d WT and Dhcr7 WT) mice along with 14 principal components (blue arrows) are shown.

Full size image

Palatine bone

We defined the anatomical landmarks on the palatine bone in the 3D μCT images (Fig. 3A and Table 1). Next, to identify anomalies in the palatine bone, we measured the length between landmarks, as well as total bone volume, using the 3D-reconstruction images of WT controls and Sc5d KO and Dhcr7 KO mice. We found bony defects in the palatine bone in both Dhcr7 KO and Sc5d KO mice, which showed reduced anterior–posterior length, medial–lateral width, and height (Fig. 3B). In agreement with our observations for the maxilla, the size of the palatine bone (represented as the length between point 1 and 4, PAdj < 0.001; width between point 3 and 5, PAdj < 0.001; and height between point 1 and 7, PAdj < 0.001; and between point 3 and 5, PAdj < 0.001) was affected in Dhcr7 KO mice (Fig. 3C, D and Table 2, and Tables S4-S6). In Sc5d KO mice, the palatal process of the palatine bone was underdeveloped (point 5 and 6), resulting in cleft palate (Fig. 3C, D and Table 2, and Tables S4-S6). Interestingly, the anterior part of the palatine bones was rotated outward; therefore, the anterior part was wider than in WT controls (width between right and left point 2 [L2‒R2], PAdj < 0.001; and between right and left point 3 [L3‒R3], PAdj < 0.001), as seen in the posterior part of the maxilla (Figs. 2C, 3C and Table 2, and Tables S4-S6). The volume of the palatine bone was reduced in both Dhcr7 KO (PAdj = 0.004) and Sc5d KO mice (PAdj < 0.001), but more pronouncedly in Sc5d KO mice (Fig. 3C, D and Table 3, and Tables S4-S6). Importantly, there was no significant difference in the anterior–posterior length of the palatine bone (length between point 1 and 4) between Dhcr7 KO (PAdj < 0.001) and Sc5d KO (PAdj < 0.001) mice, indicating that this defect was mainly due to loss of mature cholesterol.

Figure 3
figure 3

µCT analysis of the palatine bone. (A) 3D reconstruction of the palatine bone in E18.5 WT, Dhcr7 KO, andSc5d KO mice. Definitions of landmarks: 1. most anterior-lateral point of the palatine plate; 2. tip of the orbital process of the palatine bone; 3. lateral point of the palatine bone; 4. posterior point of the palatine bone; 5. posterior-medial point of the horizontal plate of the palatine bone; 6. anterior-medial point of the horizontal plate of the palatine bone; and 7. anterior superior point of the perpendicular plate. Scale bar: 1 mm. (B) Wiring trace of the palatine bone in E18.5 WT (blue), Dhcr7 KO (orange), and Sc5d KO (red) mice. Arrows indicate the missing portion in Sc5d KO mice. (C) Quantification of the size (length, width, height, right-left distance, and volume) of the palatine bone from Dhcr7 WT (green bars), Dhcr7 KO (yellow bar), Sc5d WT (blue bars), and Sc5d KO (red bars) mice. *p < 0.05; **p < 0.01; ***p < 0.001; ns, not significant. (D) Scatter plots of individual scores of PCA displaying the degree of morphological variances (length, width, height, right-left distance and volume) of the palatine in Dhcr7 WT (green dots), Dhcr7 KO (yellow dots), Sc5d WT (blue dots), and Sc5d KO (red dots) mice, shown by PC1 and PC2. Distribution in mutants (Sc5d KO and Dhcr7 KO) and control littermates (Sc5d WT and Dhcr7 WT) along with 10 principal components (blue arrows) are shown.

Full size image

Frontal bone

We defined the anatomical landmarks on the frontal bone in the 3D μCT images (Fig. 4A and Table 1). Next, to identify anomalies in the frontal bone, we measured the length between landmarks, as well as total bone volume, using the 3D-reconstruction images of WT controls and Sc5d KO and Dhcr7 KO mice and found that bone mineralization was reduced in Sc5d KO mice and increased in Dhcr7 KO mice (Fig. 4A and B). Therefore, the metopic sutures (right-left distance between point 1 [L1‒R1], PAdj < 0.001; between point 2 [L2‒R2], PAdj < 0.001; and between point 6 [L6‒R6], PAdj < 0.001) were narrower in Dhcr7 KO mice, and wider in Sc5d KO mice, compared to WT mice (Fig. 4C, D and Table 2, and Tables S4-S6). In agreement with these changes, the volume of the frontal bone in Dhcr7 KO mice was increased (PAdj < 0.001), and that in Sc5d KO mice was reduced (PAdj < 0.001), compared to WT controls (Fig. 4C, D and Table 3, and Tables S4-S6). Interestingly, the anterior–posterior length of the frontal bones (length between point 1 and 2) was not changed in Sc5d KO (PAdj = 0.192) and Dhcr7 KO (PAdj = 0.994) mice compared to WT control mice, indicating that it was not affected by any of the cholesterol metabolism aberrations (Fig. 4C, D and Table 2, and Tables S4-S6). By contrast, the height between point 3 and 4 was shorter in both Dhcr7 KO (PAdj = 0.034) and Sc5d KO (PAdj < 0.001) mice, but no difference was observed between Dhcr7 KO and Sc5d KO mice, indicating that this area of the bone was affected by loss of mature cholesterol (Fig. 4C, D and Table 2, and Tables S4-S6). Taken together, our results indicate that formation of the frontal bone may be differently regulated compared to other bones studied in this study.

Figure 4
figure 4

µCT analysis of the frontal bone. (A) 3D reconstruction of the frontal bone in E18.5 WT, Dhcr7 KO, and Sc5d KO mice. Definitions of landmarks: 1. most anterior–superior point of the frontal bone; 2. most posterior-superior point of the frontal bone; 3. most posterior-lateral intersection of the frontal bone and parietal bone; 4. most posterior-inferior point of the frontal bone; 5. most anterior-inferior point of the frontal bone; and 6. midpoint of the interfrontal suture. Scale bar: 1 mm. (B) Wiring trace of the frontal bone in E18.5 WT (blue), Dhcr7 KO (orange), and Sc5d KO (red) mice. Arrows indicate the missing portion in Sc5d KO mice. (C) Quantification of the size (length, width, height, right-left distance, and volume) of the frontal bone from Dhcr7 WT (green bars), Dhcr7 KO (yellow bar), Sc5d WT (blue bars), and Sc5d KO (red bars) mice. *p < 0.05; **p < 0.01; ***p < 0.001; ns, not significant. (D) Scatter plots of individual scores of PCA displaying the degree of morphological variances (length, width, height, right-left distance and volume) of the frontal bone in Dhcr7 WT (green dots), Dhcr7 KO (yellow dots), Sc5d WT (blue dots), and Sc5d KO (red dots) mice, shown by PC1 and PC2. Distributions in mutant (Sc5d KO and Dhcr7 KO) and control littermate (Sc5d WT and Dhcr7 WT) mice along with 9 principal components (blue arrows) are shown.

Full size image

Mandible

We defined the anatomical landmarks on the mandible in the 3D μCT images (Fig. 5A and Table 1). Next, to identify differences in the bone phenotype, we measured the length between landmarks, as well as total bone volume, using the 3D-reconstruction images of WT controls and Sc5d KO and Dhcr7 KO mice and found that Sc5d KO mice displayed severe bone defects in size at the anterior and posterior parts of the mandible (Fig. 5B). We observed a missing extension along with the incisor due to mandibular incisor agenesis (point 1 and 2) and missing condylar (point 7, 8, 9, and 10), coronoid (point 5, 6, and 7), and angular (point 10 and 11) processes in the mandible (Fig. 5A and B). Interestingly, the main body of mandibular bone was not affected in both Sc5d KO and Dhcr7 KO mice (anterior–posterior length between point 3 and 13, PAdj = 0.942; height between point 4 and 14, PAdj > 0.999; and height between point 5 and 12, PAdj = 0.48) (Fig. 5C, D and Table 2, and Tables S4-S6). These results suggest that the extension of each process of the mandible (ramus and condylar, and coronoid processes) is highly sensitive to elevated cholesterol intermediates. The volume of the mandible was therefore decreased in both Dhcr7 KO and Sc5d KO mice compared to WT controls (Fig. 5C, D and Table 3, and Tables S4-S6). Interestingly, the angle of the left and right mandible (angle R8‒1‒L8, PAdj < 0.001) was greater in Sc5d KO mice compared to Dhcr7 KO and WT control mice (Fig. 5C, D and Table 4, and Tables S4-S6), which may be due to adaptation to the widened palatine bone in Sc5d KO mice and/or due to independent changes caused by a shorter mandible length.

Figure 5
figure 5

µCT analysis of the mandible. (A) 3D reconstruction of the mandible in E18.5 WT, Dhcr7 KO, and Sc5d KO mice. Definitions of landmarks: 1. most anterior point of the mandible; 2. anterior–superior point of the mandible; 3. mental foramen; 4. molar alveolus of dentary; 5. anterior junction of the mandibular ramus and body; 6. superior tip of the coronary process of the mandible; 7. most inferior point of the mandibular notch; 8. anterior point of the condylar process of the mandible; 9. posterior point of the condylar process of the mandible; 10. superior point of the angular process of the mandible; 11. secondary cartilage of the angular process of the mandible; 12. inferior junction of the mandibular ramus and body; 13. midpoint of the external oblique ridge; 14. inferior point of the mandibular body; and 15. mandibular foramen. Scale bar: 1 mm. (B) Wiring trace of the mandible in E18.5 WT (blue), Dhcr7 KO (orange), and Sc5d KO (red) mice. Arrows indicate the missing portion in Sc5d KO mice. (C) Quantification of the size (length, width, height, angle, and volume) of the mandible from Dhcr7 WT (green bars), Dhcr7 KO (yellow bar), Sc5d WT (blue bars), and Sc5d KO (red bars) mice. *p < 0.05; **p < 0.01; ***p < 0.001; ns, not significant. (D) Scatter plots of individual scores of PCA displaying the degree of morphological variances (length, width, height, angle, and volume) of the mandible in Dhcr7 WT (green dots), Dhcr7 KO (yellow dots), Sc5d WT (blue dots), and Sc5d KO (red dots) mice, shown by PC1 and PC2. Distribution in mutants (Sc5d KO and Dhcr7 KO) and control littermates (Sc5d WT and Dhcr7 WT) along with 11 principal components (blue arrows) are shown.

Full size image
Table 4 Comparison of the angle of mandibles from E18.5 control, Dhcr7 KO, and Sc5d KO mice.
Full size table

Discussion

Loss of Dhcr7 or Sc5d leads to lack of mature cholesterol as well as accumulation of cholesterol intermediates, namely of 7-DHD/7-DHC or lathosterol/dehydrolathosterol, respectively. Considering the fact that phenotypes in Dhcr7 KO mice are partially rescued with statins, which normalize the levels of cholesterol precursors14, craniofacial bone development is more sensitive to the accumulation of cholesterol intermediates than to absence of mature cholesterol. More recently, we have reported that mice with a deficiency for Sc5d exhibit mandibular hypoplasia due to defects in osteoblast differentiation18. Sc5d KO mice are smaller than littermate controls and present developmental delay and bent shorter limbs4,18. Interestingly, while Sc5d KO mice display severe defects in both endochondral and intramembranous ossification, Dhcr7 KO mice show less severe bone defects14,18. The phenotypic differences between Dhcr7 KO and Sc5d KO mice may depend on the expression pattern of Dhcr7 and Sc5d. In addition, molecules that are modified with cholesterol may be differentially expressed at each bone and location during craniofacial development. For instance, hedgehog ligands (Sonic Hedgehog [SHH], Indian Hedgehog [IHH], and Desert Hedgehog [DHH]) and receptor Smoothened (SMO) are known to be modified with cholesterol, which is crucial for their distribution and activity21,22. For example, an absent cholesterol modification on SHH (ShhN) leads to a shorter distribution and lower activation of SHH signaling in limb buds, but no difference in its biological functions compared to the cholesterol-modified molecules23,24. Although oxysterols have a variety of biological activities and affect cell survival, apoptosis, gene expression, as well as Hedgehog signaling activity, it remains unknown to what extent oxysterols derived from 7-DHC, lathosterol, and cholesterol differ in cell toxicity25,26. In addition, the plasma membrane contains cholesterol-rich micro-domains (e.g. lipid rafts and caveolae) that act as a signaling center by assembling receptors and channels27,28, and transduction and activation of hedgehog signaling is regulated by oxysterols and binding of cholesterol on the membrane29,30,31. Therefore, a precisely controlled cholesterol synthesis process is important for cellular functions. Our results show that bone formation was differently affected in each mutant mouse model analyzed, indicating a location-specific requirement and the role of cholesterol metabolism in bone development.

This study shows how bone formation (e.g. size and volume) is affected in each craniofacial bone from Dhcr7 KO and Sc5d KO mice. We found that loss of mature cholesterol had a lower impact on bone formation than elevated levels of cholesterol intermediates. In the mandible, the extension of the process of the mandible was drastically affected in Sc5d KO mice compared to Dhcr7 KO and WT control mice. Because these areas are formed through endochondral ossification, chondrocytes may be more sensitive to elevated cholesterol intermediates compared to osteoblasts. Future studies may identify specific functions for each cholesterol intermediate in various cell types.

One of the limitations of this study is that there are some differences in the phenotypes observed in humans and mice. For instance, while DHCR7 mutations in humans causes SLOS with growth retardation, microcephaly, micrognathia, and cleft palate, Dhcr7 KO mice display a distinct cleft palate with less than 10% penetrance16. It should be noted that, in this study, we analyzed Dhcr7 KO mice without cleft palate. In addition, while a larger volume for the frontal bone was seen in Dhcr7 KO mice compared to WT littermates, Dhcr7 KO mice displayed smaller skulls at birth14. The accelerated bone formation and differentiation observed in Dhcr7 KO mice14 might lead to immature closure of sutures between bones as well as growth arrest in the growth plate of long bones. In humans, children with SLOS typically show a head circumference 2 standard deviations below the average of unaffected children. Although no three dimensional volumetric analysis was conducted in this study, overall craniofacial features seem to be conserved in Dhcr7 KO mice. Thus, the findings in mouse models need to be further evaluated in SLOS patients.