Dentoalveolar Defects of Hypophosphatasia are Recapitulated in a Sheep Knock‐In Model

ABSTRACT

Hypophosphatasia (HPP) is the inherited error‐of‐metabolism caused by mutations in ALPL, reducing the function of tissue‐nonspecific alkaline phosphatase (TNAP/TNALP/TNSALP). HPP is characterized by defective skeletal and dental mineralization and is categorized into several clinical subtypes based on age of onset and severity of manifestations, though premature tooth loss from acellular cementum defects is common across most HPP subtypes. Genotype–phenotype associations and mechanisms underlying musculoskeletal, dental, and other defects remain poorly characterized. Murine models that have provided significant insights into HPP pathophysiology also carry limitations including monophyodont dentition, lack of osteonal remodeling of cortical bone, and differing patterns of skeletal growth. To address this, we generated the first gene‐edited large‐animal model of HPP in sheep via CRISPR/Cas9‐mediated knock‐in of a missense mutation (c.1077C>G; p.I359M) associated with skeletal and dental manifestations in humans. We hypothesized that this HPP sheep model would recapitulate the human dentoalveolar manifestations of HPP. Compared to wild‐type (WT), compound heterozygous (cHet) sheep with one null allele and the other with the targeted mutant allele exhibited the most severe alveolar bone, acellular cementum, and dentin hypomineralization defects. Sheep homozygous for the mutant allele (Hom) showed alveolar bone and hypomineralization effects and trends in dentin and cementum, whereas sheep heterozygous (Het) for the mutation did not exhibit significant effects. Important insights gained include existence of early alveolar bone defects that may contribute to tooth loss in HPP, observation of severe mantle dentin hypomineralization in an HPP animal model, association of cementum hypoplasia with genotype, and correlation of dentoalveolar defects with alkaline phosphatase (ALP) levels. The sheep model of HPP faithfully recapitulated dentoalveolar defects reported in individuals with HPP, providing a new translational model for studies into etiopathology and novel therapies of this disorder, as well as proof‐of‐principle that genetically engineered large sheep models can replicate human dentoalveolar disorders. © 2022 The Authors. Journal of Bone and Mineral Research published by Wiley Periodicals LLC on behalf of American Society for Bone and Mineral Research (ASBMR).

Introduction

Hypophosphatasia (HPP) is the rare inherited error‐of‐metabolism caused by mutations in the ALPL gene that encodes tissue‐nonspecific alkaline phosphatase (TNAP/TNALP/TNSALP).^{( 1} , ^{2 )} HPP is characterized by elevated inorganic pyrophosphate (PP_i) levels leading to impaired skeletal mineralization, including rickets and osteomalacia, as well as other musculoskeletal manifestations. More than 400 HPP‐causing inactivating mutations in ALPL have been identified, including most often missense mutations, but also null alleles from nonsense and frameshift variants (https://alplmutationdatabase.jku.at/).^{( 3 )} HPP can be inherited in autosomal dominant (typically milder cases) or recessive (typically more severe cases) fashion. HPP has a broad‐ranging severity and is classified into clinical subtypes including by increasing age‐of‐onset and decreasing severity: perinatal, infantile, mild/severe childhood, adult, and odonto‐HPP (OMIM 241500, 241510, 146300). Genotype–phenotype correlations are weak and modifying mechanisms underlying the clinical heterogeneity of HPP remain speculative.

HPP is associated with an array of dental and craniofacial manifestations, including enamel defects, thin and/or hypomineralized dentin, acellular cementum hypoplasia, and craniosynostosis.^{( 4} , ⁵ , ^{6 )} Dental defects are also broad‐ranging across clinical subtypes, though premature loss of primary teeth is a hallmark of HPP and a key diagnostic criterion. Mouse models harboring Alpl knockout or mutations phenocopy many important aspects of human HPP, leading to significant insights into disease mechanisms.^{( 7} , ⁸ , ^{9 )} However, HPP mouse models also have noteworthy limitations, and do not fully recapitulate the entire spectrum of HPP skeletal, dental, craniofacial, and musculoskeletal manifestations due to differences in growth and physiology.

Benefits of large animal models for musculoskeletal research have been enumerated, highlighting the importance of convenience, relevance, and appropriateness.^{( 10} , ¹¹ , ¹² , ^{13 )} Sheep (Ovis aris) present an attractive model for many reasons: their docile disposition; lower expense than other large animals; similar size and hormone profile to humans; anatomical and biochemical features relatively similar to humans, including the musculoskeletal system; ability to take repeated measurements, scans, biopsies, and substantial blood samples over time; and the publication and detailed annotation of the ovine genome.^{( 14} , ^{15 )} Musculoskeletal disorders and bone healing have been characterized in sheep, and there are key ovine physiological similarities to humans, including diphyodont dentition (formation of primary and secondary teeth), bone organization, and osteonal (Haversian) remodeling.^{( 14} , ¹⁵ , ¹⁶ , ^{17 )} To better model HPP‐associated musculoskeletal and dental defects, we created the first gene‐engineered large‐animal sheep model of HPP by knocking in via CRISPR/Cas9 a missense mutation in the ALPL gene at the site identical to the human sequence (c. 1077C>G; p.I359M) associated with skeletal and dental defects in humans with HPP.^{( 18} , ^{19 )} We hypothesized that this ovine HPP model would mimic the dentoalveolar manifestations of human HPP, providing insights into disease mechanisms conserved across species as well as provide a translational model for studies into etiopathology and novel therapies. Our studies aimed to assess the genotype–phenotype associations in knock‐in HPP sheep with previous case reports of humans carrying the p.I359M ALPL mutation.^{( 19} , ^{20 )} A longitudinal study on dentoalveolar effects of HPP was performed in sheep using a multimodal approach including computed tomography (CT), high‐resolution micro‐computed tomography (μCT), and histology.

Materials and Methods

Animals

Gene editing of sheep by CRISPR/Cas9 with an HPP‐associated mutation (c.1077C>G; p.I359M) in ALPL exon 10 (ALPL ^I359M) has been described.^{( 18 )} Sheep analyzed in the study included: wild‐type (WT) control sheep (ALPL ^{+ / +}; n = 4), heterozygous (Het) sheep carrying a single allele with the missense mutation (ALPL ^{+ / I359M}; n = 3), homozygous (Hom) sheep carrying two mutated alleles (ALPL ^I359M/I359M; n = 1), and a single compound heterozygous (cHet) sheep with a p.I359M allele and a 178‐nucleotide deletion null allele (ALPL ^−/I359M; n = 1). Exfoliated/extracted primary incisors were analyzed from WT (n = 6), Het (n = 4), Hom (n = 4), and cHet (n = 2). Circulating alkaline phosphatase (ALP) levels were measured for blood collected at 2, 7, and 14 months. Animal procedures were approved by the Texas A&M University Institutional Animal Care and Use Committee and this work adheres to the Animals in Research: Reporting In Vivo Experiments (ARRIVE) guidelines.

CT

Longitudinal in vivo CT scans of sheep at 2, 7, and 14 months were performed under sedation using midazolam (0.4 mg/kg) and ketamine (1–2 mg/kg), intravenously (i.v.) into the jugular vein, followed by general anesthesia with 2% isoflurane under oxygen at 1–1.25 L/min administered via face mask for the duration of the scans (<15 minutes).^{( 21 )} Animals were placed in sternal recumbence for whole‐body noncontrast CT imaging (Siemens Biograph mCT; Siemens Healthineers USA, Cary, NC, USA) (128‐slice scanner, 1.5‐mm slice thickness scans with 0.6‐mm isotropic voxel reconstructions, 140 kV, and 250 mAs). Working reconstructions were performed on the Siemens Syngo workstation (Siemens Healthineers, USA). CT image files (Digital Imaging and Communications in Medicine [DICOM] images) of sheep skulls and mandibles were reconstructed and oriented in AnalyzePro 1.0 (AnalyzeDirect, Overland Park, KS, USA). Sheep skulls were oriented using the axial slice at the level of intersphenoid synchondrosis (ISS) and coronal slices at the level of posterior teeth to define the skull median position, as well as the midsagittal slice so that the maxillary/palatine bone is parallel to the axial plane as shown in Fig. S1. Three‐dimensional (3D) reconstructions of skull CT scans were displayed in sagittal and dorsal (top) views, where the skull bones were segmented above a threshold of 350 Hounsfield units (HU). Skull measurements were then performed on oriented 3D skulls, where skull length was measured using the sagittal view to define the distance between the tip of the incisive bone (INB) to the external occipital protuberance (EOP), and skull width was measured in the dorsal (top) view to define the distance between the highest points at the bases of the horn buttons, as shown in Fig. S1. Cranial volume was measured in oriented skulls of 7‐month‐old sheep by segmentation of intracranial space housing the brain (<350 HU) from surrounding skull bones, with manual corrections when necessary. Sheep mandibles were oriented to a standard orientation using incisors as landmarks. Enamel was segmented above a threshold of 1600 HU and the threshold used for bone and dentin was 350 HU, with some manual correction done as necessary to separate bone and dentin.^{( 18 )} 3D reconstruction of mandible CT scans was prepared for the labial view (Fig. 2) and two‐dimensional (2D) labial and axial views were prepared as shown in Fig. S2.

Fig. 2.

Premature loss of primary teeth and alveolar bone defects in sheep with HPP. (A,B) Intraoral photos reveal extensive premature loss of primary incisors in cHet versus WT sheep by 12 months. (B, right) Prematurely lost incisor from cHet sheep retains substantial root length. Scale bar = 1 cm. (C–E) 3D renderings of mandible CT scans at ages of 2, 7, and 14 months show primary tooth loss (red asterisks) in cHet sheep as early as 2 months, with progressive tooth loss apparent at 7 and 14 months. Red lines in C indicate the distance from the DE‐EN junction to the AB crest, which is increased in cHet versus WT and Het sheep. (E) Premature tooth loss in cHet sheep was associated with early eruption of permanent incisors (# indicates permanent central incisors). Dotted lines indicate increased distance between permanent central incisors due to distal drifting resulting from premature loss of primary teeth (red asterisks) in cHet versus WT, Het, and Hom sheep. Scale bar = 5 mm. (F) Bar graph indicates premature tooth loss in mutant sheep versus controls. (G,H) Quantitative analysis shows reduced ABP volume and density in Hom and cHet mutant sheep at 2 and 7 months, compared with controls. AB = alveolar bone; ABP = alveolar bone proper; cHet = compound heterozygous; DE = dentin; EN = enamel; Het = heterozygous; Hom = homozygous; HPP = hypophosphatasia; WT = wild‐type.

Alveolar bone proper (ABP) was identified as the alveolar bone bordering the periodontal ligament (PDL) and closest to incisor roots and therefore most invested by Sharpey's fibers and most directly involved in periodontal attachment, as described.^{( 22} , ^{23 )} In sheep, ABP was defined as alveolar bone (>350 HU) within 25 voxels (500 μm) of the root surfaces of the central incisors, in all directions (x, y, and z planes).

μCT

Sheep exfoliate their primary incisors and permanent incisors erupt into the oral cavity between 12 and 36 months, starting with central incisors and working distally.^{( 24} , ^{25 )} Exfoliated or loosely attached and therefore extracted primary incisors were collected for high‐resolution μCT scanning and analysis of dental structures. When multiple teeth were collected from the same sheep, measurements from teeth were averaged to generate a single value for inclusion in graphs and statistical analysis. After fixation in 10% formalin, sheep primary incisors were scanned in a Scanco Medical microCT 50 (Scanco, Brüttisellen, Switzerland) at 55 kVp, 114 mA, 200‐ms integration time, and 6‐μm voxel dimension. DICOM files of teeth were reconstructed, calibrated to five known densities of hydroxyapatite (mg HA/cm³), and analyzed using AnalyzePro version 1.0 (AnalyzeDirect). Orientation and analysis were performed as described.^{( 19 )} Teeth were oriented anatomically using the midsagittal slice, where the mesial and distal cementum‐enamel junctions (CEJs) were used to identify an axis that was made perpendicular to the length of the root. Enamel and dentin were segmented at 1600 mg HA/cm³ and 650 mg HA/cm³, respectively. Acellular cementum (which covers at least two‐thirds of cervical root surfaces, so would be the majority cementum type on partially resorbed primary teeth) was segmented as described.^{( 19 )} Briefly, a median filter with a kernel size of 11 was applied. Acellular cementum was then segmented between 450 and 1050 mg HA/cm³ with manual corrections to exclude softer dentin that was highlighted adjacent to the pulp. The segmentation map was then loaded back onto the original calibrated image and used as a mask to highlight any cementum under this mask with a density over 650 mg HA/cm³.^{( 19 )}

For the thickness of crown enamel and dentin, a region of interest (ROI) of 50 slices (0.5 mm) was defined 0.5 mm coronally from the CEJ. The CEJ was identified in the axial plane with a complete enamel ring being the first slice. For root dentin thickness, a region of interest of 50 slices (0.5 mm) was defined 150 μm apically from the most apical extension of enamel in the axial plane. Thickness measurements were performed as described^{( 19 )} using a cortical bone thickness algorithm for the defined ROIs. Density measurements of dentin subregions were performed on the defined root dentin ROI, where the outermost 150 μm, excluding cementum, was used for mantle dentin, innermost 150 μm was used for proximal pulpal dentin, and remaining in‐between dentin was used for circumpulpal dentin (Fig. 4A ). For tooth length, the axial plane was used to identify the first slice of incisal enamel to the most apical extension root dentin.

Fig. 4.

Altered dentin regions in sheep with HPP. (A) 2D μCT images of sheep primary incisors show regions of interest used to measure subregion thickness and densities. Scale bar = 5 mm. (B,C) Quantitative μCT analysis shows reduced crown dentin thickness in cHet incisor teeth, whereas enamel thickness did not change compared with WT and Het sheep incisors. (D–I) Density measurements reveal reduced mantle dentin density observed in cHet sheep. Acellular cementum density appears largely unaffected in Het and Hom sheep, whereas cementum is undetectable in cHet sheep. cHet = compound heterozygous; Het = heterozygous; Hom = homozygous; HPP = hypophosphatasia; WT = wild‐type.

Histology

Sheep primary incisors were decalcified in a formic acid and formaldehyde‐based solution (Polysciences, Inc., Warrington, PA, USA). Decalcified teeth were processed for paraffin embedding and sectioned at 6‐μm thickness. Histology sections were stained with toluidine blue (TB) and picrosirius red (PR) as described.^{( 26 )} Histomorphometric analysis was performed on TB‐stained images at magnification ×20 to measure acellular cementum thickness, mantle dentin thickness, and predentin thickness. Acellular cementum thickness was performed on the lingual side of the tooth root using the average of three linear measurements near the CEJ. The thickness of mantle dentin was measured throughout the root length where reported values represent the average of linear measurements. Predentin thickness was measured around the level of the CEJ using the average of three linear measurements.

Statistical analysis

Data are expressed as mean ± standard deviation (SD) in all graphs with individual animal data points indicated (Prism version 9.3.1; GraphPad Software, La Jolla, CA, USA). Statistical analysis was performed using one‐way analysis of variance (ANOVA) followed by post hoc Tukey's multiple comparison test (α = 0.05) to compare the means of all groups. Linear regression analysis of ALP activity levels at 2 months of age (most active phase of growth) was performed for each dentoalveolar endpoint measured, independent of genotype. R ² and p values are reported for phenotypes showing significant correlations.

Results

Altered skull shape in a sheep knock‐in model of HPP

Children with more severe clinical forms of HPP often present abnormal craniofacial growth due to premature closure of cranial sutures (craniosynostosis), which can lead to elevated intracranial pressure.^{( 27 )} Sheep heads scanned by CT at 2, 7, and 14 months‐of‐age underwent linear skull length and width measurements (Fig. S1 A). cHet (ALPL ^−/I359M) sheep (lowest serum ALP levels) exhibited 4%–7% reduced skull length anteroposteriorly, compared with WT (ALPL ^{+ / +}), Het (ALPL ^{+ / I359M}), and Hom (ALPL ^{I359M / I359M}) sheep (Fig. 1A–E ). Reduced skull length in cHet sheep was associated with a 2%–10% increase in skull width and 9%–14% increased skull width/length ratio over WT (Fig. 1F,G ), suggesting altered skull shape. Cranial volume at 7 months old trended downward, particularly in the cHet sheep (Fig. 1H ). However, the involvement of cranial sutures to indicate potential craniosynostosis could not be determined from the scans.

Fig. 1.

Altered skull shape in a sheep model of HPP. (A–D) Sagittal view 3D renderings of sheep skull CT scans at ages of 2, 7, and 14 months reveal reduced skull growth anteroposteriorly in cHet sheep compared to WT, Het, and Hom sheep. Blue dotted lines set a reference for the most anterior point in skulls. By 14 months, cHet sheep mandibles exhibit progressive alveolar bone loss (red arrow). Scale bar = 5 cm. (E–H) Quantitative analysis suggests reduced skull length and increased skull width in cHet sheep compared with WT, Het, and Hom sheep, resulting in increased skull width/length ratio and reduced cranial volume indicative of altered craniofacial morphology. (I–L) Circulating ALP activity levels at 2, 7, and 14 months. ALP = alkaline phosphatase; cHet = compound heterozygous; Het = heterozygous; Hom = homozygous; HPP = hypophosphatasia; WT = wild‐type.

Skull length is largely determined by the anteroposterior growth at the skull base mediated by growth plate synchondroses, including ISS and spheno‐occipital synchondrosis (SOS).^{( 28} , ^{29 )} Shortened skull length prompted us to determine if the growth at skull base was affected in HPP sheep. We examined ISS and SOS in the midsagittal plane of skull CT images (Fig. S1 B). There were no apparent differences between HPP mutant and control synchondroses (Fig. S1 C). Although our results suggest that SOS undergoes ossification before ISS, we observed a difference in the timing of SOS closure between male and female WT sheep at 7 months (sheep are reproductively mature and considered fully‐grown at 6 months). Although closed in WT females, the SOS in 7‐month‐old WT males remained patent (Fig. S1 D), suggesting an active contribution to skull growth and an observation that could explain the difference in skull length and width between male and female WT sheep (Fig. S1 E). Though the results suggest that cHet sheep had an altered skull shape due to reduced skull growth anteroposteriorly, further analysis using a histological approach will be required to determine the involvement of cranial sutures and/or skull base synchondroses.

Circulating ALP activity levels were measured at 2, 7, and 14 months‐of age, showing sustained reduction of ALP activity in HPP mutant sheep at all time points (Fig. 1I–L ).

Premature loss of primary teeth and alveolar bone defects in sheep with HPP

Healthy sheep dentition includes eight mandibular incisors that articulate with a maxillary cornified dental pad. Sheep exfoliate their primary incisors and permanent incisors erupt into the oral cavity between 12 and 36 months, starting with central incisors and proceeding distally.^{( 24} , ^{25 )} In both humans and sheep, odontoclastic resorption of tooth roots and loss of periodontal attachment leads to physiological exfoliation of primary teeth with substantially resorbed roots. However, cementum (and potentially alveolar bone) defects associated with HPP lead to poor periodontal attachment and premature primary tooth loss before roots have undergone extensive resorption, appearing as “fully rooted” teeth.

WT, Het, and Hom sheep largely retained their primary incisors at 14 months, whereas cHet sheep prematurely lost most of their primary incisors by this age (Fig. 2A,B,F ). Prematurely lost incisors that could be collected exhibited substantial root structure (Fig. 2B ). CT analysis over a time‐course of 2, 7, and 14 months revealed that cHet sheep lost incisors as early as 2 months (Fig. 2C , red asterisk) and exhibited irregular, reduced interproximal mandibular alveolar bone, compared with WT, Het, and Hom sheep. By 14 months, cHet sheep had lost most of their primary incisors and exhibited progressively severe alveolar bone loss (Fig. 1D , arrow; Fig. 2C , vertical red line; and Fig.  2D,E , red asterisks). Hom and cHet sheep had reduced ABP (the bone nearest tooth roots and most involved in periodontal attachment) volume and density at 2 and 7 months, compared with WT and Het sheep (Fig. 2G,H ; scans at 14 months were not analyzed due to complexities of the mixed dentition). Hom sheep showed a 51% and 27% reduction in ABP volume and a 16% and 5% reduction in ABP density at 2 and 7 months, respectively. cHet sheep exhibited a 39% and 32% reduction in ABP volume and a 12% reduction and a 2% increase in ABP density at 2 and 7 months, respectively. Interestingly, cHet and Hom sheep showed early eruption of permanent central incisors versus WT and Het sheep (Fig. 2E ; #). However, the erupted central incisors in cHet sheep were misaligned and separated due to loss of distal contacts with adjacent primary lateral incisors (Fig. 2E , dotted lines).

Altered dental mineralization in sheep with HPP

In addition to in vivo CT scans to monitor tooth eruption and exfoliation patterns, as well as alveolar bone properties, we employed high‐resolution ex vivo μCT to measure dental tissue material properties of prematurely exfoliated teeth. HPP has broad‐ranging effects on dentoalveolar mineralized tissues based on ALPL mutation(s), loss‐of‐function, HPP clinical type, and other modifiers.^{( 3} , ⁶ , ⁹ , ^{19 )} To understand the effect of the I359M knock‐in mutation on dental mineralization in sheep, high‐resolution μCT scanning and quantitative analysis was performed on the primary sheep incisors that could be collected, including from WT, Het, Hom, and cHet sheep. Linear measurements revealed incisor crowns trended shorter and remaining root length trended longer in accordance with genotype (Fig. 3A–C ). Incisors from cHet sheep had 11%–16% shorter crown length (not significant) and significantly longer (60%–180%) retained root structures (p = 0.0002–0.0114) compared to Hom, Het, and WT incisors. cHet incisors exhibited average changes of 27% reduced enamel volume, 41% reduced dentin volume (p = 0.0059–0.0364), and 3% reduced density compared with WT, Het, and Hom incisors, though no statistically significant differences in enamel volume and density or dentin density were identified between genotypes (Fig. 3D–G ).

Fig. 3.

Altered dental mineralization in sheep with HPP. (A) 3D μCT reconstructions of primary incisors. cHet sheep primary teeth retain substantial root length with little evidence of resorption, and Het and Hom sheep teeth have slightly more remaining root structure than WT sheep, where signs of physiological root resorption are evident. μCT analysis reveals cementum (yellow) covering WT, Het, and Hom tooth roots, whereas in cHet teeth, cementum is undetectable by the usual segmentation methods. Scale bar = 5 mm. (B,C) Linear measurements show trends of decreased crown length and longer roots in cHet, Hom, and Het teeth versus WT controls. (D–G) Quantitative analysis shows dentin volume is significantly reduced in cHet versus other genotypes. cHet = compound heterozygous; Het = heterozygous; Hom = homozygous; HPP = hypophosphatasia; WT = wild‐type.

Further ex vivo μCT analysis was performed to analyze crown and root subregions using defined regions of interest (Fig. 4A ). Although enamel thickness showed no differences between groups, crown dentin thickness was significantly lower (30%; p = 0.0002–0.0006) in cHet vs. WT, Het, and Hom sheep incisors (Fig. 4B,C ). No differences were found in crown or root dentin densities among genotypes (Fig. 4D,E ). Analysis of root dentin subregions revealed significantly reduced mantle dentin density (10%; p = 0.0019–0.0133) in cHet incisors compared with WT and Hom, whereas no differences were found between genotypes in proximal pulpal and circumpulpal root dentin regions (Fig. 4F–H ). Cementum, which could be detected and digitally segmented from dentin, showed no difference in density in WT, Het, and Hom incisors, though no cementum layer could be identified in incisors from cHet sheep (Fig. 4I ).

Acellular cementum hypoplasia and mantle dentin defects in a sheep model of HPP

Compared to the normal dentin structure in WT incisors, histological TB staining revealed abnormal interglobular dentin patterns in outer regions of cHet sheep incisors, as well as wide and erratic predentin borders in both cHet and Het sheep, marked by interglobular border patterns (Fig. 5A–D ; inset). Predentin trended toward increased thickness and variability in mutants, but no significant differences were found among groups (Fig. 5E ). Mantle dentin thickness was significantly increased (p = 0.0055–0.0212) in cHet versus WT, Het, and Hom sheep (Fig. 5F ). Acellular cementum thickness was associated with genotype. WT and Het incisors exhibited a similar layer of acellular cementum approximately 20 μm thick covering tooth roots, Hom incisors had a nearly 50% reduction in acellular cementum (49%) compared with WT and Het, and cHet incisors lacked any discernible acellular cementum (Fig. 5C,D,G ).

Fig. 5.

AC hypoplasia and mantle dentin defects in a sheep model of HPP. (A–D) TB staining reveals interglobular dentin defects (white arrow in the inset) and trending wider PD in cHet sheep incisors compared with WT, Het, and Hom sheep incisors. AC (white arrow) is observed on the root surfaces of WT and Het sheep teeth, appears reduced in Hom teeth, and is absent in cHet sheep teeth (E, red asterisk in white arrow). PR staining observed under polarized light microscopy showed Sharpey's fibers organized within the AC layer in WT and Het teeth, poorly organized Sharpey's fibers in Hom teeth, and absent AC and Sharpey's fibers in cHet sheep teeth. (E–G) Quantitative analysis suggests trending increased PD thickness across genotypes, significantly increased mantle dentin thickness in cHet sheet, and trending decreased AC thickness across genotypes. Note CC covering AC in Het sheep teeth, suggesting, unlike in human teeth, that CC can grow to the coronal third of roots in sheep teeth. AC = acellular cementum; CC = cellular cementum; cHet = compound heterozygous; Het = heterozygous; Hom = homozygous; HPP = hypophosphatasia; PD = predentin; PR = Picrosirius red; TB = Toluidine blue; WT = wild‐type.

Picrosirius red (PR) staining viewed under polarized microscopy revealed reddish‐orange birefringence in WT and Het incisors, whereas collagen fibers in cHet incisors showed less birefringence that included more green hue, suggesting collagen fiber size and/or maturation was adversely affected. There was also a notable lack of perpendicularly oriented Sharpey's fibers at root surfaces of Hom and cHet incisors, reflecting deficiency of acellular cementum.

Craniofacial and dentoalveolar defects are correlated with ALP levels

Linear regression analysis was performed to determine whether reduced ALP activity at 2 months of age (the most active growth stage) was correlated with altered craniofacial or dentoalveolar phenotypes (Fig. 6). Indeed, independent of genotype, reduced ALP activity was significantly correlated with several measurements. Skull width/length ratio at 2 months of age was correlated with ALP activity (p = 0.0324; Fig. 6A ). Increased mantle dentin thickness and longer root length were significantly correlated with reduced ALP activity levels, whereas reductions in crown length, enamel volume, crown dentin thickness, and acellular cementum thickness and density were significantly correlated with reduced ALP activity (p = 0.0047–0.0441 for all; Fig. 6B–H ). These findings support the role of ALP activity in craniofacial and dental development across the spectrum of genotypes in HPP sheep.

Fig. 6.

Craniofacial and dental endpoints correlate with reduced ALP activity in HPP sheep. (A–H) Craniofacial and dental measurements were correlated with 2 months ALP activity levels in WT (blue), Het (red), Hom (pink), and cHet (green) sheep using linear regression analysis. R ² values and p values are shown for each correlation; p < 0.05 was considered statistically significant. AC = acellular cementum; ALP = alkaline phosphatase; cHet = compound heterozygous; Het = heterozygous; Hom = homozygous; HPP = hypophosphatasia; MD = mantle dentin; Th = thickness; WT = wild‐type.

Discussion

To better understand the craniofacial and dental phenotype of the inherited error‐of‐metabolism HPP, we engineered the first large animal sheep model of HPP by knocking in via CRISPR/Cas9 a missense mutation in the ALPL gene (c.1077C>G; p.I359M) associated with HPP in humans. This approach circumvented the use of current HPP mouse models that have key limitations with regard to studying the hallmark dental and craniofacial aspects of HPP. Sheep compound heterozygous with a mutant and null ALPL allele featured craniofacial and dental defects including altered craniofacial shape, premature loss of primary incisors, dentin defects, severe cementum hypoplasia, and reduced alveolar bone volume and density. Sheep homozygous for the mutant allele showed alveolar bone effects and trends in dentin and cementum, whereas sheep heterozygous for the mutation did not exhibit significant effects, in total displaying a spectrum of severity across the loss‐of‐function genotypes, consistent with that observed in human HPP.^{( 1 )} Moreover, lifelong reduced ALP activity levels were significantly correlated with the abnormalities in skull width/length, crown and root lengths, enamel volume, crown dentin thickness, mantle dentin thickness, and acellular cementum thickness and density. Collectively, these findings validate HPP sheep as a translational proof‐of‐principle model and demonstrate the utility of the model to reproduce the range of HPP‐associated dentoalveolar defects encountered in human patients.

The sheep as a model for dentoalveolar effects of HPP

HPP shows a broad range of severity across multiple clinical subtypes.^{( 30} , ³¹ , ^{32 )} Genotype–phenotype correlations have begun to be identified, though this is complicated by the fact that inheritance of ALPL alleles associated with HPP is complex (autosomal dominant and recessive effects are described), and compound heterozygous mutations are common, making in vivo effects of individual loss‐of‐function mutations difficult to ascertain.^{( 3 )} Furthermore, the mouse knockout, conditional knockout, and knock‐in HPP models described to date have key limitations, including monophyodont dentition and other differences that affect presentation, underlying pathological mechanisms, and response to treatment.^{( 7} , ⁸ , ⁹ , ^{33 )} These factors prompted us to investigate the sheep as a model for musculoskeletal and dental effects of HPP. Musculoskeletal disorders and healing have been characterized in sheep, and there are key ovine physiological similarities to humans, including diphyodont dentition (formation of primary and secondary teeth), bone organization, and osteonal (Haversian) remodeling.^{( 14} , ¹⁵ , ¹⁶ , ^{17 )} Sheep show similar alveolar bone loss responses to humans in response to periodontal disease and ovariectomy, supporting a similar dentoalveolar and bone physiology between species.^{( 34} , ³⁵ , ^{36 )}

The sheep dentition includes eight mandibular incisors (central, middle, lateral, and corner), three premolars, and three molars per quadrant of the upper and lower jaws. Sheep primary teeth sequentially erupt into the oral cavity after birth, with replacement between 12 and 18 months and full replacement by 2 to 3.5 years old.^{( 24} , ^{25 )} The TNAP protein sequence is highly conserved in humans and sheep, with over 90% structural similarity, including conservation of the orthologous isoleucine among mammals.^{( 19 )} The sheep model reproduced the broad spectrum of HPP craniofacial and dental manifestations, including reduced circulating ALP activity, hypomineralization, and metaphyseal flaring of long bones.^{( 18 )}

Importantly, sheep have incisors similar in development and size to humans, whereas mice have a continually erupting incisor that must be analyzed with caution when inferring conclusions about human dental development. Incisors are the most frequently affected teeth by HPP among tooth types^{( 19} , ^{37 )} and were the focus of dental analysis in this report. The multimodal longitudinal analysis described herein provides a more detailed picture of the craniofacial and dentoalveolar manifestations associated with this mutation, confirming that HPP mutant sheep accurately model human disease, including altered craniofacial shape, premature loss of primary incisors, disrupted dentinogenesis reflected by reduced dentin volume, thickness, and density, novel interglobular dentin defects associated with dentin hypomineralization, hypoplasia of acellular cementum, alveolar bone defects, and malocclusion. The compound heterozygous sheep carrying a mutated and null allele displayed all of these defects, whereas homozygous and heterozygous mutant sheep manifested fewer and less severe defects, respectively, demonstrating a genotype–phenotype spectrum within this cohort that was significantly correlated with reductions in ALP activity. These features recapitulate a broad range of the craniofacial and dentoalveolar defects described over the several clinical subtypes of HPP.^{( 6} , ⁹ , ³⁸ , ³⁹ , ^{40 )} Approximately 64% of pediatric HPP patients present abnormal craniofacial growth or craniosynostosis due to premature closure of cranial sutures.^{( 41 )} This condition can raise intracranial pressure and damage the brain in early‐onset forms of severe HPP.^{( 27 )} Importantly, the dentoalveolar manifestations accurately represent those we described in a patient carrying the exon 10 c.1077C>G; p.I359M mutation, albeit in combination with another missense mutation in exon 5 (c.346G>A, p.A116T).^{( 19 )} In that subject, we described premature primary tooth loss, dentin defects, and lack of acellular cementum. The sheep is therefore the first animal model to accurately replicate premature primary tooth loss associated with HPP. A previous report described a naturally occurring autosomal recessive mutation (c.1301T>G; p.V434G) in Karelian Bear Dogs in Finland.^{( 42 )} Affected dogs exhibited early lethality, altered craniofacial shape, severe rickets, and osteomalacia. The authors reported normal appearance of teeth by histology and did not note premature tooth loss.

A question within the HPP research community has been posed as to whether alveolar bone defects precede and contribute to tooth loss, or whether alveolar bone loss is a direct effect of decreased TNAP activity or a secondary consequence due to loss of periodontal attachment. Within the limits of our study, alveolar bone defects were recognized by 2 months, though loss of teeth by this age indicates early periodontal attachment defects. Premature loss of teeth in compound heterozygous sheep was associated with alveolar bone degeneration, reduced volume, drift of teeth, and malocclusion at later ages. Malocclusion is a common problem in HPP and other conditions leading to premature primary tooth loss. Alveolar bone volume and density defects appeared to rebound by 7 months in Hom and cHet sheep, even in the face of progressive incisor loss in the latter. This observation suggests that developmental alveolar bone defects can improve with time. Secondary teeth are anecdotally less frequently lost in individuals affected by HPP, and amelioration of alveolar bone defects with age could contribute to this pattern. We noted early eruption of secondary teeth in association with premature exfoliation of primary teeth in the same positions. Although eruption abnormalities are not widely reported in the HPP literature, early eruption of permanent dentition has been noted in other conditions where deciduous dentition is extracted or lost prematurely.^{( 43 )}

The index patient preceding the sheep model carried biallelic p.A116T and p.I359M ALPL point mutations.^{( 19 )} Families carrying the p.A116T mutation alone have been described as mildly affected when heterozygous, with manifestations resembling odonto‐HPP.^{( 44} , ^{45 )} The p.I359M ALPL mutation has only been described in isolation in one case report where the proband was homozygous for the missense mutation and presented with a relatively mild form of childhood HPP, though he reportedly lost primary incisors prematurely and had development abnormalities in secondary incisors.^{( 20 )} The index patient upon which the sheep model was based presented with premature tooth loss and acellular cementum defects similar to those in cHet sheep analyzed in this report. Interestingly, we reported novel mantle dentin hypomineralization defects in the index patient^{( 19 )} that are faithfully replicated in the sheep; the magnitudes of both human and sheep mantle dentin defects are nearly identical at 10% reduced mineral density. This type of dentin defect is intriguing, as mantle dentin is adjacent to acellular cementum, and the potential developmental interactions between the tissues remain unclear, though the importance of ameloblast (enamel)‐odontoblast (dentin) interactions are necessary for proper formation of the dentin‐enamel junction in the crown, and alterations in cementum have been linked to dentin defects in previous studies.^{( 46 )} We also noted an abnormal accumulation of interglobular dentin evident by histology in the cHet HPP sheep. Interglobular dentin results from disrupted mineralization due to TNAP inactivation, resulting in unmerged calcospherites, discrete mineralization foci that normally would grow and merge into a unified mineralization front. Interglobular dentin is more commonly reported in rachitic conditions including X‐linked hypophosphatemia (XLH),^{( 47 )} and to date only a very few case reports to our knowledge have described interglobular dentin in teeth from HPP patients.^{( 44} , ^{48 )} HPP mouse models have, to date, not been useful for modeling mantle dentin defects or interglobular dentin,^{( 19 )} in part due to the extremely small size of mouse dentition, as well as rapid crown and root development with different sequences of mineralization events than in human odontogenesis.^{( 49} , ^{50 )} Time‐course analyses of dentin and cementum defects in sheep with HPP may provide insights into the nature of these tissue defects and whether there are mechanistic interconnections.

Conclusions

The sheep model presents numerous dental and craniofacial phenotypic aspects of HPP that cannot be studied in mouse models and displays genotype–phenotype associations across the spectrum of HPP defects. This study demonstrates the utility of the sheep model as a translational platform to further our understanding of the effects of HPP disease mechanisms on tooth development and an exciting model for long‐term assessment of therapeutic interventions on tooth retention. Furthermore, gene editing of sheep may be a useful approach for modeling other complex dental and skeletal conditions. Limitations of the current study include the small sample size (inherent in the use of livestock species) and that quantitative data are limited to a single tooth type, incisors. Although challenges of the sheep model include extended experimental periods due to slower development and growth and difficulty collecting all spontaneously exfoliated teeth, these are offset by major strengths including diphyodont dentition and osteonal remodeling of cortical bone as well as long lifespan and repeated biopsy potential.

The current enzyme replacement therapy (asfotase alfa; STRENSIQ™) is used for severe forms of HPP,^{( 9} , ^{51 )} though the therapeutic expense, injection site reactions, and potential for ectopic calcification has prompted research into additional approaches, including gene therapy.^{( 52 )} This model provides a unique opportunity to assess novel treatments for the many HPP musculoskeletal and dentoalveolar manifestations.

Author Contributions

Fatma F Mohamed: Formal analysis; investigation; methodology; writing – original draft; writing – review and editing. Michael B Chavez: Formal analysis; investigation; methodology; writing – review and editing. Shannon Huggins: Formal analysis; investigation; writing – review and editing. Joshua Bertels: Formal analysis; investigation; writing – review and editing. Alyssa Falck: Formal analysis; investigation; writing – review and editing. Larry J Suva: Conceptualization; formal analysis; funding acquisition; investigation; methodology; supervision; writing – original draft; writing – review and editing. Brian L Foster: Conceptualization; formal analysis; funding acquisition; investigation; methodology; supervision; writing – original draft; writing – review and editing. Dana Gaddy: Conceptualization; formal analysis; funding acquisition; investigation; methodology; supervision; writing – original draft; writing – review and editing.

Conflicts of Interest

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Peer Review

The peer review history for this article is available at https://publons.com/publon/10.1002/jbmr.4666.

Supporting information

Fig. S1. Assessment and Landmarks Used for Analyzing Skull Shape/growth in Sheep. (A, B) 3D and 2D CT reconstructions of a wild‐type (WT) sheep skull at age of 14 months to demonstrate landmarks used for measuring skull length and width (A, red lines) and used for skull orientation (B, yellow dotted lines). Scale bar: 5 cm. (C) Time‐course assessment of the patency of cranial base synchondroses: Intersphenoid synchondrosis (ISS) and spheno‐occipital synchondrosis (SOS), both of which contribute to anterior–posterior skull growth, where O: Open (highlighted in red when difference between males and females was observed), C: Closed, a: Anterior; p: Posterior. (D) 2D CT images showed patency of SOS in male (red arrowhead) and closure of SOS in female (white arrowhead) at 7 months. (E) Linear measurements revealed skull size differences between WT male and female sheep.

Fig. S2. Altered Pattern of Primary Tooth Exfoliation and Secondary Tooth Eruption in Sheep with Hypophosphatasia. 2D CT analysis of sheep primary and secondary dentition and associated alveolar bone at (A) 2, (B) 7, and (C) 14 months. Arrowheads point to alveolar bone (AB). Asterisks indicate premature loss of primary teeth and # refers to permanent central incisors.

Data Availability Statement

Data available on request from the authors