Effects of Active Vitamin D or FGF23 Antibody on Hyp Mice Dentoalveolar Tissues

Abstract

Mutations in the PHEX gene lead to X-linked hypophosphatemia (XLH), a form of inherited rickets featuring elevated fibroblast growth factor 23 (FGF23), reduced 1,25-dihydroxyvitamin D (1,25D), and hypophosphatemia. Hyp mutant mice replicate the XLH phenotype, including dentin, alveolar bone, and cementum defects. We aimed to compare effects of 1,25D versus FGF23-neutralizing antibody (FGF23Ab) monotherapies on Hyp mouse dentoalveolar mineralization. Male Hyp mice, either injected subcutaneously with daily 1,25D or thrice weekly with FGF23 blocking antibody from 2 to 35 d postnatal, were compared to wild-type (WT) controls and untreated Hyp mice. Mandibles were analyzed by high-resolution micro–computed tomography (micro-CT), histology, and immunohistochemistry. Both interventions maintained normocalcemia, increased serum phosphate levels, and improved dentoalveolar mineralization in treated versus untreated Hyp mice. 1,25D increased crown dentin volume and thickness and root dentin/cementum volume, whereas FGF23Ab effects were limited to crown dentin volume. 1,25D increased bone volume fraction, bone mineral density, and tissue mineral density in Hyp mice, whereas FGF23Ab failed to significantly affect these alveolar bone parameters. Neither treatment fully attenuated dentin and bone defects to WT levels, and pulp volumes remained elevated regardless of treatment. Both treatments reduced predentin thickness and improved periodontal ligament organization, while 1,25D promoted a more profound improvement in acellular cementum thickness. Altered cell densities and lacunocanalicular properties of alveolar and mandibular bone osteocytes and cementocytes in Hyp mice were partially corrected by either treatment. Neither treatment normalized the altered distributions of bone sialoprotein and osteopontin in Hyp mouse alveolar bone. Moderate improvements from both 1,25D and FGF23Ab treatment regimens support further studies and collection of oral health data from subjects receiving a newly approved anti-FGF23 therapy. The inability of either treatment to fully correct Hyp mouse dentin and bone prompts further experiments into underlying pathological mechanisms to identify new therapeutic approaches.

Introduction

X-linked hypophosphatemia (XLH; OMIM #307800) is the most common form of inherited rickets at 1:20,000 births (Carpenter et al. 2011). XLH is caused by inactivating mutations in PHEX (phosphate regulating endopeptidase homolog, X-linked), leading to elevated levels of the phosphaturic hormone, fibroblast growth factor 23 (FGF23). High serum FGF23 levels impair production of 1,25-dihydroxyvitamin D (1,25D) and increase urine phosphate wasting, leading to hypophosphatemia. These combined endocrine disturbances result in osteomalacia and growth plate defects that manifest as rickets. XLH also causes defects in dentoalveolar tissues. Reports on XLH-associated dental pathology highlight dentin mineralization defects, including accumulation of hypomineralized interglobular dentin and widened predentin, as well as thin dentin and cementum, enlarged pulp chambers, enamel abnormalities, and alveolar bone hypomineralization; these defects contribute to pulpal necrosis, dental abscesses, and increased prevalence of periodontal disorders (Vital et al. 2012; Foster et al. 2014; Biosse Duplan et al. 2017; Boukpessi et al. 2017; Coyac et al. 2018; Skrinar et al. 2019; Chavez et al. 2020).

Conventional treatment for XLH includes combination 1,25D and phosphate supplementation, which only partially ameliorates skeletal and dental mineralization defects (Biosse Duplan et al. 2017; Lambert et al. 2019). An FGF23-targeting monoclonal antibody (burosumab; Crysvita [Ultragenyx Pharmaceutical Inc.]) was approved in 2018 for the treatment of adult and pediatric XLH patients (Carpenter et al. 2018; Insogna et al. 2018; Imel et al. 2019). Burosumab treatment of XLH patients improves skeletal mineralization and mineral ion and hormone levels, but its effects on dental tissues remain unknown.

Hyp mice carry mutations in Phex and replicate the XLH phenotype observed in humans, including dentoalveolar defects (Eicher et al. 1976; Foster et al. 2014; Coyac et al. 2018; Zhang et al. 2020). In these studies, we aimed to compare effects of 1,25D or FGF23-inactivating antibody (FGF23Ab) treatment (both without phosphate supplementation) on dentoalveolar tissues of Hyp mice. Both optimized 1,25D monotherapy and FGF23Ab administration to Hyp mice improved skeletal microarchitecture and mechanical properties, as well as osteocyte lacunocanalicular organization, with 1,25D being superior to FGF23Ab in improving trabecular bone structure and mineralization (Liu et al. 2016; Tokarz et al. 2018). We hypothesized that 1,25D or FGF23Ab would improve dentoalveolar mineralization in the Hyp mouse model of XLH. Comparison of these therapeutic approaches will provide new insights into molecular and hormonal regulation of dentoalveolar tissues.

Materials and Methods

Mice

Animal studies were approved by the institutional animal care committee (Massachusetts General Hospital, Boston, MA) and comply with the ARRIVE checklist. Mice were housed at 22°C/72°F. Hyp and wild-type (WT) mice on a C57BL/6J background were weaned at 18 d postnatal (dpn) onto acidified water and house chow (Prolab Isopro RMH3000 irradiated, #3003219-249 [LabDiet]: 1% calcium, 0.6% phosphate). Starting at 2 dpn, male Hyp mice were subcutaneously injected daily with 1,25D (175 pg/g/d; Akorn) or thrice weekly with an FGF23-blocking antibody (35 mcg/g; Amgen) (n = 4–6 mice/group) (Appendix Fig. 1A). A total of 19 mice were used. WT and Hyp control littermates received vehicle or isotype-matched antibody (35 mcg/g). Doses were based on previous reports that determined optimized dosing of 1,25D monotherapy in Hyp mice (Marie et al. 1982; Liu et al. 2016) and demonstrated efficacy of Amgen FGF23Ab in rodents (Shalhoub et al. 2012). Serum and urine biochemistries were analyzed at study termination at 35 dpn (Liu et al. 2016). Additional details are in the Appendix. Tissues were fixed in 10% neutral buffered formalin. Effects of 1,25D and FGF23Ab on skeletal tissues in this mouse cohort were previously described (Liu et al. 2016).

Micro–Computed Tomography

Right hemimandibles were scanned in a µCT 50 (Scanco Medical) at 70 kVp, 76 µA, 0.5-Al filter, 900-ms integration time, and 6-µm voxel dimension. Analyses were performed as previously described (Zhang et al. 2020) and detailed in the Appendix for n = 4 to 6 samples per experimental group.

Histology

Left hemimandibles were demineralized in AFS solution (acetic acid, formaldehyde, sodium chloride) at 4°C with agitation for 4 to 5 wk, then embedded in paraffin for 6-µm serial sectioning. Deparaffinized coronal sections of mandibles were stained by hematoxylin and eosin (H&E), toluidine blue, or picrosirius red (Zhang et al. 2020). Silver staining was performed for osteocyte/cementocyte measurements as detailed in the Appendix. Histomorphometry for cementum and predentin and immunohistochemistry (IHC) for bone sialoprotein (BSP) and osteopontin (OPN) are described in the Appendix.

Statistical Analysis

Mean ± standard deviation (SD) are shown in graphs. Data sets were checked for normality and equal variance and analyzed using 1-way analysis of variance (ANOVA) with post hoc Tukey test (Prism version 8.4.3; GraphPad Software), whereP < 0.05 was considered statistically significant. Data reproducibility was ensured by random treatment assignments, semiautomated micro–computed tomography (micro-CT) analysis, and blinded histological measurements.

Results

Effects of 1,25D or FGF23Ab on Dentin and Alveolar Bone Defects in Hyp Mice

Our previous study demonstrated that 1,25D or FGF23Ab treatment improves skeletal mineralization and growth in Hyp mice (Liu et al. 2016), and here we investigated effects of these therapies on Hyp dentoalveolar tissues in the same cohort. Male Hyp mice were subcutaneously injected daily with 1,25D or thrice weekly with an FGF23-blocking antibody from 2 dpn to 35 dpn (Appendix Fig. 1A). Serum data are shown in Appendix Figure 1B and were previously reported (Liu et al. 2016). Normocalcemia was maintained in all groups. Treatment with 1,25D or FGF23Ab similarly increased serum phosphate in Hyp mice, although levels remained low versus WT controls. Urinary phosphate excretion was reduced by both treatments. 1,25D levels were doubled in Hyp mice administered FGF23Ab at 43.6 ± 12.2 pmol/L in Hyp versus 108.6 ± 20.1 pmol/L in Hyp with FGF23Ab (P < 0.05). Serum FGF23 levels, 8-fold higher in Hyp versus WT mice, were doubled by 1,25D treatment at 508 ± 231 pg/mL in WT versus 4,250 ± 773 pg/mL in Hyp versus 9,137 ± 1,142 pg/mL in Hyp with 1,25D (P < 0.05 between all groups).

Micro-CT showed generalized improvements in dentoalveolar mineralization from 1,25D and FGF23Ab treatments, although persistent dentin and alveolar bone defects were apparent in treated Hyp versus WT mice (Figs. 1 and 2A–D). Some FGF23Ab-treated Hyp mice showed unusual accumulation of reparative dentin. Local severe alveolar bone loss consistent with periapical lesions was noted in some untreated and FGF23Ab-treated Hyp mandibles. Enamel parameters were not different between genotypes/treatment groups, although enamel volume trended lower in all Hyp groups (Appendix Fig. 2). In Hyp mice, root dentin and cementum could not be reliably segmented due to mineralization defects, so the tissues were analyzed in aggregate. Hyp mice had reduced dentin parameters (volume, thickness, and density) compared to WT (Fig. 2E–I). 1,25D increased crown dentin volume and thickness and root dentin/cementum volume, although the majority of these were not rescued to WT levels. FGF23Ab increased crown dentin volume but failed to substantially affect other dentin parameters; these remained similar to untreated Hyp mice. Elevated pulp volume in untreated Hyp mice remained increased almost 2-fold with either treatment (Fig. 2J).

Figure 1.

Treatment effects of 1,25-dihydroxyvitamin D (1,25D) and fibroblast growth factor 23 antibody (FGF23Ab) on Hyp mouse mandibles. Three-dimensional micro–computed tomography renderings show right hemimandibles for all mice in the study (n = 4 to 6 mice per experimental group). (A) Wild-type (WT) mandibles show normal presentation of enamel (EN), dentin (DE), dental pulp (DP), and alveolar bone (AB), while (B) Hyp mice exhibit thin DE, wide DP chambers (yellow *), reduced and disorganized AB (yellow arrow), and severe regional AB loss consistent with periapical lesions (yellow #). Treatment with either (C) 1,25D or (D) FGF23Ab improves dentoalveolar mineralization in Hyp mice, although DE and AB defects remain. FGF23Ab-treated Hyp mice exhibit unusual accumulation of reparative dentin (blue arrows) in cusp tips. All images are par scaled for 1-mm scale bar, and mice in all groups are arranged in descending order from left to right based on greatest to least dentin volume.

Figure 2.

Effects of 1,25-dihydroxyvitamin D (1,25D) or fibroblast growth factor 23 antibody (FGF23Ab) on dentin and alveolar bone defects in Hyp mice. (A–D) Three-dimensional and 2-dimensional micro–computed tomography (CT) renderings show first mandibular molar (M1) enamel (EN), dentin (DE), dental pulp (DP), and surrounding alveolar bone (AB) of wild-type (WT), untreated Hyp mice and Hyp mice treated by 1,25D or FGF23Ab. Hyp mouse molars have thin DE and enlarged DP chambers (yellow *) and AB that is severely disorganized (yellow arrows). 1,25D improves the appearance of DE and AB. Effects of FGF23Ab are less clear, and regions of apparent tertiary dentin (blue arrows) are evident in some mice. Images in panels (A–D) are par scaled for 1-mm scale bar. (E–M) DE, DP, and AB parameters of volume, thickness, and density (n = 4 to 6 mice per experimental group) were measured by micro-CT and compared by 1-way analysis of variance and post hoc Tukey test. *P < 0.05. **P < 0.01. ***P < 0.001. ****P < 0.0001.

Hyp mice alveolar bone had reduced bone volume fraction (BV/TV), bone mineral density (BMD), and tissue mineral density (TMD) compared to WT (Fig. 2K–M). 1,25D improved all 3 bone readouts by 5% to 20%, although all remained lower than WT. FGF23Ab treatment did not make substantial improvements in alveolar bone parameters in Hyp mice.

Improved Dentin and Periodontal Structures in Hyp Mice with 1,25D or FGF23Ab

Histology demonstrated that Hyp molars had thin dentin, widened predentin, thin acellular cementum, reduced cellular cementum, and accumulation of osteoid in alveolar bone (Fig. 3A–D). Both treatments partially restored acellular cementum (Fig. 3B) and cellular cementum area (Fig. 3C), although substantial osteoid remained evident in all Hyp experimental groups (Fig. 3B). Notably, 1 untreated Hyp mouse showed pulpal necrosis and increased immune cells consistent with infection (star in Fig. 3C). Pulpal infection was noted in only 1 additional animal in the study, a mouse treated with FGF23Ab (data not shown). Interglobular dentin patterns and erratic predentin-dentin borders in Hyp mice were partially normalized by both treatments (Fig. 3D). Picrosirius red (PR) staining viewed under polarized light revealed highly disorganized periodontal ligament (PDL) in untreated Hyp mice, and both treatments improved functional orientation of PDL fibers in the oblique/horizontal direction between cementum and alveolar bone (Fig. 3E). Histomorphometry confirmed significant improvement but not complete rescue of predentin width to WT levels by 1,25D and FGF23Ab (Fig. 3F). Reduced acellular cementum thickness of Hyp versus WT mouse teeth was increased by 1,25D but not significantly affected by FGF23Ab treatment (Fig. 3G). Cellular cementum area was marginally improved by both treatments (Fig. 3H).

Figure 3.

Improved dentin and periodontal structures in Hyp mice treated with 1,25-dihydroxyvitamin D (1,25D) or fibroblast growth factor 23 antibody (FGF23Ab). (A–C) Hematoxylin and eosin staining shows that Hyp mice have thin dentin (DE), thin acellular cementum (AC), reduced cellular cementum (CC), and accumulation of osteoid (red *) in alveolar bone (AB). 1,25D and FGF23Ab treatments each partially restore acellular and cellular cementum, although substantial osteoid (red * in panel B) remains in all Hyp experimental groups. One untreated Hyp mouse showed signs of pulpal necrosis and inflammation (yellow # in C). Boxes labeled (B–E) in panel (A) indicate regions shown at higher magnifications in the specified panels. CC, cellular cementum; M1, first mandibular molar; PDL, periodontal ligament. (D) Toluidine blue (TB) staining reveals wide predentin (PD) and erratic PD-DE border with interglobular DE patterns in Hyp mice (black *). DE and PD are partially normalized by either 1,25D or FGF23Ab treatments, although mineralization defects persist (black *). (E) Picrosirius red (PR) staining viewed under polarized light microcopy emphasizes highly organized PDL fibers in wild-type (WT) mice (indicated by light blue dotted line). Hyp mice have loss of PDL fiber organization shown by low signal and nonuniformity of fiber directions (yellow stars), and either treatment improves functional orientation of fibers in the horizontal/oblique direction (light blue dotted lines) between AC and AB. Histomorphometry confirms (F) significant improvement but not complete rescue of PD width and significant effect of 1,25D but not FGF23Ab on (G) acellular cementum thickness and (H) cellular cementum area. Thicknesses and areas (n = 4 to 6 mice per experimental group) were measured by histomorphometry and compared by 1-way analysis of variance and post hoc Tukey test. *P < 0.05. **P < 0.01. ***P < 0.001. ****P < 0.0001.

Altered Osteocyte and Cementocyte Lacunocanalicular Parameters in Hyp Mice Are Not Corrected with Treatment

Measurements were performed to determine if either treatment improved perilacunar hypomineralization defects or abnormal canalicular network in Hyp mouse alveolar bone osteocytes or cementocytes (Tokarz et al. 2018; Zhang et al. 2020). Silver staining revealed abnormal stain accumulation in the octeocytic lamina limitans and confirmed altered lacunocanalicular organization in alveolar and mandible basal bone osteocytes and in cementocytes in Hyp mice (Fig. 4A, C, E). Densities of osteocytes and cementocytes in Hyp versus WT mice were decreased, with minimal improvement by 1,25D or FGF23Ab (Fig. 4B, D, F). Lacunar area and lacunar height and width were significantly increased in Hyp versus WT mice in alveolar and mandibular bone. Both treatments partially corrected these parameters. Canaliculi number and length were decreased in Hyp versus WT mice. Treatment with 1,25D or FGF23Ab did not improve canalicular parameters. Cementocyte lacunar parameters were highly variable and showed no trends with treatments. Visualization of the cementocyte canaliculi was inconsistent by silver staining and so was not analyzed.

Figure 4.

Altered osteocyte and cementocyte lacunocanalicular parameters in Hyp mice are not corrected with treatment. Silver staining shows normally distributed and proportioned osteocytes (Ocy; white arrowheads) and canaliculi (yellow arrows) in wild-type (WT) mice and abnormal stain accumulation in the lamina limitans of aberrant Ocy lacunocanalicular systems and short canaliculi, in (A) alveolar bone Ocy, (C) mandibular basal bone Ocy, and (E) cellular cementum cementocytes (Ccy). (B, D) Alterations in cell density, lacunar surface area, height and width, and canalicular number and length are observed in Hyp versus WT mouse Ocy, with partial recovery by 1,25-dihydroxyvitamin D (1,25D) or fibroblast growth factor 23 antibody (FGF23Ab) treatment. (F) Density of Ccy is reduced in Hyp versus WT mice, with some recovery from treatment. Cementocyte lacunar parameters are highly variable and show no trends with treatments. The cementocyte canalicular organization was not analyzed. Lacunocanalicular measurements (n = 4 to 5 mice per experimental group) were compared by 1-way analysis of variance and post hoc Tukey test. *P < 0.05. **P < 0.01. ***P < 0.001. ****P < 0.0001.

Altered Alveolar Bone Markers Persist in Hyp Mice with Treatment

Bone sialoprotein (BSP) and osteopontin (OPN) are extracellular matrix (ECM) proteins that serve as markers for cementum and bone identity and regulate tissue mineralization (Foster et al. 2015, 2018). Compared to substantial BSP localization that marks acellular cementum on root surfaces of WT teeth, the cementum of Hyp mice had inconsistent staining of BSP, indicating quantitative or qualitative defects in cementogenesis (Fig. 5A, B). Both 1,25D and FGF23Ab restored robust BSP localization on root surfaces. BSP staining in alveolar bone of Hyp versus WT mice was marked by regions of absence and excess (Zhang et al. 2020), where this abnormal distribution was not substantially changed by either treatment.

Figure 5.

Altered alveolar bone markers persist in Hyp mice with treatment. (A, B) Bone sialoprotein (BSP) is a marker of acellular cementum (AC) and alveolar bone (AB). AC defects in Hyp mice are indicated by inconsistent BSP staining (red #), whereas both 1,25-dihydroxyvitamin D (1,25D) and fibroblast growth factor 23 antibody (FGF23Ab) improve BSP localization on root surfaces (black arrows). Hyp mouse AB features areas of excess and absence of BSP staining versus wild-type (WT) mice, and this pattern is not substantially changed by either treatment. (C, D) Osteopontin (OPN) is a marker of AC and AB, and all genotypes show OPN staining on root surfaces, although Hyp mice have a more diffuse pattern (black arrows). OPN is increased in Hyp versus WT AB, and this excess persists in Hyp mice with either treatment.

WT molars have well-defined OPN staining marking the acellular cementum, while Hyp molars have a diffuse pattern of OPN expression on root surfaces (Fig. 5C, D). OPN immunolocalization was increased within Hyp versus WT alveolar bone, showing intense staining over much broader areas of bone. Excess OPN immunoreactivity persisted in Hyp mice with either treatment.

Discussion

Hyp mice carry mutations in the Phex gene and replicate dental defects associated with XLH. We compared dentoalveolar tissues of Hyp mice receiving 1,25D versus FGF23Ab treatments. Both treatments moderately improved dentin mineralization, predentin width, and PDL organization. Neither therapy significantly improved alveolar bone and mandibular osteocyte lacunocanalicular abnormalities, osteoid, or expression of bone ECM markers. 1,25D improved dentin and alveolar bone parameters as well as acellular and cellular cementum formation to a greater degree than did FGF23Ab. These findings indicate that early and consistent treatment with 1,25D or FGF23Ab can partially ameliorate developmental defects in Hyp dentoalveolar tissues.

1,25D and FGF23Ab Treatments Improve Dentoalveolar Mineralization in Hyp Mice

Conventional treatment for XLH includes 1,25D and phosphate supplementation (Glorieux et al. 1980; Harrell et al. 1985). We aimed to use optimized 1,25D monotherapy without phosphate supplementation in this study for several reasons. First, oral phosphate alone is rapidly cleared and suppresses 1,25D synthesis (Meyer et al. 1996). Second, therapy with a combination of 1,25D and phosphate can be complicated by nephrocalcinosis, where the incidence of nephrocalcinosis correlates with the dose of phosphate but not 1,25D (Verge et al. 1991). Third, 1,25D monotherapy is also advantageous over combined therapy because 1,25D alone decreases renal phosphate wasting (Liu et al. 2016; Martins et al. 2019).

Burosumab, an FGF23-inactivating monoclonal antibody, is a recently approved first-in-class treatment for children and adults with XLH. In clinical trials, burosumab increased serum phosphate, reduced osteomalacia, and improved fracture healing in adults with XLH (Insogna et al. 2018; Insogna et al. 2019; Portale et al. 2019). Preclinical murine studies of FGF23Ab did not include dental analysis (Aono et al. 2009), and clinical trial reports on burosumab have included limited data regarding treatment effects on oral health (Insogna et al. 2018; Gordon and Levine 2019; Imel et al. 2019; Insogna et al. 2019; Portale et al. 2019). Based on the ability of FGF23Ab to decrease renal phosphate excretion and increase endogenous 1,25D production, FGF23Ab-treated Hyp mice served as controls for 1,25D monotherapy, in addition to allowing us to test FGF23 inhibition as a potential therapy for improvement of dentoalveolar effects in XLH. Furthermore, the FGF23Ab has the advantage of blocking actions specific to FGF23 independent of its ability to improve serum phosphate and 1,25D. We aimed to directly compare efficacy of 1,25D monotherapy versus FGF23Ab treatments on dentoalveolar development and mineralization in the Hyp mouse model of XLH.

Dentoalveolar manifestations in Hyp mice include defects in dentin volume and mineralization, accumulation of interglobular dentin, enlarged pulp chambers, reduced cementum, osteomalacia in alveolar bone, and defective periodontal attachment and mechanical properties (Abe et al. 1989; Coyac et al. 2018; Zhang et al. 2020). Increased dietary calcium and phosphate were unable to correct interglobular dentin in Hyp mice (Abe et al. 1992; Masatomi et al. 1996). Our results show that optimized 1,25D increased dentin volume and thickness, reduced predentin thickness, improved all alveolar bone parameters analyzed, and increased cementum thickness in Hyp mice. Improvements from FGF23Ab treatment were limited to increased dentin volume and reduced predentin thickness. Both therapies resulted in similar increases in serum phosphate, allowing comparison of 1,25D and FGF23Ab effects. 1,25D outperformed FGF23Ab in attenuating Hyp dentoalveolar mineralization defects, despite increased FGF23 levels in 1,25D-treated Hyp mice. This observation underscores the importance of 1,25D in regulating dentoalveolar mineralization, also suggesting that dentoalveolar defects in XLH are not due to FGF23-specific actions independent of its role modulating serum phosphate and 1,25D levels. Further studies are needed to address this point.

Similar to tibial and calvarial osteocytes in Hyp mice (Tokarz et al. 2018), we show here and in a previous report (Zhang et al. 2020) that alveolar and mandibular bone osteocytes and cemenocytes in Hyp mice also have lacunocanalicular defects. Treatment with 1,25D or FGF23Ab significantly improved osteocyte lacunocanalicular organization in calvariae and tibiae by decreasing lacunar size and partially restoring canalicular structure (Tokarz et al. 2018). In Hyp mice treated with the same doses of 1,25D and FGF23Ab, both treatments partially decreased lacunar size but did not improve canalicular structure in alveolar and mandibular bone and cementum. Because the osteocyte lacunocanalicular network organization is associated with tissue mineralization (Dallas et al. 2013), the persistent lacunocanalicular abnormalities observed in alveolar and mandibular bone may be due to the inability of either 1,25D or FGF23Ab to normalize dentoalveolar mineralization.

PHEX is highly expressed by odontoblasts and osteocytes, with lower expression by cementocytes (Zhang et al. 2020). Although the physiologic substrates of PHEX remain uncertain, the enzyme cleaves full-length OPN and derived acidic serine- and aspartate-rich motif (ASARM) peptides that function as mineralization inhibitors (Addison et al. 2010; Barros et al. 2013; Salmon et al. 2014). Studies suggest accumulation of OPN in bone and dentin in Hyp mice and humans with XLH contributes to impaired mineralization (Zhang et al. 2010; Salmon et al. 2014; Boukpessi et al. 2017; Zhang et al. 2020). In support of this, genetic ablation of Spp1/OPN in Hyp mice partially improved vertebral and tibial mineralization despite persistent hypophosphatemia (Hoac et al. 2020). While 1,25D and FGF23Ab attenuated select downstream effects of Phex mutations (e.g., systemic disturbances in hormones regulating mineral metabolism), they did not improve local expression of OPN or BSP in the Hyp ECM. Therefore, the inability of either treatment to normalize dentoalveolar mineralization may be due to continued altered expression of OPN.

Clinical Implications for XLH Dentoalveolar Defects

Impaired mineralization (i.e., interglobular dentin, thin dentin, and widened predentin) in XLH is thought to be responsible for the high incidence of dental abscesses and necrosis and tooth extraction observed in affected individuals (Vital et al. 2012; Foster et al. 2014; Boukpessi et al. 2017; Coyac et al. 2018; Skrinar et al. 2019). Consistent with this, a clinical study demonstrated decreased severity of dental complications in adults with XLH in proportion with length of combination 1,25D and phosphate therapy (Connor et al. 2015). Surprisingly, incidence of dental abscesses nearly doubled in adults with XLH receiving burosumab relative to XLH subjects receiving placebo in a phase III trial (Insogna et al. 2018). Furthermore, a phase III trial reported that children with XLH receiving burosumab also had an increased incidence of dental caries and abscesses compared with those treated with conventional combination phosphate and 1,25D therapy (Imel et al. 2019). Although it is unclear if and why treatment with burosumab would lead to worsening dental complications, this potential increase in dental abscesses associated with burosumab is reason for concern. Dental pulp infections in Hyp mice have not been previously reported. Although our study has too small a sample size to draw conclusions, we observed periapical hypomineralized lesions and pulpal necrosis/inflammation in Hyp mice. Both 1,25D and FGF23Ab promoted consistent improvements in Hyp mouse dentin parameters. Within the limits of this study, it is unclear whether FGF23Ab is associated with increased dental abscesses, a point deserving additional murine studies and human clinical data collection.

Few reports on XLH have included cementum analysis. Reductions in acellular cementum thickness were shown in studies of the primary and secondary dentitions in individuals with XLH (Biosse Duplan et al. 2017; Chavez et al. 2020), and cellular cementum mineralization defects were also associated with XLH (Coyac et al. 2017). Hyp mice have reduced acellular cementum, increased periodontal detachment and alveolar bone osteomalacia, and significantly altered periodontal mechanical properties (Zhang et al. 2020). These periodontal defects in humans and mice with XLH may contribute to increased periodontal disease later in life in individuals with XLH (Biosse Duplan et al. 2017). We found that both 1,25D and FGF23Ab increased cementum thickness in Hyp mice. Improved PDL attachment and collagen organization accompanied thicker cementum, contributing to a more normalized periodontal structure. These data support the findings that early and consistent treatment of XLH patients with 1,25D and phosphate supplements improved acellular cementum organization and reduced the prevalence of periodontal disease (Biosse Duplan et al. 2017). Therefore, early and consistent treatment with 1,25D monotherapy or FGF23Ab treatment in patients with XLH would thus be expected to lead to improvements in periodontal mineralization, attachment, and function.

Conclusions

Both 1,25D and FGF23Ab treatments moderately improved dentoalveolar mineralization in Hyp mice, with 1,25D being superior to FGF23Ab in improving dentin and alveolar bone microarchitecture. These results support further studies, including use of combined 1,25D and FGF23Ab therapy, to determine additive or synergistic effects. Studies collecting oral health data from patients receiving burosumab, 1,25D alone, or conventional treatment are necessary to understand whether targeting FGF23 or enhancing 1,25D action makes similar improvements to dentoalveolar tissues as it does for the skeleton. The inability of either treatment to fully correct Hyp mouse dentoalveolar tissues prompts further experiments to more fully define underlying pathological mechanisms and identify new therapeutic approaches.

Author Contributions

E.J. Lira dos Santos, F.F. Mohamed, contributed to data acquisition, analysis, and interpretation, drafted and critically revised the manuscript; M.B. Chavez, M.H. Tan, T.N. Kolli, contributed to data acquisition, analysis, and interpretation, critically revised the manuscript; B.L. Foster, contributed to conception, design, data analysis, and interpretation, drafted and critically revised the manuscript; E.S. Liu, contributed to conception, design, data acquisition, analysis, and interpretation, drafted and critically revised the manuscript. All authors gave final approval and agree to be accountable for all aspects of the work.