Dental tissue changes in juvenile and adult mice with osteogenesis imperfecta

Osteogenesis imperfecta (OI), a disorder of type I collagen, causes skeletal deformities as well as defects in dental tissues, which lead to increased enamel wear and smaller teeth with shorter roots. Mice with OI exhibit similar microstructural dentin changes, including reduced dentin tubule density and dentin cross‐sectional area. However, the effects of these mutations on gross dental morphology and dental tissue volumes have never been characterized in the osteogenesis imperfecta murine (OIM) mouse model. Here we compare mineralized dental tissue measurements of OIM mice and unaffected wild type (WT) littermates at the juvenile and adult stages. The maxillary and mandibular incisors and first molars were isolated from microCT scans, and tissue volumes and root length were measured. OIM mice have smaller teeth with shorter roots relative to WT controls. Maxillary incisor volumes differed significantly between OIM and WT mice at both the juvenile and young adult stage, perhaps due to shortening of the maxilla itself in OIM mice. Additionally, adult OIM mice have significantly less crown enamel volume than do juveniles, potentially due to loss through wear. Thus, OIM mice demonstrate a dental phenotype similar to humans with OI, with decreased tooth size, decreased root length, and accelerated enamel wear. Further investigation of dental development in the OIM mouse may have important implications for the development and treatment of dental issues in OI patients.

dentinogenesis imperfecta (DI), which is characterized by dentition with significant discoloration and structural defects.The effects of DI in humans are particularly severe in the deciduous dentition, such that children with OI experience increased rates of dental attrition, enamel fractures, pulp chamber obliteration, malocclusion, and feeding difficulties.
In DI, the secretion of malformed type I collagen in the developing dentin matrix disrupts the normal regulation and organization of this process, causing issues such as hypomineralization, disorganized dentin tubule structure, and dentin hypertrophy.These abnormal structural properties result in discoloration, enamel attrition, and spontaneous dental fractures, which comprise the DI disease phenotype.
The most visible feature of DI is discoloration, which results from malformed and hypertrophic dentin beneath the enamel surface.Affected teeth may display characteristic appearances ranging from an opalescent blue-gray to an amber or yellow color.Teeth affected by DI also tend to have shorter roots and are frequently smaller than normal (Majorana et al., 2010;Retrouvey et al., 2014).
Functional issues in DI include malocclusion and increased enamel wear.Because enamel is not made from type I collagen, its development and microstructure are not directly affected in DI (Majorana et al., 2010).However, structural issues in the underlying dentin limit its functional capacity to absorb the forces of mastication, leaving the enamel vulnerable to accelerated wear at the occlusive surface (Retrouvey et al., 2014).
Notably, the expressivity of DI can be highly variable between subtypes of OI and even within individual OI patients, often affecting some teeth but not others (Malmgren & Norgren, 2002;Schwartz & Tsipouras, 1984).While both the deciduous and permanent dentition are affected, it has been noted that the deciduous teeth tend to be affected more frequently and more severely, likely due to the faster formation of the deciduous teeth (Nguyen et al., 2021;Opsahl Vital et al., 2012;Retrouvey et al., 2014).Scanning electron microscopy and histological studies have found aberrant ultrastructure of dentin tubules in teeth of OI patients even in the absence of clinically apparent DI (Hall et al., 2002;Lukinmaa et al., 1987;Lygidakis et al., 1996;Waltimo et al., 1996).
While the human DI phenotype has been wellestablished clinically, there is considerably less information available regarding the effects of type I collagen mutations on the dentition in model organisms.The effect of these mutations on the skeletal phenotype has been wellexplored in animal models, but effects on dental phenotypes have not been explored as thoroughly.In one mouse model for OI (the OIM mouse), the dentition has been shown to have marked ultrastructural abnormalities, including reduced density and biomineralization, and increased irregularity of dentin tubules (Lopez Franco et al., 2005;Phillips et al., 2000).Further characterization of developmental changes in tissue volumes in the OIM mouse model will establish support for the use of this model to investigate development of DI.
While mice have served as the most commonly used model organism in research on dental development, a few key differences between human and rodent dental development merit consideration.Rodent molars develop in a pattern generally analogous to human teeth, with growth of the tooth completely arrested after the roots are formed.Beyond this point, the size of the molars either remains the same or decreases as the enamel wears down over time at the occlusive surface.Rodent incisors, however, continue to grow throughout the lifespan from stem cells at the apical end, replenishing tissue that is lost at the occlusive end (Krivanek et al., 2020).Notably, rodents develop only a single set of teeth that must last the entire lifespan, whereas humans develop a deciduous set followed by a permanent dentition.The timeline of the development and eruption of teeth in rodents is comparable to that of the deciduous dentition in humans, while the human permanent dentition develops and erupts over a much longer period.Thus, the analogous developmental timelines of rodent and deciduous human dentition permit the use of mice for the investigation of dental development in DI, particularly considering the condition's outsized effects on the deciduous dentition in humans (Nguyen et al., 2021;Opsahl Vital et al., 2012;Retrouvey et al., 2014).
This study investigates the presence of DI in the OIM model by quantifying gross dental morphology, enamel and dentin volumes, and structural differences in the teeth.OIM and unaffected wild type (WT) mice are compared at two ages: 4 weeks, the age at which juvenile mice are weaning and transitioning to solid food, and 16 weeks, when young adult mice have reached skeletal maturity.Further validation of the OIM mouse as a model for DI may lay the groundwork for future investigators to examine the early developmental pattern of this unique phenotype.This is of particular importance because the early events of odontogenesis cannot feasibly be studied in humans directly.

| Experimental model
The osteogenesis imperfecta murine (OIM) is a mouse strain with a nonlethal recessive inherited mutation of the COL1A2 gene.This experiment utilized homozygous OIM mice (B6C3FE a/a-Col1a2 oim/oim ; hereafter "OIM"), which are a model for the human presentation of OI type III, the "severe OI" phenotype.The control group in this experiment consisted of wild type (WT; B6C3FeF1/ J) mice.

| Animal husbandry
All procedures and animal care were approved by the Indiana University School of Medicine Institutional Animal Care and Use Committee.OIM mice and wild type (WT) littermates were raised from weaning (week 4) to adulthood (week 16; Table 1).All animals were imaged in vivo using micro-computed tomography (microCT) at weeks 4 and 16 through the Indiana University Center for Musculoskeletal Health.Scans were collected using a Skyscan 1176 microCT machine (Bruker Corp, MA) with reconstructions using 0.008 or 0.017 mm 3 voxels, depending on cranial length.

| Segmentation protocol
Digital 3D models of maxillary and mandibular incisors and first molars were constructed from microCT scans of each mouse using the imaging software 3D Slicer (Kikinis et al., 2014).Right maxillary and mandibular incisors and molars were segmented out by thresholding as described in Verdelis and Salmon (2019), followed by manual separation of tissue islands (Figure 1a).Segmentation by thresholding was used to further segment the enamel and dentin components of each tooth (Figure 1b,c).

| Data acquisition
Within 3D Slicer, the Segment Statistics module was used to quantify the following variables: total volume (the hard tissue volume of the whole tooth), enamel volume, and dentin volume.To obtain the mesial root length, a mesiodistal line was drawn through the cemento-enamel junction (CEJ) in a sagittal section of the mandibular first molar.Mesial root length was then measured as the distance from the apical foramen to the CEJ line (Figure 2).F I G U R E 2 Sagittal section of right first mandibular molar demonstrating cemento-enamel junction (CEJ) line and measurement of the mesial root length.

| Data analysis
Due to the small sample size for all groups, nonparametric tests were used for all comparisons.
To assess gross morphological differences between teeth of OIM and WT mice, Mann-Whitney U tests were used to compare genotype means of dentin, enamel, and total tooth volumes of maxillary and mandibular incisors and molars of OIM and WT mice at week 4 and week 16.
To assess differences in root length between teeth of OIM and WT mice, Mann-Whitney U tests were used to compare genotype means of mesial root length for mandibular molars of OIM and WT mice at week 4 and week 16.
Different aspects of enamel attrition were also explored.For molars, enamel volumes could be compared directly.However, the increased length of the rodent incisor throughout the life course prohibits the use of raw volumes to assess enamel attrition over time.Thus, to detect differences in relative amounts of enamel and dentin, E/D (enamel/dentin) ratios were utilized.This variable was defined as follows: Thus, an E/D ratio of 1 indicates equal amounts of enamel and dentin, E/D ratios >1 indicate more enamel than dentin, and E/D ratios <1 indicate more dentin than enamel.
To assess differences in proportions of dentin and enamel between week 4 and week 16 for mice in each genotype, Mann-Whitney U tests were used to compare E/D ratios of maxillary and mandibular molars and incisors.
For analyses of volumes and mesial root lengths, comparisons of unscaled measurements were utilized to detect absolute differences between genotypes and timepoints.To account for the effect of body size on size differences between teeth, additional analyses were performed after scaling measurements to mandibular centroid size.Centroid size is a dimensionless estimate of an object's threedimensional size and is often used as a proxy for body size in geometric morphometric analyses.The centroid sizes were quantified using 32 bilateral anatomical landmarks of each mouse's mandible, calculated as the square root of the sum of squared distances of all mandibular landmarks from their centroid (Table 2).For one specimen from the week 4 OIM group and one specimen from the week 16 WT group, mandibular centroid size could not be calculated due to anatomical variations that precluded complete landmarking.In these cases, measurements were scaled to the mean centroid size for the other individuals at that genotype and timepoint (Table S1).

| Incisor volumes
Maxillary incisors of OIM mice were found to have significantly smaller total volumes both before and after adjusting for centroid size at both week 4 and week 16 (p < 0.05; Figure 3; Table 3).Thus, maxillary incisors of OIM mice were both absolutely smaller and smaller relative to body size compared with WT mice.
Mandibular incisors of OIM mice were also found to have significantly smaller total volumes at week 4 (p = 0.001; Figure 3, Table 3).However, after scaling to centroid size, this difference was not significant.Thus, OIM mice had mandibular incisors that were absolutely smaller, but not smaller relative to body size, compared with their WT littermates at week 4.At week 16, in both scaled and unscaled comparisons, there was no significant difference between OIM and WT mice in total mandibular incisor volume.

| Molar volumes
No significant difference was found in the unscaled volumes of maxillary first molars between OIM and WT mice at week 4.At week 16, OIM mice had significantly smaller unscaled maxillary molar volumes compared with WT mice (p = 0.027; Table 3).Mandibular molars of OIM mice were also significantly smaller in total volume than those of WT mice at both week 4 (p = 0.010) and week 16 (p = 0.022; Figure 4, Table 3).For both mandibular and maxillary first molars, no statistically significant differences were found between genotypes after scaling to centroid size at either timepoint (Figure 4, Table 3).

| Changes in root length
Mesial root lengths of mandibular molars were shorter in OIM mice than in their WT littermates at week 4 (p < 0.001) and week 16 (p = 0.004; Figure 5; Table 3).After scaling, mesial root lengths continued to be statistically shorter in OIM mice than WT mice at week 4 (p < 0.001) and week 16 (p = 0.009; Figure 5; Table 3).
In both absolute terms and relative to mandibular centroid size, OIM mice have shorter mandibular molar roots than WT mice at the juvenile stage, and these differences are not attenuated after progression to young adulthood.

| E/D ratios
Incisors E/D ratios of incisors were not different between genotypes at week 4.However, by week 16, mandibular incisors of OIM mice had significantly lower proportions of enamel compared with WT mice ( p = 0.017; Figure 6b, Table 3).Comparing timepoints within each genotype, WT mice demonstrated higher proportions of enamel by week 16 compared with week 4 in both maxillary (p = 0.003; Figure 6a) and mandibular incisors (p = 0.005; Figure 6b).There was no such increase in proportions of enamel in OIM mice (Figure 6a,b).
Molars E/D ratios of molars were not different between genotypes at either timepoint.However, there was a significant timewise difference in molar E/D ratios for OIM mice, demonstrating proportionately less enamel at week 16 compared with week 4 in both maxillary (p = 0.009) (Figure 7a, Table 3) and mandibular molars (p = 0.013) (Figure 7b).This indicates that while there is an increased proportion of enamel in incisors as WT mice age from juvenile to young adult, there is a concurrent decrease in the proportion of enamel in molars as OIM mice age from juvenile to young adult.

| Enamel volume comparisons
To understand what drives differences in E/D ratios, the differences in proportion of enamel between OIM and WT mice were further investigated by analyzing differences in enamel volumes.

Incisors
Both maxillary and mandibular incisor enamel volumes demonstrated no significant differences between OIM and WT mice at week 4 (Table 3).At week 16, OIM mice had lower enamel volumes compared with WT mice in both maxillary (p = 0.003) and mandibular incisors (p = 0.006).These relationships held in scaled comparisons (Figure 8).

Molars
In molars, a similar pattern was observed.Both maxillary and mandibular molar enamel volumes show no significant differences between OIM and WT mice at week 4, in both unscaled and scaled comparisons (Figure 9, Table 3).At week 16, OIM mice had lower unscaled enamel volumes compared with WT mice for maxillary (p = 0.024) (Figure 9a) and mandibular molars (p = 0.046; Figure 9b).The same relationship did not hold when scaled enamel volumes were compared (Figure 9c,d).Notably, week 16 WT mice had a larger mandibular centroid size than did week 16 OIM mice, which could account for the lack of difference in the scaled volumes.

Time-wise comparison of molar enamel attrition
A final analysis of time-related changes in molar enamel volumes was done by comparing molar enamel volumes between week 4 and week 16 mice for each genotype.There were no significant differences in molar enamel volume between week 4 WT and week 16 WT mice (Figure 9a,b; Table 3).However, week 16 OIM mice had significantly lower enamel volumes than week 4 OIM mice for both mandibular and maxillary molars, both in scaled and unscaled comparisons (p < 0.05) (Figure 9).Thus, OIM mice, but not WT mice, had significantly less molar enamel at young adulthood than at weaning, both absolutely and relative to mandibular centroid size.This timewise loss of enamel volume reinforces the finding that E/D ratios decrease in OIM mice between week 4 and week 16.

| DISCUSSION
This study demonstrates that OIM mice exhibit changes in dental morphology and dental development which are similar to those seen in humans with DI in two ways: the gross morphology of the teeth reveals small tooth size and short root length, and temporal comparisons of enamel demonstrate the pattern of enamel loss over time.

| Total volume
At both weaning and skeletal maturity, OIM mice had smaller teeth compared with their WT littermates.Small teeth are a hallmark of DI, so this finding is consistent with the human phenotype (Retrouvey et al., 2014).After scaling the raw tooth volumes to centroid size, OIM and WT mice had proportionately similarly sized first molars and mandibular incisors, but OIM mice had significantly smaller maxillary incisors than the WT mice.Prior studies have demonstrated shortening of the maxilla in adult OIM mice (Menegaz et al., 2020).Thus, the smaller maxillary incisors in OIM mice may be related to shorter maxillae, which may limit the elongation of the upper incisors.

| Root length
The molar roots in OIM mice were significantly shorter than those of WT mice, both absolutely and relative to mandibular centroid size.This difference held at both timepoints.Shorter tooth roots are a commonly reported feature of DI (Majorana et al., 2010).Thus, these findings of shorter root lengths in OIM mouse molars are consistent with the known human DI phenotype, and further validate the OIM mouse as a model organism for DI.

| Enamel loss patterns
No difference was observed between E/D ratios in weaning OIM and WT mice.However, at adulthood, the mandibular incisors of the OIM mice demonstrated proportional increases in dentin, while WT mice demonstrated proportional increases in enamel.Thus, while mandibular incisors of both OIM and WT mice had similar hard tissue compositions at weaning, these values diverged by adulthood.These relative changes in composition are consistent with two hallmark features of dentinogenesis imperfecta: dentin hypertrophy due to continuous deposition of abnormal dentin and accelerated mechanical loss of enamel due to the frailty of the underlying dentin (Majorana et al., 2010;Retrouvey et al., 2014).Comparison of enamel volumes between adult OIM and WT mice demonstrated significantly lower enamel volumes for OIM mice in the incisors and in the molar crowns, and OIM mice had significantly lower molar enamel volumes at adulthood than they did at weaning.Taken together, these findings demonstrate that OIM mice, like humans with DI, experience accelerated enamel wear during growth.

| CONCLUSIONS
This study demonstrates that OIM mice exhibit a dental phenotype that is consistent with the human DI phenotype.This is the first comparative study of these measurements in OIM mice and illustrates the extent of these phenotypic differences as well as their developmental trajectory during adolescence.These results support the use of the OIM mouse as a model for DI in future investigations.While these results demonstrate that teeth in OIM mice already have significant abnormalities at weaning, differences in developmental trajectories during the prenatal and perinatal periods remain to be characterized.Studying the early development of the DI phenotype is particularly important considering prevalence of feeding difficulties in children with DI.Such studies cannot be carried out in human subjects and will be facilitated by the establishment of the OIM mouse as a model for DI.The changes in dentin microstructure and mineralization previously found in the OIM mouse may also merit further investigation earlier in development.Finally, applying the techniques of geometric morphometrics to the teeth more rigorously would help elucidate the effects of OI on tooth and crown shape, in addition to size and volume.
T A B L E 1 Sample composition for microCT data.Digital 3D models of mandibular and maxillary incisors and first molars of the wild type mouse phenotype, showing (a) whole-tooth segmentations, (b) enamel-only segmentations, and (c) dentin-only segmentations.

F
I G U R E 4 Relative maxillary (a) and mandibular (b) molar volumes (scaled to mandibular centroid size).*p < 0.05.F I G U R E 5 Relative mesial root lengths (mm) in OIM and WT mice (scaled to mandibular centroid size).*p < 0.05.F I G U R E 6 E/D ratios in (a) maxillary and (b) mandibular incisors.*p < 0.05.F I G U R E 7 E/D ratios in (a) maxillary and (b) mandibular molars.*p < 0.05.F I G U R E 8 Relative maxillary (a) and mandibular (b) incisor enamel volumes (scaled to mandibular centroid size).*p < 0.05.

F
I G U R E 9 Raw enamel volumes from the (a) maxillary and (b) mandibular molars, compared with relative enamel volumes from the (c) maxillary and (d) mandibular molars (scaled to mandibular centroid size).*p < 0.05.
T A B L E 3 Means, standard deviations and p-values of tissue volumes, ratios, and lengths.OIM vs. WT p-value (abs) OIM vs. WT p-value (scaled)