Impact of claudin‐10 deficiency on amelogenesis: Lesson from a HELIX tooth

Abstract In epithelia, claudin proteins are important components of the tight junctions as they determine the permeability and specificity to ions of the paracellular pathway. Mutations in cldn10 cause the rare autosomal recessive HELIX syndrome (Hypohidrosis, Electrolyte imbalance, Lacrimal gland dysfunction, Ichthyosis, and Xerostomia), in which patients display severe enamel wear. Here, we assess whether this enamel wear is caused by an innate fragility directly related to claudin‐10 deficiency in addition to xerostomia. A third molar collected from a female HELIX patient was analyzed by a combination of microanatomical and physicochemical approaches (i.e., electron microscopy, elemental mapping, Raman microspectroscopy, and synchrotron‐based X‐ray fluorescence). The enamel morphology, formation time, organization, and microstructure appeared to be within the natural variability. However, we identified accentuated strontium variations within the HELIX enamel, with alternating enrichments and depletions following the direction of the periodical striae of Retzius. These markings were also present in dentin. These data suggest that the enamel wear associated with HELIX may not be related to a disruption of enamel microstructure but rather to xerostomia. However, the occurrence of events of strontium variations within dental tissues might indicate repeated episodes of worsening of the renal dysfunction that may require further investigations.


INTRODUCTION
In mammals, tooth enamel is the most mineralized structure of the organism and forms the outer layer of the dental crown. Amelogenesis is a complex process that occurs before tooth eruption. It results from a complex epithelial-mesenchymal cross-talk between the ectoderm-derived enamel organ and the neural crest-derived dental mesenchyme. [1][2][3][4] Enamel synthesis encompasses two major steps, namely, the secretory stage during which the ameloblasts, the enamelsecreting cells of the enamel organ, secrete a template of enamelspecific extracellular matrix proteins, and the maturation stage during which most of this scaffold is replaced by hydroxyapatite. [5][6][7][8][9] Two types of maturation ameloblasts are reported according to morphological criteria. 9,10 Ruffle-ended ameloblasts exhibit a distal plasma membrane with multiple invaginations, whereas smooth-ended ameloblasts display a smooth distal membrane. These two cell types alternate during the maturation process and the pH of the associated enamel matrix from 6.2 for ruffle-ended ameloblasts to 7.2 for smooth-ended ameloblasts. 9,11,12 At the secretory stage, ameloblasts display a double set of tight junctions (TJs), both at their apical and basal ends. At the maturation stage, smooth-ended ameloblasts remove their apical TJs, whereas they reform in ruffle-ended ameloblasts. 9,10,13 Claudin proteins are the main components of the TJs that are either sealing the paracellular space or forming a pore, thus determining their permeability and ion specificity. 14 The expression of several claudins has been reported in the secretory ameloblast TJs, 15 including claudin-1, -3, -16, and -19, 10,16-18 whereas claudin-16 was not found in maturation ameloblasts. 16 Claudin-10 was shown to be expressed in the enamel organ and more precisely in the stratum intermedium, a layer of epithelial cells located immediately adjacent to the basal end of the ameloblast layer. 19 Two isoforms of claudin-10 are expressed in the kidney, 20 claudin-10a and claudin-10b. The expression of claudin-10a, which is anion-selective, is restricted to the proximal tubule. 21 Claudin-10b, which is cation-selective and may determine paracellular sodium permeability, 22,23 is expressed not only in the thick ascending limb of Henle's loop in the kidney 21 but also in other epithelia. 24,25 Several genetic disorders affect the enamel structure of all the teeth from both dentitions, resulting in Amelogenesis imperfecta manifested by severe dental defects, which require complex restorations and significantly alter a patient's quality of life. 4,5,[26][27][28][29] Among them, nonsyndromic Amelogenesis imperfecta are due to pathogenic variants of genes that encode enamel-specific extracellular matrix proteins (AMELX, ENAM, and AMBN), or proteins involved in enamel maturation (MMP20, KLK4, and SLC24A4), or cell-cell and cell-matrix attachments (ITGB6, COL17A1, LAMA3, and LAMB3). 26,30 Amelogenesis imperfecta is also frequently found in patients with genetic disorders related to kidney, skin, and other organs. 16,26,30,31 A disorder of this sort was recently found to be associated with lossof-function variants of CLDN10, resulting in the autosomal recessive HELIX syndrome characterized by hypohidrosis, electrolyte imbalance, lacrimal gland dysfunction, ichthyosis, and xerostomia (OMIM 617671; prevalence: <1/1,000,000). [32][33][34][35][36][37] In addition, it has been reported that the patients with HELIX syndrome displayed a very early and severe enamel wear. 34 At the time of examination, it was difficult to determine whether this severe enamel wear mainly resulted from erosion due to the impaired salivary secretion, 32 or from enamel fragility directly related to claudin-10 deficiency, as claudin-10 is expressed in the forming tooth germ. 15,19 In the present study, the examination of the enamel of a retained third permanent molar, which was in a submucosal position and, therefore, partially exposed to the oral environment, was used to explore this question. The tooth was removed for orthodontic therapeutic reasons from a young female adult patient with HELIX syndrome. By combining microanatomical and physicochemical approaches, we showed that neither the rate of enamel formation nor its morphology, organization, and structure were significantly impacted by claudin-10 deficiency. However, we identified the occurrence of random events of strontium variations within both enamel and dentin that may reflect a disorder in strontium handling, potentially caused by the renal dysfunction.

Samples
An impacted right permanent third lower molar was collected from a 19-year-old female patient with the HELIX syndrome. This patient was already reported as patient A-IV-2. 34 This family presented a missense variation c.386C >T (NM_182848), p.S129L in claudin-10a (c.392C> T (NM_006984), and p.S131L in claudin-10b. A 3D cone beam computed tomography (CBCT) exam of the lower jaw was performed to prepare the surgery. Since this impacted third molar displayed severely curved roots, the surgeon decided to section it to limit postoperative adverse events. Three impacted age-matched third molars were gathered and were randomly used as control for the various experiments.
All teeth were extracted at the request of an orthodontist in the context of a treatment plan and were collected with the informed consent of the patients, in accordance with the ethical guidelines laid down by French law (agreement IRB 00006477 and n • DC-2009-927, Cellule Bioéthique DGRI/A5). All teeth were fixed in 70% ethanol for a week.

Preparation of the tooth sections
Sections of the crown of the HELIX patient's molar and control third molars were carried out to study the microanatomy of dental tissues by optical and scanning electron microscopy (SEM) and to perform chemical analyses by Raman microspectroscopy, energy-dispersive-Xray (EDX) microanalysis and synchrotron X-ray fluorescence (SXRF) imaging. For the HELIX molar, since the distal half of the crown came as a fragment and was, therefore, more difficult to handle although it preserved both enamel and dentin, the section was performed on that fragment, thus yielding a bucco-distal crown thin section. A buccolingual section through the mesial cusps was performed on the control lower third molars.
For preparation of the thin sections for microanatomy, Raman, and SXRF analyses, the teeth were embedded in cyanoacrylate and fixed with wax on the glass slide. We used a saw equipped with a diamond disk (Struers, Champigny-sur-Marne, France) under a continuous water spray. After the first cut, the surface of the block was polished with carbide grinding paper (Grit 600/P1200) and Chemomet paper with 1 µm aluminum powder (Bühler, Uzwil, Switzerland). This surface was glued onto the slide with Araldite 2020 (Huntsman Corporation, The Woodlands, TX, USA). The block was then sectioned into ∼300 µm slices and polished (Grit 600/P1200) to reach an average thickness of ∼160 µm for the HELIX molar and ∼60 µm for the control third molar. The polishing process was kept minimal for the HELIX tooth because of its smaller size. Finally, the sections were polished again with Chemomet paper with 1 µm aluminum powder until a completely flat surface was obtained.
For preparation of the sections for SEM and EDX analyses, 1 mmthick sections in mirror of the control and HELIX third molars were prepared with a saw equipped with a diamond disk (Struers) under a continuous water spray. For SEM analysis, after thorough polishing, surfaces were cleaned with 5% sodium hypochlorite under ultrasonic activation for 2 min, rinsed twice with distilled water, etched with 36% orthophosphoric acid (DeTrey® Conditioner 36, Dentsply Sirona, York, PA, USA) for 12 s, and then thoroughly rinsed with distilled water. For EDX analysis, after polishing, surfaces were cleaned under ultrasonic activation for 2 min and rinsed twice with distilled water.

Study of the microanatomy of dental tissues
The HELIX crown section was mounted on a glass slide for observation and analysis. The section was analyzed using incident light with a stereomicroscope Leica M8 and transmitted light with a Zeiss Universal photomicroscope. The Zeiss microscope was fitted with an Idea camera connected to a computer using Spot software (Version 5.4).
The images were processed with Nikon ViewN2 and their analysis was performed with ImageJ.
Analysis of the dental microanatomy allows the study of the daily secretion rate (DSR) of the enamel and the formation time of the crown thanks to the presence of periodical growth lines in the enamel, the cross-striations, and the striae of Retzius ( Figure S1). Cross-striations reflect the circadian variation of the enamel secretion, their spacing is indicative of the amount of enamel formed per day and yields the DSR. 38 The striae of Retzius, which correspond to longer successive steps of enamel formation, are formed at regular intervals. Their periodicity is determined by counting the number of cross-striations in between two successive striae. This periodicity is assumed to be constant during the entire crown formation time of a given tooth.
According to the arrangement of Retzius' striae in the enamel, the dental crown can be divided into a cuspal part, located at the occlusal third of the tooth, in which striae are arranged in successive arches around the dentin horn and a lateral part, which is formed subsequently and until crown completion at the cervix 39 ( Figure S1). In the lateral enamel, striae of Retzius crop out and terminate at the surface of the enamel rather than arch over the dentin horn.
The DSR was obtained in the cuspal portion of the crown near the apparent dentin horn ( Figure S1). A line running along the direction of an enamel prism between the enamel-dentin junction (EDJ) and the outer enamel surface (OES) was then divided into 100 µm-thick zones to calculate DSR changes during the course of crown formation. 40,41 In each zone, the average spacing between cross-striations was measured. This was performed several times in each zone, always across a minimum number of three cross-striations, in order to obtain an average DSR for each zone, and finally to calculate an overall average DSR for the cuspal enamel. The total cuspal enamel formation time is equal to the thickness of cuspal enamel divided by the average daily cross-striations spacing.
To describe the development of the lateral enamel, the height of the crown, taken between the cusp tip and the enamel cervix, was divided into deciles of crown height. 42 Noticeably, in the first two deciles, striae were difficult to distinguish so that the formation time was estimated by dividing the length of the prism path between the limit cuspal-lateral enamel and the first striae of the third decile by the DSR of this area.
The number of striae of Retzius was counted within each decile and multiplied by their periodicity. Periodicity was obtained in three locations. The lateral enamel formation time is equal to the total count of Retzius lines multiplied by their periodicity, that is, 8 days ( Figure 1E). The crown formation time was obtained by the sum of cuspal and lateral enamel formation times ( Figure S1).

SEM imaging and EDX analyses
For SEM imaging, thick sections of the HELIX and control third molars were coated with a thin gold layer (∼30 nm) using a CRESSINGTON 108 AUTO gold sputter coater. For EDX measurements, the thick sections were coated with a thin layer of carbon (∼ 20 nm) by evaporation using a CRESSINGTON 208C carbon coater. Imaging and analysis were performed at 15 kV at different magnifications using a Hitachi SU-70 microscope equipped with a Field Emission Gun. All the samples were evaluated for Ca, P, and Mg content (% atom) in the outer and inner layers of enamel, and at least three measurements were performed for each layer.

Raman microspectroscopy
Raman analyses were performed using a Senterra Raman microspec- respectively) was also taken into account in the X-ray mass attenuation coefficients of the hydroxyapatite phase during attenuation correction. 49 Glass slides and kapton foil substrates were included in the overall sample model as appropriate (i.e., background subtraction).
Normalization to the incoming X-ray flux was applied. In the calibrated data, SXRF concentrations are reported by mass fraction (µg.g −1 , i.e., ppm), and/or areal density (g.cm −3 ).
A multiscale scanning strategy was used to optimize efficiency. First, a fast overview scan was acquired at 100 µm (dwell time: 10 ms) to check that the tooth section is well-centered in the field of view, and assess the overall signal of the dental tissues. Then, a middle resolution to better reveal the stress pattern. To note that to denoise the images, a 2D Gauss filter was applied with a kernel size of 0.8 × 0.8.

Variations in Sr content
Since the chronology of the crown formation was established, any variation in Sr concentration could be given a time relative to the initiation of crown formation. Each significant variation (i.e., enrichment or depletion) in Sr content was allocated a letter, their distance from the EDJ was measured, and their chronological order of formation calculated. The timing of the changes in Sr content between the EDJ and the enamel surface along a transect was quantified using the same methodology as used in the cuspal enamel, that is, the cumulative length of the prism between the two reference points was divided by the average DSR of the concerned area.

Physiopathological condition and phenotype of the HELIX patient
The HELIX patient is a 19-year-old French female (patient A-IV-2) born at full term from consanguineous parents, and was raised in France. 34 She displayed xerosis of the skin with keratosis pilaris of cheeks, arms, thighs with a slight palmo-plantar keratoderma, and xerostomia. As  Figure 1A,B). The crown of the third lower molar was fully formed although retained in a submucosal position and partially erupted into the oral cavity and displayed a normal morphology (red arrow, Figure 1A,B). In spite of the relatively low resolution of the CBCT scan (200 µm), the enamel thickness indices of the HELIX molar could be calculated on a virtual 2D section taken through the mesial cusps in the developmental plane ( Figure 1C and Supplementary Information). These indices were within the range of the published values for modern human permanent third molars, 52,53 suggesting that the volume of formed enamel was not disturbed by claudin-10 deficiency. Furthermore, CBCT showed that this third molar displayed an almost completed root formation and exhibited an enamel carious lesion located in the mesial fissure of the occlusal aspect (red arrow, Figure 1A).

Daily secretion rate and crown formation time in the HELIX molar
The average DSR for each 100-µm area in the HELIX molar is presented in Table S1. As shown in Figure 1D, the DSR increased from 1.96 µm/day in the inner zone of enamel near the EDJ to 4.62 µm/day in the outer zone of enamel near the enamel surface. The values and the pattern of the DSR for this HELIX third molar were similar to those reported for any normal human tooth. [54][55][56] The total crown formation time of the HELIX tooth was then established ( Figure 1E). For this tooth, we determined that the formation time of the cuspal enamel was 564 days. The periodicity of the striae of Retzius was 8 days. The number of striae in each decile of the lateral enamel is given in Figure 1E.

Enamel microstructure of the HELIX tooth
We next investigated whether the microstructural characteristics of the enamel were affected by the HELIX syndrome. SEM observation showed that the enamel was correctly organized in HELIX when compared to control (Figure 2A,B). The enamel rods were perfectly formed and aligned in both cases ( Figure 2C-F). Quite remarkably, the cross-striations, which correspond to the circadian variation in enamel apposition, 38 were particularly well-distinguishable in the HELIX rods ( Figure 2E, black arrow-head), which may suggest a potential difference in the enamel content in HELIX. However, we cannot exclude that such variation may result from the sample processing, even if both teeth were prepared by the same operator, at the same time, and in the same conditions. 57

F I G U R E 2 Characteristics of the HELIX enamel by SEM. (A-F) Representative aspects of the HELIX and control enamel microstructure
imaged by SEM at various magnifications (A, B: ×25-30, C, D: ×300, E, F: ×1000, respectively). There is no difference between the HELIX enamel and the control enamel. Note that in the HELIX enamel, cross-striations are particularly well-distinguishable (E, black arrowhead). Abbreviation: EDJ, enamel-dentin junction.

Crystallinity and carbonatation of the mineral phase
We then explored the mineral phase composition and structure of the HELIX dental tissues by Raman microscopy. Control and HELIX enamel samples were analyzed by recording Raman mappings based on the the mineral phase of enamel is a highly crystalline hydroxyapatite with low carbonatation rate, whereas the mineral phase of dentin has a higher degree of substitution and lower crystallinity. 59 No significant difference in the evolution of both FHWM 960 ( Figure 3E) and I 1070 /I 960 ( Figure 3F) values from dentin to enamel could be evidenced in the HELIX sample compared to the control, suggesting the absence of modification of the crystallinity and carbonatation degree of the mineral phase in these two tissues.

Chemical analyses
An investigation on the chemical composition of the enamel was performed to determine if it was altered, which would indicate an abnormal maturation and contribute to explaining the rapid enamel wear observed in all the patients with the HELIX syndrome. 34 Figure S4) or EDX (data not shown).
Next, multielement analysis of the samples was performed using inner enamel is at ∼100 ppm, while the dentin is at ∼220 ppm. In the control sample, although the enamel cap is not fully preserved, Zn values at the OES peak at ∼1.0 × 10 3 ppm on the occlusal aspect of the cusps, while it is slightly less on the surface of the lateral enamel (600-800 ppm). The inner and middle zones of enamel contain ∼60 to ∼100 ppm of Zn, while the dentin is at ∼200 ppm. Zinc enrichment at the OES has been previously described as a normal feature, potentially related to the processes of enamel mineralization and maturation. 60 values ranging from 400 to 500 ppm at the OES, 60 while in Pongo, the outer layer of enamel concentrates 1.5 × 10 3 to 2.0 × 10 3 ppm of Zn. 66 Rautray et al. 67 reported Zn values on human healthy enamel yielding an average of 172.2 ppm, which is within the same order of magnitude as the present middle and inner values measured in HELIX and control.

Strontium
In the control tooth, the Sr distribution ( Figure 4, right lower panel) followed previously published observations. 60  frequency. Both HR SXRF scans at 1.5 µm confirmed that these accentuated Sr markings occurred simultaneously in enamel and dentin ( Figure S5). As the abrupt changes in the content of Sr followed the direction of the striae of Retzius ( Figure 5A), further investigation was performed to determine whether these events take place at specific periods during the crown formation, with no overprinting from subsequent Sr ingestions. 66 Within the ∼1330 µm of cuspal enamel, these several episodes of Sr variation were each calculated to last from 33 to 121 days, ( Figure 5B and Table S2) in Sr concentration as well as their timing in days did not suggest that these changes followed any regular and periodical pattern. These findings prompted us to question the HELIX patient on a particular exposure to strontium at any time of her growth, including the period corresponding to the third molar formation. She did not report either a specific diet or using any specific toothpaste enriched in strontium ions to prevent tooth hypersensitivity.

Other elements
Other elements with Kα lines in the detectable energy range, which include Cu, Fe, Mn, Ti, Cr, S, Cl, Ar, and Rb, did not yield any significant detectable differences between the HELIX and the control tooth in terms of elemental distribution or abundance ( Figure S6).

DISCUSSION
The HELIX syndrome is a very rare disorder (OMIM 617671; Prevalence: <1/1,000,000), which manifests as abnormalities in renal ion homeostasis, resulting in hypermagnesemia, hypocalciuria, and hypokalemia, in epidermal integrity and homeostasis of the ectodermal glands, including salivary glands. We previously reported that all patients with HELIX syndrome displayed early and severe enamel wear, 34 compromising their oral health and particularly their chewing capacity. The present study is based on the observation of a third molar from a patient with HELIX syndrome, which was collected prior to the full eruption of the tooth and, therefore, before the crown was fully exposed to the challenges of the environment of the oral cavity, [68][69][70] and shows that the enamel was correctly formed and displayed normal maturation.
The general pattern of enamel formation appears to be consistent with that described for normal human permanent molars. 54 Specifically, the DSR gradient increase from the EDJ to the surface in the cuspal enamel matches that in normal third permanent molars. The time taken to form lateral enamel is also within the ranges reported for human third permanent molars, 42 and the total enamel formation time is within the ranges reported in the literature. 42,52,71,72 It seems unlikely, therefore, that the HELIX syndrome has any impact on the timing of enamel formation.
The structure of enamel in both HELIX and control samples was studied both at the microscale by SEM and Raman spectroscopy.
Although these techniques were previously shown to clearly characterize other dental disorders, such as Amelogenesis imperfecta 73,74 or X-linked hypophosphatemia, 75  high detection limit (∼1000 ppm), especially for light elements. In contrast, SXRF allows for quantification below 1 ppm and was already applied for multielement analysis of trace elements in many biological tissues, 76 including teeth. 60,66 Here, among the many collected elements, Sr stood out as the only investigated element showing a clear difference in amount and distribution between the two samples. Not only its local concentration could be six times higher in the enamel of the HELIX tooth compared to the control, but it also formed well-defined zones of enrichment and depletion parallel to the striae of Retzius, attesting that this chemical signal has been integrated to the dental tissues during development. However, in contrast to these periodical growth markings, no periodicity could be found in the occurrence of these Sr bands, suggesting that Sr incorporation is not linked to a specific, regular step of amelogenesis. In fact, it is notable that Sr bands are also present synchronously in the dentin, suggesting that Sr presence is related to an overall mineral homeostasis disorder rather than to a local disturbance of amelogenesis.
Dental hard tissues contain trace elements of both dietary and environmental origin. 77 Among them, strontium ions are divalent cations and thus can easily substitute for Ca in the hydroxyapatite structure or interact with the mineral phase. 78 It has been previously reported that the substitution of Ca by Sr in hydroxyapatite alters its solubility. 79 However, this was demonstrated for Sr amount above 1%, 79 which is not the case in the present tooth analysis, where the substitution appears to occur at 0.1%. Therefore, although we cannot fully exclude a modification of apatite solubility with such a low substitution rate, it is expected to be very minor. In vivo, it was found that by oral treatment of increasing Sr dose in rats, it was possible to obtain Sr/(Sr + Ca) molar ratio > 0.01 in femur without altering the bone structure. 80  One of the main functions of the enamel organ during the maturation stage is to transport very large amounts of mineral ions, especially calcium and phosphate, from the blood vessels to the enamel matrix. This critical process is finely controlled by a great number of ion channels, transporters, and exchangers. 10,11,84,85 In the HELIX syndrome, one of the consequences of the renal dysfunction is hypocalciuria. 32 The precise role of Mg in amelogenesis remains not fully understood, yet the expression of the Mg transporter CNNM4 in ameloblast cell membranes at the transition and maturation stages supports that Mg might be removed from the enamel matrix to promote mineralization. 86 It has indeed been consistently reported that Mg content of the enamel is inversely correlated with the extent of mineralization. 87,88 Here, the magnesium content was found to be normal in the HELIX enamel, suggesting that the hypermagnesemia found in the HELIX patients 32 may not have direct consequence on enamel mineralization.
Hypokalemia is a frequent feature of the HELIX patients. 32 Potassium ions have been shown to be important for normal enamel formation as several ion exchangers or cotransporters are K +dependent. 10,11,84,89 For example, the alteration of the K + -dependent Na + /Ca 2+ exchanger isoform 4 (SLC24A4) into mouse maturation ameloblasts by an excess of fluoride in the drinking water was shown to impair amelogenesis. 89 Furthermore, patients with syndromes associated with hypokalemia, such as the Bartter's syndrome, were reported to display Amelogenesis imperfecta. 90,91 It is, therefore, likely that the hypokalemia measured in the patients with the HELIX syndrome may contribute to higher enamel fragility.
Taken together, our data show that enamel formation is not significantly impaired by CLDN10 deficiency, rather designating xerostomia as the main culprit of the enamel wear found in HELIX patients.
However, the abnormal concentrations of Sr measured in the dental mineralized tissues suggest that the tooth mineral content may reflect repeated episodes of worsening of renal dysfunction. It cannot be denied that such changes in Sr concentrations, albeit very low, may alter enamel and dentin solubility. These events urge the need to investigate more teeth from HELIX patients. Serum strontium concentration should also be monitored. Nevertheless, the study of a murine model of the HELIX syndrome may provide further insights on the direct role of claudin-10 in amelogenesis.

ACKNOWLEDGMENTS
The authors do thank David Montero from the Institut des Matéri-