Developmental expression of SLC26A4 (Pendrin) during amelogenesis in developing rodent teeth

Abstract

Ameloblasts need to regulate pH during formation of enamel crystals, a process that generates protons. Solute carrier family 26A member 4 (SLC26A4, or pendrin) is an anion exchanger for chloride, bicarbonate, iodine and formate. It is expressed in apical membranes of ion-transporting epithelia in kidney, inner ear and thyroid where it regulates luminal pH and fluid transport. We hypothesized that maturation ameloblasts express SLC26A4 to neutralize acidification of enamel fluid in forming enamel. In rodents, secretory and maturation ameloblasts were immunopositive for SLC26A4. Staining was particularly strong in apical membranes of maturation ameloblasts facing forming enamel. RT-PCR confirmed the presence of mRNA transcripts for Slc26a4 in enamel organs. SLC26A4 immunostaining was also found in mineralizing connective tissues including odontoblasts, osteoblasts, osteocytes, osteoclasts, bone lining cells, cellular cementoblasts and cementocytes. However, Slc26a4-null mutant mice had no overt dental phenotype. The presence of SLC26A4 in apical plasma membranes of maturation ameloblasts is consistent with a potential function as pH regulator. SLC26A4 does not appear critical for ameloblast functioning and is likely compensated by other pH regulators.

During the secretory stage of amelogenesis ameloblasts of the enamel organ epithelium deposit a protein-rich enamel matrix that acts as temporary scaffold to initiate and foster crystal growth. This matrix is gradually degraded and removed from the enamel space in the more advanced maturation stage (1). The thin long hydroxyapatite crystals that had formed in secretion stage will further expand their dimensions and grow out in width and thickness until completion of mineralization.

Formation and growth of hydroxyapatite crystals in the enamel space generates large numbers of protons. Without neutralization these protons will impair further crystal growth (1–3). A number of workers therefore suggested that ameloblasts also regulate pH in the fluid of the forming enamel (1–3). During secretory stage – when mineral accretion is still relatively low-the high proton-binding capacity of amelogenins is one of potential mechanisms to buffer the protons (1, 4). However, when the mineral accretion rate progressively increases during maturation stage (5) and amelogenins are removed from the enamel space, other buffering mechanisms are needed to complete crystal growth.

In recent years there is increasing evidence that maturation ameloblasts express a set of membrane-spanning and cytosolic proteins that are typical for ion-transporting and pH regulating epithelial cells found in thyroid, in non-acid secreting cells of the collecting ducts of kidney and pancreatic ducts. These molecules include several types of carbonic anhydrase (3, 6–9), the cystic fibrosis transmembrane conductance regulator (CFTR) (10–12), sodium independent anion exchanger AE2 (Solute Carrier 4A2 or SLC4A2) (9, 13, 14), and the sodium-proton exchanger (NHE1) (9), while the adjacent papillary layer expresses the electrogenic sodium bicarbonate cotransporter (NBCE1 or SLC4A4) (9, 15–17), the electroneutral sodium bicarbonate cotransporter NBCn1 and sodium-potassium-ATPase (9). Collectively, these data strongly suggests that maturation ameloblasts (along with the adjacent papillary layer) actively regulate pH by generating and secrete bicarbonates into the enamel space but the mechanism is not clear.

SLC26A4 or pendrin is an anion exchanger that transports Cl⁻, iodide, formate, OH⁻ or HCO₃⁻ across plasma membranes of ion-transporting polarized epithelia. SLC26A4 belongs to the Slc26A gene family which contains 10 different members. Pendrin is the protein product of the SLC26A4 gene (formerly called PENDRED or PDS) which is associated with thyroid goiter and congenital hearing loss (18–20). The human SLC26A4 gene codes for a 780 amino acid protein with 11 to 15 predicted transmembrane domains (18, 21, 22). SLC26A4 protein expression has been reported in apical membranes of subsets cells in thyroid (23), cells of the endolymphatic compartment of the inner ear (24) and non-acid secreting intercalated (IC) cells of the renal cortical collecting ducts (25, 26). Targeted disruption of Slc26a4 impairs iodine secretion by thyroid epithelium in vitro (21), bicarbonate secretion by renal intercalated epithelium (25), native thyroid epithelium (27), by epithelial tissues of cochlea and the vestibular labyrinth (20, 27–29). These effects result in goiter (23), hearing loss and anatomical changes in the bony structure of the inner ear (20, 24, 27–32).

In view of these reports we examined whether SLC26A4 is expressed in maturation ameloblasts, where it might be involved in pH regulation during enamel crystal formation.

Materials and methods

Animals and cells

Tissues for histology were collected from mouse, hamster and rat pups (3–14-d-old) and adult mice and rats. Erupted teeth of 11–18-d-old and 9 months old Slc26a4 null mutant mice and heterozygous littermates were used to examine gross anatomical changes. Slc26a4 null mutant and heterozygous littermates of 10–12-d-old pups served to evaluate tooth development histologically and immunohistochemically. Detailed data on raising these null mutant mice have been published elsewhere (27, 30, Fig. 1.). At the Vrije Universiteit Amsterdam, animals were euthanized by intraperitoneal injection with nembutal (125 mg/kg body weight) in compliance with National and International Standards and approved by the institutional review Board for Animal Care at the Vrije Universiteit Amsterdam. At Kansas State University, the null mutants and littermate mice were anesthetized with tri-bromo-ethanol (560 mg/kg) and euthanized by decapitation prior to tissue harvest. All procedures concerning animals were approved by the Institutional Animal Care and Use Committee at Kansas State University. Immortalized mouse ameloblast-like cell line LS8 was kindly provided by Dr. Malcom Snead (UCLA, Los Angeles, CA, USA).

Fig. 1. Domains of mouse SLC26A4 and position of the amino acid sequences used for raising antibodies.

(structure based on: NP_035997.1). The gene contains 20 exons (1st exon is non- coding) coding for a protein of 780 amino acids. The grey bar (amino acids 110–673) represents the transmembrane part of the molecule with putatively 11–15 transmembrane domains (22). The black box within this domain labeled ST is a putative sulphate transporter region. The yellow box in the ST domain represents the neocassette inserted in exon 8 (between 5^th and 6^th transmembrane domain) to functionally inactivate the gene (30). The light blue bars indicate position of the peptide sequences to which antibodies were raised, one at the N-terminal and two C-terminal end. The green bar (“PCR”) represents the part of the molecule for which primers were designed to measure the mRNA transcripts by RT-PCR (coding for amino acid 749–780).

Histological procedures

Skinned heads and kidney samples were fixed by immersion in 5% paraformaldehyde in 0.1 M phosphate buffer pH 7.3 and embedded in paraffin. Heads of rodents older than 12 d were first decalcified in 4% EDTA, pH 7.3 for 2 wk at 4°C. Sagittal serial sections 5–7 μm thick were prepared and mounted on polylysine coated glass slides.

Antibodies

Three different antibodies to SLC26A4 were purchased, each recognizing a different part of the molecule (Fig. 1). The first one was an affinity purified rabbit antibody raised to a 17 amino acid long synthetic fragment located at the N-terminal end of the human SLC26A4 (Abcam, Cambridge, UK; catalogue number 66702; Fig. 1). The antibodies were raised against a region of the protein’s N-terminus ‘within the first 100 amino acids’ (information given by the manufacturer). The second antibody was a non-purified mouse antiserum raised against a partial GST-tagged recombinant protein (amino acids 674–755) at the C-terminal end of the human SLC26A4 (Abnova, Heidelberg, Germany; catalogue number H0005172-A01). The third antibody was an affinity purified rabbit antibody raised to a large fragment (amino acids 586–780) at the C-terminal end from human SLC26A4 (Santa Cruz Biotechnology, Tebu-Bio, Heerhugowaard, The Netherlands, catalogue SC-50346, H-195). The N-terminal fragment was located upstream from the site where the neocassette was inserted to disrupt Slc26a4 whereas both C-terminal end fragments were downstream from it (Fig. 1). As negative controls were used normal rabbit control IgG (Universal rabbit negative control, Dakocytomation catalogue N1699, Dakopatt, Glostrup, Denmark) or normal mouse serum (Dakocytomation catalogue X0910) applied in the same concentrations or dilutions as the primary antibodies.

Immunohistochemistry

Paraffin sections were dewaxed in xylene and rehydrated in a descending series of ethanol and rinsed in phosphate buffered saline (PBS). Antigen retrieval was performed in either 10 mM citrate buffer (pH 6.0) at 60 °C overnight or for 20 min in microwave at 95 °C prior to staining, or by a mild predigestion with a proteinase K solution (10 μg/ml; in phosphate buffered saline) for 15 min at 37°C. After retrieval endogenous peroxidase was blocked with a peroxidase block solution (Envision kit, Dakocytomation) for 5 min. Sections were washed in 0.1 M Tris-buffered, 0.9 % NaCl pH 7.2 (TBS) containing 0.1% bovine serum albumin (BSA). Non-specific staining was blocked for 30 min with 2% BSA or 30% normal goat or normal rabbit serum. Next sections were incubated overnight at 4 °C with primary antibodies or matched non-immune IgG or serum (dilution of 1:100–1:200 for mouse antisera, and final concentration of 1 and 2 μg/ml for the affinity purified rabbit antibodies). To reduce high background staining when mouse primary antisera were used on mouse tissues the sections were treated with blocking reagents from a mouse on mouse blocking kit (Histomouse, Beat blocking kit, Invitrogen, Bleiswijk, The Netherlands) before incubation with primary antibodies. After overnight incubation at 4°C with primary antibodies sections were washed three times in TBS and incubated for 1 h at room temperature with goat-anti-rabbit IgG antibodies conjugated with peroxidase polymer or with rabbit antimouse IgG conjugated with peroxidase polymer (Envision kits). After washing the peroxidase conjugates were visualized with DAB substrate (Envision kit, for a brown end–product) or AEC substrate (Invitrogen, for a red end-product) for 10 min at room temperature according to manufacturer instructions, and counterstained with haematoxylin.

RNA isolation and qRT-PCR analysis

Total RNA was extracted from mouse enamel organ, dental pulp, kidney and the ameloblast cell line LS8 using either Trizol (Invitrogen) or the NucleoSpin RNA/protein kit (Macherey-Nagel, Düren, Germany) according to the manufacturer‘s instructions. First strand cDNA synthesis was performed in a 20 μl reverse transcription reaction containing 200 ng of total RNA using VILO kit (Invitrogen) according to the manufacturer‘s instructions. Real-time PCR analysis was performed to analyse mRNA expression of Slc26a4 (mouse-specific primers located in exon 19 and 20 corresponding to last 31 amino acids) and the house keeping gene tyrosine 3-monooxygenase (Ywhaz) with the primers sequences shown in Table 1 by using the LightCycler 480 system based on SYBR Green I dye (Roche Applied Science, Indianapolis, IN, USA). The LightCycler reactions were prepared in 20 μl total volume with 7μl PCR-H2O, 1μl forward and reverse primers (final concentration 1 pmol/μl of each primer), 10 μl LightCycler Mastermix (LightCycler 480 SYBR Green I Master; Roche Applied Science, IN, USA), to which 2 μl of 5 times diluted cDNA was added as PCR template. Controls in the real-time RT-PCR reaction included RT reactions without the reverse transcriptase (control for DNA carry over) and RT reactions without template (control for reagent contamination). With the Light Cycler software, the crossing points were assessed from a standard curve of 5 serial dilutions ranging from 10, 2, 0.4, 0.08 or 0.016 ng of a reference cDNA. PCR efficiency (E) was automatically calculated using the fit point method (E = 10–1/slope). Gene expression data were used only if the PCR efficiency was within a 1.85–2.0 range. Expression of Slc26a4 mRNA transcripts in mouse ameloblasts and odontoblasts was normalized for Ywhaz housekeeping gene. The normalized values were then related to the Slc26a4 transcript values obtained in mouse kidney. Quantified data were obtained from at least 6 mice from two separate studies.

Table 1.

Primer sequences used for real time PCR (mouse specific)

Target gene	primer	Accession number	Oligonucleotide sequence	Anneal temp °C	Size bp
Slc26a4	Forw2466	NM_011867.3	5′GACTGTAAAGACCCTCTTGATCTGA 3′	63	90
	Rev2555		5′ GGAAGCAAGTCTACGCATGG 3′
Ywhaz	Forw673	NM_003406	5′ GATGAAGCCATTGCTGAACTTG 3′	56	229
	Rev901		5′ CTATTTGTGGGACAGCATGGA 3′

^*Ywhaz = house keeping protein tyrosine 3-monooxygenase

Results

Validation of antibody specificity

Preliminary studies indicated that antigen retrieval was required to obtain positive staining for SLC26A4. After antigen retrieval all three antibodies stained a selected set of tubular duct cells in mouse renal cortex (Fig. 2a,b). The staining was concentrated at the apical plasma membrane of the cells facing the ductal lumen (Fig. 2a,b) in line with published data (25). In Slc26a4 null mutant mouse kidney both antibodies raised to the C-terminal end of the molecule failed to stain these cells as reported (25). In contrast the antibody reacting with the N-terminal end of the molecule gave a weak positive staining in these cells which was not confined to the apical membrane but distributed over the cytoplasm (Fig. 2c). this staining was interpreted as binding of the antibodies to a truncated, soluble form of the SLC26A4 consisting of the N- terminal part upstream from the site of insertion of the neomycin cassette.

Fig. 2. Validation of antibodies to SLC26A4 in kidney (Fig. 2a–c) and immunostaining for SLC26A4 during amelogenesis in developing teeth (Fig. 2d–i).

Fig. 2a,b: Mouse heterozygous tubular ducts of renal cortex stained for SLC26A4 with the Santa Cruz antibody to C-terminal end (Fig. 2a, 400x and inset x1000) or Abnova antibody to C-terminal end (Fig. 2b; x1000, AEC product). Note the apical staining near the lumen (arrow). The Abcam antibody gave an identical staining (not shown). Fig. 2c (x1000): Renal tubular ducts from null mutant mouse stained with Abcam antibody to N-terminal end. Some cells show a weak diffuse staining (arrow) but never apical staining. Fig. 2d (x100): Developing molars in a heterozygous wild type mouse (11–d-old) with first (left) and less developed second molar (right) stained with anti-N-terminal (Abcam) antibody. Positive intracellular staining is obtained in polarizing preameloblasts (second molar), secretory ameloblasts (SA) and strong apical staining (arrow) in maturation ameloblast (MA). Fig. 2e (x1000) Intracellular staining in 4- d- old hamster molar secretory ameloblasts (SA)(Abcam antibody). Fig. 2f. (x1000) Secretory ameloblasts of a mouse incisor. Staining accumulates in a diffuse band near apical end of the cell but Tomes processes (TP) are negative (arrows) (antibody Abnova). Arrow heads show some supranuclear staining. Fig. 2g (x1000): Mouse molar and (Fig. 2h. x 1000) mouse incisor maturation ameloblasts stained strongly in apical membranes (Abcam antibody). Fig. 2i. (x200): Mouse molar from a null mutant stained with C-terminal antibody (Santa Cruz). Od odontoblasts, EM enamel matrix, PL papillary layer, si stratum intermedium

Localization of SLC26A4 in developing enamel organ

All three antibodies reacted with all three rodent species and gave the same tissue distribution but their staining intensity and background varied. Most consistent strong staining was obtained with the antibody to the N-terminal end. Antibodies to the C-terminal ends gave weaker staining of which the mouse antiserum gave some overall background, particularly on mouse tissues despite attempts to block with a mouse on mouse blocking kit.

The first positive but weak staining was noted in presecretory ameloblasts and intensity increased as cells elongated and polarized (Fig. 2d). In fully secretory ameloblasts positive stained vesicles were located supranuclearly; their number and size increased towards the distal part of the cells. A positive intracellular staining concentrated in a band near the proximal part of the Tomes processes which itself were negative (Fig. 2e,f). Also the surface layer of secretory stage enamel matrix stained positive with varying degrees of intensity.

Maturation ameloblasts stained intracellularly and a prominent strong to very strong band of staining was located over the apical plasma membrane facing the enamel space (Fig. 2d,g,h). Extracellular maturation stage enamel matrix often stained but variably and either across the entire enamel width or the enamel near the dentin. Replacing primary antibodies for non-immune antibodies or staining null mutant tissues with anti-SLC26A4 antibodies did not give staining in cells (Fig. 2i). However, also some batches of non-immune IgG or normal control sera could give variable staining in the extracellular matrix.

Localization of SLC264 in mineralizing connective tissues

Odontoblasts (Fig. 2d; Fig. 3a–c), bone cells (osteoblasts, osteocytes, lining cells, osteoclasts, Fig. 3e–g), cementoblasts and cementocytes (Fig. 3h), developing periodontal ligament cells (Fig. 3h), and cells of the dental follicle (not shown) were all positive for SLC26A4. In odontoblasts, the staining was initially seen as fine material throughout the entire cytoplasma, with slightly more pronounced staining in supranuclear area (Fig. 3a). In more mature cells, staining was located intracellularly in the proximal part of the processes running into the predentin, but not in the distal part of the process located in the mineralized dentin (Fig. 3a,b,d). In fully differentiated odontoblasts the staining formed a thin band in the ‘shoulder’ of adjacent cells, over the distal plasma membrane of the cell body near the base of the process (Fig. 3a,b). Cross-and oblique sections through the odontoblast layer suggested that immunostaining in the proximal part of the process and the distal cell body was also located in the plasma membranes (Fig. 3c). The distal part of the processes in the older (deeper layers of) dentin were not immunostained (Fig 3d.).

Fig. 3. Immunostaining for SLC26A4 in mineralizing connective tissues.

Fig. 3a,b: In odontoblasts (Od) strong staining is located in supranuclear area (white asterisk), in the proximal part of the process (arrows heads) running in the predentin (Pd)(incisor, adult mouse). Fig. 3b: In more advanced stage staining forms a band in the distal membrane of the odontoblast cell body (arrows, hamster molar, 5 d-old). Fig. 3c,d (incisor adult mouse; x1000): In obliquely cut sections staining is located in the membrane (Fig. 3c, arrows) and in the proximal part of the process near the predentin (Pd) surface. Fig. 3d: Proximal part of the process is positive (arrows on the right side), whereas distal part of the processes are negative in deeper layers of dentin (De) (arrows on the left side). Fig. 3e–g. Alveolar bone (x1000): Positive staining in osteoblasts (Ob), osteocytes (Oc) (Fig 3e), and osteoclasts (Ocl; Fig. 3f.)(hamster, 8-d), seen as fine particulate material. Fig. 3g. Positive bone lining cells (Blc) in adult mouse jaw bone. Fig. 3h. (5 wk old mouse; x 400). Staining of cells in and along cellular cementum (Cc) and developing periodontal ligament cells (Pdl). All stainings with N-terminal antibody (Abcam).

RT-PCR for Slc26a4 mRNA transcripts

PCR analysis using mouse-specific primers confirmed that mRNA transcripts were detectable in enamel organ, dental pulp, jaw tissue, kidney and the murine ameloblast-like cell line LS8 (Fig. 4). The mRNA expression level in mouse enamel organ and mouse pulp was 16% ± 9% and 30% ± 15% (p≤ 0.05, n=6), respectively, of the level in mouse kidney, an established site of expression.

Fig. 4.

mRNA- transcripts for Slc26a4 in various mouse tissues detected by RT- PCR. Primary ameloblasts and pulp cells were harvested from two different mice indicated by different numbers.

Anatomy and histology of dental tissues of Slc26a4 null mutant mice

In null mutant mice teeth erupted normally as seen by visual inspection at postnatal day 11 and day 18. Enamel of null mutant mouse incisors (up to 9 months of age) was yellow orange as in wild type and heterozygous mice. Visual inspection of null mutant incisors at 10x magnification showed a smooth surface without any roughness, pitting or surface changes.

Histological analysis of sagittal sections of incisors and developing molars from 10–12-d-old null mutant mice did not reveal overt structural changes in any of the developing dental tissues or surrounding jaw bone tissues.

Discussion

SLC26A4 in maturation ameloblasts

We found a strong positive immunostaining for SLC26A4/pendrin in the apical membranes of maturation ameloblasts facing the enamel space. Each of the three different antibodies was directed to a different part of SLC26A4 but regardless which one was used the staining pattern was the same. The specificity of both antibodies reacting with the C-terminal portion was validated on mouse null mutant tissues. The antibody to the N-terminal end gave the same staining pattern in wild type mice as both validated C-terminal antibodies; hence the weak staining of this antibody in null mutant tissues was attributed to its binding to a truncated form of the protein upstream from the mutation site. We furthermore could detect Slc26a4 mRNA transcripts in enamel organs of wild type mice. Collectively, the data show that wild type maturation ameloblasts transcribe and translate Slc26a4 and incorporate the translated protein into their apical plasma membranes.

The location of SLC26A4 in apical membranes of maturation ameloblasts is consistent with the concept that this exchanger could be involved in pH regulation by secreting bicarbonates to neutralize protons released into the enamel space during crystal growth. Paine and coworkers speculated that it is AE2 in apical plasma membranes of maturation ameloblasts that mediates bicarbonate secretion (15, 16). However, several groups including ours could not localize AE2 in apical membranes but in basolateral membranes of maturation ameloblasts (9, 13, 14). This suggests that in apical membranes other ion exchangers will operate to secrete bicarbonate into enamel space. SLC26A4 is a possible candidate to fulfill such function.

We noticed positive staining in the maturation stage enamel in areas where enamel matrix was still present. Matrix staining was occasionally also seen in control sections with using normal sera or non-immune IgG. Whether this matrix staining is due to non-specific binding to residual crystals or represents SLC26A4 or its immunoreactive fragments released into the extracellular space is unclear at present.

The absence of changes in dental structures in Slc26a4 null mutants indicates that pendrin is not critical for enamel mineralization. Ameloblasts seem to function normally in the null mutants and the enamel appears to mineralize to completion. After targeted disruption of Slc26a4 in mice also several other tissues that normally express this protein were reported not to show any major changes, e.g. there were no signs of systemic hypothyroidism (27, 30) or overt changes in renal function or structure in vivo (30). However, when challenged in vitro or in vivo, renal function in these null mutants appear to be different from wild types. Slc26a4-deficient mice secrete less bicarbonate in isolated perfused cortical collecting ducts during metabolic acidosis (25). In addition, stimulation by aldosteron analogues, which causes hypertension and weight gain in wild type mice, does not affect null mutants (22). Hence, in ameloblasts of wild type mice (as in renal tubular epithelium) the role of SLC26A4 is likely limited, and its function redundant, possibly to become only apparent in Slc26a4 deficient mice after challenge.

In Fig. 5 we present a simplified hypothetic model how SLC26A4 might function in maturation stage ameloblasts as part of the pH regulatory machinery, similar as in pancreatic tubular epithelium (33, 34). Pendrin along with another (yet unknown) exchanger transports cytosolic bicarbonate across the apical membrane in exchange for Cl⁻ from the enamel fluid. This Cl⁻ in enamel fluid is in turn replenished by efflux of Cl⁻ through CFTR. High intracellular levels of Cl⁻ are sustained by influx of Cl⁻ at the basolateral membranes mediated by basolateral AE2 in exchange for outward directed bicarbonate (9, 13, 14).

Fig. 5. Simplified working model for bicarbonate transport into the enamel space by maturation ameloblasts.

In green indicated are the pH regulator-molecules identified and located immunohistochemically by various groups (9, 12, 13, 15, 36). SLC26A4/Pendrin (PD) is located in apical membrane and exchanges bicarbonate for chloride. The apical membrane may also contain another yet unknow exchanger (indicated by ?) that compensates for PD. Gap junctions (grey) between papillary layer and ameloblasts (9) might enable transfer of bicarbonates from papillary layer into ameloblasts, assuming that NBCe1 mediates influx of bicarbonates into papillary layer. Apical black bars: tight junctions, NHE1: Na- H-exchanger-1; PD pendrin; AE2: anion exchanger-2; NBCe1: electrogenic sodium bicarbonate cotransporter e-1; CF Cystic Fibrosis Transmembrane Conductance Regulator; CAR carbonic anhydrase. Not drawn for sake of clarity: vacuolar-H-ATPase in apical membrane of maturation ameloblasts (9), Na-K-ATPase and the electroneutral sodium bicarbonate cotransporter NBCn1 in papillary layer (9).

SLC26A4 in secretory ameloblasts

SLC26A4 is also present in secretory ameloblasts and concentrates in the cytosol of the distal part of the ameloblast cell body near the terminal web. Its function in secretory ameloblasts is unknown. As the immunostaining indicates an association with intracellular vesicles, SLC26A4 may have other, intracellular functions. It perhaps functions as anion exchanger for Cl⁻ across membranes of transport-or resorption vesicles during secretion or resorption of matrix.

SLC26A4 in mineralizing connective tissues

Connective tissue cells that form mineralizing tissues such as dentin, bone and cementum are likely exposed to pH changes and all these cell types appear to express SLC26A4. Odontoblasts strongly express carbonic anhydrase enzyme activity (6, 35), and immunostain for SLC26A4 (this study), AE2 (12), CFTR (14) and NBCe1 (17). This suggests that also odontoblasts have the potential to produce and secrete bicarbonates. Odontoblasts deposit a collagen-rich unmineralized matrix, predentin, that shortly later quite abruptly turns into mineralized dentin by formation and deposition of apatite crystals at the predentin-dentin junction. Dentin in deeper layers will acquire far less mineral per time unit than when dentin is formed at the mineralization front at the predentin-dentin junction. Conceivably the major crystal growth and hence the major potential release of protons will happen at the mineralization front. This local acidification might be perceived by the distal plasma membrane of the cell body and the membranes of the proximal part of the odontoblast process which could explain the staining for SLC26A4 in these structures. Absence of SLC26A4 however does not influence dentinogenesis and suggests that pendrin is not essential for odontoblast function.

Bone cells also expressed SLC26A4 and structural changes have been reported in the bony structure of the inner ear in Slc26a4 null mutant mice (18–20). Intramembranous and enchondral ossification and mineralization around the inner ear of Slc26a4 null mice are delayed by 1–4 d, but not in phalangeal and metatarsal bones (27). We found no changes in the structure of the jaw bones in the null mutant mice. This is in agreement with the view that the effects on intramembranous and endochondral ossification in Slc26a4 mutants are not systemic (by less well functioning thyroid) or due to absence of SLC26A4 in bone cells, but due to local effects around the expanding inner ear (27). Local changes in pH in or pressure by endolymphatic fluid perceived by null mutant epithelial cells of the inner ear may be transferred to connective tissue which may influence bone formation around these sites (27–29).

In short, apical membranes of ameloblasts highly express pendrin during amelogenesis potentially reflecting part of the machinery to regulate pH during crystal formation. SLC26A4 may also have similar but redundant functions in formation of other mineralizing tissues including dentin, bone and dental cementum.