Na⁺/Ca²⁺ exchange in enamel cells is dominated by the K⁺-dependent NCKX exchanger

Calcium (Ca²⁺) is involved in many fundamental cellular processes and its regulation requires precise molecular control. Cytosolic calcium (_cCa²⁺) is maintained low, ∼100 nM, compared with the significantly higher concentration outside the cell (∼2 mM) and, therefore, small changes in _cCa²⁺ have important effects on cell function (Strehler, 1990). However, a rise in _cCa²⁺ must be transient to prevent deleterious effects including mitochondrial overloading, activation of proteases, activation of DNA-fragmenting enzymes, and cell death (Guerini, 1998). Cells not only use Ca²⁺ as a signaling messenger in mineralizing cells, such as the enamel forming ameloblasts, Ca²⁺ is also needed to build the enamel crystals composed of carbonated hydroxyapatite.

Enamel forms in two stages, termed secretory and maturation (Smith, 1998). The enamel space where the enamel crystals grow de novo is adjacent to the apical/distal pole of the cells. This space is surrounded by the ameloblasts and hence these cells exert an important modulatory control on crystal growth (Lacruz, 2017; Lacruz et al., 2017; Nurbaeva et al., 2017). Ameloblasts are tall columnar cells forming a relatively tight sheet with cells ranging in height from ∼60 to 70 μm in the secretory stage to ∼35 μm in the maturation stage and maintaining a 5-μm diameter (Smith, 1998). Ameloblasts undergo cyclic changes from a distal ruffled-ended morphology characterized by membrane infoldings to a smooth-ended appearance devoid of these infoldings. The vast majority of maturation ameloblasts are in the ruffled-ended type (Smith, 1998), and this morphology has been linked with the ion transport mode of the ameloblasts, specifically Ca²⁺ transport (Ashrafi et al., 1989; Reith and Boyde, 1979, 1981). Secretory and ruffled-ended ameloblasts are bound distally by tight junctions sealing the paracellular space. The junctions near the basal/proximal pole are leaky (Sasaki et al., 1982), and because of this arrangement, it is widely considered that ion transport is primarily transcellular (Hubbard, 2000; Lacruz, 2017; Smith, 1998). The majority of Ca²⁺ transport occurs in the maturation stage as enamel crystal growth increases considerably and as protein fragments and water are removed (Smith, 1998). A previous evaluation of the volume of Ca²⁺ acquired during maturation reported that ∼86% is incorporated during that stage (Smith, 1998), which suggests that the molecular components of the Ca²⁺ extrusion system of maturation stage ameloblasts should also be enhanced to efficiently contribute to crystal growth (Lacruz, 2017; Lacruz et al., 2017). Ca²⁺ influx in ameloblasts is largely regulated by the store-operated Ca²⁺ entry (SOCE) pathway (Eckstein et al., 2017, 2019; Lacruz, 2017; Nurbaeva et al., 2015a, 2017). In line with the increased capacity for ion transport during maturation, the molecular components of SOCE and the magnitude of Ca²⁺ influx mediated by SOCE increase considerably during that stage (Aulestia et al., 2020; Nurbaeva et al., 2015a, 2018; Souza Bomfim et al., 2020). However, addressing Ca²⁺ homeostasis in enamel cells following the activation of SOCE remains under investigation. In most mammalian cells, Ca²⁺ extrusion is dominated by the activity of Ca²⁺ pumps and Na⁺/Ca²⁺ exchangers (Carafoli et al., 2001). In enamel cells, the functional role of the plasma membrane Ca²⁺-ATPase (PMCA) pumps was reported recently (Bomfim et al., 2023). PMCAs were found to be functional in both stages, but their role was more substantial when ameloblasts experienced low to moderate elevations in _cCa²⁺, i.e., the secretory stage (Bomfim et al., 2023). However, the functional contribution of the Ca²⁺ exchangers in enamel formation is poorly defined and could be important given the difference in kinetics between the pumps and the exchangers. PMCAs have a high affinity for Ca²⁺ (K_d ∼0.2 μM) but low transport capacity, whereas the Na⁺/Ca²⁺ exchangers have lower Ca²⁺ affinity (K_d 1–20 μM) but have higher transport capacity, estimated to be ∼10 times faster than the PMCAs in tissues such as the heart (Carafoli et al., 2001).

The Na⁺/Ca²⁺ exchangers can be broadly separated into two main families sharing ∼20% homology (Lytton et al., 2002): the Na⁺/Ca²⁺ exchanger (NCX) and the K⁺-dependent Na⁺/Ca²⁺ exchangers (NCKX). The NCX family (coded by the SLC8 gene) came to light in the late 1960s, known primarily from studies in the contractility of the heart. The first member of the family of NCX (NCX1) was cloned in 1990 and since then the expression of NCXs has been described in most cells of higher animals (Blaustein and Lederer, 1999; Philipson and Nicoll, 2000). In the late 1980s, Na⁺/Ca²⁺ exchange in retinal rod photoreceptors was shown to require and transport K⁺ (Cervetto et al., 1989; Schnetkamp et al., 1989) and was later shown to represent the first member of the distinct SLC24 gene family (Reiländer et al., 1992). Both NCX and NCKX are predominantly involved in Ca²⁺ extrusion that requires the energy of the Na⁺ gradient produced by Na⁺ pumps; however, both are also capable of bidirectional exchange, e.g., influx and efflux (Blaustein and Lederer, 1999). In fact, the Ca²⁺ influx mode of operation in which Na⁺ moves out in exchange for the simultaneous influx of Ca²⁺ and K⁺ is widely used to study the function of NCKXs (Altimimi et al., 2013; Jalloul et al., 2016a; Li and Lytton, 2014; Lytton et al., 2002; Paillart et al., 2007; Schnetkamp et al., 1991). Two important features distinguish their function, the coupling ratio of NCX is 3 Na⁺:1 Ca²⁺, whereas NCKX exchanges 4 Na⁺:1 Ca²⁺ + 1 K⁺ (Schnetkamp et al., 1989, 1991). Further, the transport rates for NCXs are faster than those reported for NCKXs (Baazov et al., 1999; Hilgemann, 1996; Jalloul et al., 2016b). In enamel cells, beyond expression data for several of the SLC8 and SLC24 genes or proteins, only one study has previously addressed, by electrophysiology, the functional expression of NCX in ameloblasts (Okumura, et al., 2010). However, that study used ameloblasts derived from young mice which likely prevented the characterization of maturation-stage ameloblasts. Mutations in the SCL8 genes have not been associated with enamel defects to date. In 2012, we first identified the expression of Slc24a4 (coding for NCKX4) in the enamel organ by genome-wide transcript profiling (Lacruz et al., 2012a) and determined its localization in ameloblasts (Hu et al., 2012). Subsequent studies recognized that human mutations in SLC24A4 caused severe enamel defects (Parry et al., 2013; Wang et al., 2014) and that this isoform is considered essential in enamel mineralization (Hu et al., 2012; Lacruz et al., 2012a). More recently, it was reported that the introduction of three of the SLC24A4 mutations causing enamel defects (known as amelogenesis imperfecta) results in single-residue substitutions in the mutant NCKX4 proteins, and, when these mutant NCKX4 proteins were expressed into HEK293 cells, it resulted in the complete abolition of NCKX4-mediated exchanger activity, as measured in the Ca²⁺ influx mode (Jalloul et al., 2016a). These studies suggest that ameloblasts express both NCX and NCKX families of Ca²⁺ exchangers, raising the question about the necessity to have a dual system that is independent of the PMCA pumps.

To address the functional role of NCXs and NCKXs in ameloblasts and to determine whether the kinetics of Ca²⁺ exchange differ between the secretory and maturation stages that define enamel formation, we analyzed rat secretory and maturation stage ameloblasts to quantitate Ca²⁺ exchanger function in the Ca²⁺ influx mode. We first analyzed the expression profiles of the SCL8 and SLC24A genes at each stage and confirmed that SLC24A4 is the most highly upregulated transcript in maturation, as also further validated by Western blot. Ameloblasts of both stages expressed functional Ca²⁺ exchange that was more prominent in maturation. By comparing NCX and NCKX function using K⁺-free solutions, a pharmacological inhibitor of NCX, and ameloblasts of NCKX4-deficient mice, we determined that the contribution of NCKX is greater than that of NCX in ameloblasts. Our results are the first functional evidence that ameloblasts primarily rely on K⁺-dependent Na⁺/Ca²⁺ exchange.

All procedures employed in this study were conducted in accordance with guidelines approved by the Institutional Animal Care and Use Committee (IACUC) of New York University College of Dentistry (protocol # IA16-00625). Experiments were carried out in male and/or female (100 ± 10 g) Sprague–Dawley rats ∼4.5- to 5.0-week-old obtained from Charles River Laboratories. The genetically engineered C57BL/6 mice harboring a homozygous Slc24a4^−/− deletion were kindly provided by Dr. Haiqing Zhao (Stephan et al., 2011).

We used primary enamel organs to dissect ameloblasts because of the limited value of ameloblastcell lines (Klein et al., 2017). Secretory and maturation enamel organ cells were isolated from the rat lower incisors, as has been described in detail following anatomical landmarks in the lower jaw (Lacruz et al., 2012a; Nurbaeva et al., 2018; Smith and Nanci, 1989; Souza Bomfim et al., 2020). Moreover, the rat lower incisors are a well-established system for studying enamel development (Josephsen, 1974; Lacruz et al., 2017; Smith and Nanci, 1989). To obtain single ameloblasts populations, isolated cell clumps from secretory and maturation stages were transferred to Eppendorf tubes containing 450 μl of Hanks’ balanced salt solution (cat #14065-056; Thermo Fisher Scientific) with 1% antibiotic-antimycotic (cat #15240-062; Thermo Fisher Scientific). Cells were digested (0.25 mg/ml cat #414,654; Liberase, Roche) for 35 min at 37°C in a 5%-CO₂ incubator and manually pipetted before stopping the enzymatic reaction by adding 2 ml of X-Vivo 15 (cat #m04418Q; Lonza Bioscience) cell media containing 10% FBS and 1% penicillin–streptomycin. Cells were centrifuged (1,500 × g for 5 min), washed with X-Vivo media, and plated on 25-mm borosilicate cover glass coated with Corning Cell-Tak (cat #CB40240; Thermo Fisher Scientific). Cells were used within 24 h after dissection. Single-cell imaging enabled us to increase the purity of the ameloblast population sampled by eliminating potential contamination of fibroblast/connective tissue cells using a fibroblastic marker, a PE-conjugated monoclonal anti-rat CD90 antibody (cat #202523; Biolegend), and by analyzing gene markers for secretory and maturation stage as we have reported (Bomfim et al., 2023; Lacruz et al., 2012b). In Slc24a4^−/− mice, we did not separate between stages because of the smaller size of the incisors, which resulted in a significantly reduced number of ameloblasts compared with rats, and because the sampling method in mice has not been widely tested (Yin et al., 2014). NCKX4 protein expression in ameloblasts was detected by immunohistochemistry (IHC). Briefly, the lower jaw of 10-d-old mice was isolated and fixed in 4% PFA overnight, washed, embedded in paraffin, and sectioned. Tissue sections were stained with primary rabbit polyclonal anti-SLC24A4 (NCKX4) antibody 1:500 (cat #ab136968; Abcam) and detected by biotinylated anti-goat IgG secondary antibody 1:200 (cat #BA-1000-1.5; Vector Laboratories) and Pierce High Sensitivity NeutrAvidin-HRP 1:500 (cat #31030; Thermo Fisher Scientific). To increase the chromogenic signals and detection sensitivity, a rapid colorimetric detection system using diaminobenzidine (DAB) was used and a semiautomatic quantitative analysis of IHC staining was performed using IHC-ImageJ analysis toolbox.

Total RNA from secretory and maturation enamel organ tissues were isolated using RLT lysis buffer from the RNeasy Mini Kit (cat #74106; Qiagen) with β-mercaptoethanol (1:100), as indicated by the manufacturer, followed by reverse transcription using the iScript cDNA Synthesis Kit (cat #1708891; Bio-Rad). For qRT-PCR, we used the SsoAdvanced Universal SYBR Green Supermix (cat #1725271; Bio-Rad), and experiments were performed in a CFX Connect Optics Module thermocycler (Bio-Rad). β-Actin was used as the housekeeping gene. Primer sequences for relative quantification are shown in Table S1. Expression levels were calculated relative to the secretory stage using the delta CT methods for a general comparison. For Western blot analysis, total protein was extracted from secretory and maturation enamel organ tissue and NCKX4^−/− heart tissue using Pierce RIPA Lysis and Extraction Buffer (cat #89900; Thermo Fisher Scientific) with added 1:100 Halt Protease Inhibitor Cocktail (100×; cat #1861277; Thermo Fisher Scientific) and 1:1,000 phenylmethylsulfonyl fluoride (PMSF). Samples were homogenized manually with a pestle 8–10 times and then sonicated with a VIRTIS VirSonic 300 Ultrasonic Cell Disruptor (10–15% intensity, 10 s on and 10 s off; VIRTIS Co.). Samples were then cleared by centrifugation (15,000 g, 15 min, 4°C). Proteins were quantified using Pierce BCA Protein Assay Kit (cat #23227; Thermo Fisher Scientific). Protein samples were prepared by gentle heating at 37°C for 20 min and 20 μg of protein was equally loaded onto 10% SDS–PAGE resolving gels. Proteins were then transferred to a PVDF membrane and blocked with 5% milk in TBST for 1 h before loading with primary antibody overnight. The following day, membranes were washed three times for 15 min with TBST and loaded with goat anti-rabbit HRP conjugated secondary antibody for 75 min (cat #1705046; Bio-Rad). Membranes were washed three times for 15 min with TBST and imaged with SuperSignal West Pico PLUS Chemiluminescent Substrate (cat #34577; Thermo Fisher Scientific). For loading, control membranes were stripped (cat #21059; Thermo Fisher Scientific), reblocked with 5% milk in TBST for 1 h, and reloaded with a primary antibody overnight. Antibodies against GAPDH 1:1,000 (14C10 rabbit mAb, cat#2118S; Cell Signaling Technology), SLC24A4 (NCKX4) 1:500 (polyclonal, cat #136968; Abcam), and SLC8A3 (NCX3) 1:200 (polyclonal, cat #ANX-013; Alomone Labs) were used.

Cytosolic Ca²⁺ measurements were performed as previously described (Nurbaeva et al., 2015a, 2018; Souza Bomfim et al., 2020). Cells were incubated for 1 h at room temperature with 1 µM of the ratiometric Ca²⁺ probe Fura-2 AM (cat #F1221; Thermo Fisher Scientific). To record the optical membrane potential imaging, preparations of secretory and maturation cells were loaded for 30 min at room temperature with FluoVolt Membrane Potential Kit (cat #F10488; Thermo Fisher Scientific). As indicated by the manufacturer, the fresh FluoVolt loading solution was prepared by adding 10 μl of FluoVolt dye 1,000× and 100 μl of 100× PowerLoad concentrate to 10 ml of regular Ringer’s solution (in mM): 155.0 NaCl; 4.5 KCl; 2.5 CaCl₂; 1.0 MgCl₂; 10 D-glucose; and 10 HEPES, pH 7.4. Cytosolic Na⁺ concentration was determined using the ratiometric Na⁺ probe SBFI-AM cell permeant (cat #S1264; Thermo Fisher Scientific). Secretory and maturation ameloblasts were loaded with 10 µM SBFI-AM plus Pluronic F-127 0.05% wt/vol (cat #P6866; Thermo Fisher Scientific) for 1.5 h at room temperature. SBFI ratios were calibrated to indicate _c[Na⁺] as reported (Chabbert et al., 1995). Briefly, SBFI-loaded ameloblasts were perfused with an ionophore cocktail solution containing nigericin 6 µM (cat #SML1779; Sigma-Aldrich), monensin 6 µM (cat #M5273; Sigma-Aldrich), and gramicidin 2 mg/ml (cat #G5002; Sigma-Aldrich) to produce transmembrane pores and equilibrate internal monovalent cations with external concentrations. Sodium-free, 5, 10, 20, and 50 mM of NaCl (cat #S7653; Sigma-Aldrich) solutions were prepared replacing NaCl with N-methyl-D-glutamine (NMDG; cat #66930; Sigma-Aldrich) and maintaining the osmolarity and equimolar amounts of NMDG and NaCl. Steps to calibrate the SBFI probe fluorescence ratios followed the equations and procedures detailed in Chabbert et al. (1995). Changes in membrane potential or cytosolic Ca²⁺ levels were elicited by different isotonic high KCl concentrations (10, 40, or 75 mM, cat #P9541; Sigma-Aldrich). In the nominally Ca²⁺-free Ringer’s solution, CaCl₂ was omitted and the osmolarity was maintained by adding 3.5 MgCl₂. All reagents were obtained from Sigma-Aldrich. Ca²⁺, Na⁺, or optical membrane potential transients of secretory and maturation cells were measured at room temperature (24 ± 2°C) using a polycarbonate 260 μl chamber (cat #RC21BR; Harvard Bioscience Inc.) mounted on an inverted microscope coupled to a perfusion system electrically controlled. Fluorescence recordings were obtained using the Nikon Ti2-E Eclipse inverted light microscope, equipped with an objective (Nikon S Fluor × 20; numerical aperture: 0.75) and a digital SLR camera (DS-Qi2; Nikon) controlled by computer software (NIS Elements version 5.20.01). Cells were continuously perfused by a six-way perfusion system (VC-8 valve controller) at 5–6 ml per minute with a common outlet 0.28-mm tube driven by controlled valves (Harvard Bioscience Inc.). The Ca²⁺ probe Fura-2 AM and Na⁺ probe SBFI-AM were excited alternatively at 340 and 380 nm using a Lambda LS xenon-arc lamp (Sutter Instrument) and/or pE-340 Fura (Cool Led). Emitted fluorescence was collected through a 510-nm emission filter. Excitation/emission of FluoVolt dye was collected using standard GFP settings. All fluorescence images were generated at 5-s intervals, normalized, and the ratio values were calculated using Image J (1.53J).

To address the possible effect of replacing KCl solutions on Ca²⁺ transients, we monitored the latter in Fura-2 AM–loaded cells during the addition or removal of solutions containing low (10 mM KCl) or high (40 mM KCl) K⁺. Osmolarity was maintained by adjusting the NaCl levels. Exchanger activity was monitored in the Ca²⁺ influx mode elicited by replacing external Na⁺ (122 mM NaCl) with N-methyl-D-glucamine (119.5 mM NMDG) to maintain osmolarity in the presence of external Ca²⁺ (2.5 mM CaCl₂) and 40 mM KCl. To differentiate between NCX and NCKX function, the activity of the exchangers in Ca²⁺ influx mode was elicited by replacing Na⁺ (NaCl) with Li⁺ (LiCl) in the presence of Ca²⁺ (2.5 mM CaCl₂), with or without 10 K⁺ (10 mM KCl), to selectively activate NCKX function. To inhibit the Ca²⁺ influx mode of NCX, we used the blocker KBR7943 (30 min, 5 μM, cat #K4144; Sigma-Aldrich) in a K⁺-free solution.

All data, mathematical analyses, and graphs were analyzed and/or generated using the GraphPad Prism software version 9.5.0 and/or 10.0 (GraphPad Inc.), as previously described (Bomfim et al., 2020; Souza Bomfim et al., 2020; Souza Bomfim et al., 2021). Basal Ca²⁺ and Na⁺ levels were calculated averaging the values from 0 to 60 s of each independent experiment. The Ca²⁺ transients were analyzed by measuring the Ca²⁺ influx mode calculated in each individual trace and fitted by the nonlinear regression curve. The plateau followed by one phase association equation has the following function: Y = IF(X < X0, Y0, Y0 + [Plateau-Y0]*{1 − exp[−K*(X − X0)]}). The delta (∆) Ca²⁺ peak parameter was calculated by the subtraction of maximum fluorescence from the basal/minimum fluorescence intensity, expressed in the same unit of Y-axis values. Data represent the mean ± SEM of the minimum of three independent experiments. Differences between the means of the group data that fit a normal distribution were analyzed using one-way ANOVA followed by Tukey’s multiple comparisons post-hoc test or two-tailed unpaired Student’s t test. The limit of significance was established *P < 0.05; **P < 0.01; ***P < 0.001; and n.s., non-significant, as indicated in each figure legend.

Fig. S1 shows gene expression levels of SLC8 and SLC24A genes in secretory and maturation stage ameloblasts. Fig. S2 shows a Western blot analysis of NCX3 and NCKX4 in secretory and maturation. Fig. S3 shows immunolocalization of NCKX4 in ameloblasts. Fig. S4 shows the calibration of SBFI. Fig. S5 shows changes in cytosolic [Ca²⁺] in response to the removal of extracellular K⁺. Table S1 shows the primer uses to detect gene expression by qRT-PCR.

The relative abundance of the Ca²⁺ exchangers determines their functional relevance in various cell types (Philipson and Nicoll, 2000). The SLC8 gene family includes three members coding for NCX1–3 (Khananshvili, 2013) and five members are recognized in the SLC24 gene family coding for NCKX1–5 (Schnetkamp, 2004). Their expression patterns in enamel cells at the mRNA and protein level had been previously reported by us and others (Hu et al., 2012; Lacruz et al., 2012a, 2012b; Robertson et al., 2017), although results have been contradictory (Lacruz et al., 2012b; Robertson et al., 2017). Therefore, we reanalyzed mRNA expression levels and showed that NCX1 and NCX3 are the most abundant products in ameloblasts with NCX3 being more highly expressed in secretory ameloblasts (Fig. S1 A). NCKX1 and NCKX4 were upregulated in the maturation stage with the latter being the most abundant product showing a ∼50-fold increase (Fig. S1 B). Given that NCX3 and NCKX4 are the most abundant transcripts in each family, we further determined protein expression levels by Western blot. We show that NCXK4 is highly expressed in the maturation stage while NCX3 expression decreased at that stage (Fig. S2). We also confirmed the cellular localization of NCKX4 in the distal pole of maturation stage ameloblasts (Fig. S3) as previously reported, being absent in secretory stage ameloblasts and in cells surrounding the ameloblasts in the enamel organ (Fig. S3; Hu et al., 2012; Jalali et al., 2017; Wang et al., 2014). NCKX4 signals were also notably stronger in ruffled-ended ameloblasts than in smooth-ended cells (Fig. S3). This subcellular localization supports previous models suggesting a critical role of NCKX4 in the extrusion of Ca²⁺ out of the distal region of maturation stage ameloblasts and into the enamel space (Hu et al., 2012; Jalloul et al., 2016a).

The NCX and NCKX families are capable of bidirectional exchange and can function in the Ca²⁺ influx mode if the Na⁺/Ca²⁺ gradients are reversed (Blaustein and Lederer, 1999). Because Na⁺ levels can affect the kinetics of this exchange, we first analyzed intracellular _c[Na⁺] in secretory and maturation stage ameloblasts. Using the ratiometric Na⁺ probe, SBFI-AM calibrated as reported (Chabbert et al., 1995), we showed that secretory and maturation have similar _c[Na⁺]_I, averaging ∼9 mM (Fig. 1 A and Fig. S4), similar to levels found in other cells (Amorino and Fox, 1995; Fleysher et al., 2013). Because the study of NKCX function in the Ca²⁺ influx mode involves the addition and removal of extracellular [K⁺] and because it is well-established that extracellular [K⁺] can alter the resting membrane potential, we next analyzed if the addition of various concentrations of K⁺ in the solutions depolarized the enamel cells and if this could lead to an increase in _cCa²⁺. To investigate membrane depolarization, we loaded the cells with the voltage-sensitive probe FluoVolt, and to address changes in _cCa²⁺, cells were loaded in a separate experiment with Fura-2 AM. In the presence of external Ca²⁺ (2 mM), secretory and maturation stage ameloblasts were exposed to 10, 40, and 75 mM K⁺ (KCl). We show that perfusing the cells with 10 and 40 mM of KCl had no effect on membrane depolarization in either cell type (Fig. 1, C and E). These concentrations of KCl failed to evoke changes in _cCa²⁺ (Fig. 1, D and F). However, the perfusion of 75 mM of KCl depolarized both cell types but without increasing _cCa²⁺ (Fig. 1, G and H). These data indicate that membrane depolarization does not increase _cCa²⁺. Moreover, because the experimental approach used involved the removal of solutions containing KCl, we addressed whether this affected intracellular Ca²⁺. We show that removing K⁺ (10 mM, 40 mM) had no effect on _cCa²⁺ (Fig. S5). A potential caveat in our studies is that the concentration of solutes that ameloblast face at the basal and distal poles in vivo appear to be different (Aoba and Moreno, 1987), which may have a bearing in our in vitro studies where the cells are surrounded by the same solutions.

Next, to address the combined function of the two Na⁺/Ca²⁺ exchanger families in the Ca²⁺ influx mode, we loaded secretory and maturation stage ameloblasts with Fura-2 AM in a Ca²⁺ free Ringer solution containing 140 mM NaCl and 10 mM KCl to allow for a raise of intracellular Na⁺ levels (Kraev et al., 2001). This solution was replaced by a Ca²⁺-containing Ringer solution in the presence of K⁺, but Na⁺ was replaced by NMDG to maintain osmolarity. This experiment shows significantly higher Fura-2 AM fluorescence in maturation stage ameloblasts compared with secretory cells (Fig. 2 A), with the velocity of Ca²⁺ influx and the peak being significantly higher in maturation (Fig. 2, B and C). These results demonstrate that the combined function of NCX and NCKX is more dominant in the maturation stage.

To further dissect the relative contribution of NCX and NCKX function in ameloblasts, we next stimulated secretory and maturation stage ameloblasts in the presence or absence of external K⁺ that is required for NCKX exchange (Cervetto et al., 1989; Schnetkamp et al., 1989). We showed above that perfusing ameloblasts with K⁺ containing bath solutions or removing K⁺ did not cause changes in _cCa²⁺ (Fig. 1 D and Fig. S5) and, therefore, this approach could be used to differentiate between NCKX and NCX function. We perfused the cells with a Ca²⁺-free Ringer solution containing external Na⁺ before replacing the perfusate with Li⁺ to reverse the Na⁺ gradient, which is now high inside the cells (Blaustein and Lederer, 1999; Schnetkamp, 2004). We show that Fura-2 AM fluorescence is significantly higher in secretory and maturation stage ameloblasts in the presence of external K⁺ (Fig. 3, A and C, dark blue and red traces, respectively), required for NCKX activity, compared with the signals generated in the absence of K⁺, which rules out the involvement of NCKX (Fig. 3, A and C, light blue and red, respectively), as quantitated in Fig. 3, B and D. These data highlight an important distinction between NCX and NCKX function.

To further support that the Fura-2 AM signals generated in the K⁺-free bath solution were ascribed only to NCX, we reasoned that perfusing the cells with a specific NCX blocker would further decrease or abolish _cCa²⁺ transients. Therefore, we maintained ameloblasts in a K⁺-free bath solution and treated them with an NCX-specific blocker, KB-R7943 (Iwamoto et al., 2001). This experiment showed that the Fura-2 AM signals were nearly abolished in secretory and maturation stage ameloblasts (Fig. 3, B and D, black traces in both). Additional validation resulted from exposing the cells to a K⁺-containing solution in the presence of KB-R7943, which elicited a decrease in the Fura-2 AM signal (Fig. 3 A and C, gray traces), reflecting the inhibition of NCX but not NCKX. These data confirm the functional participation of NCX in ameloblasts despite the lower mRNA expression levels of the SLC8A genes compared with the SLC24A genes. Based on the data shown in Fig. 3, B and D, we calculated the percentage of Fura-2 AM ratios generated in the presence or absence of K⁺, i.e., with/without NCKX involvement. We show that in both ameloblast stages, NCKX-mediated function was more pronounced compared with that of NCX-mediated function (Fig. 3 E). For example, in the maturation stage, NCKX accounted for 65% of Na⁺/Ca²⁺ exchanger function (Fig. 3 E).

We showed that NCKX-mediated function was more prominent than NCX-mediated function (Fig. 3 E). In addition, expression data indicate that NCKX4 is by far the dominant isoform of all the Na⁺/Ca²⁺ exchangers expressed in ameloblasts (Fig. S1). Therefore, it is expected that genetic deletion of SLC24A4 in ameloblasts would significantly reduce Na⁺/Ca²⁺ exchange. To address this and to determine the contribution of the NCKX4 isoform alone, we obtained the Slc24a4^−/− mice described in Stephan et al. (2011), kindly shared by Dr. H. Zhao. The isolation of secretory and maturation stage ameloblasts in mice is more difficult than in rats because of the smaller size of the incisors in mice and because the delineation of secretory and maturation regions is not as well defined. Therefore, instead of separating cells by stage, we pooled secretory and maturation stage ameloblasts, as we have previously done in other KO mouse models (Eckstein et al., 2017, 2019), to gain an overall picture of the changes in Na⁺/Ca²⁺ exchange in the absence of NCKX4 alone. We show that loss of Slc24a4 in ameloblasts reduced Na⁺/Ca²⁺ exchange by ∼62% (Fig. 4, A and B). This percentage is similar to the reduction in Na⁺/Ca²⁺ exchange shown in Fig. 3 E (65%), confirming that the bulk of NCKX-mediated function in ameloblasts is driven by NCKX4.

We have analyzed the expression and function of members of the Na⁺/Ca²⁺ exchanger families NCX and NCKX in rat ameloblasts. By far the most highly expressed gene in ameloblasts representing both exchanger families is Slc24a4 (coding for NCKX4; Figs. S1 and S2). Immunolocalization of NCKX4 in mouse molars and incisors showed that in the enamel organ, this protein is only found in ameloblasts, predominantly in the distal pole of maturation-stage ameloblasts near the enamel space where enamel crystals form (Fig. S3), a similar localization to what we and others have previously reported (Bronckers et al., 2015; Eckstein et al., 2017; Hu et al., 2012; Wang et al., 2014). A previous study had shown that NCX3, the most highly expressed gene of the NCX family in enamel cells (Fig. S1 A), was localized at the distal pole of the ameloblasts (Okumura et al., 2010). This subcellular localization of NCKX4 and NCX3 supports the notion that these exchangers participate in the extrusion of Ca²⁺ distally, providing the necessary Ca²⁺ required for crystal growth in the extracellular space. Because of the similar localization of both NCX3 and NCKX4, we next investigated protein expression levels by Western blot and showed that while NCX3 is relatively abundant in the secretory stage, its expression declines in the maturation stage (Fig. S2). Conversely, the expression of NCKX4 is difficult to detect in the secretory stage and vastly increases in maturation, the stage when the Ca²⁺ transport machinery is upregulated as crystals experience significant growth in thickness (Costiniti et al., 2021; Hubbard, 2000; Lacruz, 2017; Lacruz et al., 2017; Smith, 1998).

Next, to address whether the NCX and NCKX exchangers are in fact functional proteins in enamel cells, we isolated rat ameloblasts from the secretory and maturation stages. The isolation of single cells has been a very successful approach to investigating Ca²⁺ dynamics in ameloblasts, including the analysis of SOCE (Aulestia et al., 2020; Eckstein et al., 2017, 2019; Nurbaeva et al., 2018; Souza Bomfim et al., 2020), the modulation of cytosolic Ca²⁺ by mitochondria (Costiniti et al., 2022) as well as PMCA function (Bomfim et al., 2023). Although the isolated ameloblast preparations do not appear to maintain the morphology of the tall columnar, polarized cells in situ, they do retain the properties of the membrane channels they express, such as their ability to sustain Ca²⁺ influx via the ORAI1 channel, which is expressed at the basolateral membrane (Eckstein and Lacruz, 2018), as we have shown (Eckstein et al., 2017, 2019; Nurbaeva et al., 2015a). Thus, we aimed to determine if ameloblasts would also retain properties associated with Na⁺/Ca²⁺ exchange.

NCX and NCKX primarily mediate Ca²⁺ removal out of cells based on the Na⁺ gradient, the Ca²⁺ gradient, and the membrane potential (Philipson and Nicoll, 2000). However, this can be reversed in physiological conditions during membrane depolarization, although the physiological role of the Ca²⁺ influx mode is unclear (Philipson and Nicoll, 2000). In our study, to determine Na⁺/Ca²⁺ exchange, we analyzed Fura-2 AM fluorescence generated byboth exchangers in the Ca²⁺ influx mode by modifying ionic concentrations in the bath solutions as this is a commonly used method validated by several groups (Cervetto et al., 1989; Jalloul et al., 2016a, 2016b; Kraev et al., 2001; Lytton et al., 2002; Schnetkamp, 1986; Schnetkamp et al., 1991). First, we found that perfusing ameloblasts with a K⁺-containing bath solution or removing this perfusate had no effect on Fura-2 AM fluorescence in secretory or in maturation stage ameloblasts (Fig. 1 and Fig. S5). This is indicative of a lack of an apparent voltage-gated Ca²⁺ channel function in these epithelial cells. Our analysis showed that both secretory and maturation stage ameloblasts are capable of Na⁺/Ca²⁺ exchange, but maturation stage ameloblasts showed higher capacity (Fig. 2).

To further dissect possible differences in the contributions made by NCX and NCKX, we analyzed the Ca²⁺ influx mode in the absence of K⁺, to eliminate NCKX function, and in the presence of the NCX blocker KBR7943, to eliminate NCX function. In the absence of K⁺, there was a reduction of ~65% in Na⁺/Ca²⁺ exchange function in both stages, suggesting that the contribution of NCKX function is greater than NCX (Fig. 3). In conditions when the cells were exposed to no K⁺ in the presence of KBR7943, the Fura-2 AM signals were almost completely abolished (Fig. 3). Further, we also observed that ameloblasts of Scl24a4^−/− mice showed a reduction in _cCa²⁺ of ∼66% confirming the primacy of NKCX4 isoform in ameloblasts. This likely explains why patients with mutations in SLC24A4 suffer from severe enamel hypomineralization, clinically referred to as amelogenesis imperfecta (Parry et al., 2013; Wang et al., 2014), while to date, no mutations in the NCX coding genes have been linked to enamel defects. It is interesting to note that, based on the severe effects caused by NCKX4 mutations, it would appear that neither PMCA nor NCX3 can replace NCKX4 in proper ameloblast function.

Na⁺/Ca²⁺ exchange operates independently of the PMCA Ca²⁺ pumps. However, many cells express all three systems including the ameloblasts (Bomfim et al., 2023), posing the question of why this replication in Ca²⁺ clearance mechanisms might be required. A distinction in the Ca²⁺ clearance dynamics has been suggested as an obvious cause (Blaustein and Lederer, 1999). PMCA pumps have a high Ca²⁺ affinity but a Ca²⁺ extrusion capacity of ∼150 Ca²⁺/s, significantly lower than the exchangers (Brini et al., 2012; Cali et al., 2017; Baazov et al., 1999; Hilgemann, 1996; Jalloul et al., 2016b). We recently demonstrated that PMCA function in ameloblasts was dependent on _cCa²⁺ availability whereby the pumps are more dominant at low levels of _cCa²⁺ elevation and were more effective during the secretory stage (Bomfim et al., 2023). While data on ameloblasts and other cells reinforce the notion that PMCA is efficient at clearing moderate levels of _cCa²⁺, the requirements for utilizing two families of Na⁺/Ca²⁺ exchangers remains an interesting research area.

Both NCKX4 and members of the NCX family are expressed at the distal pole of the ameloblasts (Hu et al., 2012; Okumura et al., 2010; Wang et al., 2014), the site that would be most efficient to extrude Ca²⁺ out of the ameloblasts into the enamel space. However, whether these exchangers cooperate synergistically to extrude Ca²⁺ is unknown. Although NCKX dominates Ca²⁺ extrusion function in ameloblasts, our data show that NCX also contributes to this task. This is evidenced by the presence of Fura-2 AM signals in ameloblasts of Slc24a4^−/− mice or in rat ameloblasts after removal of external K⁺, indicative of a K⁺-independent exchanger, i.e., NCX. Perhaps the lack of reports on enamel defects associated with loss of NCX function might indicate redundancy of the NCX isoforms, whereas the expression of the NCKX4 isoform in maturation stage ameloblasts is so dominant that it prevents functional compensation by other members of the NCKX family.

A question that remains is why the contribution of NCKX appears critical when both NCKX4 and NCX3 are expressed in the ameloblasts and contribute to the Ca²⁺ fluxes described here. There are differences in Ca²⁺ kinetics and transport stoichiometry between these two exchanger families. Recently, the rate of transport of members of the NCKX family was described by one of us showing that NCKX4 had transport rates ∼560 Ca²⁺/s when expressed in HEK-293 cells (Jalloul et al., 2016b), significantly higher than PMCAs but lower than NCXs, which has been reported to be ∼5,000 Ca²⁺/s (Juhaszova et al., 2000). But why do ameloblasts predominantly use NCKX4 for _cCa²⁺ clearance rather than the NCX proteins? We suggest that this is linked to the cellular demands of the ameloblasts. For instance, NCXs are well known for their role in cardiac excitation–contraction (Eisner et al., 2017; Ottolia et al., 2013), responding to the action of voltage-gated Ca²⁺ channels capable of generating very short inward fluxes of several tens of thousands of Ca²⁺/s (Catterall, 2011). By contrast, Ca²⁺ influx in ameloblasts relies on the SOCE mediated by the ORAI channels (Eckstein et al., 2019; Nurbaeva et al., 2015a, 2015b, 2018), whose unitary conductance is several orders of magnitude lower than the voltage-gated channels but results in more sustained Ca²⁺ fluxes (Prakriya and Lewis, 2015). This may be more like the well-understood role of NCKX1 in retinal rod photoreceptors where a sustained influx of Ca²⁺ entering the rod outer segment via the cGMP-gated channels in darkness under conditions of a depolarized membrane potential is extruded via NCKX1 as discussed (Cervetto et al., 1989; Schnetkamp et al., 1989, 1991) or via an NCKX2:NCKX4-mediated response as reported more recently for cone photoreceptors (Vinberg et al., 2017). Because enamel crystallites precipitate from a supersaturated solution in which the degree of saturation of the enamel fluid is regulated, with Ca²⁺ playing a central role in the thermodynamic properties of the process (Simmer and Fincham, 1995), we suggest that the sustained Ca²⁺ fluxes of relatively low magnitude by SOCE may have been coopted for with the slower extrusion rates by NCKX4 to meet the cellular demands of ameloblasts facilitating the precipitation of ions in a manner that supports the thermodynamic regulation of enamel mineralization. However, we cannot discount the possibility that the equilibrium potential of [Na⁺] and [K⁺] in the distal membrane of the maturation stage ameloblast may influence the energy required to extrude Ca²⁺ into the extracellular fluid such that the stoichiometry of NCKX4 provides the required energy to drive this exchange. However, the ionic concentrations in the extracellular space in the maturation stage are unknown, although we suspect that [Ca²⁺] should be higher than in secretory stage enamel fluid given the overall increase in Ca²⁺ influx and extrusion in the former.

Although the coupling ratio of the exchangers is the same in the influx and efflux modes (Blaustein and Lederer, 1999), a potential caveat in our interpretation is our assumption that the Ca²⁺ influx rates of NCX and NCKX are similar to the outward rates. While this could be true in ameloblasts, it is reported that Ca²⁺ binding in NCX is asymmetrical in the internal and external surfaces of the exchanger (Almagor et al., 2014; Blaustein and Lederer, 1999; Giladi et al., 2016). Studies in pig cardiomyocytes, squid and rat neurons, and all excitable cells, there are documented differences between the apparent inward- and outward-facing Ca²⁺ binding sites (DiPolo and Beaugé, 1988; Fontana et al., 1995; Miura and Kimura, 1989). Despite this, the dominant expression of NCKX4 in ameloblasts and the fact that this exchanger is essential for enamel mineralization (Wang et al., 2014) lends support to our observation that Ca²⁺ extrusion in ameloblasts is largely regulated by this exchanger. Therefore, it is plausible that the differences in the inward Ca²⁺ kinetics we observed between NCX and NCKX in these cells may reflect apparent differences in their physiological roles.

Data shown here contributes to our understanding of the Ca²⁺ signaling toolkit of ameloblasts and how this system contributes to enamel mineralization. Previous research highlighted that the bulk mineralization of the enamel crystals takes place during the maturation stage as mineral uptake rates and Ca²⁺ acquisition rates increase during that stage (Smith, 1998; Smith et al., 2005). Recent evidence indicates that the expression and function of the molecular components of SOCE are upregulated during the maturation stage (Aulestia et al., 2020; Nurbaeva et al., 2015a, 2018; Souza Bomfim et al., 2020). The Ca²⁺ extrusion toolkit of most cells includes the PMCA and Na⁺/Ca²⁺ exchange (Brini and Carafoli, 2011; Cali et al., 2017). In ameloblasts, PMCA function operates in both secretory and maturation stages but dominates Ca²⁺ extrusion during the secretory stage when Ca²⁺ requirements are low (Bomfim et al., 2023). By contrast, maturation stage ameloblasts have higher Na⁺/Ca²⁺ exchanger activity, with NCKX4 as the predominant exchanger function. Lack of data on mutations involving PMCA or NCX coding genes resulting in enamel disease contrasts with the severe hypomineralization defects in patients and mice with defective NCKX4 (Parry et al., 2013; Wang et al., 2014). Therefore, functional studies support the dominance of the K⁺-dependent NCKX proteins as important contributors to ameloblast Ca²⁺ extrusion and hence to enamel mineralization.

The data that support the findings of this study are available in the methods and/or supplementary material of this article.

David A. Eisner served as editor.

This work was funded by the National Institute for Dental and Craniofacial Research grants DE025639 and DE027679 to R.S. Lacruz.

Author contributions: G.H. Souza Bomfim, P.P.M. Schnetkamp, and R.S. Lacruz designed the research studies. G.H. Souza Bomfim and E. Mitaishvili performed the experiments. G.H. Souza Bomfim, E. Mitaishvili, P.P.M. Schnetkamp, and R.S. Lacruz analyzed the data. G.H. Souza Bomfim, P.P.M. Schnetkamp, and R.S. Lacruz wrote the paper. All authors contributed to the final preparation of the manuscript.