Extrarenal Signs of Proximal Renal Tubular Acidosis Persist in Nonacidemic Nbce1b/c-Null Mice

Significance Statement

Recessive SLC4A4 mutations are a cause of proximal renal tubular acidosis (pRTA), a rare but devastating disorder associated with loss of electrogenic sodium bicarbonate cotransporter 1 (NBCe1) function in kidney and other organs. Alkali therapy is the preferred treatment for pRTA, which is characterized by acidemia, developmental impairment, and vision loss, and often enamel hypomineralization. However, which nonrenal findings are secondary to acidemia is poorly understood. The authors describe the phenotype of a line of transgenic mice in which NBCe1 expression is blocked in all tissues except the proximal tubule. These mice are not acidemic but still exhibit many of the extrarenal signs associated with pRTA, revealing the potential limitations of pH correction by alkali therapy in pRTA and the need to develop novel therapies.

Visual Abstract

Abstract

Background

The SLC4A4 gene encodes electrogenic sodium bicarbonate cotransporter 1 (NBCe1). Inheritance of recessive mutations in SLC4A4 causes proximal renal tubular acidosis (pRTA), a disease characterized by metabolic acidosis, growth retardation, ocular abnormalities, and often dental abnormalities. Mouse models of pRTA exhibit acidemia, corneal edema, weak dental enamel, impacted colons, nutritional defects, and a general failure to thrive, rarely surviving beyond weaning. Alkali therapy remains the preferred treatment for pRTA, but it is unclear which nonrenal signs are secondary to acidemia and which are a direct consequence of NBCe1 loss from nonrenal sites (such as the eye and enamel organ) and therefore require separate therapy. SLC4A4 encodes three major NBCe1 variants: NBCe1-A, NBCe1-B, and NBCe1-C. NBCe1-A is expressed in proximal tubule epithelia; its dysfunction causes the plasma bicarbonate insufficiency that underlies acidemia. NBCe1-B and NBCe1-C exhibit a broad extra-proximal-tubular distribution.

Methods

To explore the consequences of Nbce1b/c loss in the absence of acidemia, we engineered a novel strain of Nbce1b/c-null mice and assessed them for signs of pRTA.

Results

Nbce1b/c-null mice have normal blood pH, but exhibit increased mortality, growth retardation, corneal edema, and tooth enamel defects.

Conclusions

The correction of pRTA-related acidemia should not be considered a panacea for all signs of pRTA. The phenotype of Nbce1b/c-null mice highlights the physiologic importance of NBCe1 variants expressed beyond the proximal tubular epithelia and potential limitations of pH correction by alkali therapy in pRTA. It also suggests a novel genetic locus for corneal dystrophy and enamel hypomineralization without acidemia.

Electrogenic sodium bicarbonate cotransporter 1 (NBCe1; encoded by the SLC4A4 gene) is the archetypal and the most extensively studied of the five Na⁺-coupled bicarbonate transporters of the SLC4 solute carrier family.^1,2 SLC4A4 expresses five NBCe1 variants, designated A, B, C, D, and E, from two distinct promoters (Figure 1A).^3,4 Each is composed of an approximately 400 amino-acid cytosolic amino-terminal domain (Nt), 14 transmembrane spans, and an approximately 100 amino-acid cytosolic carboxy-terminal sequence (Ct). NBCe1-A is expressed in the basolateral membranes of renal proximal tubule epithelia and is a critical part of the mechanism that supplies blood plasma with newly generated and reabsorbed HCO₃⁻ to maintain whole-body pH.^5,6 NBCe1-B and NBCe1-C (hereafter collectively referred to as NBCe1-B/C) are expressed in a wide variety of epithelial and excitable cells where they import HCO₃⁻, either to maintain intracellular pH (e.g., in neurons) or to support transepithelial anion/fluid secretion (e.g., in gastrointestinal epithelia).^7–9 NBCe1-A and NBCe1-B/C differ not only in their distribution but also in their Nt and Ct sequences; NBCe1-A includes a unique 41 amino-acid autostimulatory domain (encoded by the orange section of exon 4 in Figure 1A), whereas NBCe1-B/C include in its place an 85 amino-acid autoinhibitory domain (encoded by exons 2 and 3; blue in in Figure 1A).¹⁰ The consequence of these alternative Nt sequences is that NBCe1-A exhibits a greater conductance than NBCe1-B/C. However, under the influence of the autoinhibitory domain–binding, secretagogue-activated protein IRBIT, NBCe1-B/C can achieve a similar magnitude of conductance to NBCe1-A.^10–12 The consequence of the Ct differences between NBCe1-B and NBCe1-C (46 amino-acid domain in NBCe1-B is replaced by a different 61 amino-acid domain in NBCe1-C) have yet to be fully understood. Of the two variants, NBCe1-B appears to exhibit a broader distribution than NBCe1-C, with expression of the latter being predominantly limited to neurons and glia.^8,13 NBCe1-D and NBCe1-E are a subset of NBCe1-A and NBCe1-B transcripts with apparently overlapping expression profiles that, because of alternative splicing, omit a nine amino-acid sequence from the amino-terminal domain of the cognate protein.¹⁴ Little is known of their unique physiologic contribution, but it is to be noted that any disruption of NBCe1-A should also disrupt NBCe1-D and any disruption of NBCe1-B should also disrupt NBCe1-E.

Figure 1.

Nbce1b/c is specifically disrupted in Nbce1b/c-null (KO_b/c) mice. (A) Cartoon depicting the structure of the Slc4a4 gene (not to scale) and its major products Nbce1a, Nbce1b, and Nbce1c (minor products Nbce1d and e are produced by choice of an alternate 5′ splice site within exon 6 of a and b).¹⁴ Numbered gray boxes are exons. Exons that are included in each gene product are reproduced below the gene cartoon: Noncoding exons are colorless, exons common to all variants are colored dark gray, exons encoding sequence unique to individual exons are colored orange, blue, or red. Exons 7–22 are included in all variants and are omitted to highlight the variable regions. (B) Genomic DNA and encoded protein sequence around the 5 bp deletion in the described strain of Nbce1b/c-null mouse. (C) Example of a genotyping gel showing the shift in molecular weight of the amplicon caused by the 5 bp deletion. (D) Western blots showing the effect of the 5 bp deletion upon the abundance of Nbce1 in lysates prepared from the kidneys, brains, and hearts of Nbce1b/c-null mice. Each blot is representative of results obtained from lysates prepared from three pairs of wild-type versus KO_b/c littermates and probed with the anti-Slc4a4 antibody. Four further paired kidney lysates were probed with the anti–NBCe1-A/B antibody αrb1NBC (a gift from Walter Boron at Case Western Reserve University⁸) with similar results. WT, wild-type.

Recessive inheritance of mutations in SLC4A4 causes the rare but systematically devastating disease proximal renal tubular acidosis (pRTA).^15,16 All of the 15 affected individuals who have been described to date exhibit acidemia (low plasma [HCO₃⁻] due to a renal reabsorption defect), growth retardation, and one or more of the following ocular pathologies: band keratopathy, cataracts, and glaucoma.^15,17–28 The other signs and sequelae of pRTA exhibit variable penetrance and include intellectual impairment, dental abnormalities, and a variety of other neurologic and neuromuscular features including migraine, muscle weakness, and bilateral calcification of the basal ganglia (reviewed in Parker and Boron¹). Nbce1a/b/c-null mice model several aspects of the human disease, namely acidemia, growth retardation, and dental abnormalities, as well as several other signs that have not been observed with human pRTA, such as corneal edema, impacted intestines, and greater mortality.^24,29,30 No strain of Nbce1-null mice has been described to exhibit any of the neurologic signs that are common to individuals with pRTA.

The cause of the acidemia in pRTA is logically considered to be NBCe1-A loss, but it is unclear which of the other signs or sequelae of pRTA are secondary to acidemia, specifically due to NBCe1-B/C loss, or are the combined results of acidemia and NBCe1-B/C loss. The current treatment for pRTA is the same as for any form of metabolic acidosis: alkali administration.^31–33 Although this approach, if well administered, can correct acidemia, it would not be expected to ameliorate those signs that follow NBCe1-B/C loss.

We have generated a novel mouse model to discover the consequence of Nbce1b/c loss in the absence of acidemia. This mouse was designed to model pRTA with a perfectly administered alkali therapy and highlight those signs that require additional therapies.

A preliminary report of this work has appeared in abstract form.³⁴

Methods

Ethical Statement

All procedures involving mice were approved by and performed in accordance with the rules and recommendations of the Institutional Animal Care and Use Committee of the University at Buffalo.

Generation of Nbce1b/c-Null Mice

Slc4a4-targeted founder mice were generated by the Gene Targeting and Transgenic Shared Resource at the Roswell Park Comprehensive Cancer Center (Buffalo, NY) using CRISPR/Cas9 technology with a guide RNA (5′-GGACCCCAATGTAGATCGTGNGG-3′) that targets the antisense strand of the second coding exon of Slc4a4. The guide was designed, tested, and generated by the Genome Engineering and iPSC Center at Washington University in St. Louis (St. Louis, MO). A female C57BL/6J mosaic mouse that harbored a 5-nt deletion in the targeted exon was generated via pronuclear microinjection. The female was then bred with a wild-type male C57BL/6J mouse (Jackson Laboratory, Bar Harbor, ME) to confirm germline transmission. Heterozygous pups from this union were designated as generation F1 and thenceforth backcrossed onto the C57BL/6J genetic background. All experimental animals were produced from heterozygote×heterozygote crosses using animals backcrossed to F₃–F₁₀ generation.

Genotyping

Genomic DNA was prepared from mouse tail snips using a DNeasy Blood and Tissue Kit (QIAGEN Inc., Germantown, MD) and used as a template for PCR. PCR primers (5′-CATGGGTTGAAATGACCGTTG-3′ forward, 5′- CCTTGTGCCCAGCCTTCCTC-3′ reverse) flanking the site of the deletion were used to amplify a 125-bp product for wild-type mice or a 120-bp product for Nbce1b/c-null mice. The products were resolved by electrophoresis on 6% TBE-Urea gels (Thermo Fisher Scientific, Waltham, MA) alongside a Track-It 10bp ladder (Thermo Fisher Scientific; Figure 1C). Initially we confirmed the identity of these PCR products and the presence of the 5-nt deletion by subcloning into a TOPO vector (Thermo) followed by DNA sequencing (Eurofins Genomics, Louisville, KY: data not shown).

Blood Gas Measurements

Blood drawn by cardiac puncture from isoflurane anesthetized mice (5%, inhaled) was immediately transferred into a blood gas electrolyte and metabolite test card and analyzed using an epoc reader according to the manufacturer’s instructions (both Siemens Medical Solutions USA Inc., Malvern, PA).

Western Blotting

Organs were homogenized in 1 ml of ice-cold homogenization buffer (100 mM NaCl, 25 mM HEPES, 250 mM sucrose, pH 7.4) plus cOmplete Protease Inhibitor Cocktail (Sigma-Aldrich, St. Louis, MO). Then, 5 µg/lane of organ lysate was resolved on a 3%–8% Tris-Acetate gel and transferred onto a PVDF membrane (Thermo Fisher Scientific). The PVDF was incubated overnight in TBS containing 0.1% Tween-20 and 5% milk powder and probed using an anti-SLC4A4 rabbit polyclonal antibody ESAP14635 (#E-AB-14348; Elabscience Biotechnology Inc., Houston, TX; characterized in this study by single band of appropriate molecular weight by Western blotting, as well as loss of immunoreactivity from Nbce1b/c mice) followed by an HRP-conjugated goat–anti-rabbit secondary antibody (MP Biomedicals, Solon, OH). Immunoreactive protein bands were disclosed using ECL2 reagent and the chemiluminescent signal was imaged using a myECL imager (Thermo Fisher Scientific). Densitometry was performed using FiJi software.^35,36 Further details are provided in the Supplemental Material.

Bone Measurement

Mice were euthanized by isoflurane overdose followed by cervical dislocation, and then vacuum-sealed into plastic bags. The sealed bags were incubated in a 50°C water bath overnight and the limb bones were removed and defleshed using a mascara brush. The bones were air-dried overnight and their lengths measured using a dissecting microscope equipped with an OMAX digital camera and Toupview software (Amscope.com).

Immunofluorescence

Sections of paraformaldehyde-fixed eyes were mounted on slides and blocked with Rodent Block M (Biocare Medical, Pacheco, CA). NBCe1 immunoreactivity was probed using the anti-NBCe1 antibody ESAP14635 (Elabscience Biotechnology Inc.) followed by an Alex488-conjugated goat–anti-rabbit secondary antibody A-11034 (Thermo Fisher Scientific). The labeled tissue was fixed beneath a coverslip using Vectashield mounting medium with DAPI (Vector Laboratories, Burlingame, CA). Images were acquired using an Axio Imager 2 fluorescence microscope (Carl Zeiss Inc., Thornwood, NY). Further details are provided in the Supplemental Material.

Confocal Reflection Microscopy

A freshly enucleated eye was held in a putty ring containing mineral oil such that the cornea was apposed to the cover glass of a glass-bottomed microwell dish (MatTek Corps., Ashlan, MA). z-stacked images (0.24 µm spacing) were obtained using a White Light Laser Confocal Microscope (TCS SP8; Leica Microsystems, Wetzlar, Germany) with a 488 nm laser emission and a band-pass filter set at 480–496 nm to capture reflected light. From z-stack projections, the stroma was selected as an area of interest and pixel intensity profiles were generated from grayscale images using FiJi.³⁵ Density (pixel intensity) medians and variances were calculated using Microsoft Excel 1997 using the formulae SUMPRODUCT ([“pixel intensity” array], [“number of pixels” array])/SUM([“number of pixels” array] for medians and SUMPRODUCT (([“pixel intensity” array]−“calculated median”)^2, [“number of pixels” array])/SUM([“number of pixels” array]−1) for variance.

Pachymetry

A drop of proparacaine hydrochloride anesthetic solution (Akorn, Lake Forest, IL) was applied to each eye of a scruffed mouse before the measurement of in vivo corneal thickness using a handheld Pachette-3 ultrasonic pachymeter with flap option (DGH Technology Inc., Exton, PA). The probe was applied for several seconds perpendicular to the center of the cornea to record the mean of 25 consecutive measurements.

Scanning Electron Microscopy and Energy-Dispersive X-Ray Spectroscopy

Extracted lower incisors were air dried, embedded in acrylic resin, and polished to reveal their cross section in the transverse plane. The polished surface was coated with evaporated carbon and the tissue imaged using a Hitachi SU70 field emission SEM operated at 20 KeV. The images collected were backscattered electron images in which the pixel intensities are related to atomic number, in this case showing differences in mineral density in the tissues. EDS spectra were collected with beam energy of 20 KeV and live collection time of 60 seconds. Pixel intensities from electron micrographs were analyzed in the same way as the corneal confocal images.

Statistical Analyses

Statistical analyses were performed in Microsoft Excel 1997 (paired and unpaired t tests assuming equal variances) or MiniTab version 17 (ANOVA with Tukey post hoc analysis). Paired t tests were performed in instances where data were taken from sex-matched littermates.

Results

Expression of Nbce1 Protein in Wild-Type versus Nbce1b/c-Null Mice

The genome of the Nbce1b/c-null mouse strain reported in this study (hereafter referred to as “KO_b/c”) lacks 5 bp in exon 3 of Slc4a4 (Figure 1, B and C), a frameshifting deletion that is predicted to result in the expression of the prematurely truncated Nbce1b/c polypeptide p.His27Aspfs40X. Figure 1D shows Western blots of lysates produced from wild-type (WT) versus KO_b/c littermate organs showing that NBCe1 immunoreactivity (the approximately 240 kD band is consistent with the predicted molecular weight of an NBCe1 dimer) persists in the (predominantly NBCe1a-expressing) kidneys of Nbce1b/c-null mice, but is lost from the other (predominantly Nbce1b/c-expressing) organs such as the brain and heart. We detected no significant difference in the intensity of Nbce1 immunoreactivity in kidney lysates from WT versus KO_b/c mice; the WT normalized NBCe1-immunofluorescence signal intensity was 108±13% in KO_b/c kidney preparations (n=7; P=0.96; result of a paired, two-tailed t test assuming equal variance).

Blood Gas and Chemistry Measurements from KO_b/c Mice

Table 1 shows the results of measurements of blood parameters among WT, heterozygous (+/−), and KO_b/c mice. None of the measured parameters was significantly different between wild-type and +/− mice (each parameter was individually assessed by ANOVA among the groups). Although KO_b/c mice have normal blood pH and no deficit in plasma [HCO₃⁻], they exhibit an increased pCO₂ (+7 mm Hg) with increased [HCO₃⁻] (+3 mmol/L) and decreased pO₂ (−20 mm Hg), which are signs of a chronic (compensated) respiratory acidosis.

Table 1.

Blood analysis data

Parameter	WT	+/−	KOb/c	Unit
Blood gas
pH	7.36±0.01	7.35±0.01	7.33±0.01
pCO2	36.4±2.3	36.5±2.2	44.3±2.0a	mm Hg
pO2	65.8±7.7	70.5±6.1	46.2±4.8a	mm Hg
HCO3− (calculated)	20.2±1.0	20.1±1.0	23.2±0.6a	mmol/L
Blood chemistry
Na+	143±1	142±1	142±1	mmol/L
K+	4.85±0.27	4.69±0.14	4.54±0.22	mmol/L
Ca2+	1.25±0.02	1.26±0.02	1.29±0.02	mmol/L
Cl−	111±1	111±1	108±1	mmol/L
Blood metabolites
Glu	284±11	271±6	283±16	mg/dL
Lac	4.85±0.47	5.78±0.39	5.92±0.49	mg/dL

Blood gas, chemistry, and metabolite data from wild-type (n=11), heterozygous (n=13), and Nbce1b/c-null (n=11) littermate mice. Note that these data correspond to mixed arterial venous blood collected by cardiac puncture from anesthetized mice.

^aSignificantly different from WT by ANOVA with Tukey post hoc analysis.

KO_b/c Mice Have Increased Mortality

We examined KO_b/c mice for the increased mortality that is exhibited by Nbce1a/b/c-null mice. The average size of litters born to heterozygous couples was 6±1 pups (n=38). Statistically fewer KO_b/c pups were born to each pair than WT pups (P=0.01; one-tailed, paired t test). Of 232 pups, 63 were WT, 120 were +/−, and 45 were KO_b/c; a Mendelian deficit that indicates that as many as 25% of Nbce1b/c-null mice could have died pre/perinatally. Figure 2A tracks the postnatal mortality of 16 KO_b/c mice. A total of 50% of this cohort of KO_b/c mice died within 6 weeks of birth. The cause of death is undetermined. The oldest KO_b/c mouse in this cohort was euthanized at 67 days. All eight WT and ten +/− littermates included in this study survived to 70 days or longer before being euthanized. Even at 6 months, there was no mortality among a separate cohort of eight heterozygous mice (F₆–F₉ generation) that were kept for breeding purposes (data not shown).

Figure 2.

Nbce1-b/c-null (KO_b/c) mice exhibit reduced survival and impaired growth. (A) Mortality of 16 KO_b/c mice. Eight wild-type (WT) and ten +/− littermates served as controls and survived beyond the 70 day period. (B) The weight of WT and KO_b/c mice at weaning (21 days) and at 28 days. Males and females are plotted on separate axes. Mean weights±SEM at each time point are shown as filled circles and the growth trajectories of individual mice are represented as lines. Statistics are results of unpaired, one-tailed t tests between WT and KO_b/c data at 21 or 28 days. (C) Representative photographs of limb bones from six WT mice and sex-matched KO_b/c littermates (four male wild-type versus KO_b/c pairs, two female wild-type versus KO_b/c pairs, 20–43 days old). Scale bar, 5 mm. (D) The length of the six KO_b/c bones normalized to their WT counterparts are shown in the graph. Statistics are results of paired, one-tailed t tests between prenormalized WT and KO_b/c data for each bone type.

KO_b/c Mice Exhibit Lower than Normal Weight and Smaller Limb Bones

We examined KO_b/c mice for signs of failure to thrive that are exhibited by Nbce1a/b/c-null mice. KO_b/c mice are noticeably smaller than their WT littermates. Two pieces of data speak to a growth defect in KO_b/c mice. Figure 2B shows the weight of WT and KO_b/c mice at weaning (i.e., day 21) and at 1 week after weaning (i.e., day 28: 7 days on solid chow.) At both time points, male and female KO_b/c mice weigh less on average than their WT counterparts. For both males and females, this difference is significant at 21 days and 28 days (one-tailed, unpaired t test with Bonferroni correction for multiple comparisons, setting significance at P=0.025=0.05/2). The difference at 28 days is exacerbated in males by a significantly smaller increase in weight for KO_b/c mice over the 7 day interval (+4 g versus +6 g; P<0.01, two-tailed, unpaired t test). Interestingly, for females, the ability to gain weight during the first postweaning week is not compromised (+4 g versus +4 g; P=0.23, two-tailed unpaired t test). Figure 2C shows example photographs of the long bones isolated from age- and sex-matched WT and KO_b/c mice. Figure 2D reports their relative lengths. The humeri, ulnae, femora, and tibia-fibulae of KO_b/c mice are consistently approximately 10% shorter than those isolated from comparable WT mice and in our samples, the difference is significant in all cases except the humeri. (one-tailed, paired t test with Bonferroni correction for multiple comparisons, setting significance at P=0.013=0.05/4).

KO_b/c Mice Have Thickened Corneas

We examined KO_b/c mice for signs of the corneal edema that is exhibited by Nbce1a/b/c-null mice. Figure 3A shows the NBCe1 immunoreactivity in corneal sections from wild-type and KO_b/c mice. Nbce1 immunoreactivity is detected in the basolateral membrane of corneal endothelial cells of WT mice, but is absent from the corneal endothelial cells of KO_b/c mice. Figure 3B shows the average in vivo corneal thickness of WT, +/−, and KO_b/c mice, as measured by ultrasonic pachymetry. The corneas of KO_b/c mice are significantly thicker than those of WT or +/− mice. A similar degree of thickening was evident in KO_b/c mice of that were 25 (the youngest examined), 40, and 47 (the oldest examined) days old (not shown). Although slit-lamp examination did not reveal obvious signs of corneal clouding in KO_b/c mice (not shown), the corneal stroma of KO_b/c mice reflected more light, consistent with an increase in opacity (Figure 3, C and D).

Figure 3.

Nbce1b/c-null (KO_b/c) mice exhibit corneal swelling and increased corneal opacity. (A) Light micrograph of a hematoxylin and eosin stained wild-type (WT) mouse eye section showing the approximate locations highlighted in the accompanying fluorescence microscopy images of the corneas from WT and KO_b/c mice. Note the absence of Nbce1 immunoreactivity (using the anti-Slc4a4 antibody) from the corneal endothelia of KO_b/c mice (representative of images gathered from eye sections from three pairs of mice). (B) The in vivo (left eye) corneal thickness of WT mice (two males, six females), together with data gathered from their heterozygous (+/−; four males, two females) and KO_b/c (four males, five females) littermates. *P<0.05: significantly different from WT thickness according to ANOVA with Tukey post-hoc analysis. (C) Representative z-axis projections of confocal reflection micrographs of the corneas of freshly enucleated WT and KO_b/c mouse eyes (the endothelium is visible as a bright band, the epithelium could not be imaged because of its proximity to the highly reflective coverslip) together with a distribution plot of the pixel intensities of the stroma. (D) The median stromal pixel intensities for a larger number of WT and KO_b/c mice.

KO_b/c Mice Have Defective Tooth Enamel

We examined KO_b/c mice for signs of the weakened dental enamel that is exhibited by Nbce1a/b/c-null mice. Figure 4A shows representative photographs of the teeth of WT, +/−, and KO_b/c mice (more than ten per group). As most clearly seen for the lower incisors, the enamel of KO_b/c mice, unlike the normal appearance of the enamel of wild-type and heterozygous mice, is markedly chalky in appearance, chipped at the biting edge, and lacks iron pigmentation. EDS revealed no gross defects in the mineral composition of the enamel from KO_b/c mice (Figure 4B), although analysis of electron micrographs (such as shown in Figure 4C) showed differences in the medians and variance of the mineral density (Figure 4D). Specifically, although both wild-type and KO_b/c mice incisors have enamel that is denser than dentin in the same tooth (Figure 4E), the difference between enamel and dentin density is significantly smaller in KO_b/c mice (Figure 4F) and the enamel has a substantially more heterogeneous composition (Figure 4G).

Figure 4.

Nbce1b/c-null (KO_b/c) mice exhibit enamel defects. (A) Photographs of the incisors of wild-type (WT), +/−, and KO_b/c littermates (representative of more than ten mice each). (B) The results of EDS analysis of the enamel of a greater number of incisors taken from preparations such as those shown in the following panel. Note that a parallel analysis of dentin also showed no significant difference in composition between WT and KO_b/c mice (data not shown). (C) Representative electron micrograph of polished cross-sections of incisors from a pair of WT and KO_b/c littermates. (D) Representative plots of pixel intensity for dentin and enamel taken from electron micrographs of a WT and KO_b/c incisor. Note that the pixel intensities can only be compared between dentin and enamel in the same tooth and not between teeth as the intensity depends on the distance between the electron detector and the sample surface, which was not constant among samples [an exception being the images in (C)]. (E) A comparison of the median pixel intensity of dentin and enamel from a greater number of electron micrographs of WT and KO_b/c incisors (one micrograph per tooth). (F) A comparison of the difference in median pixel intensities between dentin and enamel from the same images. (G) Variation of pixel intensity between dentin and enamel from the same images.

Discussion

KO_b/c mice exhibit many of the signs exhibited by Nbce1a/b/c-null mice (increased mortality, growth retardation, corneal edema, and enamel defects) despite a normal plasma pH and no deficit in plasma [HCO₃⁻]. This observation reveals, for the first time, those signs of pRTA that are not secondary to metabolic acidosis, are not amenable to alkali therapy, and which therefore require treatment beyond alkali therapy.

Increased Mortality

The effect of NBCe1 mutations on human mortality is unknown: cases of pRTA are rare and most probands are still young at the time of report. The oldest individual was 50 years of age.²⁸ However, there is some anecdotal evidence among this small cohort of increased childhood mortality. The sister of one proband died at 1 day of age of unknown cause, but her pRTA status is unknown.³⁷ One of two brothers with pRTA died at age 4 years after an inappropriately severe reaction to a mild head trauma, although the genetic cause of pRTA was not verified in either brother.³⁸ What is certain is that NBCe1a/b/c-null mice exhibit a severe increase in perinatal mortality, with none surviving beyond 30 days.^24,29 Studies in which alkali therapy was administered to Nbce1a/b/c-null mice led to an imperfect correction of acidosis but a substantial increase in survival: the majority of prenatally treated mice survived beyond 60 days and one mouse survived to 120 days.³⁹ However, expectations that perfect plasma pH correction might further enhance survivability are at odds with the poor survival of KO_b/c mice; although their mortality is not as great as that of Nbce1a/b/c-null mice, few KO_b/c mice survive beyond 2 months. Furthermore, acidemic Nbce1a-null mice have a normal lifespan, suggesting that acidemia is not a prime cause of mortality.^40,41 Taken together these data suggest that Nbce1b/c loss is a key factor in early mortality of Nbce1-null mice, but that acidemia secondary to Nbce1a loss worsens the outlook. Thus, although alkali therapy might enhance survival, it is not sufficient to normalize lifespan.

Growth Defects

Individuals with pRTA are typically several standard deviations below average height and weight, often with an inappropriately young bone age.^17,37 A study of one girl with pRTA indicated that early administration of alkali therapy could improve height velocity.²⁶ Similarly, alkali-supplemented Nbce1a/b/c-null mice exhibited an improvement in weight gain.³⁹ Together these data indicate that acidosis is a major factor in growth retardation; indeed, acidosis would be expected to impair bone growth. However, even KO_b/c mice exhibit a significant growth retardation, suggesting that alkali therapy, even if perfectly administered, would not be sufficient to completely correct the growth defects in pRTA (except perhaps in instances where the genetic defect specifically affects NBCe1-A²⁰). We are unaware of a specific role for Nbce1b/c in bone formation, but recent studies indicate that Nbce1b/c loss from the intestinal epithelia impairs electrolyte and nutrient absorption,⁴² which would contribute toward a general failure to thrive. An unusual feature of Nbce1a/b/c-null mice that has not been noted in individuals with pRTA is severe impaction of the intestines. Nbce1b/c supports intestinal fluid secretion and so it might be predicted that its loss would contribute to such a phenotype.^29,42 However, it has also been noted that ENaC is upregulated in the colons of Nbce1a/b/c-null mice, secondary to hypovolemia and acidemia; hyperabsorption would also contribute to the phenotype.²⁹ We did not see evidence of the impacted intestines in postmortem studies of Nbce1b/c-null mice, further supporting a complex cause.

Corneal Edema

The ocular phenotype of pRTA has variable expressivity and penetrance. Almost all individuals with pRTA exhibit band keratopathy (or increased corneal opacity^18,23), glaucoma (or at least elevated intraocular pressure²²), and cataracts. The cause of these signs is not well understood, although an individual with a mutation that affects only NBCe1-A exhibited glaucoma but no other ocular pathology, whereas aged Nbce1a-null mice develop elevated intraocular pressure.^20,43 Among the other signs, only corneal opacity has been reported in Nbce1a/b/c-null mice and histologic study of their corneas revealed corneal edema.²⁴ Nbce1b has long been suggested to play a critical role in the corneal HCO₃⁻ pumping mechanism.^44–46 Our data show that the in vivo baseline corneal thickness in KO_b/c mice is increased to an extent that would be considered clinically significant in humans.⁴⁷ The greater stromal opacity that we observed by confocal microscopy in KO_b/c mice is similar to that observed by in vivo imaging of edematous human corneas from individuals with Fuchs dystrophy.^48,49 These observations demonstrate a direct and critical role for Nbce1b in maintaining corneal health. These data also implicate the region that encompasses promoter 1 and exons 1–3 of SLC4A4 as a novel candidate locus for cases of recessively inherited corneal dystrophies whose genetic origin remains obscure.^50,51

Enamel Defects

Dental abnormalities are frequently noted in individuals with pRTA,^18,25 and easily chipped, hypomineralized enamel is characteristic of the teeth of Nbce1a/b/c-null mice.^29,52 The apparently normal mineral composition of KO_b/c enamel combined with its unusually heterogeneous structure are consistent with hypomineralization. The relative contribution of Nbce1a versus Nbce1b/c loss to this phenotype has been controversial. On the one hand, Nbce1b/c is expressed in the ameloblasts of the enamel organ where it is hypothesized to support enamel remodeling.⁵³ On the other hand, wild-type and Nbce1a/b/c-null molars developed similarly when mandibles from those mice were explanted and cultured in the kidney capsules of wild-type mice, suggesting that correction of acidemia could prevent hypomineralization.⁵⁴ The exhibition of defective enamel in KO_b/c mice establishes the importance of Nbce1b/c for enamel formation and demonstrates that the enamel defect in pRTA cannot be prevented by normalizing blood pH. With regard to the contrary explant data, we note that neither the wild-type nor the Nbce1a/b/c-null explanted teeth mineralized their enamel as well as in situ teeth, and development of the explanted teeth may have been sufficiently slow to allow even disabled ameloblasts to perform adequately.

Chronic Respiratory Acidosis

The elevation of pCO₂ in KO_b/c mice is not a feature shared with Nbce1a/b/c-null mice. In fact, metabolic acidosis typically results in hyperventilation producing a compensatory reduction of pCO₂, which may mask such a defect. However, inappropriately high pCO₂ (≥35 mm Hg, consistent with a confounding respiratory acidosis) has been reported in one case of pRTA¹⁸ and can be calculated from the pH and [HCO₃⁻] values provided in two other cases.^22,23 The cause is unknown but we note that Nbce1b is expressed in lung epithelia, myocytes, and neurons,¹ any and all of which can influence the set point of pCO₂.

The Role of Renal Nbce1b

Finally, in reference to a recent report of Nbce1b expression in the renal medulla,⁹ the ability of the kidneys to maintain an apparently appropriately elevated plasma [HCO₃⁻] does not reveal any gross renal acid-base handling defect in KO_b/c mice.

Future Directions

Although alkali therapy is considered beneficial for treatment of many kidney diseases, the phenotype of Nbce1b/c-null mice reveals the limitations of alkali therapy as a catch-all treatment for the signs of pRTA, highlights the need to develop separate therapies for the ocular, dental, and growth defects, and indicates that mutations specific to the NBCe1-B/C–encoding region of SLC4A4 could underlie unusual cases of corneal dystrophy and enamel hypomineralization. Pending the development of conditional Nbce1-null mice, further work will need to be done to establish a cause of chronic respiratory acidosis and death in Nbce1b/c-null mice, knowledge of which may be useful to prolong their lifespan and increase their usefulness as a model for developing novel therapies.

Disclosures

Dr. Patel reports grants from NIH during the conduct of the study. Dr. Parker reports grants from NIH, grants from ASN Foundation during the conduct of the study.

Supplemental Material

This article contains the following supplemental material online at http://jasn.asnjournals.org/lookup/suppl/doi:10.1681/ASN.2018050545/-/DCSupplemental.