Expression of the regulated isoform of the electrogenic Na⁺/HCO₃⁻ cotransporter, NBCe1, is enriched in pacemaker interstitial cells of Cajal

Abstract

Interstitial cells of Cajal (ICCs) generate electrical slow waves, which are required for normal gastrointestinal motility. The mechanisms for generation of normal pacemaking are not fully understood. Normal gastrointestinal contractility⁻ and electrical slow-wave activity depend on the presence of extracellular HCO₃⁻. Previous transcriptional analysis identified enrichment of mRNA encoding the electrogenic Na⁺/HCO₃⁻ cotransporter (NBCe1) gene (Slc4a4) in pacemaker myenteric ICCs in mouse small intestine. We aimed to determine the distribution of NBCe1 protein in ICCs of the mouse gastrointestinal tract and to identify the transcripts of the Slc4a4 gene in mouse and human small intestinal tunica muscularis. We determined the distribution of NBCe1 immunoreactivity (NBCe1-IR) by immunofluorescent labeling in mouse and human tissues. In mice, NBCe1-IR was restricted to Kit-positive myenteric ICCs of the stomach and small intestine and submuscular ICCs of the large intestine, that is, the slow wave generating subset of ICCs. Other subtypes of ICCs were NBCe1-negative. Quantitative real-time PCR identified >500-fold enrichment of Slc4a4-207 and Slc4a4-208 transcripts [“IP3-receptor-binding protein released by IP3” (IRBIT)-regulated isoforms] in Kit-expressing cells isolated from Kit^creERT2/+, Rpl22^tm1.1Psam/Sj mice and from single GFP-positive ICCs from Kit^tm1Rosay mice. Human jejunal tunica muscularis ICCs were also NBCe1-positive, and SLC4A4-201 and SLC4A4-204 RNAs were >300-fold enriched relative to SLC4A4-202. In summary, NBCe1 protein expressed in ICCs with electrical pacemaker function is encoded by Slc4a4 gene transcripts that generate IRBIT-regulated isoforms of NBCe1. In conclusion, Na⁺/HCO₃⁻ cotransport through NBCe1 contributes to the generation of pacemaker activity in subsets of ICCs.

NEW & NOTEWORTHY In this study, we show that the electrogenic Na+/HCO₃⁻ cotransporter, NBCe1/Slc4a4, is expressed in subtypes of interstitial cells of Cajal (ICCs) responsible for electrical slow wave generation throughout the mouse gastrointestinal tract and is absent in other types of ICCs. The transcripts of Slc4a4 expressed in mouse ICCs and human gastrointestinal smooth muscle are the regulated isoforms. This indicates a key role for HCO₃⁻ transport in generation of gastrointestinal motility patterns.

INTRODUCTION

The underlying pattern of rhythmic contractions in the gut originates from interstitial cells of Cajal (ICCs), which generate electrical pacemaker activity that propagates to the smooth muscle cells (1, 2). ICCs are mesoderm-derived mesenchymal cells that are distributed across the thickness of the gastrointestinal tunica muscularis (3), which also serve to mediate types of neuromuscular neurotransmission, act as mechanosensors, and set smooth muscle membrane potential gradients (4, 5). Therefore, ICCs are necessary for normal gastrointestinal function, and loss of ICCs has been associated with gastrointestinal motility disorders including slow transit constipation and gastroparesis (6, 7).

Pacemaker activity is generated by ICCs in the myenteric region (ICC-MY) between the circular and longitudinal muscle layers of the stomach from the corpus through the antrum, and in the small intestine (1, 2, 8). In the large intestine, slow waves depend on an intact interface between the circular smooth muscle and the submucosa, and these are generated by submuscular ICCs (ICC-SMP) in this region (some authors refer to these cells as ICCs of the submucosal plexus, ICC-SM) (9–11). Since the identification of ICCs as pacemaker cells, the mechanism of electrical pacemaker generation has been a subject of extensive investigation. It is clear that release of Ca²⁺ from inositol 1,4,5-trisphosphate (IP3)-sensitive intracellular stores, Ca²⁺ influx through voltage-gated, plasma membrane ion channels, and activation of the Ca²⁺-activated Cl⁻ channel, Ano1 (also known as TMEM16A), are necessary for the generation of normal electrical slow wave activity (12–14). Other ion transporters in the plasma membrane including NKCC1 (Slc12a2) contribute to the setting of the reversal potential for Cl⁻ (15), and NCX3 (Slc8a3), which regulates the movement of Na⁺ and Ca²⁺ ions (16). Mitochondrial function is also critical for normal pacemaker activity (17). Many of these molecules, linked to generation of pacemaker activity, are expressed in all subtypes of ICCs, but the key factors that are specific to the pacemaker cells and that ensure normal pacemaking remain to be elucidated.

One approach to this question has been the comparison of transcriptional profiles of ICC-MY versus ICC-DMP in the tunica muscularis of mouse small intestine (15, 18). We observed that one of the differentially expressed genes was the solute carrier family 4 member 4 (Slc4a4) encoding the electrogenic Na⁺/HCO₃⁻ cotransporter 1 (NBCe1) (18). This protein regulates absorption and secretion of bicarbonate and carbonate, and contributes to intracellular pH homeostasis (19). Normal activity of NBCe1 is essential for life (20); it mediates critical roles in acid-base homeostasis through HCO₃⁻ absorption in the renal proximal tubule (19) and also contributes to physiological function in other tissues (21) including the central nervous system (22, 23), eye (24), secretory epithelia of exocrine glands (25), gastrointestinal tract (26), and urogenital system (27). In many of these tissues, NBCe1 is responsible for bulk transport of Na⁺ and HCO₃⁻, but it should also be noted that NBCe1 regulates neuronal and cardiac myocyte excitability (28, 29). It is established that normal gastrointestinal contractility and slow wave activity depend on the presence of extracellular HCO₃⁻ (30, 31), but it is not known how HCO₃⁻ transport is mediated in gastrointestinal tunica muscularis.

There are at least three main Slc4a4 transcripts that encode functional NBCe1 proteins in mouse and human, which are generated by transcription from two different promoters and due to alternative splicing of the RNA (19). In mice, the NBCe1-A variant (encoded by Slc4a4-203—ENSEMBL nomenclature) contains 41 unique amino acids at the NH₂ terminus that arise from an alternative promoter in intron 3 of Slc4a4 (32). This variant is constitutively active and predominantly expressed in the kidney (33, 34). The NBCe1-B variant (encoded by Slc4a4-207) is identical to NBCe1-A except for 85 NH₂-terminal amino acid residues that replace the 41 NH₂-terminal residues of the A variant (33, 35). NBCe1-C (encoded by Slc4a4-208) is identical to B except for 61 unique COOH-terminal residues due to splicing out of a 97-base-pair exon near the 3′ end of the gene. This replaces 46 residues in the COOH-terminus of NBCe1-B (21, 36). The NH₂-terminal peptide sequences of NBCe1-B and NBCe1-C form an auto-inhibitory domain that results in low baseline Na⁺/HCO₃⁻ transport activity and make NBCe1-B/C sensitive to activation by the “IP3-receptor binding protein released by IP3” (IRBIT) molecule (37). Thus, identification of the specific transcripts of NBCe1 in ICCs and other cells of the gastrointestinal tunica muscularis is important to understanding the role of this protein in gastrointestinal contractility.

The primary aim of this study was to test the hypothesis that the Na⁺/HCO₃⁻ electrogenic cotransporter, NBCe1, is expressed in pacemaker ICCs owing to the requirement for HCO₃⁻ when recording normal electrical slow wave activity and owing to the enrichment in small intestinal, pacemaker ICC-MY of Slc4a4 mRNA, a gene that encodes NBCe1, which is expressed in other excitable cells. We set out to establish whether NBCe1 protein is expressed in some or all subtypes of ICCs in the mouse and human gastrointestinal tract. We also determined the Slc4a4 transcripts expressed in ICCs from mouse and human gastrointestinal tract. We found that NBCe1 expression is restricted to ICCs that have pacemaker function in the stomach, small intestine, and large intestine of the mouse and that the Slc4a4 transcripts encoding the IRBIT-regulated isoforms of NBCe1 were the dominant transcripts in the tunica muscularis of the mouse and human small intestine.

MATERIALS AND METHODS

Ethical Approval

This study was approved by the Institutional Review Board (IRB) for human tissues (Approval numbers 13–008138 and 26–73) and the Institutional Animal Care and Use Committee of the Mayo Clinic (IACUC, Rochester, MN) under protocol number A00002343-16.

For RNA extraction, normal human jejunal tissue was waste material obtained from patients undergoing surgery for gastric bypass or jejunal resection (see Table 1 for details). Tissues were collected immediately after surgery and brought to the laboratory for further processing. The tunica muscularis was separated from the mucosa and submucosa by sharp dissection and then snap frozen in liquid nitrogen for storage at −80°C until further processing for RNA extraction. An additional specimen of deidentified human small intestinal tissue from a male patient aged 50 yr was obtained from the Mayo Clinic Biobank under Mayo Clinic IRB# 19–006544. This tissue was a formalin-fixed, paraffin-embedded (FFPE) sample taken from the normal margins of a surgical, small intestinal resection.

Table 1.

Patient details for human small intestinal tissues used for gene expression studies

Age	Sex	Region	IRB	Procedure
69	M	Jejunum	13-008138	Jejunal resection
45	F	Jejunum	13-008138	Gastric Bypass Revision
48	F	Jejunum	26-73	Gastric Bypass Revision
40	F	Jejunum	26-73	Gastric Bypass Revision
35	F	Jejunum	26-73	Gastric Bypass Revision
50	F	Jejunum	26-73	Gastric Bypass Revision

IRB, Institutional Review Board.

Animals

Adult (6–8 wk old) C57BL/6J and Kit^creERT2, Rpl22^tm1.1Psam/J (Ribotag), neonatal Kit^tm1Rosay (Kit^copGFP), and 14- to 21-day-old Slc4a4^tm1Ges (Slc4a4 knockout) mice of either sex were used. Food and water were provided ad libitum. Mice were humanely killed by CO₂ inhalation followed by cervical dislocation. The Kit^creERT2, Rpl22^tm1.1Psam/J mice were generated in house to express the Rpl22^HA allele (38) in response to activation by tamoxifen of Cre-mediated recombination in Kit-positive ICCs (39). The Slc4a4 knockout mice received daily intraperitoneal injections of 7.5% NaHCO₃ 15 mL/g body weight once a day from day 10 after birth to correct for, otherwise lethal, systemic acidosis caused by deletion of the gene (20).

Tissue Preparation

Mouse tissues (gastric fundus, corpus and antrum, small intestine, and proximal and distal colon) were excised, placed in ice-cold calcium-free Hanks' balanced salt solution (Corning Manassas, VA), and opened along the lesser curvature of the stomach or the insertion of the mesentery for small intestine and colon, and their contents were washed away. The tissue was pinned out flat with the serosa facing upwards on petri dishes lined with Sylgard elastomer (Dow Corning Corp., Midland, MI); the external muscle layers were carefully peeled off the mucosa. The mucosa was discarded, and only the whole thickness preparations of tunica muscularis with intact myenteric ICCs (ICC-MY), deep muscular plexus ICCs (ICC-DMP), and submuscular ICCs (ICC-SMP) were used for immunohistochemical (IHC) experiments. For cross-sectional histological analysis, small intestine and colon were dissected as a tube and placed immediately into a cryomold with OCT, and the mold was placed into 2-methylbutane cooled with dry ice and left to solidify. Tissues were stored at −80°C until sectioned.

Antibodies

For details about the primary and secondary antisera and the concentrations used, see Table 2. We did the following negative controls for each antibody: incubating the sections with secondary antibodies but no primary antibodies, by applying secondary antibodies directed against IgG from a species that was not the host for raising the primary antibody, and by examining singly labeled tissues under illumination with the filter sets designed for the wrong fluorophore. Also, pre-absorption controls and tissue from Slc4a4 knockout mice were used. Sections of kidney and specific labeling of the mucosal epithelium were used as positive controls.

Table 2.

Antisera used in this study

Catalog No./Clone	Supplier	RRID	Target	Host Species	Use	Final Concentration
Primary Antibodies
11885-1-AP	Proteintech	AB_2191458	Slc4a4	Rabbit	IHC	0.3 µg/mL (mouse)0.75 µg/mL (human)
AF1356	R&D system	AB_354750	Kit	Goat	IHC	0.2 µg/mL
AF1062	R&D system	AB_2236897	PDGFRα	Goat	IHC	0.2 µg/mL
HuC/D-14_Lennon	V.A. Lennon, Mayo Clinic	AB_2813895	HuC/D	Human	IHC
DOG-1-L-CE/K9	Leica Biosystems	AB_10555293	ANO1	Mouse	IHC	0.24 mg/mL
901533	Biolegend	AB_2801249	HA	Mouse	RiboTag	0.025 mg/mL
MABF1081Z	Millipore-Sigma	AB_2828024	Negative Control	Mouse	RiboTag	0.025 mg/mL
Secondary Antibodies
711-165-152Cy3	Jackson ImmunoResearch	AB_2307443	Rabbit-IgG	Donkey	IHC	1.88 µg/mL
705-545-147Alexa Fluor 488	Jackson ImmunoResearch	AB_2336933	Goat-IgG	Donkey	IHC	3.75 µg/mL
711-605-152Alexa Fluor 647	Jackson ImmunoResearch	AB_2492288	Rabbit-IgG	Donkey	IHC	2.5 µg/mL
709-165-149Cy3	Jackson ImmunoResearch	AB_2340535	Human-IgG	Donkey	IHC	1.88 µg/mL
AP192SAG Alexa Fluor 647	Millipore-Sigma	AB_2687879	Mouse-IgG	Donkey	IHC	0.002 mg/mL

Immunolabeling

For mouse tissues, we cut 12-μm sections using a freezing microtome and fixed in cold 4% paraformaldehyde (PFA) and 0.1 M phosphate buffer (4°C, 10 min). After fixation, we washed the sections three times for 5 min each and incubated with 10% normal donkey serum (NDS, Jackson Immunoresearch Laboratories, West Grove, PA) and 0.3% Triton X-100 (Sigma) in 0.1 M phosphate-buffered saline (PBS) for 1 h to minimize nonspecific antibody binding, incubated with primary antibody at the corresponding dilutions (10% NDS) overnight at 4°C. Next, we rinsed the sections in PBS and incubated in the dark for 1 h at room temperature with the proper secondary antibody. The labeled tissues were mounted using SlowFade Gold (Thermo Fisher, Waltham, MA), a mounting medium with 4′,6-diamidino-2-phenylindole (DAPI) as a nuclear counterstain.

Mouse whole mounts were stretched over the surface of a Sylgard-coated petri dish then fixed with cold 4% PFA (4°C, overnight), washed with PBS (4°C, 8 times), and blocked with 1% bovine serum albumin (BSA, Sigma-Aldrich, St. Louis, MO) in PBS plus 0.3% Triton X-100 in PBS (4 h at room temperature). Primary antibodies were added at the indicated dilutions (Table 2) in 1% BSA, PBS plus 0.3% Triton X-100, and incubated for 48 h at 4°C. The tissue was washed with PBS eight times for 30 min each. Then, it was incubated overnight in the dark with secondary antibodies, washed, and counterstained with DAPI dilactate (Invitrogen, Carlsbad, CA) to label nuclei. The labeled tissues were mounted using Diamond SlowFade mounting medium (Thermo Fisher).

Human FFPE tissues were cut into 5-µm sections and doubly immunolabeled for ANO1 to identify ICCs and NBCe1 as follows. Slides were baked at 60°C for 1 h, deparaffinized in xylene, and rehydrated in serial dilutions of ethanol. For antigen retrieval and unmasking, slides were submerged in Tris-EDTA buffer (10 mM, pH 8.0), heated to boiling point for 2 × 5 min at full power in a microwave, and then cooled down at room temperature on the bench. We washed the slides 2 × 5 min in Tris-buffered saline (0.1 M, 1X TBS), tapped off excess liquid, and marked the section with a hydrophobic pen. Tissues were permeabilized (0.1% Triton X-100 in TBS for 10 min) and blocked with 3% BSA in 1X TBS plus 0.05% Tween-20 in a humidified chamber for 1 h at room temperature to minimize nonspecific antibody binding. We incubated the tissues with primary antibody at the dilutions listed in Table 1 in blocking buffer overnight at 4°C in a humidified chamber and then rinsed in TBS three times for 5 min. For detection, secondary antibodies were added and incubated in the dark for 1 h at room temperature. We washed the tissues in 1X TBS, then nanopure water 2 × 3 min, and incubated those with DAPI dilactate (0.04 mM in water) for 15 min to label nuclei. Finally, sections were rinsed with water 3 × 5 min and mounted in Diamond SlowFade mounting medium for imaging.

Image Acquisition

To show immunoreactivity of NBCe1, images of immunolabeling were collected using an Olympus FV1000 laser scanning confocal microscope. Confocal images were collected with either ×20 0.95 NA; ×25 1.05 NA or ×60 1.2 NA water immersion objectives using confocal apertures and Z step dimensions set to obtain the optimal image resolution for the wavelengths and objectives used. Image stacks were collected through the regions of the tissues shown and flattened into 2-D maximum intensity projection images for display. YZ projections were generated by digital sectioning through the 3-D image stack then flattening along the y-axis to generate 2-D projections. All images were prepared for individual figures using Adobe Photoshop CS. No 3-D reconstructions, deconvolution, surface or volume rendering, or gamma adjustments were performed for the final images. The following lasers and emission filters were used to visualize the labeled structures and collect images: multiline Ar laser at 488 nm (used for the excitation of Alexa Fluor 488); emission filter 535 ± 15 nm; 543 nm HeNe laser (used for Cy3); emission filter 575–630 nm; and 633 nm HeNe laser (used for Alexa Fluor 647); emission filter HQ660 nm (650–700 nm).

Ribosome Immunoprecipitation

Tunica muscularis was extracted from mouse small intestine, flash-frozen in liquid nitrogen, and stored at −80°C until use. Samples were homogenized on ice in ice-cold homogenization buffer (50 mM Tris, pH 7.4, 100 mM KCl, 12 mM MgCl₂, 1% Nonidet P-40, 1 mM DTT, 1:100 protease inhibitor (Sigma), 200 U/mL RNasin (Promega), 1 mg/mL heparin, and 100 µg/mL cycloheximide (Sigma)) with a Dounce homogenizer (Sigma) until the suspension was homogeneous. To remove cell debris, 1 mL of the homogenate was transferred to an Eppendorf tube and was centrifuged at 10,000 g at 4°C for 10 min. Supernatants were transferred to a fresh Eppendorf tube on ice, and then, 100 μL was removed for “input” analysis and 350 μL was added to anti-HA antibody (H9658, Sigma) or 5 μL (= 1 µg) of mouse monoclonal IgG1 antibody (Sigma, Cat. No. M5284) coupled to magnetic beads was added to the supernatant, followed by incubation overnight with slow rotation at 4°C. After not more than 12 h, samples were washed three times with high-salt buffer (50 mM Tris, 300 mM KCl, 12 mM MgCl2, 1% NP-40, 1 mM DTT, 1:200 protease inhibitor, 100 U/mL RNasin, and 0.1 mg/mL cycloheximide in RNase-free nanopure water), 5 min per wash in a cold room on a rotator. After the final wash, high-salt buffer was removed and beads were resuspended in 400 μL of supplemented Qiagen lysis (RLT) buffer (10 μL β-mercaptoethanol/10 mL of RLT buffer) from the RNeasy Plus Micro kit (Qiagen, Germantown, MD) and vortexed vigorously. RLT buffer containing the immunoprecipitated RNA was removed from magnetic beads before RNA purification using RNeasy kit according to manufacturer’s protocol. Equally, 350 μL of RLT was added to input fractions before RNA isolation using RNeasy kit. Both input and RiboTag-IP samples were eluted in 15 μL of water. RNA integrity numbers were determined using High Sensitivity RNA ScreenTape on a 2200 TapeStation (Agilent, Santa Clara, CA).

Single-Cell RT-PCR

ICCs were dissociated from the small intestines of neonatal Kit^copGFP mice as described previously for generating primary cultures of ICCs (40). After dissociation of mouse jejunal smooth muscle strips, each single GFP-labeled ICC cell was aspirated using a patch clamp pipette and deposited into a tube, and then flash-frozen on dry ice. Single-cell RT-PCR was performed using the Single Cell-to-CT kit (Life Technologies). For the RT-PCR, the cell was incubated for 5 min at room temperature in the presence of 10 μL of single-cell lysis solution containing DNase I. The lysis reaction was stopped by adding 1 μL of stop solution and incubating at room temperature for 2 min. cDNA was then produced by adding 4.5 μL of RT reaction mix and performing reverse transcription in a thermal cycler by incubating for 10 min at 25°C, 60 min at 42°C, and 5 min at 85°C. A pre-amplification reaction was done by adding to the cDNA 4.5 μL of PreAmp mix and 6 μL of a solution containing pooled primers for all the genes of interest diluted in 1X TE (pH 8.0) at a final concentration of 80 femtomoles. This reaction was held at 95°C for 10 min for enzyme activation and then cycled 14 times for at 95°C for 15 s (denaturation) and 60°C for 4 min (annealing/extension) and held again at 99°C for 10 min for enzyme deactivation. The preamplified DNA was diluted 1:20 in TE, and 5 μL was used for PCR (25 μL total) in the presence of Platinum Taq DNA Polymerase (Invitrogen) and primer sets specific for the isoform of interest (Table 3) at a final concentration of 0.4 μM. Primer assays specific for detecting total Slc4a4, Slc12a2 (NKCC1), and Tacr1 (NK1 receptor) were purchased from Qiagen. The program used for PCR was the following: preincubation at 94°C for 2 min (for the activation of the enzyme and DNA denaturation); the amplification was obtained by cycling 30 times between 95°C for 30 s, 60°C for 30 s, and 72°C for 30 s, followed by cooling at 4°C. The bands were resolved in a 4% agarose gel. For all assays, primers for detecting variants of Slc4a4 were designed based on the ENSEMBL annotations of mouse and human SLC4A4 transcripts and purchased from IDT.

Table 3.

Sequences of primers used in this study

	Primers 5′-3′
Target	Forward	Reverse	Product Size, bp
Single-cell PCR
mA	CCTCAGGGTTTTCCAGCCAA	CCACCCTGCTCCACTTTCTC	222
mBC	GAGGGGCTTCCTTCCTTAAACAT	GAGGGGTTTGAGGATACTGCTG	227
mABC	GCACGACCTCAGCTTCCTT	ATGTGTGGCGTTCGAGGAAT	245/148
qPCR from RiboTag
mBC	GTCCATGTGCCCAAGAGCTA	CGAGTTCCGTGAAGAGCTGA	241
mAB	TCCTCAAGTCAACAGTGGCTG	ACCTTTTCTGAGTATGGGCAGTC	229
mC	CTCAAGTCAACAGTGGCTGC	GGATCTTTCTCATCGTCATTGTCGC	212
Normalization primers for mouse qPCR
mABC-5′	AGCCTGTTTGAGCTGAGGAC	ATGGCTGGGCTGCTGCCATTAT	275
mABC-3′	CTGGATCTGTTTTGGGTCGC	CAAGGGTTCCAGTGACTCGT	187
qPCR human samples
hA	AGGGTTGTCCAGCCAATGTTT	TCCCCACCCTGTTCCACTTT	221
hBC	CCAAACTGGAGGAGCGACG	AACTCTTCGGCACATGGACTC	153
hAB	ACAGCAACCTTTCCTAAGCGA	TGATGTGTGGCGTTCAAGGA	76
Normalization primers for qPCR human samples
hABC-5′	GGTGGGGAAAGATGGAGCAA	GGTGCTTCCGGAGCAAAGTA	217
hABC-3′	GGTGTGCAGTTCATGGATCG	AGTGAACAGGTGGACTCTGC	108

Quantitative RT-PCR

From the ribosome precipitation sample, cDNA for quantitative RT-PCR expression analyses of the samples was synthesized from total RNA using the SuperScript Vilo cDNA Synthesis kit (Thermo Fisher). Quantitative PCR was performed with SYBR Green I Master Mix (Roche, Basel, Switzerland) on the LightCycler 480 (Roche) according to standard procedures, using primers specific for all genes of interest purchased from Integrated DNA Technologies IDT (Table 3; see maps in Figs. 10A and 11A). RNA expression levels were normalized to total Slc4a4 using primers that amplify all isoforms.

Fig. 10.

NBCe1 immunoreactivity colocalizes with ANO1 in myenteric ICC of the human small intestine. Flattened projection of high-resolution confocal image stack of immunolabeled thin sections (5 µm) from human small intestine. Images are mosaics of high-resolution tiles of confocal stacks collected with a ×60 1.2 NA objective. A: NBCe1 immunoreactivity was detected in epithelial cells of the mucosa. B: NBCe1 colocalizes with ANO1 in ICC-MY (yellow arrow). The green arrow points to septal ICC fibers that are not immunoreactive for NBCe1. Myenteric neurons are NBCe1-positive but negative for ANO1. Scale bar= 50 µm. ICC, interstitial cells of Cajal; ICC-MY, interstitial cells of Cajal in the myenteric region.

Quantitative RT-PCR for Human Samples

Total RNA was extracted from tunica muscularis of human small intestine using RNA-Bee (Tel-Test Inc. Friendswood, TX) according to standard procedures, followed by an in-column clean-up with an RNeasy kit (Qiagen). Quality and concentration of the RNA was assayed using the Agilent TapeStation system. 1 µg of total RNA was applied for reverse transcription using the SuperScript VILO kit from Invitrogen. Quantitative PCR was performed with SYBR Green I Master Mix (Roche, Basel, Switzerland) on the LightCycler 480 (Roche) according to standard procedures, using primers specific for all genes of interest purchased from IDT (Table 3; see map in Fig. 13A). RNA expression levels were normalized to total SLC4A4 using primers that amplify all isoforms.

Fig. 13.

Quantification of the abundance of SLC4A4 transcripts in human small intestinal tunica muscularis. A: map of primer sequences chosen to specifically amplify variants identified by ENSEMBL nomenclatures. B: qRT-PCR analysis of SLC4A4 transcripts expressed in tunica muscularis of human small intestine. The different isoforms of SLC4A4 expressed were normalized to total (all isoforms) SLC4A4 levels using the ΔΔCt method. The transcripts analyzed by qRT-PCR were as follows: SLC4A4-202 (variant A), SLC4A4-201 (variant B), and SLC4A4-204 (variant C). **P < 0.05; ***P < 0.0005, Friedman test for multiple comparisons. Values are the means ± SE; n = 6. qRT-PCR, real-time quantitative reverse transcription polymerase chain reaction.

RESULTS

Presence of NBCe1-IR in Subtypes of ICCs in the Mouse Gastrointestinal Tract

We know that NBCe1 is robustly expressed in the basolateral membrane of the kidney proximal tubules, and since the antibody used for our immunolabeling experiments does not discriminate between the different splice variants of NBCe1, adult mouse kidney was used as a positive control. As expected, the characteristic pattern of NBCe1-IR in the basolateral membrane of the proximal tubules was observed (Fig. 1) (33).

Fig. 1.

NBCe1 immunoreactivity in sections of mouse kidney NBCe1 immunoreactivity (red) is restricted to epithelial cell, basolateral membranes of proximal tubules in mouse kidney. Nuclei are counterstained with DAPI (blue). Scale bar = 100 µm.

In thin cross sections of mouse small intestine, NBCe1 immunoreactivity was detected in mucosal epithelial cells, on the serosal surface, and in Kit-positive ICC-MY (white arrows, Fig. 2A) but not in ICC-DMP (yellow arrowheads, Fig. 2A) of the tunica muscularis. The specificity of the antisera was demonstrated by use of small intestine from a Slc4a4 knockout mouse, which had no evidence of NBCe1-IR in any type of cell (Fig. 2B). Additionally, preabsorption of the antiserum with the antigenic peptide reduced the NBCe1-IR to background levels (Fig. 2C). In further control experiments, there was no labeling of cells by any of the secondary antibodies in the absence of the primary antisera and there was no labeling of cells when the secondary antisera were applied to tissues incubated with the primary antisera raised in a different species. No signal was observed when singly labeled tissues were examined using the excitation and emission filters for the wrong fluorophore.

Fig. 2.

In the mouse small intestine, pacemaker myenteric interstitial cells of Cajal (ICC-MY) are NBCe1 positive. A: images of Kit (green) and NBCe1 (red) immunoreactivity in adult C57Bl6/J mice small intestine. NBCe1 immunoreactivity was detected in epithelial cells of the mucosa as well as ICC-MY of small intestine but not in ICC-DMP. Kit-negative cells on the serosal surface were also positive for NBCe1 (indicated by yellow star). B: NBCe1 immunoreactivity is absent in Slc4a4 KO mice (PND 12), but Kit expression was not altered in ICCs. C: antigen preabsorption blocks all NBCe1 immunoreactivity, while Kit immunoreactivity was preserved. ICC-MY (white arrow). ICC-DMP (yellow arrowhead). Scale bar = 20 µm. Images representative of similar data from n = 3 mice. ICC-DMP, deep muscular plexus ICC; ICC-MY, interstitial cells of Cajal in the myenteric region.

Having established that the antisera are specific for NBCe1, further immunofluorescence experiments were done using whole mounts of the tunica muscularis of the whole gastrointestinal tract, starting with the proximal gastrointestinal regions. In mouse stomach, NBCe1-IR was restricted to neuron-like cells in the myenteric region of the gastric fundus (arrows, Fig. 3A). ICC-IM in the fundus, identified by morphology, location, and labeling for Kit, were not positive for NBCe1-IR (Fig. 3A). However, in the corpus and antrum of the stomach, NBCe1-IR colocalized with Kit in ICC-MY (Figs. 3C and 4A), which are responsible for slow wave generation in those regions. ICC-IM in the circular muscle of the corpus and antrum were not positive for NBCe1-IR (Figs. 3B and 4B). Mast cells in the gastric smooth muscle layers, identified by bright Kit-IR and round, process-free cell bodies, were also negative for NBCe1-IR (yellow arrowheads, Fig. 3B). NBCe1-positive neuron-like cells were rarely observed in the gastric corpus and antrum, but NBCe1-IR was observed in flat, Kit-negative cells on the serosal surface with a tile-like morphology characteristic of mesothelial cells (orange arrows, Fig. 3B).

Fig. 3.

NBCe1 expression localizes in ICC-MY networks in mouse stomach. Confocal image stacks of whole mounts of adult mouse stomach. A: NBCe1 immunoreactivity (red) was not present in Kit-positive ICC-IM (green) of the mouse fundus but was detected in Kit-negative cells that were Hu C/D positive (cyan) with neuronal morphology (white arrows). In the corpus of the stomach, in the myenteric region (B), NBCe1 immunoreactivity colocalized with Kit in ICC-MY, Kit-negative mesothelial cells were also NBCe1-positive on the serosal surface (orange arrows), and a YZ projection confirms that the cells are on the serosal surface of the tissue (shown at right side of the X-Y). In the circular muscle layer (C), NBCe1 immunoreactivity was absent from ICC-IM, Kit-positive mast cells were also negative for NBCe1 (yellow arrow). Scale bar = 50 µm. Images representative of similar data from n = 4 mice. ICC-IM, intramuscular Interstitial cells of Cajal; ICC-MY, interstitial cells of Cajal in the myenteric region.

Fig. 4.

NBCe1 expression localizes in ICC-MY networks in stomach antrum. Whole mounts of adult mouse gastric antrum. A: NBCe1 immunoreactivity (red) colocalized with Kit (green) in ICC-MY of the mouse gastric antrum. B: in the circular muscle layer, NBCe1 immunoreactivity was absent from ICC-IM. Scale bar = 30 µm. Images representative of similar data from n = 4 mice. ICC-IM, intramuscular Interstitial cells of Cajal; ICC-MY, interstitial cells of Cajal in the myenteric region.

In the small intestine, NBCe1-IR fully colocalized with Kit in ICC-MY (Fig. 5A), but not in ICC-DMP (Fig. 5B) in whole mount preparations, which strengthened our previous observations from thin cross sections. In the proximal and distal colon, a different pattern was observed, NBCe1-IR was not observed in Kit-positive cells in the myenteric region (Figs. 6A and 7A) or in the circular muscle layers (Figs. 6B and 7B). However, in submuscular region, on the submucosal boundary, NBCe1-IR colocalizes with Kit in ICC-SMP (Figs. 6C and 7C), which generate electrical slow waves in the colon (11). In the myenteric and submuscular plexuses of the colon, Kit-negative cells with a neuronal morphology appeared to express NBCe1 (arrows, Figs. 6A and 7A). We confirmed the identity of these cells by labeling the tissues with type I antineuronal nuclear antibody (ANNA-1), which recognizes neuron-specific HuC/D expression (41). We observed that NBCe1-IR in the myenteric and submucosal plexuses was restricted to ANNA-1-positive neurons but not all neurons were NBCe1-positive (Fig. 8).

Fig. 5.

NBCe1 expression localizes in ICC-MY networks in mouse small intestine. Confocal image stacks of whole mounts of adult mouse small intestine. A: NBCe1 immunoreactivity (red) colocalized with Kit (green) in ICC-MY of the mouse small intestine. Kit-negative cells with neuronal morphology (white arrows) were detected. B: NBCe1 immunoreactivity is not present in Kit-positive ICC-DMP of the mouse small intestine. Scale bar = 50 µm. Images representative of similar data from n = 4 mice. ICC-DMP, deep muscular plexus ICC; ICC-MY, interstitial cells of Cajal in the myenteric region.

Fig. 6.

NBCe1 immunoreactivity localizes in ICC-SMP networks in mouse proximal large intestine. Confocal image stacks of whole mounts of adult mouse proximal colon. A: NBCe1 immunoreactivity (red) was not present in Kit-positive ICC-MY (green) of the mouse proximal colon but was detected in Kit-negative cells with neuronal morphology (white arrows). B: NBCe1 immunoreactivity was not present in Kit-positive ICC-IM of the mouse proximal colon. C: NBCe1 immunoreactivity colocalized with Kit in ICC-SMP of the mouse proximal colon. Scale bar = 50 µm. Images representative of similar data from n = 4 mice. ICC-IM, intramuscular Interstitial cells of Cajal; ICC-MY, interstitial cells of Cajal in the myenteric region; ICC-IM, intramuscular Interstitial cells of Cajal; ICC-MY, interstitial cells of Cajal in the myenteric region.

Fig. 7.

NBCe1 immunoreactivity localizes in ICC-SMP networks in mouse distal large intestine. Confocal image stacks of whole mounts of adult mouse distal colon. A: NBCe1 immunoreactivity (red) was not present in Kit-positive ICC-MY (green) of the mouse distal colon but was detected in Kit-negative cells with neuronal morphology (white arrows). B: NBCe1 immunoreactivity was not present in Kit-positive ICC-IM of the mouse distal colon. C: NBCe1 immunoreactivity colocalized with Kit in ICC-SMP of the mouse distal colon. Scale bar = 30 µm. Images representative of similar data from n = 4 mice. ICC-IM, intramuscular Interstitial cells of Cajal; ICC-MY, interstitial cells of Cajal in the myenteric region; ICC-IM, intramuscular Interstitial cells of Cajal; ICC-MY, interstitial cells of Cajal in the myenteric region.

Fig. 8.

NBCe1 immunoreactivity localizes in Hu C/D-positive cells in myenteric region of mouse large intestine. Confocal image stacks of the myenteric region in whole mounts of adult mouse large intestine. NBCe1 immunoreactivity (red) was detected in a set of Hu C/D-positive myenteric neurons (cyan) of the mouse large intestine (white arrows). Scale bar = 30 µm. Images representative of similar data from n = 3 mice.

No Overlap of NBCe1 and PDGFRα+- Immunoreactivity

The morphology of NBCe1-positive cells in the tunica muscularis of the gastrointestinal tract resembles both ICC and PDGFRα+/“fibroblast-like cells” (13). To determine whether NBCe1 is restricted to ICCs, we labeled for PDGFRα and NBCe1 and imaged the immunofluorescent signal by high-resolution confocal microscopy. We observed no significant overlap between PDGFRα-IR and NBCe1-IR in any part of the tunica muscularis of the stomach, small intestine, and large intestine (Fig. 9 shows a representative image from the small intestine).

Fig. 9.

cNBCe1 immunoreactivity does not colocalize with PDGFRα-IR in mouse small intestine. High-resolution confocal image stack of whole mounts of adult small intestine. NBCe1 immunoreactivity (red) was not present in PDGFRα-positive cells (green). Inset shows these two different populations of cells are very close, but PDGFRα-positive cells do not express NBCe1. Scale bar = 10 µm. Images representative of similar data from n= 3 mice. PDGFRα, platelet-derived growth factor-α-positive fibroblast-like cells.

NBCe1-IR Colocalizes with ANO1 in ICCs of the Myenteric Region of Human Small Intestine

NBCe1-IR in thin sections of human small intestine was detected in the epithelial cells of the mucosa (Fig. 10A), as expected, in some myenteric neurons (white arrows) and also in ANO1-positive ICCs of the myenteric region (yellow arrows, Fig. 10B). In human small intestine, colocalization of NBCe1-IR and ANO1-IR was not as complete as that observed for the mouse small intestine with some ICC-MY appearing to be negative for NBCe1. Septal ICCs were NBCe1-IR negative (green arrows).

Enrichment of IRBIT-Regulated Slc4a4 Variants in Mouse Small Intestinal ICCs and in Tunica Muscularis of Human Small Intestine

Our immunolabeling results clearly establish that NBCe1 in ICCs is restricted to ICCs responsible for slow wave generation. Knowing that at least three functionally different variants of NBCe1 are generated from the Slc4a4 gene (19), we investigated the identity of the isoform expressed in ICCs. We took two approaches to address this question. Initially, we determined the relative enrichment of actively translated mouse Slc4a4 transcripts from Kit-positive cells in mouse small intestinal tunica muscularis using Kit^creERT2/+, Rpl22^{tm1.1Psam/SjJ} mice (38). RNA immunoprecipitated using anti-HA antibodies from the gastric tunica muscularis of tamoxifen-treated mice was 15- and 53-fold enriched for the ICC-specific markers Ano1 and Kit, respectively, relative to the levels in the RNA isolated from the whole tissue (“input”). Primers directed against all variants of Slc4a4 also identified significant enrichment of Slc4a4 in Kit-expressing cells compared with expression in the input (Fig. 11B). Using primers targeted against the junctions between alternatively spliced exons (see Fig. 11A, Table 3), we found that transcript, Slc4a4-203 (variant A), was significantly less abundant than transcripts, Slc4a4-207 and Slc4a4-208 (variants B and C), in the RNA pulled down from Kit-expressing cells (Fig. 11B). The specific enrichment of the Slc4a4-207 (B) and Slc4a4-208 (C) variants in the small intestine was confirmed in identified pacemaker ICC-MY by RT-PCR amplification from single, dissociated GFP-positive ICCs from Kit^tm1Rosay mice using slightly modified primers (Fig. 12A, Table 3). PCR products derived from the Slc4a4-208 (C) variant mRNA but not Slc4a4-203 (A) or Slc4a4-207 (B) variants were amplified from ICC-MY identified by the expression of Kit and Slc12a2 (NKCC1) but not Tacr1 (neurokinin-1 receptor) (Fig. 12B). NKCC1 has been previously identified as a marker of ICC-MY, and Tacr1 is enriched in nonpacemaker ICC-DMP in mouse small intestine (15, 42, 43). These studies indicate that the variants of NBCe1 that are not constitutively active but regulated by IRBIT are enriched in ICC-MY.

Fig. 11.

Quantification of the abundance of Slc4a4 transcripts in Kit-positive cells of the mouse small intestinal tunica muscularis. A: map of primer sequences chosen to specifically amplify variants identified by ENSEMBL nomenclatures. B: qRT-PCR analysis of Slc4a4 transcripts expressed in interstitial cells of Cajal after immunoprecipitation (IP) of polysomes from Kit^creERT2/+, Rpl22^{tm1.1Psam/SjJ} mice small intestine. Relative expression of the Slc4a4 variants in Kit-positive cells was determined by normalizing the abundance of the specific isoforms to total Slc4a4 expression in the pulldown samples and comparing this to the abundance in the total isolated tissue using the ΔΔCt method. The immunoprecipitated RNA samples were compared to the input sample in each case. The ENSEMBL-annotated transcripts analyzed by qRT-PCR were as follows: Slc4a4-203 (variant A), Slc4a4-207 (variant B), and Slc4a4-208 (variant C). **P < 0.005; *P < 0.05. Friedman test for multiple comparisons. Values are the means ± SE; n = 4. qRT-PCR, real-time quantitative reverse transcription polymerase chain reaction.

Fig. 12.

Expression of Slc4a4 isoforms in pacemaker, Kit-positive, ICC-MY. A: schematic of the Slc4a4 gene with highlighted differences between variants at the 5′-end and 3′-end. The arrows indicate the location of the primers used for RT-PCR experiments. Primers were designed to amplify the regions in exon 4 specific to variant A, the boundary between exons 23 and 25 in variants A/B and C and across the boundary between exons 23 and 24 in variants A and B. B: agarose gel showing the expression of only isoform C of Slc4a4 in pacemaker (expressing Slc12a2, NKCC1 but not Tacr1, the NK1 receptor), Kit-positive, ICC. ICC, interstitial cells of Cajal; ICC-MY, interstitial cells of Cajal in the myenteric region; qRT-PCR, real-time quantitative reverse transcription polymerase chain reaction.

Transcriptional variants of the human SLC4A4 gene have also been identified (19); therefore, we determined the relative abundance in normal human jejunal tunica muscularis tissue of the constitutively active SLC4A4-202, variant A, and the regulated isoforms, SLC4A4-201, variant B, and SLC4A4-204, variant C. Using primers targeted to amplify these variants (Fig. 13A, Table 3), we detected low levels of the SLC4A4-202 (A) variant and ∼83- and 336-fold higher levels of the SLC4A4-201 (B) and SLC4A4-204 (C) variants, respectively (Fig. 13B, n = 4, BC and AB different from A, P < .05, Friedman test). These data indicate that NBCe1 is enriched in the small intestinal tunica muscularis of humans as well as mice.

In summary, our data indicate enrichment of the NBCe1, electrogenic Na⁺/HCO₃⁻ cotransporter in subsets of ICCs that are identified as pacemaker cells in the mouse gastrointestinal tract based on RNA and protein expression.

DISCUSSION

This study demonstrates that the expression of the electrogenic Na⁺/HCO₃⁻ cotransporter protein, NBCe1 encoded by the gene Slc4a4 (34), is enriched in subsets of ICCs in the stomach, small intestine, and colon of mice. Specifically, ICCs known to generate electrical slow wave activity in the stomach, small intestine, and colon (10, 44, 45) expressed levels of NBCe1 protein that were detectable using specific, selective antisera, whereas other subtypes of nonpacemaker ICCs were not NBCe1-immunopositive, which indicates a significant role for NBCe1 in electrical pacemaking of the gastrointestinal tract. Furthermore, mouse ICCs expressed higher levels of the Slc4a4-207 and Slc4a4-208 (variants B and C) transcripts of the Slc4a4 gene that generate IRBIT-regulated isoforms of NBCe1 (37), when compared to expression of the Slc4a4-203 (variant A) transcript, which generates a constitutively active isoform of the transporter (46). In the tunica muscularis of human small intestine, NBCe1-IR colocalized with ANO1 in myenteric ICCs, and the IRBIT-regulated isoform of SLC4A4, SLC4A4-201 (variant B) was expressed at significantly higher levels than the other two major transcripts of the SLC4A4 gene.

Although it was previously shown that Slc4a4 RNA is enriched in ICC-MY in mouse small intestine compared with ICC-DMP (18), and the expression of the cognate NBCe1 protein has been reported in epithelial cells of duodenum (26), the broader distribution of this protein in other regions of the gastrointestinal tract has not been reported. There have been no previous reports of SLC4A4 gene expression in human gastrointestinal tunica muscularis.

The distribution pattern of NBCe1 is significant because it matches with the distribution of ICCs that have been determined to generate the electrical slow wave or pacemaker potential in all regions of the gastrointestinal tract. For example, NBCe1 is expressed in ICC-MY of the gastric body and antrum and in the small intestine but not in ICC-MY of the large intestine, whereas NBCe1 is absent in ICC-DMP of the small intestine but present in the ICC-SMP of the large intestine. NBCe1 was not detected in ICC-IM of any region of the GI tract. In the small intestine and stomach, it is well established that the ICC-MY are responsible for the generation of the electrical slow wave (1, 2, 8), whereas ICC-DMP and ICC-IM are linked with neuromuscular neurotransmission (47–50). In the large intestine, large, low-frequency slow waves are generated by the ICC-SMP (9–11), whereas ICC-MY in the large intestine appear to generate much higher frequency, short-duration oscillations in membrane potential that do not resemble classical slow waves (51, 52). In small rodents, ICC-IM in all regions of the gastrointestinal tract play a role in neuromuscular neurotransmission (49, 53) and mediate regenerative potentials (8, 54) but are not thought to generate slow wave activity. NBCe1-IR was absent from all ICC-IM of mouse gastric antrum in our study, and this suggests that the regenerative potentials in this subset of ICCs are not reliant on NBCe1 activity.

This variety of ICC function and distribution highlights the diversity of ICC types and their contributions to gastrointestinal motility. Another way to define ICCs is by differences in the developmental regulation of subtypes of ICCs. However, the expression of NBCe1 does not correlate with, for example, differences in the dependence of ICCs on stem cell factor/Kit receptor tyrosine kinase as indicated by differential sensitivity to reduced Kit signaling in mutant rodents with loss-of-function mutations in Kit (3), and the effects of Kit-neutralizing antibodies on ICCs survival in vivo (55). NBCe1-positive ICC-MY are absent, whereas NBCe1-negative ICC-DMP are conserved in the small intestine of W/Wv mice, which have impaired Kit signaling (1,2). In the colon, the ICC-SMP found to be NBCe1-positive in our study are conserved in the W/Ws rat (56). This indicates that the transcriptional regulation of ICCs development by Kit signaling is separate from the transcriptional regulation of ICCs development that leads to pacemaker function and that the diversity of ICCs function is a mixture of lineage and local environmental factors.

The mechanism for generation of slow waves is not fully understood, but it is known that it relies on Ca²⁺ release from IP3 receptor-gated intracellular Ca²⁺ stores, normal mitochondrial function, cation influx through nonselective ion channels, and Ca²⁺-activated Cl⁻ transport by Ano1 (24, 44, 49). However, the proteins responsible for this function are expressed at similar levels in pacemaker and nonpacemaker ICCs, as indicated by differential display RNA expression studies (15, 18, 42). Slc4a4 is one of several ion channels and ion transport proteins that are differentially upregulated in ICC-MY compared with ICC-DMP (18), suggesting that it, along with NKCC1 (Slc12a2) (15) and the T-type Ca²⁺ channel (Cacna1h) (57), is a part of molecular apparatus that combines to cause pacemaker activity.

The expression of Slc4a4 in pacemaker ICCs is consistent with the well-established requirement for extracellular HCO₃⁻ for recording normal slow wave activity (31) and also indicates a central role for regulation of intracellular pH homeostasis in the generation of the electrical slow wave. Electrical activity of other excitable cells including neurons and cardiac myocytes is enhanced in the presence of HCO₃⁻ and modified by NBCe1 activity as well as other transport proteins that regulate intracellular pH homeostasis (23, 29). NBCe1 is localized to the T-tubules of rat cardiac myocytes, and NBCe1 activity is linked to shortening of the cardiac action potential duration and changes in resting membrane potential in cat and rat cardiac myocytes (58, 59). Although the mechanism by which NBCe1 may contribute to pacemaker activity remains to be established, changes to intracellular pH are known to modify key steps in generation of pacemaker activity including IP3 receptor-mediated Ca²⁺ release (60, 61) and mitochondrial function (62). It is also known that increased intracellular pH enhances intercellular coupling between ICC-MY in small intestine (63), so it is likely that local or global changes in intracellular pH are significant factors in determining the presence and nature of pacemaker activity in ICCs.

The isoforms of NBCe1 present in ICCs indicate an additional potential link between the role of IP3-receptor activation and generation of pacemaker activity. Both the proteins encoded by mouse Slc4a4-207 and Slc4a4-208 (variants B and C) are activated by IRBIT (37), which will be released from the IP3 receptor in the presence of IP3 (64). This can potentially lead to a feed-forward amplification of IP3-receptor signaling due local HCO₃⁻ influx, alkalinization, and increased Ca²⁺ release, potentially resulting in sustained Ca²⁺ transients and consequently pacemaker activity. This model requires close spatial association of the proteins in a subcellular, pacemaker unit of ICCs. Slc4a4-208 (variant C) transcripts encode a protein that contains a type I PDZ domain, which is responsible for the association of the NBCe1 protein with scaffolding proteins in membrane domains of polarized cells (46). This PDZ domain can also result in NBCe1 protein being restricted to membrane regions with other key molecules for generating pacemaker activity.

NBCe1-IR was highly enriched in ICCs, but was not detected in platelet-derived growth factor-α-positive fibroblast-like cells (PDGFRα⁺/FLCs/telocytes), which reside in similar anatomical niches, have similar morphologies, and are closely associated with ICCs (65, 66). Some NBCe1-IR was observed in subset of enteric neurons, most notably in the large intestine but also in other regions. These cells could be clearly distinguished from ICCs by their morphology and ANNA1 immunoreactivity, and NBCe1 expression is consistent with a role for intracellular pH regulation in neuronal excitability in general (23).

The specific function of NBCe1 in ICCs physiology and development remains to be determined. The differential presence of NBCe1 in pacemaker ICCs in mice and the known properties of HCO₃⁻ transport and pH regulation point toward a role in normal pacemaker function of ICCs, although the nature of that role is not clear. Our studies on human tissue confirmed that SLC4A4 transcripts for the regulated isoforms of NBCe1 were enriched in the tunica muscularis and that NBCe1 colocalizes to a subset of ICC-MY of the human small intestine. This colocalization was not complete, possibly because of species differences or because the human tissues used for immunolabeling were de-identified but appeared to be derived from the distal small intestine, whereas the mouse studies were done on proximal small intestine. A prospective study on the distribution and role of NBCe1 in human gastro-intestinal function using material from clearly identified tissues is merited based on our results from mice and these initial studies on human tissues. Mutations in the SLC4A4 gene that affect NBCe1 function associate with severe renal dysfunction and glaucoma in humans (24), and our studies are consistent with these and other mutations in SLC4A4 potentially contributing to disruption of gastrointestinal motility. The penetrance of loss-of-function mutations in membrane protein genes has been observed to vary in different tissues; for example, loss-of-function mutations in a voltage-gated Na⁺ channel, SCN5A, associate with irritable bowel syndrome and cardiac long QT syndrome, but some mutations present with a severe gastrointestinal phenotype but not a cardiac phenotype and vice versa (67, 68). Mutations that alter SCN5A function consistently result in cardiac arrhythmias in male patients but less commonly or have less severe phenotypes in females, whereas mutations in SCN5A that have been identified in patients with IBS were more common in female patients (67).

In conclusion, the electrogenic Na⁺/HCO₃⁻ cotransporter, NBCe1, encoded by the gene Slc4a4, is selectively expressed in ICCs with electrical pacemaker functions. The gene transcripts that encode regulated isoforms of NBCe1 are enriched in ICCs. These observations point toward a dynamic role for NBCe1 function in the generation of pacemaker activity in the gastrointestinal tract.

GRANTS

This work was supported by an NIH R01 to S. J. Gibbons and G. Farrugia—DK-57061 and by the Mayo Clinic Center for Cell Signaling in Gastroenterology P30DK084567.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

M-G.C.A., G.F., and S.J.G. conceived and designed research; M-G.C.A., A.M., S.T.E., and P.R.S. performed experiments; M-G.C.A. and A.M. analyzed data; M-G.C.A. and S.J.G. interpreted results of experiments; M-G.C.A. prepared figures; M-G.C.A. and S.J.G. drafted manuscript; M-G.C.A., A.M., S.T.E., P.R.S., C.E.B., H.L.H., M.F.R., G.F., and S.J.G. edited and revised manuscript; M-G.C.A., A.M., S.T.E., P.R.S., C.E.B., H.L.H., M.F.R., G.F., and S.J.G. approved final version of manuscript.