The Slc9a family of Na + /H + exchangers (NHEs) plays a critical role in neutral sodium absorption in the mammalian intestine as well as other absorptive and secretory epithelia of digestive organs. These transport proteins mediate the electroneutral exchange of Na + and H + and are crucial in a variety of physiological processes, including the transepithelial Na + and water absorption, fine tuning of intracellular pH, cell volume control and systemic electrolyte, and acid-base and fluid volume homeostasis. They also secondarily affect other cellular transport mechanisms as well as cell survival, motility, adhesion, and repair mechanisms. In this chapter, we review the role of the Na + /H + exchange mechanism as it relates to the physiology of organs and cells involved in nutrient intake and absorption, and we describe physiological and molecular aspects of individual isoforms, including their structure, tissue-, and subcellular distribution, as well as their regulation by physiological stimuli at the transcriptional and posttranscriptional levels. Consequences of gene-targeted mutation of individual isoforms are discussed in the context of the physiology of digestive organs. Where available, we also provide a review of pathophysiological states related to aberrant expression and/or activity of NHEs within the confines of the digestive system.
KeywordsSLC9, Na/H exchanger gene family, Antiporter, Na absorption, Intracellular pH, Epithelium, Membrane transport
NIH grant R01DK041274 is acknowledged for financial support.
The first physiological observations suggesting the existence of a Na + /H + exchange mechanism in mammalian membranes were made by Mitchell and Moyle and by Brierley et al. in rat liver and cow heart mitochondria, respectively. These observations were soon followed by similar findings in prokaryotic plasma membrane by Harold and Papineau in Streptococcus faecalis and by West and Mitchell in Escherichia coli. In 1987, Goldberg et al. cloned the first Na + /H + antiporter from E. coli , later termed nha A. Its mammalian counterpart was cloned shortly after by Sardet et al. and reported in 1988 and 1989. Cloning of this Na + /H + exchanger (NHE), initially described as growth factor-activatable Na + /H + antiporter, and later termed NHE-1, initiated the explosion of knowledge about function, structure, and regulation of what was soon determined to be a family of proteins involved in membrane Na + /H + antiport mechanism. Families of NHE were identified in bacteria, yeast, plants, and animals.
Although the basic principle of Na + /H + exchange is consistent among species, there are also some considerable differences. In vertebrate animals, the ubiquitous plasma membrane Na + /K + -ATPase mediates the efflux of 3Na + and the influx of 2K + , a process coupled to the hydrolysis of ATP. This electrogenic Na + /K + exchange establishes a Na + gradient across the plasma membrane that is used by the cell for a number of physiological functions. These processes are largely dependent on Na + /H + exchange, a mechanism critical in the regulation of intracellular and systemic pH, cell volume, absorption, and reabsorption of sodium in the intestinal tract and kidney, respectively. This physiological need dictates the polarity of the exchange, with extracellular sodium being exchanged for intracellular proton. On the other hand, in lower organisms such as bacteria, plants, and yeast, the ability to withstand extreme hypersaline or hyperalkaline conditions results from developed Na + /H + antiport mechanisms leading to the net uptake of protons and net loss of sodium. In addition to differences in Na + /H + exchange polarity, its stoichiometry also varies. For example, while all the described mammalian NHE are electroneutral, E. coli bacteria exhibit electrogenic Na + /H + exchange, with a stoichiometry of 1Na + /2H + and 2Na + /3H + for NhaA and NhaB isoforms, respectively (see Fig. 56.1 ).
In mammalian cells, tight pH i regulation is crucial for cell function and survival, as its subtle changes may significantly affect cellular physiology. Variations in pH i mediated by NHE mechanism have been associated with basic biological processes such as proliferation, cell adhesion and migration, cell volume regulation, and transepithelial electrolyte and water transport. This chapter focuses primarily on the structure, function, and regulation of NHE expressed in the mammalian gastrointestinal (GI) tract. The reader is referred to the following publications for an overview or original articles on plants, Caenorhabditis elegans , and procaryotic Na + /H + antiport, respectively. Comprehensive reviews on mammalian Na + /H + exchange were published by Orlowski and Grinstein, Zachos et al., Donowitz et al., and by Gurney et al. with a particular emphasis on intestinal Na + /H + exchange. Phylogenetic analysis of evolutionary relations among NHE sequences from all phyla, including those sequences well characterized as well as those electronically annotated based on sequence similarity, has been published by Brett et al.
Mammalian Monovalent Cation Proton Antiporter Superfamily
As described by Brett et al., > 550 sequence entries could be identified as putative NHE by automated annotation projects on the basis of sequence conservation within 10 transmembrane domains TM3–TM12 within the NH 2 -terminal half of proteins. The classification of membrane transport proteins remains an evolving process and is periodically updated with the new and improved evolutionary bioinformatics tools. Based on the most recent classification (Ref. www.tcdb.org /), monovalent cation proton antiporter (CPA) superfamily is now divided into three CPA families: CPA1 (TC# 2.A.36), CPA2 (TC# 2.A.37), and CPA3 (TC# 2.A.63). According to the classification current at the time of writing this chapter, CPA2 and CPA3 families include primarily bacterial, fungal, and plant transporter proteins, with one exception, human transmembrane and coiled-coil domain 3 (TMCO3) and a member of CPA2 family expressed in expressed in the human cornea, lens capsule, and choroid-retinal pigment epithelium. CPA1 family is the largest of the three and includes proteins derived from Gram-positive and Gram-negative bacteria, blue-green bacteria, archaea, yeast, plants and animal, including the nine mammalian NHE paralogs from the SLC9A subfamily (NHE1–9), NHA1, and NHA2 in SLC9B subfamily, and two sperm-specific NHE’s (NHE10, NHE11) from the SLC9C subfamily, which were previously categorized in the NaT-DC family. CPA2 and CPA3 families of the monovalent CPAs, as well as NHE10 (SLC9C1) and the putative NHE11 (SLC9C2) will not be discussed in this chapter. Within the CPA1 family, NHA1 (Slc9b1) was identified in the spermatozoa where it controls sperm motility beyond cauda epididymis. NHA2 (or NHE Domain-Containing protein 2, NHEDC2, Slc9b2), was described as an amiloride-insensitive Li + -NHE. In the sperm, along NHA1, it contributes to its motility, although its mouse tissue expression pattern is broad and includes the GI tract, liver, pancreas. In transfected polarized renal epithelial cells, it was primarily localized to the apical membrane. Kondapalli et al. showed recently that the endogenous NHA2 was expressed in the distal convoluted and connecting tubules where it was upregulated in response to high-Na + diet and thus postulated to play a role in salt tolerance and in the pathogenesis of essential hypertension. In the pancreas, NHA2 is expressed in insulin-producing β-cells, where it was found to play permissive roles in sulfonylurea- and secretagogue-induced insulin secretion. Beyond 5 months of age, NHE2-deficient mice develop glucose intolerance, which is further exacerbated by a high-fat diet. NHA2 was also found in osteoclasts, where it was upregulated by RANK ligand. The same group localized NHA2 expression in the osteoclast mitochondria and showed that its knockdown reduced osteoclast differentiation and resorptive function. However, with improved methods of NHA2 detection, Hoffstetter et al. showed that NHA2 colocalizes with late endosomal and lysosomal markers, is highly enriched at the basolateral membrane of polarized osteoblast, and is not expressed in the mitochondria. While upregulation by RANKL was confirmed, NHA2-deficient mice did not demonstrate any skeletal defects, and the authors concluded that NHA2 is dispensable for osteoclast differentiation and bone resorption in mice. Although NHA2 protein has been detected in the stomach, jejunum, and colon, its roles in the mammalian gut remain unknown. Intriguingly, in Drosophila gut epithelium, both Nhe1 and Nha2 were required for the protection against high luminal Na + load, despite their distinct transport properties (as H + -Cl − cotransporter and an NHE, respectively), thus implicating its role in response to osmotic stress and Na + tolerance. Since, with the exception of NHA2, all of the NHE expressed in the GI tract belong to the CPA2 family, NHE1–9 isoforms will remain the primary focus of this chapter’s discussion.
Following the cloning of NHE1, later considered a prototypical mammalian NHE, eight other isoforms from diverse species have been described by various methodological approaches. The nine proteins belonging to the SLC9A subfamily of mammalian NHE demonstrate considerable variation in their amino acid sequence, ranging from under 12% (hNHE1 vs. hNHE9) to over 70% identity (hNHE6 vs. hNHE7). Eight of them have been detected in the GI tract ( Fig. 56.2 ) with segmental differences, crypt-villus gradients of expression, and different cellular localizations, all of which determines their proven or alleged functions. An additional isoform of chloride-dependent NHE has been cloned from the rat colon. Its sequence shared 100% homology of the 375 N-terminal amino acids (aa) with NHE1, and a divergent 63aa sequence in the C-terminal portion of the protein, suggestive of an alternative splice variant of the NHE1 gene. However, while the rNHE1 gene is located on chromosome 5q36, the cDNA coding for the novel C-terminal 63aa aligns fully with chromosome 17p14, thus representing a chimeric noncolinear transcript. This cDNA may represent either a product of trans -splicing of pre-mRNA or a cloning artifact and has not been independently verified.
Membrane Topology and Functional Domains
Although all the nine NHEs have not been systematically studied in terms of their secondary structure and membrane topology, modeling algorithms predicting hydrophobic and hydrophilic regions of the proteins suggest the same general arrangement. According to the modeling, approximately 60% of the amino-terminal of the protein is amphipathic and contains 10–12 membrane spanning α-helices, which are relatively conserved among different isoforms. Much more hydrophilic and less conserved carboxyl-terminus faces the cytoplasm (see Fig. 56.3 ). This region of NHE proteins, as determined empirically and through prediction sequence analyses, contains multiple phosphorylation sites and sites responsible for interaction with accessory proteins, all of which are believed to serve regulatory functions.
Of the nine NHEs cloned to date, two-dimensional structure of NHE1 and NHE3 has been most extensively studied utilizing a variety of experimental approaches including cysteine substitution and accessibility, C-terminal truncation, identification of glycosylation sites, proteolytic cleavage, and epitope immunolocalization. Not all of these studies confirm this general model of NHE membrane topology, particularly with regard to the C-terminal region. The notion of entirely cytoplasmic location of this domain has been contended both in NHE1 and NHE3, by Khan and Biemesderfer et al., respectively. Both studies argued that truncations or amino acid substitution may alter the natural structure and folding of the protein, and suggested that epitope immunolocalization of native proteins indicates at least some extracellular epitopes located with the C-terminal tail. This observation seems to be supported by the observations with Nhx1, a yeast homologue of mammalian NHEs, which has been shown to be glycosylated on two asparagine residues located within the C-terminus. However, considering the fact that the C-terminal domain of NHE1 or NHE3 does not appear to have hydrophobic regions of sufficient length to form a membrane-spanning domain (15–20 amino-acids), and functional analyses of regulatory domains within this region suggest multiple interactions with cytoplasmic and cytoskeletal factors, it remains unclear whether the controversy over the topology of NHE C-terminus is a result of methodological differences or whether it reflects actual variations in membrane orientation.
Due to the highly variable amino acid sequence among various NHE isoforms, the N-terminus may contain a cleavable signal peptide. In NHE1, for example, the cleavage would occur before the experimentally confirmed glycosylation site in the first extracellular loop ( Fig. 56.3 ). If this hypothetical scenario were true, the predicted topology of NHE1 and NHE3 would change to an 11-membrane-spanning domain configuration. Experimental evidence from studies on NHE1 protein utilizing cysteine substitution suggests, however, that the N-terminus is retained in the mature protein and remains in the cytosol. On the other hand, experiments with in vitro translated NHE3 point to cleavage of the N-terminal signal peptide during processing in microsomes. Based on the results of these studies, membrane topology predictions should not be generalized to all members of NHE family and secondary structures of each of the remaining NHE proteins would have to be experimentally addressed. Some of the different models of NHE1 membrane topology have been reviewed by Kemp et al.
NHE1 and NHE3 have been demonstrated to form homodimers in the cytoplasmic membrane. The formation of dimers appears to depend on protein-protein interaction within the amphipathic membrane-spanning region of the proteins, but individual subunits were found to function independently within the complex. Kinetic analyses of renal brush-border membrane Na + /H + exchange also suggested the presence of dimers and tetramers, dependent on the outside Na + concentration.
Based on the secondary structure of NHE proteins and their membrane topology, two major domains can be distinguished: an N-terminal amphipathic region including all 10–12 membrane-spanning domains, and a C-terminal hydrophilic tail. The latter segment is believed to be intracellular and confers regulation of NHE activity by direct or indirect interactions with kinase, cytoskeletal, and other proteins. Within the amphipathic region, transmembrane domain 4 (TM4) is crucial for NHE1 function, with the residues Phe161, Phe162, Leu163, and Gly173 affecting affinity for Na + and/or its resistance to inhibitors. Slepkov et al. showed that within this transmembrane helix, Pro167 and Pro168 are critical in NHE1 activity, expression, and membrane targeting. Within the seventh transmembrane domain (TM7), Glu262 and Asp267 are indispensable for NHE1 activity, with their charge and acidity being the most critical. Transmembrane domain 9 (TM9) contains a sequence conferring sensitivity to antagonists. A hybrid NHE1, in which this transmembrane helix has been replaced with analogous segment of the amiloride-resistant NHE3, was resistant to amiloride, ethylisopropylamiloride, HOE694, and cimetidine. Within this transmembrane domain, His349 may be one of the critical moiety bestowing sensitivity to amiloride compounds, as described by Wang et al. Mutation of amino acids Tyr454 and Arg458 within the eleventh transmembrane domain (TM11) has shown that both amino acids are essential in targeting NHE1 to the cell surface. TM11 and its neighboring intracellular loop (IL5) have also been implicated in pH sensing. Mutation of Gly455 or Gly456, although it did not affect the protein’s affinity for Na + or H + ions, resulted in an alkaline shift in NHE1 pH-dependence. In contrast, mutation of Arg440 in IL5 had the opposite effect. Based on these data, it has been postulated that both Arg440 in IL5 and glycine residues in the conserved segment of TM11 constitute the putative pH i sensor in NHE1. Other NHE isoforms have not been studied in such detail.
The long, hydrophilic cytosolic domain of NHE1 (amino acids 500–815), which regulates the activity of the amphipathic N-terminal domain, is a target for phosphorylation by protein kinases and participates in binding of regulatory proteins. Several protein kinases are thought to phosphorylate NHE1, including Erk1/2, p90rsk, p160ROCK, p38, and a Nck-interacting kinase. By changing NHE1 pH dependence, phosphorylation increases its activity in a more alkaline pH. A number of regulatory proteins have been demonstrated to bind to the cytosolic tail of NHE1, including calmodulin, calcineurin homologous protein (CHP), tescalcin, and carbonic anhydrase II. The reader is referred to review articles published by Putney et al., Slepkov and Fliegel, Baumgartner et al., Malo and Fliegel, and references therein for more detailed descriptions of modes of NHE1 activity regulation. The role of the cytoplasmic C-terminal domain of NHE3 functioning as a scaffold which binds multiple regulatory proteins and links NHE3 to the cytoskeleton has been reviewed in detail by Donowitz et al. and is described in a later section on posttranscriptional regulation of NHE3 activity.
Transport Characteristics and Pharmacology
The steady-state velocities of most NHE isoforms show a saturating, first-order dependence on the outside Na + o concentration (with K Na values 3–50 mM), consistent with simple, saturating, Michaelis-Menten kinetics, which is indicative of a single binding site. The kinetics of NHE4 isoform seems more complicated. Under unstimulated isotonic conditions, NHE4 appears inactive; however, under hypertonic stress, the kinetics of NHE4-mediated Na-uptake follow a sigmoidal rather than hyperbolic curve with increasing concentrations of extracellular Na + , a phenomenon characteristic of allosteric regulation. Most plasmalemmal NHE isoforms are specifically transporting Na + in exchange for H + , with much lower efficiency toward Li + or NH 4 + and essentially no affinity for K + . NHE4 is an exception to this rule, since it shows similar exchange rates for Na + and Li + , and an even higher rate for K +, suggesting that this isoform functions as a nonspecific cation/H + exchanger. Similarly, NHE8, which is described as a both intracellular and plasmalemmal isoform, is capable of K + /H + exchange when reconstituted in artificial liposomes. NHE7, an exclusively organellar isoform localized in the trans-Golgi, network, also functions primarily as a K + /H + exchanger, a feature likely shared by other organellar NHE isoforms.
Contrary to a typical first-order dependence of exchange kinetics for Na + , hydrogen concentration dependence does not follow a simple Michaelis-Menten equation which assumes no cooperativity, but rather displays characteristics typical of allosteric effect, with more than one binding site for H + . The kinetic analysis of Na + /H + exchange appears to be in agreement with structural data suggesting a presence of a H + i sensor in the 11th transmembrane domain, and an element regulation pH set point located in the intracellular loop 5 (see Section 56.2 ). The allosteric regulation of Na + /H + exchange by intracellular protons may not be a universal feature of all NHEs. As an example, NHE5 displays a simple first-order dependence on the H + i concentration, when expressed in fibroblasts, suggesting no cooperativity.
All well-characterized mammalian NHE have been classified as electrochemical-potential-driven transporters, and they have been categorized into the monovalent cation proton antiporter family (CPA1; 2.A.36), according to classification developed by Busch and Saier and endorsed by the Nomenclature Committee of the International Union of Biochemistry and Molecular Biology (NC-IUBMB). This classification is based on the fact that cation fluxes via NHE mechanism are driven exclusively by the transmembrane gradients of substrates and are only secondarily dependent on ATP. Although NHE proteins do not bind or consume ATP directly, cellular ATP depletion results in marked inhibition of NHE1, NHE2, and NHE3 activities, despite a maintained transmembrane H + gradient. ATP depletion affects NHE1 and NHE2 by reducing their sensitivity to intracellular pH (H + i ), while NHE3 is characterized by both impaired H + i sensing and reduced maximal velocity of transport ( V max ). Although the precise mechanism of these phenomena is not clear, a role for plasmalemmal phosphatidylinositol 4,5-bisphosphate (PIP 2 ) has been postulated. Since binding of PIP 2 to the C-terminus of NHE1 is critical in maintaining NHE1 activity and ATP depletion results in a decline in plasma membrane PIP 2 content, this could represent a potential mechanism for ATP-dependence of NHE1. It remains to be determined whether ATP depletion impairs NHE2 and NHE3 via similar mechanisms.
Chloride-dependent Na + /H + exchange has been functionally identified in the epithelium of the rat distal colon. This phenomenon was postulated to be the result of functional coupling of the chloride channel to a novel NHE isoform, later cloned as a putative Cl − -dependent NHE. Two other studies in the mouse, however, demonstrated no Cl − -dependent NHE in the colonic crypts. One explanation for this discrepancy is a possibility that not all Cl − replacements are equally inert, but this is not likely as two different anion substitutions with different permeabilities (nitrate or gluconate) produced the same results. Other explanations include the possibilities that the volume changes caused by Cl − depletion affect different cell types differently, or that coupled Cl − /anion exchange may be necessary for dissipation of the intracellular gradient in some cells. These discrepancies may also be due to species differences (rat vs. mouse), site of analysis (crypt base or midcrypt region), or simply the use of different methodological approaches. It has also been demonstrated that NHE1–3 isoforms can be Cl − dependent to a certain extent. This phenomenon is incompletely understood, but may involve impaired H + i sensing with depletion of intracellular Cl − . Overall, the described inconsistencies in descriptions of Na + /H + exchange Cl − dependence suggest that it may not be a ubiquitous function of colonic crypt epithelia.
Over the years, a number of Na + /H + exchange inhibitors have been developed, initially as an effort to inhibit NHE1 in cardiac ischemia/reperfusion injury, and later as potential adjuncts in anticancer therapy. The pharmacology of NHE inhibitors was reviewed by Masereel et al. Amiloride, a K + -sparing diuretic, was the first described NHE inhibitor, which could also inhibit a electrogenic Na + channels and the Na + /Ca ++ exchanger. NHE1 and NHE2 isoforms are the most sensitive to amiloride inhibition, whereas NHE3 and especially NHE4 are amiloride-resistant isoforms. The sensitivity of NHE8 and NHE9, two of the most recently cloned NHE isoforms, to currently known inhibitors has not been evaluated, and only limited information on sensitivity of NHE4 and NHE7 is available. Development of several pyrazine or phenyl derivatives of amiloride increased their potency toward NHEs, particularly NHE1, and more importantly increased their selectivity by eliminating the inhibitory potency toward the Na + channel and Na + /Ca 2 + exchangers. Of these molecules, DMA, EIPA, HOE-694, and HOE-642 are the most frequently used in experimental settings.
Several NHE inhibitors based on a bicyclic template have been introduced, such as zoniporide, SM-20550, BMS-284640, T-162559, or TY-12533. Other compounds not related to amiloride have also proven useful, especially S-3226, as the first NHE3-specific inhibitor. Cimetidine, clonidine, and harmaline, although not frequently used, have also been reported to act as weak and nonspecific inhibitors of Na + /H + exchange. More recently, ligustrazine (2,3,5,6-tetramethylpyrazine) and its analogs, along with several other compounds (e.g., KR-32560, KR32570, or KR-33028) have been introduced as NHE1 inhibitors. It is important to point out that the reported IC 50 values have been frequently derived from studies with forced expression NHEs in NHE-deficient fibroblasts, and significant differences in sensitivities of endogenous NHEs may not be uncommon.
Intestinal Na + absorption mediated by NHE3 has more recently become a target of commercial investigation, with the hope of developing novel treatment for hypertension and/or constipation-predominant irritable bowel syndrome (IBS-C). Two oral nonabsorbable NHE3 inhibitors have been developed, SAR218034 (SAR) and tenapanor. Pharmacokinetics studies indicated that they did not cross the intestinal barrier in biologically active doses. Tenapanor was well tolerated in phase I clinical study and both inhibitors increased fecal and reduced urinary Na + concentrations in rodents and humans. Not surprisingly, both drugs caused an increase in luminal fluid resulting from increased Na + , leading to loose stools. In a spontaneously hypertensive rat model, SAR in conjunction with NaCl-laden drinking water markedly reduced systolic blood pressure. In a rat model of chronic kidney disease model associated with hypertension, hypervolemia, cardiac hypertrophy, and arterial stiffening (salt-fed 5/6 nephrectomized rats), tenapanor reduced extracellular volume expansion, albuminuria, and blood pressure, in addition to promoting protective cardiorenal effects such as reducing left ventricular hypertrophy. Both drugs show enhanced effects if administered in conjunction with an angiotensin-converting enzyme inhibitor, which was deemed important in cases in which hypertension could not be controlled by the administration of a single medication. However, at the time of writing this chapter, the development of tenapanor and SAR, and the general concept of inhibition of intestinal Na + /H + exchange as an antihypertensive strategy, appears to have been abandoned by Ardelyx (Fremont, CA) and Sanofi (Paris, France), respectively. Ardelyx continues the investigation into the use of tenapanor and its close analogs in patients with IBS-C and for the treatment of hyperphosphatemia in end-stage renal disease patients on dialysis. In a phase 2, randomized, placebo-controlled efficacy and safety trial, 50 mg of tenapanor twice a day, response rate for IBS-C symptoms such as pain, discomfort, bloating, cramping, and fullness, was significantly higher in the tenapanor than placebo group, with diarrhea as the most frequent adverse effect. The concept behind the use of NHE3 inhibition in hyperphosphatemia is based on the described increased fecal P i excretion and reduced urinary P i excretion in NHE3 inhibitor-treated rats with chronic kidney disease with vascular calcification, in which tenapanor markedly reduced ectopic calcification and protected renal function. The mechanism of this phenomenon is not clear. However, in hemodialysis patients, tenapanor indeed moderately but significantly lowered serum phosphate level, albeit with diarrhea affecting as much as 68% of participants. It is not yet evident whether targeting NHE3 would be more efficacious and have less adverse effects that the P i binders currently used clinically.
Gastrointestinal NA + /H + Exchangers
NHE1, the first cloned mammalian NHE, remains the most extensively studied NHE isoform, although the preponderance of information on NHE1 expression, activity, and regulation comes from systems other than the GI tract, and as such may or may not be applicable to GI physiology and pathophysiology. Since the large body of knowledge about this isoform precludes us from including it in this chapter, the reader is referred to more focused reviews discussing various aspects of the biology of NHE1. Only a basic overview of this isoform is presented here, with a particular emphasis on its role in the physiology of the digestive tract.
Mammalian NHE1 is an 813–822 amino acid protein with a calculated molecular mass of ~ 91 kDa. NHE1 contains consensus sequences for both N- and O -linked glycosylation, and there is evidence that Asn-75 in the first extracellular loop of NHE1 is glycosylated, explaining the appearance of the mature 110 kDa form of NHE1 in Western blotting. Its membrane topology has been extensively studied and is schematically represented in Fig. 56.3 .
Tissue Distribution and Cellular Localization
NHE1 is expressed ubiquitously in almost all mammalian cell types where it resides exclusively on the plasma membrane. Dependent on the cell type, NHE1 tends to accumulate in distinct membrane domains. In polarized epithelial cells, NHE1 is expressed on the basolateral membrane ; in cardiac myocytes, it is concentrated around the intercalated disks and t-tubules ; while in fibroblasts, it is found along the border of lamellipodia. In the rat small-intestinal epithelium, no detectable difference in segmental expression of NHE1 mRNA has been described, with only a minor decrease in expression along the crypt-villus axis in the jejunum. Similarly, no longitudinal differences in NHE1 expression have been detected in the human intestine. This relatively uniform expression of NHE1 is consistent with its perceived role as a “housekeeping isoform” participating in the regulation of intracellular pH and volume.
NHE1 serves primarily to regulate intracellular pH, and its activation is associated with a number of downstream cellular events. The transient increase in pH i induced by NHE1 participates in cell proliferation and promotes transit through the G2-M checkpoint of the cell cycle. This finding may be related to the role NHE1 plays in proliferative responses of hepatocytes and hepatic stellate cells (HSCs) as described later in this chapter. NHE1 also appears to regulate cell differentiation, since deletion or inhibition of NHE1 has been shown to impair differentiation pathways. A role for NHE1 in apoptosis regulation has also been postulated, since high NHE1 activity confers resistance to proapoptotic stimuli. Additionally, NHE1 function is important in cytoskeletal organization and cell migration. The cytoplasmic tail of NHE1 acts as an anchor for actin filaments via binding of ezrin, radixin, and moesin (ERM) proteins, and disruption of these interactions or inhibition of NHE1 activity results in inhibition of cell migration and of formation of focal adhesions. NHE1 knockout mice are viable, although they have stunted growth and reduced survival rates. They also exhibit severe neurological defects (slow wave epilepsy, ataxia, and neuronal degradation) and present with abnormalities in gastric histology (see Section 56.5.4 ).
It is not clear whether involvement of NHE1 in these mechanisms is equally critical in the cells of the GI tract. No intestinal defect was demonstrated in otherwise rapidly renewing intestinal epithelium in NHE1 −/− mice, suggesting that the role of NHE1 in the intestinal crypt cell proliferation is minor. Also the described role of NHE1 in cellular differentiation may not represent a ubiquitous mechanism since its expression along the crypt-villus axis did not correlate with the differentiation status of enterocytes.
Regulated expression of NHE1 mRNA has been described in various systems, but especially in myocardium. Human, mouse, rabbit, and pig NHE1 gene promoter have been cloned and characterized to a various extent, with mouse NHE1 promoter analyzed in more detail than other species. The activity of this promoter is largely dependent on AP-2-like transcription factors as well as a poly(dA:dT) region of the promoter interacting with yet unidentified nuclear protein. Serum and growth factors have been shown to stimulate promoter activity in cardiomyocytes and fibroblasts through more distal elements of the promoter (0.8–1.1 kb) interacting with COUP transcription factors ; however, these in vitro findings do not correlate with data obtained from transgenic mice-bearing NHE1 gene promoter reporter construct. In the latter studies, crossing these mice with AP-2α or COUP-TFI knockout mice did not change the reporter gene expression in embryonic mouse tissue.
Regulation of NHE1 gene expression in GI tissues has not been extensively studied, although a limited amount of available data suggests that, consistent with its housekeeping role, NHE1 is not regulated at the mRNA level in situations where other NHE isoforms are. Examples of such circumstances are metabolic acidosis, microvillous inclusion disease, small bowel resection, glucocorticoid administration, or postnatal development.
The cytoplasmic C-terminal regulatory domain is associated with a number of functionally distinct signaling molecules, including phosphatidylinositol 4,5-bisphosphate (PIP 2 ), calcineurin homologous protein (CHP)1, and actin-binding proteins of the ezrin, radixin, moesin (ERM) family. In the distal C-terminal region, NHE1 contains a number of serine residues phosphorylated by ERK-regulated kinase p90RSK and Ste20-like Nck-interacting kinase (NIK) upon activation of growth factor receptors and by Rho kinase 1 (ROCK1) upon activation of integrin receptors and G protein-coupled receptors for thrombin and lysophosphatidic acid (LPA). Phosphorylation of C-terminal serine results in increased NHE1 activity, whereas phosphorylation of Ser703 by p90RSK promotes direct binding of the multifunctional adaptor protein 14-3-3, which conceivably serves as a focal point for the assembly of other signaling molecules. Additional proteins such as calmodulin (CaM), heat shock protein HSP70, and carbonic anhydrase II have also been shown to bind to this regulatory domain of NHE1. The latter interaction is particularly intriguing, as it may explain the ultimate changes in NHE1 exchange activity observed upon phosphorylation. It has been postulated that serum-induced phosphorylation within the last 178 amino acids of the C-terminus facilitates binding of carbonic anhydrase II, which through catalysis of CO 2 hydration causes local acidification resulting in the increase in NHE1 activity. The signaling molecule scaffolding at the C-terminus of NHE1 has been exhaustively overviewed by Baumgartner et al., and the reader is referred to this article and the references therein for more detail. It is not known whether all the described mechanisms of posttranslational modifications of NHE1 activity are ubiquitous to all cell types, and whether they have functional consequences in the digestive tissues. Na + /H + activity at the basolateral membrane of enterocytes increases with age, despite unchanged expression of NHE1 mRNA. This may represent age-dependent changes in NHE1 activity mediated by one or more of the abovementioned mechanisms, especially since another potential basolateral isoform NHE4 has not been detected in the small-intestinal epithelium.
Two studies implicated NHE1 in the pathophysiology of inflammatory bowel disease (IBD). In a rat model of acetic acid or trinitrobenzenesulfonic acid (TNBS)-induced colitis, Khan et al. described an induction of NHE1 mRNA in the colonic mucosa. Also in vitro, in Caco-2 and HT-29 human intestinal epithelial cells, inhibition of Na + /H + exchange with amiloride and other unrelated NHE inhibitors has been shown to reduce IL1β-, TNFα-, and LPS-stimulated IL-8 production, IL-1β-induced NF-κB activation, and phosphorylation of ERK (extracellular signal-regulated kinase). In the latter study, amiloride administered in vivo to DSS-treated rats resulted in attenuated symptoms of colitis and decreased neutrophilic infiltration in the colonic mucosa. The interpretation of these results is complicated by that fact that IC 50 for inhibition of IL-8 production by amiloride was ~ 30-fold higher than IC 50 required to inhibit NHE1 and NHE2, the two isoforms likely to be expressed in the selected cells used under culture conditions. Plasma concentration of amiloride in DSS-treated rats was not evaluated. It is possible, therefore, that the observed effects may represent nonspecific effects of the selected inhibitors. Moreover, recent analysis of NHE1 gene expression in human IBD is not consistent with the data obtained from rodent models of colitis. Khan et al. demonstrated a decreased mRNA expression in colonic biopsies from Crohn’s disease (CD) and ulcerative colitis (UC) patients, as compared to healthy colon. Similarly, Sullivan et al. found the expression of NHE1 to be decreased in the sigmoid mucosa of CD and UC patients. Consistently, Magro et al. described acute or chronic inhibitory effects of IFNγ on the activity of NHE1 in Caco-2 cells. Contrary, a recent study by Farkas et al. showed elevated expression and activity of NHE1 in the colons of patients with active UC. Therefore, the involvement of NHE1 in the pathogenesis of IBD and its potential as a therapeutic target is still unclear. In a rat model of necrotizing enterocolitis (NEC), the observed decrease in expression and activity of NHE1 was ascribed to cellular acidification, which was postulated to participate in the failure of the epithelial barrier and consequently in the pathogenesis of NEC. The plausibility of NHE1 involvement in liver cirrhosis through activation of stellate cells, as well as in hepatic tumorigenicity, is discussed in Section 56.5.2 .
In the rabbit and rat esophagus, NHE1 is the only plasma membrane NHE and is allosterically activated by reduced pH i in a protein kinase C (PKC)-dependent mechanism. The same mechanism of NHE1 activation along with Ca 2 + /calmodulin-dependent pathway mediates the cytoprotective effects of salivary epidermal growth factor (EGF) in acid-exposed cells. Loss of this mechanism in patients with low salivary EGF levels increases susceptibility to severe esophageal damage in gastroesophageal reflux disease (GERD) and contributes to the overall risk for the development of Barretťs esophagus. NHE1 expression is increased in GERD patients and in Barretťs esophagus, where it likely represents a cellular defensive mechanism to manage the acute and chronic acid overload. Bile acids present in reflux chyme reduce the ability of the cells to control their pH i by nitric oxide-mediated NHE1 inhibition, thus leading to increased DNA damage and potentially to mutations and cancer progression.
However, NHE1 has diverse physiological roles extending well beyond pH i and cell volume control, including cell proliferation, growth, migration, and apoptosis, and contributes to pathologic processes such as cancer cell invasion and heart failure. In a Barretťs adenocarcinoma cell line, acid pulse-induced NHE1 activity correlated with increased proliferation, which could be reduced by inhibition of NHE1 or PKC. This finding leads to a somewhat paradoxical proposal that while pathophysiological NHE1 inhibition in GERD may be detrimental to cell function, genomic integrity, and progression to Barretťs esophagus, pharmacological inhibition of NHE1 may be of therapeutic value in preventing progression from Barretťs to cancer, or in esophageal cancer therapy.
NHE2 was first cloned from rat and rabbit intestinal cDNA libraries by Collins et al. and Wang et al. and by Tse et al., respectively. Human NHE2 was cloned by Ghishan et al. and later corrected by Malakooti et al. Among the members of the human Scl9a family of NHE, NHE2 protein shares the most similarity with NHE4, especially within the cytoplasmic C-terminus. Interestingly, in the human, rat, and mouse, the Slc9a2 and Slc9a4 genes cosegregate on chromosomes 2, 9, and 1, respectively. The adjacent chromosomal location of the two NHEs in all three species strongly suggests that they arose by gene duplication early in the evolution. The predicted molecular weights of NHE2 protein in rat, rabbit, and human are ~ 91 kDa, although its mobility on SDS-PAGE gels does not confirm these calculations. Mature rabbit NHE2, when expressed in PS120 fibroblasts, was shown to be an O -linked sialoglycoprotein. In these studies, neuraminidase shifted the mobility of NHE2 protein from 85 to 81 kDa, and O -glycanase further shifted the mobility of the 81 kDa protein to 75 kDa. Incubation of PS120/NHE2 cells with an O-glycosylation inhibitor benzyl N -acetyl-alpha-D-galactosaminide reduced the size of the 85 kDa protein to 81 kDa, although this was without consequence for the initial rate of Na + /H + exchange in these cells.
Tissue Distribution and Cellular Localization
NHE2 is expressed in the epithelia of all digestive organs, with particularly high expression in the proximal colon. Outside of the GI tract, NHE2 activity and/or expression has been described in the kidney (cortical thick ascending limb of the nephron, macula densa, distal convoluted tubules, and connecting tubules), endometrium and placenta, chondrocytes, inner ear, heart, testes, and adrenal glands. Expression of NHE2 in the individual digestive organs is described below in a later section on physiological roles of Na + /H + exchange in the digestive tract. With the exception of gastric epithelium, NHE2 was unambiguously demonstrated on the apical membrane of polarized epithelial cells. Since NHE2 has an exclusive ability to be activated by elevated extracellular pH (pH o ), it has been speculated that NHE2 may be the NHE isoform that mediates Na + /H + exchange activation by an increase in interstitial HCO 3 – concentration during acid secretion in gastric epithelium, a hypothesis that assumes basolateral localization of this isoform. Immunohistochemical evidence for this assumption is lacking as yet. In the intestinal epithelium, expression of NHE2 along the crypt-villus axis shows some species-dependent differences. In rabbits, NHE2 is present in the brush-border of the entire villus of the small intestine, in colonic surface cells, and in the apical membrane of the upper half of the crypt. In the mouse colon, however, NHE2 is predominantly expressed in the crypt cells, suggesting a role for this isoform in crypt pH i and volume homeostasis.
Despite a relatively wide expression of NHE2, its physiological role remains elusive. NHE2 stably transfected in NHE-deficient Chinese hamster ovary (CHO) cells (AP-1) showed a relatively high affinity for amiloride and its analogues, with potencies in decreasing order of EIPA (IC 50 = 79 nM) > DMA (IC 50 = 250 nM) > amiloride (IC 50 = 1.4 μM) > benzamil (IC 50 = 320 μM). Nonamiloride compounds also inhibited NHE2 with the following order of potency: clonidine (IC 50 = 42 μM) > harmaline and cimetidine (both with IC 50 = 330 μM). Kinetic analyses showed that NHE2 Na + o dependence followed simple, saturating Michaelis-Menten kinetics with an apparent affinity constant for Na + (K Na ) ~ 50 mM. Intracellular H + activated NHE2 by a positive cooperative mechanism with an apparent half-maximal activation value of p K 6.90. Li + and H + acted as competitive inhibitors of NHE-mediated Na + influx, while extracellular K + had no effect on NHE2 activity.
The information provided by the analysis of NHE2 −/− mice suggests a role in muscarinic stimulation of salivary secretion, as well as in gastric physiology (see discussions below). The involvement of NHE2 in gastric parietal cell homeostasis seems particularly significant since NHE2 gene ablation leads to a reduced number of parietal and chief cells, loss of net acid secretion, and progressive inflammation in the form of diffuse corporal gastritis. Other roles for NHE2 in the physiology of digestive organs, presumed from the expression and/or functional studies, are discussed in more detail in Section 184.108.40.206 of Na + /H + Exchange in the Digestive Tract. Overall, the results of the available reports suggest that NHE2 plays a negligible role in net Na + or fluid absorption in the mouse digestive tract. The disparity between these results and the functional studies demonstrating contribution of NHE2 to various cellular functions (especially intestinal Na + absorption) remains unresolved; however, unidentified compensatory mechanisms may help explain the significance of this gene in the physiology of intestinal and renal epithelium.
Rat and human NHE2 promoters have been cloned and characterized. Both proximal promoters lack canonical TATA and CAAT boxes, are highly GC rich, and share about 59% homology with a number of conserved, predicted, regulatory elements. Only rudimentary analysis of the human NHE2 promoter has been performed, with prediction analyses indicating putative binding sites for the following trans -acting factors: Sp1, AP-2, Egr-1, p300, NF-κB, Oct-1, zinc finger protein-1, MyoD, two caudal-related homeobox (Cdx) family members, CdxA and Cdx-2, glucocorticoid receptor (GRE), thyroid hormone receptor, a CACCC site, and several polyoma viral enhancer 3 sites. Of all these sites, only Sp1, AP-2, CACCC, NF-κB, and Oct-1 were conserved in human and rat NHE2 promoters. A minimal promoter of the rat NHE2 was identified and found to be regulated by Sp transcription factors, with Sp1 acting as an activator, and Sp3 and Sp4 playing inhibitory roles in transfected renal epithelial cells. Regulation of both rat and human NHE2 promoter in intestinal epithelial cells also involves Sp1 and Sp3 transcription factors, although both of them appear to be stimulatory. NHE2, like NHE1, can be activated by serum and by EGF. There also appears to be a transcriptional component to the mechanism by which NHE2 is activated by EGF. In suckling rats, parenterally administered EGF increased expression of NHE2 mRNA in the small intestinal epithelium but not in the kidney. This finding was confirmed in EGF-treated RIE cells, which also showed activation of the rat NHE2 promoter in transient transfection experiments. In adult mice, however, neither exogenous EGF nor salivarectomy affected NHE2 mRNA expression in the small intestine, suggesting that this regulation may be species and/or age dependent. The response to growth factors induces the production of 1,2 diacylglycerol (DAG), an activator of protein kinase C (PKC). Phorbol 12-myristate 13-acetate (PMA), a DAG structural analog activates NHE2 activity (see Section 220.127.116.11 ), but also stimulates its mRNA expression. PMA was later shown to trigger phosphorylation of nPKCδ, activation of extracellular signal-regulated protein kinase-1 and -2 (ERK1/2), and subsequent nuclear translocation of Egr-1 transcription factor. Egr-1 directly binds to human NHE2 promoter and promotes NHE2 gene transcription. It remains unknown whether this mechanism of NHE2 transcriptional regulation is related to the differential effects of EGF during growth/aging.
Similar bimodal regulation of NHE2 by osmolarity has been described. In PS120 cells, hyperosmolarity inhibited NHE2 activity, but other reports showed activation of NHE2 in mouse inner medullary collecting duct (mIMCD-3) cells, AP-1 cells, and colonic crypt cells. In the case of renal mIMCD cells, mRNA expression was also induced by hyperosmotic stress. A TonE-like element and a novel cis -element, termed OsmoE, were identified in the rat NHE2 promoter as being responsible for the increased transcription of the NHE2 gene induced by hyperosmolarity, with both elements acting in concert to provide maximal transcriptional induction. The transcription factors interacting with these elements have not been identified, and at this point it is also not known whether the same mechanism is present in the colonic crypts.
During postnatal development, expression and activity of NHE2 in the rat small intestinal epithelium dramatically increase around the time of weaning. This increase is due to transcriptional activation of NHE2 gene, as shown by nuclear run-on assay and by reporter gene analysis in transgenic mice bearing − 2.4 kb of the rat NHE2 promoter (Kiela et al., unpublished observations). Interestingly NHE2 expression in the rat kidney follows a reciprocal pattern, with highest expression in the suckling period and a decline toward adulthood, implying tissue-specific mechanisms regulating postnatal changes in NHE2 expression.
NHE2 protein has a relatively short half-life (~ 3 h) compared to other NHE isoforms (NHE1—24 h, NHE3—14 h) and is subject to lysosomal degradation, as determined in PS120 fibroblasts and Caco-2 cells. This suggests that changes at the level of gene transcription or translation may be more critical for NHE2 regulation than for other isoforms with long half-lives. NHE2 is a residual plasma membrane protein and unlike NHE3 does not undergo endosomal recycling. Glycosylation of NHE2 may affect its cellular localization, since unglycosylated 75 kDa rabbit NHE2 was found predominantly intracellularly, although it is not clear whether this represents a regulatory mechanism or is simply related to the maturational stage of NHE2 protein synthesis. Of the two well-characterized apically expressed NHE isoforms, NHE2 and NHE3, NHE2 activity is considered relatively stable and is not regulated by many factors. Extracellular alkalinization activates NHE2, which is believed to propel increased proton extrusion in gastric parietal cells during secretagogue-stimulated acid secretion (see Section 56.5.4 ). The maximal rate of exchange ( V max ) mediated by NHE2 was shown to be stimulated by serum, fibroblast growth factor (FGF), and protein kinase C activator PMA in PS120 fibroblasts. Intracellular ATP depletion reduced the NHE2 activity by a dramatic decrease in H + affinity as well as V max , with virtual elimination of the allosteric effect of H + . ATP depletion also eliminated the stimulatory effect of serum, suggesting that growth factor-stimulated NHE2 activity is mediated via its pH-sensing mechanism. Thrombin increased NHE2 V max without altering the Hill coefficient, although it is not clear if this could be attributed to increased intracellular Ca ++ ascribed to thrombin-treated fibroblasts. In the same study, thrombin also increased NHE3 activity, whereas it was shown later that elevation of intracellular Ca ++ by thapsigargin in Caco-2/bbe cells inhibited NHE3.
NHE2 has a particularly well-documented role in the gastric epithelium, although alterations in NHE2 expression or activity in gastric disorders have not been documented. Downregulation of NHE2 activity and gene expression has been documented in rats and Caco-2/bbe cells treated with interferon γ, implicating a role for NHE2 in inflammation-associated diarrhea. The lack of absorptive defect in the intestine of NHE2 −/− mice, however, suggests that cytokine-mediated changes in NHE2 function may not be critical for electrolyte absorption in the inflamed intestinal mucosa. Surprisingly, enteropathogenic E. coli invasion of intestinal epithelial cells significantly increased NHE2 activity via a PKC ε -mediated mechanism, while it inhibited activities of NHE3 and Cl − /OH – exchange. The authors speculated that NHE2 activity might represent a potential compensatory response to increased luminal fluid resulting from inhibition of NHE3 activity, disruption of tight junctions, inflammatory response, or alterations in anion exchanger activity. On the other hand, TNF inhibits expression of NHE2 through an NF-kB-dependent mechanism, a phenomenon postulated to contribute to inflammation-associated diarrhea in IBD. Although our group has not demonstrated any differences between wild-type or NHE2 −/− mice in their susceptibility to DSS-induced mucosal injury, Moeser et al. showed that NHE2-deficiency prolongs recovery from mesenteric ischemia with increased mucosal permeability and disruption in the localization of the tight junctions proteins occludin and claudin-1.
Rabbit and rat NHE3 were first cloned by Tse et al. and Orlowski et al. respectively. The second report also included partial cloning of a human ortholog, later fully cloned by Brant et al., and mapped to chromosome 5p15.3. The open reading frame of NHE3 mRNA codes for an 831–834 amino acid protein with a calculated molecular weight of ~ 93 kDa, with highest homology to NHE5 (51.3%) and NHE2 (33.4%). Based on the presence of potential N -glycosylation sites in the NHE3 protein of all three species, it was at first believed to be a glycoprotein. It appears, however, that glycosylation of NHE3 may be species specific. Rabbit and pig renal NHE3 was shown to be glycosylated and sensitive to glycopeptidase F and general N -linked glycosylation inhibitor, tunicamycin, while glycosylation of rat or canine NHE3 was not detected. The functional consequences of glycosylation are not clear. In vivo inhibition of N -glycosylation in tunicamycin-treated LLC-PK cells significantly decreased NHE3 activity, as measured by pH-dependent 22 Na uptake and by Na-dependent pH i recovery from an acid load. This decrease in NHE3 function in tunicamycin-treated cells was accompanied by an intracellular accumulation of seemingly unglycosylated forms of the protein, and a conceivably compensatory threefold increase in NHE3 mRNA. Based on these studies, it has been postulated that glycosylation of porcine NHE3 plays a role in membrane trafficking and ultimately in NHE3 activity. On the other hand, deglycosylation of rabbit renal brush-border protein did not impact acid-stimulated, amiloride-sensitive 22 Na influx into the vesicles. Analogous in vivo experiments with NHE1 with mutated N -glycosylation sites, and with inhibition of O-glycosylation in NHE2, did not translate into detectable functional changes of the respective isoform. Therefore, the physiological significance of NHE3 glycosylation is still unclear. The secondary structure of NHE3 follows the general model for all members of the Slc9a family and is discussed in more detail in the earlier section on Membrane Topology.
Tissue Distribution and Cellular Localization
The range of NHE3 gene expression in various tissues has been found to differ among species. Rabbit and rat NHE3 is consistently expressed at high levels in the absorptive epithelia of kidney cortex, colon, and small intestine, with lower levels detected in the stomach, brain, and heart, whereas human NHE3 is also expressed in relatively high levels in testes, ovaries, prostate, and thymus. Rat NHE3 is expressed in both acinar and ductal cells of the salivary glands, although its role in salivary secretions appears to be negligible (see Section 56.5.1 ). Expression of NHE3 in cholangiocytes and gallbladder epithelium implies a role for this isoform in bile formation, and possibly in pathogenesis of gallstones. Expression, cellular localization, and functional relevance of NHE3 in the gastric epithelium are somewhat controversial and are described in more detail in Section 56.5.4 .
Expression of NHE3 appears to be higher in the ileum than in other intestinal segments in both rabbits and humans. In humans, NHE3 mRNA levels are higher in the jejunum and the colon. In one report, NHE3 was present at approximately the same levels in both the ascending and descending colon, while more recent analysis of NHE3 mRNA in mucosal biopsies showed a caudally decreasing gradient of expression from cecum to rectum. In the small intestinal and colonic epithelium, NHE3 may be considered a marker for the absorptive epithelial cells, since it is expressed only in the villus or surface epithelium, and not in the crypts. Atypical expression of NHE3 has been demonstrated in the colonic crypts of NHE2 −/− mice, where it is believed to play a compensatory role in regulation of crypt cell volume and pH i . In polarized intestinal epithelial cells, the majority of NHE3 protein is localized to the apical membrane, where it can be found both on the microvilli as well as in the intervillus clefts. In Caco-2 cells, about 20% of total NHE3 protein is localized to a diffuse subapical pool, and recycling between plasma membrane and this endosomal compartment represents a mode of regulation of NHE3 by endocytosis/exocytosis. Similar observations have been made in the renal proximal tubule epithelium. When expressed in AP-1 fibroblasts, ~ 90% of NHE3 protein was found in the juxtanuclear endomembrane vesicles, a pool further increased by inhibition of phosphatidylinositol 3′-kinase (PI3-K). It has been suggested that the constitutive activity of PI3-K is important in the maintenance of the steady state level of NHE3 on the plasma membrane. The extent to which endosomal recycling participates in NHE3 regulation in native cells is discussed in more detail in Section 18.104.22.168 .
In the brush-border membrane of rabbit ileal enterocytes, NHE3 is equally split between a detergent-soluble and detergent-insoluble fractions, and a part of the latter fraction is present in cholesterol-enriched lipid microdomains (lipid rafts). Li et al. demonstrated that the lipid raft pool and its association with actin cytoskeleton play an important role in regulation of NHE3 activity through endocytosis.
Kinetics and regulation of NHE3 expressed in fibroblasts and intestinal epithelial cells were described by Levine et al. and by McSwine et al. According to these two reports, NHE3 mediated Na + /H + exchange, with evident cooperativity by intracellular H + , and with simple Michaelis-Menten kinetics for extracellular Na + ( K m ~ 17 mM), and was inhibited by phorbol esters and ATP depletion. NHE3 is relatively insensitive to amiloride and its analogues, with the highest affinity to the newly developed S3226 inhibitor (IC 50 = 0.02 μM). In addition to its well-defined role in the absorptive epithelium of renal proximal tubules (for a recent review, see Burckhardt et al. and Moe ), NHE3 plays a prominent role in epithelial sodium absorption in the small and large intestine, as evidenced by studies in mice with targeted disruption of Slc9a3 gene. The intestinal absorptive defect observed in these mice is described in more detail later in this chapter, in Section 56.5.5 .
Rat NHE3 gene promoter was cloned at about the same time by Kandasamy and Orlowski and by Cano. Our group later resolved a discrepancy in the transcriptional start site in those two reports by showing that the atypical TATA box located at bp-26/-31 (numbers according to the major transcription start site mapped in Ref. ) was not necessary and even detrimental for promoter activity, and that a − 20/+8-bp fragment represents a functional, albeit atypical, initiator. Within the − 81-bp upstream region, three Sp transcription factor binding sites were critical because their mutation drastically reduced promoter activity. The roles of Sp1 and Sp3 were further demonstrated by electromobility shift assay and by transactivation of the NHE3 promoter in SL2 cells by forced expression of Sp1 and Sp3. Both of these transcription factors were found to act synergistically with GATA-5 bound to a GATA box in exon 1 (+ 20/+23 bp). These studies demonstrated that rat NHE3 promoter is initiator driven and controlled mainly by Sp1 and Sp3, which functionally interact with GATA-5. This interaction may represent a regulatory mechanism participating in a gradient of intestinal gene expression along the crypt-villus axis. Cloning of the human NHE3 5′-regulatory region defined a maximal promoter activity in the − 95/+5 nt region, a sequence with very high homology with the proximal promoter of the rat NHE3, with overlapping and functional regulatory elements for Sp and AP-2 transcription factors. The following discussion of examples of long-term regulation of NHE3 involves reports describing changes in NHE3 mRNA expression that in only some instances have been confirmed to represent transcriptional regulation.
The response of NHE3 to glucocorticoid hormones is biphasic, involving nontranscriptional activation of the protein mediated by serum- and glucocorticoid-induced protein kinase (SGK1) discussed later in this section, as well as transcriptional activation of NHE3 gene. Changes in NHE3 mRNA with glucocorticoid treatment were shown in the rabbit small intestine and colon by Yun et al. In adult rats, dexamethasone increased NHE3 mRNA expression in the ileum and proximal colon but not in the jejunum or distal colon, while conversely, adrenalectomy reduced NHE3 expression in the rat ileum and proximal colon but not in the jejunum. This regulation turned out to be not only segment specific but also age dependent. Glucocorticoid responsiveness in the proximal small intestine was greatest in suckling rats and decreased with age to no detectable change in adults, while ileal NHE3 was induced by methylprednisolone only in adults. This age- and segment-specific responsiveness to glucocorticoid treatment correlated with expression and ligand binding capacity of the glucocorticoid receptor in the enterocytes. Transcriptional regulation of the NHE3 gene by glucocorticoids was demonstrated in cells transiently transfected with rat NHE3 promoter reporter constructs by Cano and Kandasamy and Orlowski.
Short-Chain Fatty Acids
Short-chain fatty acids (SCFAs) are potent stimuli of sodium and water absorption in the colon, with butyrate being the most effective. It has been speculated that the SCFA-mediated increase in Na + absorption is due to the coupling of two exchange mechanisms, Na + /H + and SCFA − /Cl − exchange (see Section 22.214.171.124 ). The use of amylase-resistant starch as an additive to oral rehydration solution proved effective in reducing diarrheal stool output in cholera patients, thus showing that SCFAs can be potent antidiarrheal agents. Part of the NHE regulation by SCFA appears to be mediated by transcriptional induction of NHE3 gene expression. In rats fed 5% pectin-supplemented diet for 2 days, NHE3 mRNA, protein and activity increased in the colonic epithelial cells. Similar results were obtained with Caco-2/bbe cells treated with SCFAs in vitro. Our group has also shown that rat NHE3 promoter reporter construct, when transiently transfected into Caco-2 cells, is significantly induced by SCFAs, especially butyrate. The mechanism of this induction involves Ser/Thr kinase activity with a likely permissive role for PKA, as the activation of the promoter by butyrate was abrogated by H-7, Rp-cAMPS, and H-89 inhibitors, as well as by overexpression of a dominant-negative mutant form of the regulatory subunit of PKA. Subsequent studies determined that butyrate induced phosphorylation of Sp1 and acetylation of Sp3 and a shift in their interaction with the proximal NHE3 promoter in favor of Sp3. Since Sp3 is a significantly more potent inducer of NHE3 gene transcription, such shift would result in increased NHE3 promoter activity.
Serotonin (5-hydroxytryptamine) plays an important role in regulating GI motility, secretion, and absorption. Increased serotonin levels have been implicated in the pathophysiology of diarrhea secondary to carcinoid syndrome, ulcerative colitis, and irritable bowel syndrome. Gill et al. shown that serotonin acutely decreases NHE3 activity via 5-HT4 receptors in human intestinal epithelial cells. The same group also demonstrated a transient transcriptional inhibition of NHE3 expression by serotonin mediated by PKCα-dependent decrease in association of Sp1 and Sp3 with the proximal NHE3 gene promoter.
Chronic metabolic perturbations in systemic acid-base balance can affect Na + absorptive functions of the gut. Metabolic acidosis induced in rats by 5% NH 4 + Cl − in drinking water induced ileal expression of NHE2 and NHE3 mRNA, protein, as well as their activities. Transcriptional regulation of NHE3 was confirmed in opossum kidney cells (OKP) transfected with NHE3 promoter construct and subjected to prolonged (24 h) exposure to acidified media. The precise mechanism of this induction has not been described.
As an adaptive response to enhance the intestinal absorptive capacity, rat small intestinal Na + /H + activity was shown to increase, primarily in the segment distal from the resection. It was later shown that this increase was associated with ~ threefold elevation of NHE3 mRNA and protein expression after a 50% massive proximal small bowel resection in rats. The increase was again observed only in the ileal segment distal from the anastomosis site, suggesting that dietary rather than humoral factors might be responsible. Similar results were obtained in enterectomized mice.
The vast majority of knowledge on acute regulation of NHE3 activity comes either from heterologous cell expression systems or from renal epithelial cells, or colon cancer cells as models of intestinal epithelia. However, the described mechanisms are likely ubiquitous and will, perhaps with certain exceptions, apply to the epithelial cells of the digestive tract. These mechanisms were reviewed in more detail by He and Yun, Zachos et al., Donowitz et al., and Alexander and Grinstein. The key elements responsible for NHE3 activity, interaction with other proteins, membrane localization are located within the C-terminal regulatory domain, which contains distinct regions required for specific aspects of acute regulation of NHE3 activity. The C-terminus of NHE3 binds to several scaffolding proteins to form higher-order multiprotein complexes, and in some respect, NHE3 itself can be considered as a scaffold. Consequently, NHE3 exists physiologically in large multiprotein complexes which range from 400 kDa in the intracellular pool to ~ 1000 kDa at the plasma membrane. These complexes are dynamic and are influenced by physiological stimuli participating in acute NHE3 regulation. Within the long C-terminus, a small, putative α-helical domain of NHE3, between amino acids 586–605 was postulated to act as a “switch domain” that associates with at least seven other proteins to both activate or inhibit NHE3. This dynamic assembly, association with cytoskeletal proteins, endosomal recycling, and protein phosphorylation events, all act in concert to provide highly regulated turnover and activity of NHE3 protein.
The C-terminal domain of NHE3 contains numerous putative phosphorylation sites for a number of kinases. Deletion of this domain renders NHE3 activity constitutive, but with only partially preserved transport activity. Domain swapping experiments have shown that regulatory characteristics of one NHE isoform can be transferred to another by the cytoplasmic domain of the first. For example, replacement of the C-terminal cytoplasmic tail of NHE1, an isoform that is largely cAMP insensitive, with an analogous domain of NHE3, transfers cAMP-mediated inhibition to the hybrid molecule. The experiments strongly suggested the existence of functionally relevant phosphorylation sites located within the cytoplasmic C-terminal tail of NHE3. In response to elevated intracellular cAMP, protein kinase A (PKA) phosphorylates NHE3 on multiple sites in the intact cell. While several putative cAMP consensus sequences in the murine NHE3 have been reported, including Ser 330 , Ser 514 , Ser 552 , Ser 576 , Ser 605 , Ser 662 , Ser 691 , Ser 692 , and Ser 805 , only Ser 552 and Ser 605 (or Ser 554 and Ser 607 in the extensively studied rabbit NHE3 protein), participate in NHE3 regulation by PKA in vitro, and only Ser 605 has been shown to be phosphorylated in vivo. In this study, phosphorylation of Ser 552 was also shown to participate in the NHE3 response to cAMP, although in another study by Kurashima et al., Ser 552 was not functionally important. The functional consequence of PKA-mediated NHE3 phosphorylation is its reduced V max , decrease in the surface amount, presumably due to increased endocytosis and decreased exocytosis. As suggested by Kocinsky et al., subcellular trafficking may be the primary consequence of C-terminal phosphorylation: the temporal dissociation between phosphorylation at Ser 552 and Ser 605 and NHE3 inhibition by PKA indicates that phosphorylation at these residues does not directly alter NHE3 activity.
The recruitment of PKA to the C-terminus of NHE3 involves a multiprotein complex including NHE regulatory factors NHERF1 (SLC9A3R1) and NHERF2 (E3KARP, SLC9A3R2) and a scaffolding protein ezrin. NHERF1 and NHERF2 proteins contain two 80–90 amino acid PDZ ( p ostsynaptic density 95, disk large, and z onula occludens-1) domains consisting of GLGF repeats mediating physical interaction with short peptide sequences located at the C-terminus of interacting proteins. Both NHERF1 (initially cloned and described as NHE3 Kinase A Regulatory Protein, E3KARP ) and NHERF2 reconstitute PKA-dependent NHE3 inhibition when expressed in NHERF-deficient cells. NHERF1 and NHERF2 interact though their C-terminal 29 amino acids with cytoskeleton-associated ezrin, which functions as A Kinase anchoring protein (AKAP). Phosphorylation of NHE3 by PKA is therefore facilitated by bringing the catalytic subunit of PKA to the vicinity of the NHE3 cytoplasmic tail by a protein complex containing either of the two NHERF factors and cytoskeleton-associated AKAP protein, ezrin (see Fig. 56.4 A ). PDZ-binding protein PDZK1 (NHERF3/CAP70/PDZ-dc-1) was also demonstrated as necessary for the cAMP- and Ca 2 + -mediated NHE3 regulation in mouse colonocytes. NHE3 is also acutely inhibited by elevated cGMP and the cGMP-dependent, type II protein kinase cGKII (PRKG2). Chen et al. showed that cGMP/cGKII-mediated rapid inhibition of rabbit NHE3 was associated with decreased plasma membrane abundance and required phosphorylation of all three serines (Ser 554 , Ser 607 , and Ser 663 , equivalent to mouse Ser 552 , Ser 605 , and Ser 659 ).
Ser 719 in the rabbit NHE3 was predicted to be the target of phosphorylation by casein kinase 2 (CK2). Sarker et al. confirmed that CK2-mediated phosphorylation at Ser 719 is required for ca. 40% of basal NHE3 activity and that mutation of this residue led to decreased NHE3 exocytosis and its reduced abundance at the plasma membrane. In a follow-up study, Sarker et al. showed that phosphorylation at Ser 719 is part of the PI3-K/AKT-dependent pathway. Mutation of Ser 719 led to a reduced NHE3 complex size, reduced expression in lipid rafts, increased brush-border mobile fraction of NHE3, and reduced binding of multiple proteins to the C-terminus, including calcineurin homologous protein (CHP), the NHERF family proteins, and SNX27 (related PDZ domains).
Ca 2 + /calmodulin-dependent protein kinase II (CaM KII) binds to the same α-helical domain as CK2, though it results in inhibition of basal NHE3 activity in both fibroblasts and colonic epithelial cells. Physical interaction of CaMKIIγ subunit with NHE3 was inversely related to intracellular Ca 2 + concentration. It occurred between aa 586 and 605 in the NHE3 C terminus, but phosphorylation occurs downstream of aa 690. CaM KII-mediated NHE3 inhibition was not associated with a change in NHE3 plasma membrane expression but, rather via decreased NHE3 turnover number.
LPA, which stimulates NHE3 via G protein-coupled LPA 5 receptor, has also been shown to require phosphorylation at Ser 719 . Alternative or complementary pathway for LTA stimulation was also proposed whereby LPA transactivates EGFR, which results in the parallel activation of two pathways: RhoA-Rho-associated kinase-proline-rich tyrosine kinase 2 (Pyk2) cascade and the MEK-ERK pathway. The results of this study suggested that Pyk2 and ERK1/2 may converge on the same effector, which was later confirmed to be the p90 ribosomal S6 kinase RSK2. RSK2 was shown to physically interact with NHE3 C-terminus in LPA-treated cells, result in phosphorylation at Ser 663 (human NHE3), and increased brush-border membrane delivery of NHE3.
Somewhat similar scaffolding mechanism is involved with glucocorticoid-stimulated NHE3 activity. In this case, however, the mediating kinase (serum and glucocorticoid inducible kinase, SGK1) interacts directly and specifically with NHERF2, acting as a bridge between the kinase and NHE3, to stimulate activity of the latter. It is the second PDZ domain of NHERF2 that binds both NHE3 (aa 585 and 660) and SGK1. A model facilitating this assembly was proposed, in which NHERF2 dimerizes as depicted in Fig. 56.4 B. It has also been postulated that the mechanism of posttranscriptional regulation of NHE3 by glucocorticoids is biphasic, with an initial phase involving phosphorylation of the preexisting membrane NHE3, and a later phase in which SGK1 and NHERF2 facilitate translocation of the newly synthesized NHE3 to the cytoplasmic membrane. A conserved Ser 663 of NHE3 appears to be the major site of phosphorylation by SGK1 and its mutation blocks the effect of dexamethasone. Ser 663 phosphorylation precedes the changes in NHE3 activity, which is associated with an increased amount of NHE3 proteins at the surface membrane. However, in vivo, SGK1 or NHERF2 deletion, while significantly attenuated the effect of dexamethasone on NHE3 activity, did not completely abolish the stimulation. The authors showed that SGK3 isoform also contributes to the end result at the endosomal level. Dexamethasone activated SGK3 and NHE3 activities via a mechanism dependent on phosphoinositide 3-kinase (PI3K) and phosphoinositide-dependent kinase 1 (PDK1). Dexamethasone-induced translocation of PDK1 to endosomes, the primary location of SGK3; Arg 90 mutation of SGK3 disrupted its endosomal localization and delayed NHE3 activation.
Regulation of NHE3 does not require phosphorylation in all cases. Stimulation of NHE3 activity by FGF or PKC-activating phorbol esters did not coincide with detectable changes on the phosphorylation status of NHE3. An indirect mechanism of action has been suggested, mediated via phosphorylation of associated regulatory factors indirectly affecting NHE3 activity.
Association With Cytoskeleton
The NHERF-mediated link with ezrin suggests the association of NHE3 with cytoskeleton as a likely mechanism controlling NHE3 activity. Consistent with this notion, NHE3 was found to cosediment with F-actin, and pharmacological disruption of cytoskeleton induced a profound inhibition of NHE3 activity. Inhibition of two kinases controlling cytoskeletal assembly, RhoA and ROK, also inhibited NHE3 activity in CHO cells stably transfected with dominant-negative mutants of a respective kinase without altering NHE3 abundance in the cytoplasmic membrane. This mechanism may at least in part account for the inhibitory effect of cAMP on NHE3 activity. Elevated PKA activity inhibits RhoA, resulting in altered cell morphology with disruption of the microfilament actin network. By analogy, expression of constitutively active forms of RhoA and ROK kinases attenuate PKA-mediated NHE3 inhibition by stabilizing actin filaments. Similarly, disruption of actin cytoskeleton by hyperosmotic stress may be responsible for the associated decrease in NHE3 activity. However, the role of ezrin in maintaining NHE3 plasma membrane expression and activity was questioned in a study by Hayashi et al., who showed that mice lacking ezrin in epithelial cells had unaltered NHE3 activity. In a more reductionist in vitro model of MDCK II cell, which lack the two other ERM (ezrin/radixin/moesin) proteins, ezrin knockdown did not affect basal activity, apical membrane localization, actin anchoring, or cAMP-mediated phosphorylation of NHE3, but blunted the inhibitory effects of elevated cAMP. This study argued that ezrin/NHE3 interaction is not an exclusive mechanism responsible for the retention of NHE3 at the apical plasma membrane or the interaction between NHE3 and the actin cytoskeleton.
As mentioned earlier in this chapter, in addition to being present at the cell surface, NHE3 is detectable in intracellular vesicles of the juxtanuclear compartment consistent with recycling endosomes. NHE3 remains in a state of dynamic equilibrium between the cell surface and the intracellular compartment: it undergoes internalization via clathrin-coated vesicles and is exocytozed back to the cytoplasmic membrane in a phosphatidylinositol 3-kinase-dependent manner. PI3-kinase inhibition leads to decreased NHE3 activity correlating with depletion of the plasma membrane pool of NHE3 protein, while constitutively active PI3-kinase or AKT transfected into NHE3-expressing PS120 cells stimulates the exchanger and increases the percentage of NHE3 present on the plasma membrane. EGF and FGF growth factors have been show to stimulate NHE3 activity by increasing the surface protein pool in a PI3-kinase-dependent manner. Other factors increasing the apical pool of NHE3 include LPA and endothelin-1. Conversely, decreased NHE3 surface expression has been associated with inhibition of the transporter by PKC, by parathyroid hormone, and by dopamine. Collectively, these studies strongly suggest that redistribution of NHE3 between subcellular compartments is an effective means of transport regulation.
Our group has recently published a comprehensive review on the pathophysiology of intestinal Na + /H + exchange, which discusses perturbations in the function and respective roles of NHE isoforms in esophageal and gastric pathology, diarrheal diseases, and intestinal inflammation. Here, we will only briefly discuss the roles NHE3 plays in GI pathologies.
Holmberg and Perheentupa and Booth et al. described a form of congenital secretory diarrhea (CSD) due to defective sodium/hydrogen exchange (OMIM %270420). Based on a close phenotypical resemblance between this rare disease and symptoms displayed by NHE3 −/− mice, NHE3 became a likely candidate for linkage. However, homozygosity mapping and multipoint linkage analysis studies in four candidate regions known to contain NHE1, NHE2, NHE3, and NHE5 genes have shown that CSD is an autosomal recessive disorder unrelated to mutations in the NHE1, NHE2, NHE3, and NHE5 genes. Since location of the human NHE2 gene is most likely adjacent to NHE4, that latter gene can also be excluded as a candidate for CSD. It would seem, therefore, that another NHE isoform or a regulatory factor may be directly responsible for the loss of Na + /H + exchange in this disease. More recent genomic studies revised the idea of reduced NHE3 activity as a contributor to the classic or nonsyndromic form of CSD. Using a cohort of 18 patients from 16 families, Janecke et al. determined that a variety of mutations (point, missense, and truncation) in the NHE3-encoding SLC9A3 gene occurred in half of the studied CSD cases. The identified SLC9A3 mutations included 1 whole-gene deletion, 1 splicing, and 2 frame-shift mutations, all of which were expected to abolish protein production. Four missense mutations/variants, p.Arg382Gln, p.Ala311Val, p.Ala269Thr, and p.Ala127Thr, were tested in vitro, and all but p.Ala127Thr (benign variant) conferred decreased basal Na + /H + exchange activity. In a recent genome-wide SNP analysis of syndromic CSD patients, Heinz-Erian et al. identified loss-of-function mutations in SPINT2 gene encoding a Kunitz-type serine-protease inhibitor. These mutations are believed to be responsible for a third of CSD cases described as syndromic CSD form. The physiological target(s) of SPINT2, the molecular pathology caused by SPINT2 deficiency, and its relationship with Na + /H + exchange remains unknown. In addition, an activating mutation in the catalytic domain of the guanylate cyclase 2C gene ( GUCY2C , which also serves as a heat-stable enterotoxin receptor) may account for the additional 20% of sporadic CSD cases. It is thought that this mutation is mechanistically linked to hyperactivation of the cystic fibrosis transmembrane conductance regulator (CFTR) and to inhibition NHE3 function via increased intracellular cGMP and a cGKII kinase-dependent mechanism.
Altered expression and activity of NHE3 in primary and diabetes-related hypertension have implied a potential role for this isoform in the pathogenesis of high blood pressure. Spontaneously hypertensive rats (SHR) have elevated NHE3 activity in the ileal brush-border membranes and in renal proximal tubules, suggesting that increased intestinal sodium absorption and decreased renal reabsorption may contribute to systemic sodium retention and the pathogenesis of hypertension. In both streptozotocin-induced diabetes and in BB/W autoimmune diabetic rats, renal cortex brush-border Na + /H + exchange (presumably mediated by NHE3) was significantly induced, likely due to acidosis and not hypoinsulinemia. The effect of diabetes on intestinal Na + /H + exchange is not well described. In streptozotocin-induced diabetes in rats (used as a model for secondary vitamin D deficiency), ileal NHE3 mRNA was induced twofold. It is likely, however, that the observed difference was due to vitamin D deficiency, as repletion of diabetic mice with 1,25(OH) 2 D3 brought NHE3 mRNA expression down to control levels.
Diarrhea observed commonly in IBDs is a direct result of perturbations in colonic absorptive and secretory processes. Proinflammatory mediator IFNγ downregulates NHE3 mRNA and protein expression both in vivo and in vitro. In interleukin-2 knockout mice with a disease resembling human ulcerative colitis, a drastic reduction in colonic transepithelial net Na + flux was paralleled by a reduction in electroneutral NaCl absorption and decreased NHE3 mRNA and protein expression in the proximal colon, and by an abrogated aldosterone-stimulated electrogenic Na + transport with decreased ENac expression in the distal colon. This strongly suggests an involvement of NHE3 in the pathogenesis of diarrhea in ulcerative colitis. Indeed, two recent studies with UC patients demonstrated inhibition of NHE3 activity measured in colonic biopsies. Interestingly, in one of these reports, NHE3 inhibition did not coincide with changes in NHE3 expression or cellular localization, a result inconsistent with an earlier observation by Sullivan et al. who showed decreased expression of NHE3 in the mucosal biopsies of UC and CD patients and in DSS-induced colitis in mice. In IL10 −/− mice, NHE3 activity, measured in the apical enterocytes within isolated colonic crypts, was significantly decreased although NHE3 expression and localization were preserved. In parallel, expression of two key NHE3-regulatory proteins, SLC9A3 regulator 2 (SLC9A3R2/NHERF-2) and PDZ domain containing 1 (PDZK1), has been decreased, and in a follow-up study, Yeruva et al. demonstrated a causal link between PDZK1 downregulation and NHE3 activity.
In an in vivo model of diarrhea mediated by anti-CD3 monoclonal antibody-induced T-cell activation, Clayburgh et al. showed that in the jejunum, tumor necrosis factor induces NHE3 internalization via PKCα activation and leads to NHE3 internalization and the resulting Na + malabsorption. On the other hand, in a rat model of NEC, ileal NHE3 activity was not affected, implying differential involvement of inflammatory mediators in regulation of NHE3 expression and activity. Although it is clear that the NHE3-mediated apical Na + /H + exchange and epithelial Na + absorption are inhibited in IBD, the exact mechanism remains unclear and may depend on the specific disease, segment involved, and/or the severity of inflammation.
Chronic NHE3 inhibition in intestinal inflammation may have consequences going beyond alteration is ion transport. There is a significant upregulation of IFNγ and IFNγ-inducible genes in the jejunal mucosa of NHE3-deficient mice. We identified the infiltrating CD8 + and asialoGM1 + NK cells as the primary sources of the cytokine ( Fig. 56.5 A ). Microarray, histological, and immunological analyses of the colonic mucosa also showed that NHE3 −/− mice spontaneously develop distal, primarily neutrophilic colitis mediated by the colonic bacteria ( Fig. 56.5 B). Moreover, NHE3-deficient mice are extremely susceptible to DSS-induced mucosal injury: concentrations of DSS insufficient to induce measurable histological changes in wild-type mice are lethal in their NHE3 −/− littermates ( Fig. 56.5 C). According to the genome-wide microarray analysis of gene expression, NHE3 status was an important determinant of the inflammatory response to acute DSS exposure in both small intestine and colon, thus suggesting that NHE3 may serve as a modifier gene determining the extent of the gut innate inflammatory responses in the face of intestinal injury. Moreover, in parallel with spontaneous colitis, NHE3-deficient mice develop colonic microbial dysbiosis reminiscent of that described in IBD patients. NHE3 deficiency exacerbated colitis in IL10 −/− mice, a model known for being microbially driven. Moreover, in Rag2/NHE3 double knockout mice, NHE3-deficiency and dysbiosis were associated with dramatically accelerated and exacerbated disease in response to adoptive naïve CD4 + T-cell transfer, with rapid body weight loss, increased mucosal T cell and neutrophil influx, increased mucosal cytokine expression, increased permeability, and expansion of CD25 − FoxP3 + Tregs ( Fig. 56.5 D). We hypothesized that inhibition of NHE3 and NHE8 (discussed later in this chapter) in the course of IBD may be one of the factors contributing to microbial dysbiosis and exacerbated inflammatory response ( Fig. 56.6 ).
NHE4 was first cloned from rat stomach cDNA library by Orlowski et al. The predicted primary structure of rat NHE4 is 717 amino acids with a calculated molecular weight of ~ 81.5 kDa. Among the nine members of the NHE gene family, NHE4 protein shares the highest homology with NHE2 (56.5% hNHE2 vs. hNHE4). Western blotting with polyclonal antibodies against fusion protein containing amino acids 393–625 of rat NHE4 protein detected a band of ~ 100 kDa, suggesting posttranslational modifications, conceivably glycosylation, in stably transfected PS120 fibroblasts. On the other hand, monoclonal antibodies raised against a similar fragment (565–675 AA) of rat NHE4 reacted with a predominant band of ~ 65–70 kDa and two minor bands at 45–50 kDa, ~ 75 kDa. These described discrepancies in NHE4 molecular weight remain unresolved. Hydropathy plot analyses have indicated membrane topology similar to the general model for NHE, although this issue was not experimentally addressed.
Tissue Distribution and Cellular Localization
NHE4 exhibits fairly limited tissue distribution, with the highest expression in the stomach, and lower levels in the kidney, pancreas, salivary glands, hippocampus, and endometrium. Although NHE4 has been reported to contribute to pH regulation in T84 colonic carcinoma cells, its expression in the small and large intestine is uncertain. An initial cloning report described detectable NHE4 transcript in these tissues, although later studies with more specific cDNA probes found no expression of this isoform in the rat jejunum or colon. Arena et al. showed Na + /H + exchange activity in the rat and human colonic crypts that was insensitive to 7 μM ethylisopropylamiloride or 400 μM amiloride and attributed it to NHE4. In the reported cases, NHE4 protein was localized on the basolateral membrane, with the exception of pancreatic acinar cells, where NHE4 was also detected on zymogen granule membrane.
Unlike other plasma membrane NHEs, NHE4 lacks sodium-specificity and acts also as a K + /H + and Li + /H + exchanger. Under basal conditions, NHE4 exhibits very low cation/H + exchange activity, which is further stimulated by hyperosmolarity or treatment with stilbene derivative, DIDS. Another distinguishing feature of NHE4 is its relative insensitivity to amiloride and EIPA inhibition. The inhibition constant for NHE4 in stably transfected LAP(−) cells was over 800 μM for amiloride (5.3 and 309 μM for NHE1 and NHE3, respectively), with even larger differences observed for EIPA. HOE-642 inhibition studies later revealed that NHE1 and NHE2 are 11,000- and 180-fold more sensitive to inhibition by the compound than is NHE4, which exhibits sensitivity to HOE-642 similar to that of NHE3. It has been hypothesized and recently confirmed that in the kidney, where NHE4 is predominantly expressed on the basolateral membrane of the thick ascending limb epithelial cells, this isoform may be specifically involved in ammonium transport. Its role in the GI tract may be the most significant in the stomach, since targeted disruption of the NHE4 gene in mice results in profound phenotypic changes. The role of NHE4 in gastric morphology and acid secretion is discussed in Section 56.5.4 . In the rat and human colonic crypts, Arena et al. postulated that NHE4-like activity contributes to regulation of the intracellular pH and is stimulated by aldosterone and elevated cAMP but not by hyperosmolarity.
Little is known about the role of NHE4 in gut-related pathologies, or its gene expression and protein activity. Beltran et al. demonstrated that exposure of T84 cells to the STa enterotoxin from enteroxigenic strain of E. coli reduced NHE4 activity and the cell capacity to recover from intracellular acidification. The authors concluded that such inhibition may promote diarrhea, although recovery requiring cAMP, but not cGMP. STa effect results in a causal phenomenon (STa/increased, cAMP/increased PKA activity/reduced NHE4 activity) ending with intracellular acidification that could have consequences in the GI cells function promoting human diarrhea, although no further evidence from in vivo studies has been provided thus far.
Klanke et al. identified a cosmid clone for NHE5 (SLC9A5) by low stringency hybridization with a rat NHE2 cDNA probe. Two exons of SLC9A5 were sequenced and compared to the other family members. The sequence most closely matched that of NHE3. Northern blots showed that SLC9A5 is expressed in the brain, testes, spleen, and skeletal muscle. No expression in the organs of the GI tract has been described to date. NHE5 is enriched in neuronal tissues, where it contributes to regulating pH of the endosomal compartment and to both neurite growth and dendritic spine formation. Human NHE5 resides in the EKD2 region of chromosome 16, linked to the pathogenesis of familial paroxysmal kinesigenic dyskinesia. Although no mutations in the coding region, intron/exon boundaries, or the 5′ and 3′ untranslated regions of the gene were identified, possible sequence variations in introns or the regulatory region could not be ruled out.
By sequencing random cDNAs corresponding to relatively long transcripts from the human myeloid cell line KG-1, Nagase et al. identified a cDNA, which they called KIAA0267, that encoded an incomplete sequence, which was later identified by Numata et al. as NHE6 (SLC9A6). Human NHE6 protein shares the highest homology with NHE7 and NHE9 (71% and 61%, respectively) and has 669 amino acids. Its predicted membrane topology is similar to that reported for other NHEs, with 12 putative membrane-spanning domains followed by a hydrophilic C terminus.
Tissue Distribution and Cellular Localization
Northern blot analysis detected an approximately 5.5-kb SLC9A6 transcript with the most abundant expression in mitochondrion-rich tissues such as brain, skeletal muscle, and heart, and a smaller expression level in the liver and pancreas. Expression of NHE6 has not been confirmed in other GI organs, although it is conceivable that its ubiquitous pattern of expression includes a majority of cell types and organs. NHE6 has a putative mitochondrial inner membrane targeting signal at its N terminus. This fact and initial fluorescence microscopy studies with GFP-tagged NHE6 expressed in HeLa cells costained with mitochondrial stain MitoTracker Red, as well as organelle fractionation, initially suggested that NHE6 represents the putative mitochondrial NHE regulating intramitochondrial Na + and H + levels as suggested by Garlid. Later evidence, however, provided evidence that the positive-charge-rich segment is not a mitochondrial-targeting sequence, and that NHE6 accumulates not in the mitochondria but in the sorting and recycling compartment of the endoplasmic reticulum where it may contribute to establishing of organelle-specific pH and in ion homeostasis. NHE6 has also been shown to transiently appear on the cytoplasmic membrane, as hemagglutinin tag introduced into the predicted EL1 extracellular loop of NHE6 was detected on the surface of live cells at a temperature nonpermissive to endocytosis (4°C). The physiological relevance of this phenomenon remains unclear.
Physiological and Pathophysiological Roles
NHE6 and its transport properties have not been fully evaluated. Its physiological function has been extrapolated from its cellular localization and homology to other member of the NHE gene family. Its high homology to NHE7 suggests that it too may represent a nonselective cation/H + exchanger and may participate in organellar pH and volume regulation. It has been speculated that NHE6 may participate in vesicle fusion and lysosomal biogenesis, and by regulating pH in the early and recycling endosomes, NHE6 may participate in regulation of surface receptor recycling. Basolaterally expressed NHE6 has been recently implicated in regulation of intracellular pH in mineralizing osteoblasts, whereas genetic linkage analysis pointed to mutations in SLC9A6 gene as an underlying mechanism for X-linked mental retardation, microcephaly, epilepsy, and ataxia, a phenotype mimicking Angelman syndrome. A recent study with HepG2 hepatoma cells also implicated the endosomal NHE6.1 splice variant in securing the polarized distribution of membrane lipids at the apical surface and maintenance of cell polarity and apical bile canaliculi. Chen et al. recently described expression of NHE6 in Caco-2 cells and showed its downregulation (along with NHE1) in cells infected with rotavirus. This report did not identify the subcellular localization of NHE6, thus it remains unclear whether iťs reduced expression contributes in any way to rotavirus diarrhea or other aspects of epithelial cell function.
This isoform was cloned through a combination of computational and molecular biology approaches by Numata and Orlowski. NHE7 is a 725 amino acid, ~ 80 kDa transmembrane protein with over 70% homology to NHE6 and ~ 58% homology with NHE9. Hydropathy analysis predicts membrane topology similar to that of other members of the NHE family, with 12α-helical transmembrane domains, followed by a hydrophilic cytoplasmic C terminus.
Tissue Distribution and Cellular Localization
Human NHE7 gene is expressed ubiquitously, but most prominently in the putamen and occipital lobe of the brain, in skeletal muscle, and in a number of secretory tissues such as prostate, stomach, pancreas, pituitary gland, adrenal gland, thyroid gland, salivary gland, and mammary gland. It is also expressed in the liver, small intestine, and colon, and this broad pattern of expression implies that NHE7 serves a “housekeeping” function. Inducible expression of hamagglutinin-tagged NHE7 in CHO cells followed by Western blot analysis identified two diffuse bands ~ 180 and 80 kDa, suggesting formation of moderately stable homodimers and a possibility of glycosylation. Dual labeling experiments revealed that NHE7 accumulates predominantly in a juxtanuclear compartment partially overlapping-mannosidase II-positive medial and trans-cisternae of the Golgi apparatus, but has been further localized primarily to the trans-Golgi network (TGN) and mid-trans-Golgi stacks.
Compared to control cells, NHE7-overexpressing CHO cells demonstrated ~ 75% higher rates of 22 Na + influx into intact intracellular membrane compartments after permeabilization of the plasma membrane with saponin. This NHE7-mediated 22 Na + uptake was pH gradient-sensitive, as H + -specific ionophore carbonyl cyanide m -chlorophenylhydrazone (CCCP), rapidly dissipating the organellar transmembrane H + gradient, significantly reduced NHE7-mediated 22 Na + influx. Alkalinization of endomembrane compartments by sustained exposure to NH 4 Cl, led to decreased 22 Na + influx, while rapid acidification of the intracellular compartment by pretreatment and rapid removal of NH 4 Cl increased 22 Na + uptake in both control and NHE7-overexpressing cells about threefold. Taken together, these data indicate the existence of an endogenous organellar Na + influx pathway that depends on the transmembrane H + gradient and that is upregulated in NHE7 HA -overexpressing cells. NHE7 has also been determined to be a relatively nonselective monovalent cation/H + exchanger, able to transport Na + , K + , Li + , and Rb +. Numata and Orlowski postulated that since K + is the main intracellular cation, NHE7 serves primarily as a K + /H + exchanger and its roles in the homeostasis of the trans -Golgi include providing a pathway for H + efflux and participating in controlling the organelle’s volume through transmembrane K + flux. The role of this NHE isoform within the alimentary tract has not been studied.
NHE8 isoform was cloned from a mouse kidney cDNA library by Goyal et al. The characterized sequence encodes a 576 amino acid protein, sharing 96% identity with its likely human ortholog (NM_015266). Human NHE8 protein shares < 24% homology with other known NHE isoforms. Hydropathy analysis predicts membrane topology similar to the general models for all NHEs with 10–12 transmembrane domains in the N-terminal portion of the protein, followed by a relatively short (~ 100 amino acids) hydrophilic C-terminal tail. The molecular mass of NHE8 detected by Western blotting is ~ 85 kDa, significantly higher than the 64 kDa predicted from the length of open reading frame. Consistent with the prediction of four N -glycosylation sites, inhibition of glycosylation with tunicamycin reduced the size of detected protein, confirming the posttranslational modification of NHE8.
Tissue Distribution and Cellular Localization
Initial expression analysis indicated ubiquitous expression of NHE8 in mouse tissues with predominant expression in the liver, skeletal muscle, kidney, and testes. In the kidney, NHE8 mRNA was localized by in situ hybridization primarily to the proximal tubules of the outer stripe of the outer medulla and to a lesser extent to the renal cortex, while NHE8 protein copurified with brush-border membranes. Further immunolocalization studies determined that renal NHE8 is expressed on both microvillar surface membranes and the coated pit regions in the epithelial cells of proximal tubules. The colocalization of NHE8 with megalin in the intermicrovillar coated pits and subapical tubules suggest that similar to NHE3, NHE8 may be regulated by endocytic retrieval and recycling. Plasma membrane localization of NHE8 has not been confirmed in the heterologous expression system of human NHE8 in COS-7 cells, where NHE8 was convincingly shown to be expressed in the mid- to trans-Golgi compartments and not on the plasma membrane. Human NHE8 is expressed in the liver, small intestine, and colon. There are segmental differences in NHE8 expression and they don’t closely overlap in mice and humans. Higher NHE8 expression was seen in stomach, duodenum, and ascending colon in human, while higher NHE8 expression was seen in jejunum, ileum, and colon in adult mouse. Moreover, the expression level of NHE8 is much higher in the stomach and jejunum in young mice compared with adult mice. In the stomach, NHE8 is expressed on the apical membrane in the epithelial cells of fundic and pyloric glands. Xu et al. described cloning of rat NHE8 cDNA and development of polyclonal antibodies raised against both N- and C-terminal peptides which detected NHE8 in the apical membrane of jejunal enterocytes both by western blotting and immunohistochemistry. NHE8 is broadly expressed in the colon, including the goblet cells, where it participates in the regulation of mucin 2 (Muc2) expression.
NHE8 colocalizes with ER, Golgi and intracellular vesicles in the retinal pigment epithelium of the eye are necessary for the survival of photoreceptor cells. It is also expressed on the plasma membrane of the epithelial cells in the conjunctiva, the cornea, and the lacrimal glands in human and mice. The apparent discrepancies between subcellular localization of NHE8 in vitro and in vivo are unresolved. They may represent true variation in protein trafficking, or may stem from methodological differences or differences in protein distribution in polarized epithelial cells and nonpolarized cell lines.
Human NHE8 protein was overexpressed, purified from Saccharomyces cerevisiae where it primarily associated with endoplasmic recticulum, and reconstituted in artificial liposomes for transport studies. Under outwardly directed pH gradient, proteolysosomes showed progressive intravesicular alkalinization, as well as 22 Na uptake, consistent with the action of NHE8 as a sodium/proton exchanger. Interestingly, K + also stimulated alkalinization of proteolysosomes, suggesting that NHE8 acts as a nonspecific cation/H + exchanger, similar to other organellar NHEs, for example, NHE7. Consistent with these observations, overexpression of NHE8 dissipated the acidic pH of the Golgi complex and increased the pH by about 0.78 pH unit from ~ 6.5 to pH 7.28, indicating an active role of NHE8 in mediating intraorganellar pH. In another study with stably transfected NHE-deficient PS120 fibroblasts, NHE8 conferred sodium-dependent proton exchanger activity with a Km for intracellular H + at about pH 6.5 and a Km for Na + at 23 mM. NHE8 activity was insensitive to low (1 μM) HOE694 concentration, but was significantly inhibited by 10 μM HOE694 and 80 μM S3226.
Analysis of NHE8 gene expression and brush-border protein abundance in preweaning and adult rats showed an age-dependent decrease in NHE8 expression in the small intestinal epithelium. This trend is opposite to that shown for the other two major epithelial NHE, NHE2, and NHE3, with which expression and activity increase around weaning and may point to a possible role for NHE8 in small intestinal Na + absorption during early stages of postnatal development.
Jejunal expression of NHE8 was inhibited in trinitrobenzene sulfonic acid (TNBS) colitis and in LPS-treated rats. This decrease could be reproduced in TNF-treated human Caco-2 colonic carcinoma cells, where a transcriptional mechanism mediated by decreased Sp3 interaction with the human proximal NHE8 promoter was proposed as a mechanism. Although the contribution of NHE8 to the intestinal Na + absorption is still not clear, it was postulated that this inhibition might contribute to the inflammation-associated diarrhea. While NHE8 −/− mice do not show symptoms of impaired Na + absorption or diarrhea, increased NHE2 and NHE3 expression in the small intestine may compensate for the loss of NHE8 activity in this model. Similar transcriptional mechanism has been also implicated in EGF-induced inhibition of NHE8 expression in intestinal epithelial cells.
The antidiarrheal effects of somatostatin analogs have also been recently studied in the context of NHE8 function. Wang et al. demonstrated that octreotide (somatostatin analogue) increased brush-border membrane NHE8 expression in the small intestine and in Caco-2 cells. This effect was mediated by SSTR2 receptor and required p38 MAPK activity. A follow-up study showed that NHE8 expression is reduced in patients with ulcerative colitis and in murine DSS colitis. Octreotide as well as specific agonists of somatostatin receptor 2 (SSTR2) and somatostatin receptor 5 (SSTR5) (seglitide and L-817,818, respectively) restored NHE8 expression in colitic mice.
NHE8 may also contribute to mucosal homeostasis through its emerging roles in the goblet cell function. In experimental colitis, NHE8 is downregulated, and its deletion confers higher susceptibility with an enhanced T-helper 2-like response. Reduced expression or absence of NHE8 also leads to decreased Muc2 mRNA expression in goblets cells and decreased mucosal expression of antimicrobial peptides. Reduced Muc2 expression in NHE8 −/− mice could be reproduced in colonic organoids from this strain (Xu et al., unpublished data), which is likely responsible for the reduced inner mucous layer and closer proximity of luminal bacteria to the brush-border membrane in these mice. Consistent with this observation, deletion of NHE8 in vivo led to higher adhesion of Salmonella typhimurium , with similar effects seen in vitro in epithelial cells with siRNA-mediated NHE8 knockdown. Interestingly, NHE8 knockdown did not affect the adhesion of a probiotic strain Lactobacillus plantarum JDM1, thus suggesting at least some degree of selectivity that could not be explained by mucus production alone.
NHE9 isoform was identified by de Silva et al. as one of two genes affected by a pericentric inversion of chromosome 3, 46N inv(3)(p14:q21), associated with attention deficit hyperactivity disorder (ADHD). Computational analysis indicated homology with mammalian NHE, with highest similarity to human NHE6 (61.2% at the protein level). Phylogenetic analysis indicates that NHE9 belongs to the same clade as other organellar isoforms NHE6 and NHE7 ( Fig. 56.2 ). The SLC9A9 gene spans approximately 470 kb of genomic DNA, has 16 exons, and its cDNA is approximately 3.4 kb in length. The predicted SLC9A9 protein has 645 amino acids, has a molecular weight of 72.6 kDa, and its sequence analysis suggests a membrane protein with 10 or 12 transmembrane domains.
Tissue Distribution and Cellular Localization
Northern blot analyses indicate that NHE9 expression is fairly widespread, with the highest expression in the heart, skeletal muscle, and brain. In the GI tract, NHE9 was detected in the liver and in smaller quantities in the small intestine. No detectable levels were observed in the colon. The homology of NHE6 and NHE9 proteins suggested that both of these isoforms are located in the intracellular compartments. Coimmunolocalization studies in COS-7 cells combined with rhodamine-labeled transferrin chase imply that NHE9 is primarily expressed in late recycling endosomes, with a smaller proportion of NHE9 localizing to EEA1-positive early endosomes.
Although initial identification of NHE9 was purely computational, functional studies confirm its identity as a NHE. In COS-7 cells, overexpression of NHE9 resulted in luminal alkalinization to near cytosolic pH in the compartments in which NHE9 resided [from pH 6.73–7.14, as indicated by pH-sensitive GFP mutant (EYFP) fusion with NHE9]. Since overexpression systems (in this case, expression close to 100-fold higher than that of endogenous NHE9) may tend to overestimate the physiological role of NHE9 in regulating intraorganellar pH, more data from gene targeting or knockdown experiments will be needed to validate these results. The NHE9 gene has been cloned relatively recently, and there is very limited or no published data pertinent to transcriptional and posttranscriptional modes of regulation of its expression and activity.
The role of NHE9 in the physiology of the GI tract has not been elucidated. A recent genome-wide scan of ADHD affected families identified marker D3S1569 (lod 1.37, P 0.0060) located in intron 5 of the SLC9A9 gene. Although the lod score did not exceed the QTL threshold for significance, its location within the NHE9 gene suggests that it could play a role in the development of the ADHD phenotype. The expression of NHE9 in neuronal tissues (primarily medulla and spinal cord) and more recent genetic studies seem to supports this hypothesis. Three genetic variants (L236S, S438P, and V176I) were also associated with autism, were shown in astrocytes to represent loss function mutations, and Kondapalli et al. suggested that these mutations may contribute to the autistic phenotype by modulating synaptic membrane protein expression and neurotransmitter clearance.