Ca2+ is an essential ion in all organisms, where it plays a crucial role in processes ranging from the formation and maintenance of the skeleton to the temporal and spatial regulation of neuronal function. The Ca2+ balance is maintained by the concerted action of three organ systems, including the gastrointestinal tract, bone, and kidney. An adult ingests on average 1 g Ca2+ daily from which 0.35 g is absorbed in the small intestine by a mechanism that is controlled primarily by the calciotropic hormones. To maintain the Ca2+ balance, the kidney must excrete the same amount of Ca2+ that the small intestine absorbs. This is accomplished by a combination of filtration of Ca2+ across the glomeruli and subsequent reabsorption of the filtered Ca2+ along the renal tubules. Bone turnover is a continuous process involving both resorption of existing bone and deposition of new bone. The above-mentioned Ca2+ fluxes are stimulated by the synergistic actions of active vitamin D (1,25-dihydroxyvitamin D3) and parathyroid hormone. Until recently, the mechanism by which Ca2+ enter the absorptive epithelia was unknown. A major breakthrough in completing the molecular details of these pathways was the identification of the epithelial Ca2+ channel family consisting of two members: TRPV5 and TRPV6. Functional analysis indicated that these Ca2+ channels constitute the rate-limiting step in Ca2+-transporting epithelia. They form the prime target for hormonal control of the active Ca2+ flux from the intestinal lumen or urine space to the blood compartment. This review describes the characteristics of epithelial Ca2+ transport in general and highlights in particular the distinctive features and the physiological relevance of the new epithelial Ca2+ channels accumulating in a comprehensive model for epithelial Ca2+ absorption.
The maintenance of the extracellular Ca2+ concentration is of utmost importance for many vital functions of the body. In face of large variations in Ca2+ in- and output, the organism is equipped with a set of regulatory systems to keep the plasma Ca2+ levels around a value of 2.5 mM. To this end, the Ca2+ fluxes between the extracellular compartment and several organ systems are tightly controlled.
Ca2+ absorption occurs in epithelia, including kidney, intestine, placenta, mammary glands, and gills. In mammals, the small intestine and kidney constitute the influx pathways into the extracellular Ca2+ pool, while fish have an additional specialized organ for Ca2+ uptake, the gills (426). In addition, during reproductive events, Ca2+ transport in the placenta and mammary glands can seriously influence the Ca2+ balance (29, 345). For instance, during lactation, plasma Ca2+ levels can significantly drop due to Ca2+ excretion into the milk. Furthermore, during pregnancy Ca2+ transport from the mother to the fetus takes place across the placenta and challenges the plasma Ca2+ concentration. The Ca2+ demand is great in children during skeletal growth and decreases gradually with advancing age. Finally, factors as pH, extracellular Ca2+ concentration, and various hormones have been shown to influence the Ca2+ movement across epithelia (37, 129, 175).
In general, Ca2+ transport is mediated by a complex array of transport processes that are regulated by hormonal, developmental, and physiological factors. The distinct processes by which Ca2+ can be absorbed across epithelial tissues include paracellular and transcellular pathways. The paracellular pathway allows the direct exchange of Ca2+ between two compartments, while the transcellular route involves transport across at least two plasma membrane barriers (Fig. 1). The vitamin D metabolite 1,25-dihydroxyvitamin D3 [1,25-(OH)2D3] and parathyroid hormone (PTH), among others, are prominent hormones controlling the Ca2+ balance.
Numerous methods have been utilized to study the process of Ca2+ absorption. Traditionally, metabolic balance, micropuncture, and radioactive tracer techniques were employed to estimate epithelial Ca2+ fluxes in humans and animal models. More recently, freshly isolated epithelia and established cell models were used to measure Ca2+ transport by radioisotopes of Ca2+ and fluorescent indicators. Finally, molecular biological tools were applied to identify and characterize the individual transport proteins and to develop new animal and cell models to further delineate the process of Ca2+ transport. As a consequence, current research is beginning to define the molecular mechanisms, the regulation, and coordination of the epithelial Ca2+ transport mechanisms (35, 125). Here, we review our present understanding of these processes in detail and discuss potential implications for further research.
II. PARACELLULAR TRANSPORT
A fundamental function of epithelia is to separate different compartments within the organism and to regulate the exchange of substances between them (151, 425). Epithelia consist of a continuous layer of individual cells, and the intercellular spaces between the epithelial cells are very narrow, but nonetheless allow the diffusion of small molecules and ions (152). This route is called the paracellular pathway, which must be regulated for the epithelium to remain selectively permeable (Fig. 1). Depending on the functional requirements of an epithelium, there may be small or large amounts of solutes flowing passively through this path. The tight junction constitutes the barrier to the passage of ions and molecules through the paracellular pathway (5). In general, the importance of this route has not been as thoroughly investigated as the role of the transepithelial pathway. However, recent studies identifying the molecular components of the paracellular transport route as causally related to particular inherited diseases, including familial hypomagnesemia (353), hypertension (429), and autosomal recessive deafness (428), confirm the importance of paracellular transport.
A. Tight Junction
The tight junction is a specialized membrane domain at the most apical region of polarized epithelial cells that not only creates a primary barrier to prevent paracellular transport of solutes, but also restricts the lateral diffusion of membrane lipids and proteins to maintain the cellular polarity. Tight junctions are intercellular structures in which the plasma membranes of adjacent epithelial cells come into very close contact. These structures consist as linear arrays of integral membrane proteins, which include occludin, claudins, and several immunoglobulin superfamily members, such as the junctional adhesion molecule (98, 152, 248). The claudin family consists of at least 20 related integral membrane proteins with four transmembrane (TM) domains and functions as major structural components of the tight junctional complex, while occludin is an accessory protein involved in the tight junction formation of which three isoforms have been described (262, 378, 384).
It has been hypothesized that tight junctions behave similar to conventional ion channels, while others have suggested a function as water-filled channels with no ion selectivity (5, 384, 434). Recently, it was shown by applying new technology that the tight junction complex manifests biophysical properties of ion channels including ion and size selectivity, concentration-dependent ion permeability, competition between permeable molecules, anomalous mole-fraction behavior, and sensitivity to pH (378). This latter study suggests that discrete ion channels are being inserted in the tight junction to facilitate paracellular ion transport. This is further supported by the disease mutations documented in claudins (434). Mutations in claudin-16, which are associated with a renal Mg2+ wasting syndrome, implicate this particular claudin in paracellular resorption of Mg2+ and Ca2+, but not monovalent ions (353). Mutations in claudin-14 cause nonsyndromic recessive deafness, and this tight junction protein is essential to maintain the electrochemical gradient between the endolymph and its surrounding tissues (428).
Movement of ions through the tight junctions is a passive process, which largely depends on the concentration gradient of the permeable ions and the electrical gradient across the epithelium. Hormones and factors affecting the electrochemical gradient across the epithelium, therefore, indirectly influence the passive fluxes through the tight junctions. The tight junction permeability itself is dynamically regulated under various physiological conditions (152, 299). Tight junctions undergo modulation by growth factors, cytokines, bacterial toxins, hormones, and other factors (30, 134, 153, 419). Some studies suggest that phosphorylation of tight junction proteins plays a role in tight junction assembly and function. Modulation of protein kinase C (PKC) disrupts the cell-cell junctions of epithelial cells, whose action is supposedly mediated by mitogen-activated protein kinase (419). Recently, the serine-threonine kinase WNK4 has been identified as part of the junctional complex, and mutations in this protein cause pseudoaldosteronism type II, a Mendelian trait featuring hypertension (429). Furthermore, a serine-threonine-dependent phosphorylation of tight junction proteins has been shown to enhance the paracellular permeability in rabbit nasal epithelium (294). These studies underline the essential role of serine-threonine kinase in the regulation of the paracellular route, but the precise mechanism remains to be established.
III. TRANSCELLULAR TRANSPORT
Transcellular Ca2+ transport is a multistep process, comprised of the transfer of luminal Ca2+ into the enterocyte or renal epithelial cell, the translocation of Ca2+ from point of entry to the basolateral membrane, and finally active extrusion from the cell into the circulatory system (Fig. 1).
A. Calcium Entry
Ca2+ is postulated to enter the epithelial cell via Ca2+-selective channels at the luminal membrane under the influence of a steep, inwardly directed electrochemical gradient. The molecular nature of the apical entry mechanism has remained obscure for a long time. In the past, the mechanism responsible for the luminal entry of Ca2+ has been extensively characterized in renal and intestinal cells, but these studies have not led to the definite identification of this Ca2+ transporter. A variety of studies implicated Ca2+ channels in mediating Ca2+ entry in absorptive epithelia (206, 228, 443). From these studies it is likely that several distinct Ca2+ channels, which resemble the classified voltage-dependent Ca2+ channels in part, are present in cells of the distal part of the nephron (443). Previous studies indicated the presence of the pore-forming subunit α1A or α1 of a Ca2+ channel expressed in the Ca2+-transporting renal cells. In general, voltage-sensitive Ca2+ channels are heteromeric, multisubunit proteins, as has been well demonstrated for the skeletal muscle (69) and brain (431) channels, and their properties are not only determined by the α1-subunit itself, but also depend on the presence of regulatory subunits. Of these, the β3-subunit might be an interesting subunit, which can influence such fundamental channel properties as current amplitude (355), kinetics (226, 355, 400), voltage dependence (355), and regulation by cyclic nucleotide-dependent protein kinases (216). Stimulation of the Ca2+ influx channel in the renal distal tubule by the calciotropic hormone PTH is thought to be mediated by both protein kinase A (PKA) and PKC (124, 170). Comparable studies in rat duodenum pointed to acute effects of 1,25-(OH)2D3 on Ca2+ influx in isolated enterocytes (112–114, 144, 152). This calciotropic hormone significantly increased 45Ca2+ uptake within 1–10 min in a dose-dependent manner (249). The effects of 1,25-(OH)2D3 were mimicked by the Ca2+ channel agonist BAY K 8644 and completely abolished by nifedipine and verapamil. Incubation of duodenal cells with 1,25-(OH)2D3 rapidly (1–5 min) increased cAMP levels. Forskolin caused a rapid increase in Ca2+ uptake by enterocytes that was similar to the action of the hormone. Moreover, pretreatment of cells with the specific cAMP inhibitor Rp-cAMPS suppressed the changes in 45Ca2+ influx induced by 1,25-(OH)2D3. These results suggested the involvement of a Ca2+ channel activation through the cAMP/PKA-pathway by 1,25-(OH)2D3 in mammalian intestinal cells (249). However, the physiological significance of these postulated Ca2+ channels with respect to Ca2+ influx during Ca2+ (re)absorption remains unclear, since most of the studies were performed in single cell systems rather than in polarized epithelia derived from kidney or duodenum epithelia. With the use of these nonpolarized cell preparations, it was not feasible to discriminate between apical versus lateral Ca2+ influx. Until now, two polarized confluent epithelial cell systems representing active duodenal and renal Ca2+ (re)absorption have been studied. First, Caco-2 cells were employed, which represent an intestinal cell line derived from a human colorectal carcinoma that spontaneously differentiates under standard culture conditions in a tissue that exhibits functional duodenal transport processes (144). These cells become polarized columnar epithelial cells, form tight junctions and domes, and express several markers that are unique to differentiated small intestinal epithelium (e.g., high sucrase-isomaltase mRNA and protein levels) (72). Caco-2 cells exhibit saturable apical-to-basolateral Ca2+ transport kinetics, net transport is positive in the apical-to-basolateral direction, and the rate of transport can be increased by pretreatment with 1,25-(OH)2D3 (114, 143). Vitamin D-induced upregulation of Ca2+ transport requires transcriptional events (114) and is modulated by changes in the vitamin D receptor (VDR) content of the cell (342). Second, the use of primary cultures and immortalized cell lines originating from the renal distal tubular cells greatly facilitated our understanding of Ca2+ influx in these cells and how PTH and 1,25-(OH)2D3 regulate transepithelial Ca2+ transport. Two groups, Bindels and co-workers (36–39, 170, 172, 177, 320, 389–391) and Gesek and Friedman and co-workers (12, 120, 124, 126–128, 139, 138, 246, 427), have used immunodissected cell lines from rabbit and mouse kidney, respectively, to investigate PTH-stimulated Ca2+ transport. These studies verified that PTH increases transepithelial Ca2+ transport and suggested that both PKA and PKC participate in this process (124, 170). In addition, Bindels and co-workers demonstrated that the primary cultures of rabbit connecting tubule (CNT) and cortical collecting duct (CCD) cells exhibit many characteristics of the original epithelium, including 1,25-(OH)2D3- and vasopressin-stimulated Ca2+ reabsorption (36–39, 50, 170, 172, 177, 218). The stimulatory effect of cAMP/PKA on apical Ca2+ influx is in agreement with a report of Tan and Lau (377), describing a 25-pS, cAMP-sensitive apical Ca2+ channel in rabbit kidney CNT cells, that is inhibited by dihydropyridine agonists and depolarization. However, luminal administration of a variety of Ca2+ channel antagonists, including dihydropyridines, failed to affect Ca2+ absorption in primary cultures of these CNT tubules, suggesting that Tan and Lau (377) did not study the apical entry Ca2+ mechanism.
Previous studies indicated that over a wide range of transepithelial Ca2+ transport rates, the Ca2+ influx at the apical membrane is correlated in a 1:1 fashion with the apical to basolateral 45Ca2+ flux (320). The mechanism underlying this tight coupling could in principle be located at three distinct Ca2+ transport positions, namely, influx, cytosolic diffusion, or efflux (175, 320). Ca2+ influx across the apical or luminal membrane could be the rate-limiting step for transcellular Ca2+ transport. This would imply that the availability of cytosolic Ca2+ for the basolateral Ca2+ efflux pumps controls the rate of Ca2+ efflux. This is in line with the finding that apical H+ directly inhibits the Ca2+ influx pathway and consequently decreases the intracellular Ca2+ concentration, which in turn limits the Ca2+ extrusion (37). In addition, extracellular H+ are known to inhibit voltage-gated Ca2+ channels in excitable tissues by changing their conformations and modifying their gating properties.
To unravel the molecular identity of the apical Ca2+ influx protein in Ca2+-transporting epithelia, several different experimental strategies have been employed. The classic approach of purification of the Ca2+ transporter followed by amino acid sequencing of the isolated peptide has been hindered by the lack of a rich source of channel protein. In another strategy, a calbindin-D28K affinity column has been applied to purify associated proteins potentially involved in transepithelial Ca2+ (re)absorption (191). Previous studies indicated that the intracellular Ca2+ concentration is a critical determinant for Ca2+ influx. In theory, the Ca2+ buffer calbindin, which reaches submillimolar concentrations in these epithelial cells, might interact with this putative influx channel to buffer and bind Ca2+ in a rapid way to facilitate an efficient Ca2+ transport. This attempt, however, did not unveil the molecular identify of the Ca2+ influx mechanism. Alternatively, investigators have focused on homology-based cloning strategies using sequences of previously described voltage-gated Ca2+ channels (21, 22, 397, 443), but also this approach did not result in the identification of the apical Ca2+ entry transporter.
Subsequently, functional expression cloning using a rabbit primary CNT/CCD cDNA library in Xenopus laevis oocytes was applied by Hoenderop et al. (178). This technique has previously been successful for the cloning of several epithelial transporters (66). The rationale to use the functional expression cloning approach was threefold: 1) mRNA isolated from primary cultures of rabbit CNT tubules, when injected in oocytes, induced a 45Ca2+ uptake two to three times above background; 2) transcellular Ca2+ transport in primary cultures of rabbit CNT tubules was not affected by voltage-gated Ca2+ channel blockers, which allowed us to distinguish between voltage-gated Ca2+ channels and the Ca2+ transporter involved in transepithelial Ca2+ transport; and 3) the absence of a substantial endogenous Ca2+ influx in Xenopus laevis oocytes. Based on these characteristics, an expression cloning strategy was established to identify the apical Ca2+ entry channel. First, a cDNA library from poly(A)+ RNA isolated from primary cultures of rabbit kidney CNT and CCD was generated and subsequently screened for Ca2+ uptake activity in Xenopus laevis oocytes in the presence of a cocktail of known voltage-gated Ca2+ channel antagonists, including nifedipine, verapamil, and Ba2+. After an extensive screening procedure, a single transcript was isolated encoding for a novel epithelial Ca2+ channel, named ECaC1 and recently renamed as the transient receptor potential channel TRPV5 (178, 257).
To study the functional characteristics of this new Ca2+ channel, TRPV5 was initially heterogeneously expressed in Xenopus laevis oocytes and later in human embryonic kidney (HEK293) cells. Functional data on TRPV5 include 45Ca2+ uptake, electrophysiology using voltage-clamp and patch-clamp experiments, and fluorimetric measurements (see sect. iv). In short, the channel was inhibited in order of potency by La3+ > Cd2+ > Mn2+. Permeability to Na+ was negligible in the situation where Ca2+ and Na+ were both present, whereas Ba2+ and Sr2+ did not affect the Ca2+ influx. These experiments unequivocally demonstrated that TRPV5 exhibits the defining properties of Ca2+ influx in Ca2+-transporting epithelia. Subsequently, Hediger and co-workers (257, 304) applied the functional expression cloning technique and identified the Ca2+ transporter 1 (CaT1 or recently renamed as TRPV6) from rat intestine, which shares 80% amino acid identity with TRPV5 (note: TRPV5 is also known in the literature as ECaC, ECaC1, and CaT2, whereas TRPV6 has been named previously CaT1, ECaC2, and CaT-like). Electrophysiological studies demonstrated that the characteristics of TRPV6 are comparable to those measured for TRPV5, but its expression pattern is more ubiquitous (see sect. iv). The functional properties of TRPV5 and TRPV6 are in line with those of epithelial Ca2+ (re)absorption, providing the first evidence that these transporters are the anticipated Ca2+ influx proteins initiating the process of transcellular Ca2+ transport, which is described in detail in section iv (Fig. 1) (175, 302).
B. Cytosolic Diffusion
Epithelial cells involved in transcellular Ca2+ transport are continuously challenged by substantial Ca2+ traffic through the cytosol, while simultaneously maintaining low levels of cytosolic Ca2+. To date, two models have been proposed to explain transport of Ca2+ across these cells. First, the facilitated diffusion model in which the basal rate of Ca2+ uptake and Ca2+ extrusion from the absorptive epithelial cell is proposed to be sufficient to accommodate the elevated rate of transport observed after vitamin D stimulation. In contrast, mathematical modeling predicts that intracellular diffusion is the rate-limiting step (58, 357). There are two major subclasses of vitamin D-dependent Ca2+-binding proteins, calbindin-D9K and calbindin-D28K. These cytosolic proteins have been proposed as shuttles that can bind Ca2+ and facilitate the Ca2+ diffusion between the apical and basolateral surfaces of the cell. Calbindin-D28K is highly conserved during evolution and present in kidney, small intestine (only birds), pancreas, placenta, bone, and brain, and calbindin-D9K is present in highest concentrations in small intestine as well as in kidney (only mouse). The expression level of these calbindins in kidney and intestine is closely correlated with the efficiency of Ca2+ (re)absorption and, therefore, these proteins play a central role in the facilitated diffusion model. Second, a vesicular model was proposed in which the absorptive cells use lysosomes to sequester Ca2+ and facilitate its movement to the basolateral membrane (230). Formation of Ca2+-enriched vesicles is initiated by influx of Ca2+ through Ca2+ channels in the apical or luminal membrane. The rapid increase in Ca2+ concentrations in close vicinity to the apical membrane disrupts the actin filaments near the Ca2+ channels and initiates the formation of endocytic vesicles. Ca2+ bind to calmodulin (CaM) associated with myosin I or alternatively CaM-associated with the Ca2+ channels, which leads to inactivation of the channels. Inactivation of the Ca2+ channels causes a decrease in the free Ca2+ levels close to the apical membrane, and the actin filament network can be restored. The formed Ca2+-containing vesicles are transported by microtubules and fuse with lysosomes (230). While calbindins have been found to associate with lysosomes, the role of these Ca2+-binding proteins in this latter model is less clear. Experimental evidence suggests, however, an important role in epithelial Ca2+ transport, which is best described by the first model in which calbindin-D9K and calbindin-D28K facilitate the cytosolic diffusion of Ca2+ from the apical influx to the basolateral efflux sites and acts as cytosolic Ca2+ buffer to maintain low intracellular Ca2+ levels during changes in transcellular Ca2+ transport (35, 107, 108). Interestingly, calbindin-D9K may directly enhance plasma membrane Ca2+-ATPase (PMCA) activity (417). Due to the relatively slow binding kinetics of these Ca2+-binding proteins, Ca2+ signaling can occur independently of transcellular Ca2+ movement mediated by calbindin-D9K and calbindin-D28K (218). Depending on the vitamin D state, the cytosolic calbindin-D concentration can reach values in the submillimolar range, which is indeed sufficient to fulfill the above-mentioned functions (see sect. viA) (38, 106).
Previously, a homozygous mutant of calbindin-D28K gene-knockout mice was generated by gene targeting which developed normally (3, 23, 361, 362). These animals exhibited a two times higher urinary Ca2+ excretion compared with wild-type littermates, but no significant differences in serum Ca2+, PTH, or in serum and urinary Mg2+ and phosphate were observed (361). This suggests that the hypercalciuria induced by calbindin-D28K deficiency is compensated by, for instance, increased intestinal absorption of Ca2+. Renal calbindin-D9K expression was not affected in these knockout mice. The cytosolic Ca2+ transport function of calbindin-D9K remains to be confirmed, since calbindin-D9K knockout mice have not been generated until now.
The calbindins, like CaM, belong to a group of intracellular proteins that bind Ca2+ with high affinity and undergo structural changes upon binding (33). Each calbindin is encoded by a separate gene, and there is no direct association between the two genes. In fact, nature has produced a wide variety of EF-hand Ca2+-binding proteins (i.e., calbindin-D9K, calbindin-D28K, CaM, parvalbumin, S100, troponin C) that display minor overall sequence homology (307). An important functional feature of CaM and troponin is their ability to interact with and regulate the function of voltage-gated Ca2+ channels in a Ca2+-dependent fashion. The ubiquitously expressed CaM directly interacts with an IQ motif present in the carboxy termini of these channels where it functions as a Ca2+ sensor (358). This IQ motif is, however, not present in TRPV5 and TRPV6. At present, it is unknown whether calbindin-D9K or calbindin-D28K could fulfill a similar Ca2+ sensor function for which a specific interaction with TRPV5 and/or TRPV6 would be required. Freud and Christakos (118) reported an interaction of calbindin-D28K with microsomal membranes in kidney (118), whereas Shimura and Wasserman (351) showed that calbindin-D28K is associated with purified chicken intestinal brush-border membranes. Together with the striking colocalization of calbindin-D9K and/or calbindin-D28K in all TRPV5/6-expressing tissues, this suggests a functional interaction between these two proteins. Future experiments are needed to delineate whether the function of calbindin is restricted to its buffer capacity maintaining low Ca2+ concentrations in close vicinity of the channel mouth or whether a physical interaction between calbindin and TRPV5/6 is needed to exert a direct regulatory function.
C. Extrusion Mechanisms
The efflux of Ca2+ occurs against a considerable electrochemical gradient, and two Ca2+ transporters have been located in the basolateral membrane of absorptive cells to extrude Ca2+, i.e., a Na+/Ca2+ exchange mechanism (NCX) and a Ca2+-ATPase (PMCA).
1. The Na+/Ca2+ exchanger
To date, three genes for NCX, designated NCX1, NCX2, and NCX3, have been identified in mammals. Similarities between these proteins include a homology of ∼70% sequence identity, the presence of an amino-terminal signal sequence, two sets of multiple transmembrane α-helices near the ends of the protein, and a large intracellular loop (46, 340). Splicing of RNA transcripts is a general characteristic of the NCX genes in mammals to generate diversity. Reilly and Shugrue (325) published the sequence of the rabbit kidney NCX1, and in kidney the expression of this transporter is restricted to the distal part of the nephron where it is predominantly localized along the basolateral membrane (40, 171, 240). NCX1 is widely distributed in many different mammalian tissues, whereas NCX2 and NCX3 are only expressed in brain and skeletal muscle (236, 275). Unfortunately, specific inhibitors of NCX1 are not available to substantiate the relative importance of this exchanger for overall Ca2+ reabsorption. Bindels and co-workers (39, 391) demonstrated that NCX1 is the primary extrusion mechanism, whereas only a minor amount of Ca2+ in the distal tubular cells is extruded by the plasma Ca2+ pump. NCX1 is also expressed in the basolateral membrane of the enterocytes (164, 210, 387). In fish enterocytes, NCX appears to be the main mechanism by which transcellular Ca2+ fluxes are extruded from the cells at the basolateral surface, whereas in mammals PMCA is the predominant extrusion mechanism (115, 164, 337, 387). Together these functional studies suggest that in kidney, basolateral Ca2+ extrusion is mainly carried out via NCX1, whereas Na+/Ca2+ exchange seems of minor importance in the small intestine. Recently, it was demonstrated that targeted deletion of NCX1 results in NCX1-null embryos that do not have a spontaneously beating heart and die in utero (219, 327). Therefore, this animal model is, unfortunately, not suitable to verify the importance of NCX1 in renal epithelial Ca2+ transport.
In addition, several K+-dependent Na+-Ca2+ exchangers (NCKX) have been described (46, 308). Northern blot analysis demonstrated that some isoforms [i.e., NCKX4 (235) and NCKX6 (65)] of this family are expressed in epithelia including small intestine and kidney. Ubiquitous expression of these exchangers in various tissues suggests a key role in regulating intracellular Ca2+ homeostasis in mammalian cells. It remains to be established whether these transporters play a role in epithelial Ca2+ transport.
2. Regulation of NCX
Kimura et al. (212) were the first to directly measure a NCX current. Their experiments were carried out in guinea pig cardiac myocytes and provided evidence that the NCX exchanger actually generates a measurable membrane current in which the exchanger can translocate a net positive charge across the plasma membrane (100). The stoichiometry of NCX has been investigated by several groups, and calculations varied from 3 Na+:1 Ca2+ (15) to 4 Na+:1 Ca2+ (266, 267), indicating that this transporter is electrogenic. Comprehensive functional studies in oocytes and mammalian cell systems indicated that NCX is regulated by several factors including the membrane potential, PKC activation, protons, nucleotides, and calciotropic hormones (46).
Interestingly, it has been shown that NCX1 is regulated by PTH. Functional data demonstrated that PTH markedly stimulates Ca2+ reabsorption in the distal part of the nephron primarily by augmenting NCX1 activity via a cAMP-mediated mechanism.
However, the exact mechanism by which PTH activates the exchanger remains controversial. Initial experiments suggested that the kidney isoform of NCX1 is poorly activated by PKA (161). In another study, it was found that PTH does not affect the intracellular Ca2+ concentration (312), indicating that the activation of NCX1 cannot be explained by increased substrate availability. Moreover, it was postulated that PTH increases Cl− conductance in DCT cells leading to decreased intracellular Cl− activity and membrane hyperpolarization (137). Hyperpolarization of the basolateral membrane will increase the Na+/Ca2+ exchange rate. Therefore, coactivation of apical Ca2+ entry through TRPV5 occurs under conditions that are favorable for basolateral Ca2+ extrusion (see also sect. iv). Other studies on isolated basolateral membrane vesicles implied that the PTH-induced rise in the intracellular Ca2+ concentration is not required for the stimulatory effect of PTH (195). Whether PTH stimulation affects expression of the NCX1 mRNA or the protein itself is not known. In addition to PTH, the calciotropic hormone 1,25-(OH)2D3 also regulates the renal expression of NCX1. Studies in vitamin D-deficient knockout models showed an impressive downregulation of NCX1 mRNA that could be normalized by 1,25-(OH)2D3 supplementation (sect. vi) (169). In these animal models there was no significant downregulation of PMCA in line with a primary role of NCX in Ca2+ extrusion. These findings point to a supportive role of the exchanger in vitamin D-stimulated Ca2+ reabsorption.
PMCAs are high-affinity Ca2+ efflux pumps present in virtually all eukaryotic cells, wherein they are responsible for the maintenance and resetting of the resting intracellular Ca2+ levels (45). Four genes encode separate isoforms designated PMCA1–4. In addition, alternative splicing of the transcripts yields a large variety of splice variants differing mainly in their carboxy-terminal amino acid sequence (365, 369). PMCAs are a universal system for the extrusion of Ca2+ in cells. In kidney, PMCA is, in contrast to NCX, present in all nephron segments with highest expression in the basolateral membrane of cells lining the distal part of the nephron. Compared with other nephron segments, the distal convoluted tubule (DCT) possesses the highest Ca2+-ATPase activity (95) and exhibits the strongest immunocytochemical reactivity for PMCA protein expression (48, 49, 245). Studies at the transcript level using RT-PCR and advanced tissue microdissection techniques indicated that all four PMCA isoforms are distinctively expressed in the kidney, and variable abundance of the individual isoforms along the different regions of the nephron has been documented (369). Other studies using RT-PCR on whole kidney RNA as well as studies at the protein level have been controversial but suggested that PMCA1 and PMCA4 are the major isoforms expressed in the kidney, while PMCA2 and PMCA3 may be minor components (246). Based on the fact that PMCA1 and PMCA4 are widespread, while PMCA2 and PMCA3 are more tissue specific, it has been suggested that PMCA1 and PMCA4 are housekeeping isoforms involved in the maintenance of cellular Ca2+ homeostasis (365). At variance with this conclusion is the observation that PMCA1b transcripts were definitely observed in rabbit CNT and CCD, whereas expression of the PMCA2 isoform was not (171, 213). In addition, Kip and Strehler (213, 214) demonstrated in Madin-Darby canine kidney (MDCK) cells that PMCA4b plays a significant role in basolateral Ca2+ extrusion. Furthermore, PMCA1b is the predominant isoform and abundantly expressed in the small intestine, where NCX1 is expressed at a low level. These data suggest indirectly that PMCA1b is the principal Ca2+ extrusion mechanism in intestinal Ca2+ absorption.
4. Regulation of PMCA
In general, there is only limited data available regarding the regulation of PMCA by hormones or signaling mechanisms. Several studies indicated that PMCA is positively regulated by 1,25-(OH)2D3 in the intestine to increase Ca2+ absorption. Cai et al. (64) addressed the effect of vitamin D on the synthesis of the chicken intestinal PMCA mRNA. Northern blot analysis indicated that repletion of vitamin D-deficient chickens with vitamin D increases PMCA mRNAs in the duodenum, jejunum, ileum, and colon. After injection of 1,25-(OH)2D3 intravenously in these deficient chickens, duodenal PMCA mRNA tended to increase by 2 h, reached a maximum at ∼16 h, and returned to baseline levels at 48 h (64). These results were confirmed by Johnson and Kumar (199), who demonstrated that 1,25-(OH)2D3 causes an increase in abundance of the PMCA and stimulates Ca2+-pumping activity. Kip and Strehler (214) showed that 1,25-(OH)2D3 upregulates the expression of the PMCAs (mainly PMCA4b) in MDCK cell lysates in a time- and dose-dependent manner. Interestingly, 1,25-(OH)2D3 caused a decrease of the PMCAs in the apical plasma membrane fraction and a concomitant increase of the number of pumps in the basolateral membrane. Functional transport assays demonstrated that transcellular 45Ca2+ flux from the apical-to-basolateral compartment was significantly enhanced by 1,25-(OH)2D3. These findings confirm that 1,25-(OH)2D3 is a positive regulator of the PMCAs in MDCK cells.
Prince and colleagues (90) demonstrated a stimulatory effect of estrogen and dihydrotestosterone on PMCA activity measured in isolated vesicles from Madin-Darby bovine kidney cells (MDBK) with a magnitude comparable to that of 1,25-(OH)2D3. Unlike 1,25-(OH)2D3, which stimulates PMCA protein expression in MDBK cells, neither estrogen nor dihydrotestosterone increased PMCA protein expression (90). This suggests that these latter hormones regulate Ca2+ transport by increasing PMCA activity, rather than by increasing PMCA protein abundance. Furthermore, PMCA activation is dependent on CaM, and inhibition of CaM is in turn known to prevent PMCA stimulation. In MDCK cells, trifluoperazine and calmidazolium, two inhibitors of CaM, decreased the Ca2+ transport by 45 and 33%, respectively (213). Kip et al. (213) used a variety of agents known to inhibit the PMCAs in these cells to determine the contribution of the Ca2+ pump to transcellular Ca2+ flux. Interestingly, ∼30% of this flux was due to PMCA and the remaining flux depended on a Na+/Ca2+ exchange process. These results are in line with the dominant role of NCX1 in transcellular Ca2+ flux studies in cultured rabbit kidney cells isolated from CNT and CCD (39).
IV. EPITHELIAL CALCIUM CHANNELS
A. Molecular Diversity
TRP channel proteins constitute a large and diverse family of proteins that are expressed in many tissues and cell types (74, 256). The large functional diversity of TRPs is reflected in their diverse permeability to ions, activation mechanisms, and involvement in biological processes ranging from pain perception to male aggression. Mammalian homologs of the Drosophila TRP gene encode a family of at least 20 ion channel proteins (Fig. 2). They are widely distributed in mammalian tissues, but most of their specific physiological functions are largely unknown. The molecular structure that is conserved among all members of the TRP family is a channel subunit, containing six TM spanning domains and a pore region loop, that most probably assemble into tetramers to form unique ion channels allowing the influx of cations into cells (Fig. 3A) (182). The TRP channels can be divided by sequence homology into at least six subfamilies, designated TRPC (canonical or classical), TRPV (vanilloid), TRPM (melastatin), TRPP (polycystins, PKD-type), TRPA (for ankyrin), TRPML (for mucolipin), and the more distant subfamily TRPN (N for “nomp” no mechano-receptor potential). ANKTM1, a new member of the TRP ion channel family recently implicated in the detection of noxious cold, is sufficiently different from members of the other TRP subfamilies to warrant assigning this protein to a new subfamily, referred to as TRPA (201, 368). While there is only one mammalian TRPA member, there are four in flies and two in worms (74, 77). To maintain consistency with the rest of the TRP nomenclature, it is recommended to refer to ANKTM1 as TRPA1 in future studies as suggested and approved by the HUGO nomenclature committee. Polycystic kidney disease is an autosomal-dominantly inherited disease causing progressive development of cysts in the kidney and liver (268, 269). Nearly all cases result from mutations in either PKD1 or PKD2 genes, both of which encode proteins of the TRP family. PKD1 and PKD2 are usually lumped together in the TRPP branch, but the two groups have a markedly different molecular architecture and represent distinct phylogenetic branches of the superfamily. PKD1 is a very large protein of 12 TM segments, whereas PKD2 has the predicted topology of the TRP channel consisting of 6 TM spanning domains and clearly conduct ions (156). Importantly, TRPP channels are mainly expressed in ciliated epithelial cells (270).
Sequence alignments revealed that TRPC, TRPM, and TRPV, but not the TRPP channels, contain a characteristic six-amino acid motif located in the carboxy terminus (41). An intriguing subfamily within the TRP superfamily is the TRPV subfamily consisting of six members. This group of channels includes TRPV1–4, which respond to heat, osmolarity, odorants, and mechanical stimuli, whereas epithelial Ca2+ channels TRPV5–6 have been implicated in maintaining the body Ca2+ balance by facilitating Ca2+ (re)absorption in kidney and small intestine (175, 180, 302).
The epithelial Ca2+ channel family is restricted to two members, and genomic cloning demonstrated that these Ca2+ channels are transcribed from distinct genes rather than being splice variants (181, 301). TRPV5 and TRPV6 are juxtaposed on the human chromosome 7q35 with a distance of only 22 kb, which suggests an evolutionary gene duplication event. The same situation was observed in the mouse genome in which TRPV5 and TRPV6 are localized on chromosome 6 in a region that is syntenic to human chromosome 7q33–35 (180, 301). There is one other example found as a direct repeat of two homologs TRP channels. Comparable to the TRPV5/6 couple, TRPV3 is located juxtaposed the TRPV1 gene at a distance of 7.5 kb (41). To date, TRPV5 and TRPV6 have been cloned from many species including rabbit, rat, mouse, and human (181). The identified sequences exhibit an overall homology of ∼75%. Strikingly, several domains in TRPV5/6 are completely conserved within these species including the core structure of the protein consisting of 6 TM segments and the pore region, ankyrin repeats, PDZ motifs, and putative PKC phosphorylation sites, of which three are conserved among all identified TRPV5/6 channels (Fig. 3A). Detailed sequence analysis of TRPV5 and TRPV6 revealed the identification of several putative phosphorylation sites including PKC, PKA, and cGMP-dependent kinase (89). For instance, in rabbit TRPV5, two combined putative phosphorylation sites for PKA and cGMP-dependent kinase (S669 and T709) were originally identified (178). However, these predicted phosphorylation sites are not conserved in other species or in TRPV6. This is in contrast to the putative PKC phosphorylation sites of which three are conserved within the complete TRPV5/6 subfamily (Fig. 3A) (181). However, the physiological relevance of the conserved PKC sites is not clear, since mutations of these putative regions did not affect channel activity in HEK293 cells (Hoenderop and Bindels, unpublished data). To date, no information is available about the phosphorylation of TRPV5 and TRPV6. In addition, TRPV5 and TRPV6 contain PDZ motifs and ankyrin repeat domains in the amino-terminal region, which are also present in a diverse range of receptors and ion channels including the TRP superfamily. PDZ motifs are recognized by PDZ domains that are modular protein interaction domains playing a role in protein targeting and protein complex assembly (292). There is evidence that they can regulate the functions of their ligands in addition to serving as scaffolds. Although binding to carboxy-terminal motifs appears to be the typical mode of interaction, PDZ domains could also interact with internal motifs that are present in TRPV5/6. In general, ankyrins link transporters and cell adhesion molecules to the spectrin-based cytoskeletal elements in specialized membrane domains (197). Neural-specific isoforms of ankyrin have been demonstrated to participate in the maintenance and targeting of ion channels to subcellular regions in cells (217). So far, no PDZ domain- or ankyrin domain-interacting proteins have been identified.
B. Ion Selectivity/Gating Mechanisms
1. Basic biophysical properties
The conspicuous biophysical hallmarks of TRPV5, which are also representative for TRPV6, are shown in Figure 4 (181, 444). In HEK293 cells heterogeneously expressing TRPV5, currents through TRPV5 can be activated under conditions of high intracellular buffering of Ca2+ by hyperpolarizing voltage steps. In the absence of divalent cations, large inward currents are observed, which show a typical time-dependent increase (“gating”) and are constant over more than 1 min (Fig. 4A). Outward currents are extremely small, indicating that the channel is nearly completely inwardly rectifying. In the presence of Mg2+, but the absence of Ca2+, the initial currents are large and rapidly inactivate due to block by extracellular Mg2+, which is driven into the open pore by the hyperpolarizing voltage. In the presence of Ca2+, but in a nominally Mg2+-free solution, currents through TRPV5 increase again when the extracellular Ca2+ concentration is elevated. If the extracellular Ca2+ concentration is changed from 0 to 30 mM, then 1) increasing extracellular Ca2+ up to 100 μM reduces the current amplitudes through TRPV5/6, and 2) an increase from an extracellular Ca2+ concentration of 100 μM up to 30 mM or higher enhances the current again. This finding is reminiscent of the anomalous mole fraction behavior described previously for L-type voltage-gated Ca2+ channels (4, 85, 163). In analogy to L-type Ca2+ channels, permeation through TRPV5/6 can by described by “repulsion” pore models considering a pore consisting of two high-affinity binding sites, whereby the double occupation of the two binding sites by either Na+ or Ca2+ provides the “drive” for Ca2+ conduction due to mutual repulsion of the two cations (403, 406) or by a three binding-site model in which two low-affinity sites flank a high-affinity binding site for Ca2+ (181, 403, 406). The current through TRPV5/6 is carried exclusively by Ca2+ at extracellular Ca2+ concentrations exceeding 10 μM.
Figure 5 shows single-channel data for Na+ currents through TRPV5 in inside-out patches. Current-voltage relationships show inward rectification also at the level of single channels. Conductance between −200 and −100 mV is ∼75 pS, whereas between −100 and +20 mV values were obtained of∼35 pS (Fig. 5A). So far, no reliable single-channel measurements have been performed in the presence of extracellular Ca2+. Ensembled currents show a similar time course as macroscopic currents including a gating phase at the beginning of the hyperpolarizing voltage step (Fig. 5B). The range of single-channel conductances for monovalent cations are all in a similar range between 40 and 70 pS and are obviously, although not compared in detail under identical condition, similar for TRPV5 and TRPV6 (285, 402, 444). These values are important in regard to the controversy of whether TRPV6 might form the CRAC pore (see below). Another typical feature of currents through TRPV5/6 is the inactivation at negative potentials (Fig. 6). This inactivation is nearly complete if Ca2+ is the charge carrier and is delayed when Ba2+ substitutes Ca2+ (Fig. 6A). Currents of monovalent cations through TRPV5/6 do not inactivate. This feature is typical for a Ca2+-dependent component of inactivation. Repetitive stimulation of TRPV5/6 currents by short hyperpolarizing pulses consequentially result in a decay of the current (Fig. 6B). The respective current-voltage relationships are shown in Figure 6, C and D. The slow recovery from inactivation in divalent cation-free solution after the Ca2+-dependent decay is a further typical hallmark of TRPV5/6 channels. The time to 50% recovery from inactivation, dependent on the size of the Ca2+ inward currents, is in the range of 1–2 min for TRPV5 and between 40 and 60 s for TRPV6.
Although all the above-described structural aspects and basic electrophysiological properties of TRPV5 and TRPV6 are rather similar, some differences remain which concern permeability of divalent cations, kinetics of Ca2+-dependent inactivation, and recovery from inactivation (Fig. 7). Interestingly, these typical differences could be explained by sequence differences between TRPV5 and TRPV6. Ba2+ permeates TRPV5 better then TRPV6, e.g., the current ratio IBa/ICa is ∼0.9 for TRPV5, but only 0.4 for TRPV6. TRPV6 clearly shows a fast component of inactivation, which is less obvious for TRPV5, e.g., time to 10% inactivation at hyperpolarizing steps to −100 mV is ∼125 ms for TRPV5 but only ∼40 ms for TRPV6. More dramatic are the kinetic differences when Ba2+ is the charge carrier. Fast inactivation is about twofold prolonged when Ca2+ is substituted by Ba2+ for TRPV5, but ∼20 times for TRPV6 (Fig. 8) (281). Interestingly, the structural determinants of these differences seem not to be located in either the amino or carboxy terminus, but in the TM2-TM3 linker. Swapping of the TM2-TM3 linker of TRPV6 to TRPV5 confers the kinetic and permeation phenotype of TRPV6 to TRPV5, whereas swapping of the amino or carboxy termini is ineffective (Fig. 7, A and B) (281). Interestingly, this first intracellular loop is entirely encoded by one exon and could, therefore, provide an example of how a single exon may alter function (302). Furthermore, functional differences concern the recovery from Ca2+-dependent inactivation, which is about three times slower for TRPV5 compared with TRPV6 (181). Changes in the inactivation phase of TRPV6 have been reported by mutations of the histidine residue H587 in the carboxy terminus close to TM6 of TRPV6, which is not present in TRPV5 (375). This residue might be involved in the fast inactivation of TRPV6 (302, 375). Also, some striking pharmacological differences are present, e.g., ruthenium red is a 100-fold more potent blocker for TRPV5 than TRPV6 (IC50 ∼9 μM for TRPV6 but ∼100 nM for TRPV5) (Fig. 7C). TRPV5 is about four times more sensitive to block by Cd2+ than TRPV6 (IC50 ∼70 nM for TRPV5 but ∼260 nM for TRPV6) (182).
2. Pore properties
TRPV5 and TRPV6 are so far the only known highly Ca2+-selective channels in the TRP superfamily. This unique permeation property is also not conserved in the TRPV subfamily, which shares the highest homology with TRPV5/6. TRPV1–4 are, however, all Ca2+ and Mg2+ permeable, but discriminate much less between divalent and monovalent cations, e.g., the relative selectivity for Ca2+ and Mg2+ over Na+ is between PCa/PNa = 1–10 (TRPV1-TRPV4) and PMg/PNa = 2–3 (only measured for TRPV4) (32, 155, 287, 411). TRPV5/6 display PCa/PNa values of >100. Permeation of monovalent cations is Na+ > Li+ > K+ > Cs+ (181, 285, 286, 404, 406). This refers to an Eisenman sequence X for a strong field-strength binding site (181, 285, 286, 406). For TRPV6, a sequence of K+ > Na+ > Li+, e.g., Eisenman V or VI (304), has been reported which, however, seems to be unlikely given the identical pores of TRPV5/6. For divalent cations, a permeation sequence of Ca2+ > Ba2+ > Sr2+ > Mn2+ has been reported (181, 305, 404, 406).
A striking and important feature of TRPV5/6 channels is the open pore blockage by intracellular Mg2+ (Fig. 9) (409, 410). In the absence of extracellular Ca2+, hyperpolarizing voltage steps (prestep, Fig. 9A) activate inward currents with a slowly rising phase (“gating”) (Fig. 9B). If after complete opening of the channels at negative potentials steps are applied to different test potentials, inward, but not outward, currents can be observed, indicating that TRPV5/6 channels nearly completely rectify. The typical current-voltage relationships, shown in Figure 9D, indicate that at negative potentials the intracellular Mg2+ concentration is ineffective and also an intrinsic rectification remains in the absence of Mg2+. The plot of the instantaneous initial current after the negative test potential against the preceding variable prepotentials defines the fraction of open channels (or the apparent open probability, Po) at the end of the prepotentials. Stepping back from the test potential to the prestep potential unmasks clear voltage dependence: at less negative potentials, partially blocked channels open time dependently due to unblock (voltage dependence). The number of available channels is decreased in a Boltzmann-type voltage dependence, which completely disappears in Mg2+-free intracellular solutions (Fig. 9E). These three features (i.e., gating, rectification, and voltage dependence) only appear in the presence of intracellular Mg2+. In the absence of intracellular Mg2+, gating and voltage dependence disappear, whereas rectification is still present, but diminished (Fig. 9C). An open pore blockage by intracellular Mg2+ explains the following findings: at depolarizing potentials, Mg2+ moves towards the pore, thereby plugging the permeation pathway for monovalent ions. Unblock occurs at hyperpolarizing voltages. At very large depolarization, Mg2+ is pushed through the pore, which results in a partial unblock of the channels (increased apparent Po, Fig. 9E). These results are crucial to understand the pore properties of TRPV5/6.
Significant progress in the identification of the molecular determinants of TRP channel pores and the understanding of the high selectivity for Ca2+ has been particularly achieved for TRPV5 and TRPV6 channels (286, 404, 406, 409, 410). Structural differences in the channel pore explain the striking permeation differences in the TRPV subfamily. Figure 10A shows an amino acid sequence alignment of the putative pore regions of the three mammalian TRPV channels. Based on structural similarity with the selectivity filter of the potassium channel KcsA (“signature sequence” TXXTXGYGD) (96), the structural determinant of the low Ca2+-selective TRPV1–4 channels is the GM(L/M)GD motif in TRPV1–4 (287, 411, 446). The sequence similarities may indicate conserved pore structures for these cation channels. Importantly, this motif is missing in TRPV5/6. Instead, we have demonstrated that the molecular determinants of the Ca2+ selectivity and permeation of TRPV5/6 reside at a single aspartate residue (TRPV5-D542 and TRPV6-D541) present in the pore-forming region (Fig. 10B) (286, 409). Neutralization of these residues not only affects the high Ca2+ selectivity of TRPV5/6, but also drastically reduces their sensitivity to extracellular Cd2+ and abolishes Mg2+-dependent voltage-dependent gating of these channels (409, 410, 412). It thus appears that high Ca2+ selectivity in TRPV5 and TRPV6 depends on a ring of four aspartate residues in the channel pore, similar to the ring of four negative residues (aspartates and/or glutamates) in the pore of voltage-gated Ca2+ channels (Fig. 10B). Likely, D542/D541 is the narrowest part of the selectivity filter with a diameter of ∼5.2 Å and forms part of the high-affinity binding site for Ca2+ and Mg2+ (408). A detailed analysis of the structure of the TRPV5 and TRPV6 pores has now been published (94, 407). To obtain insight in the pore architecture of TRPV6, a pore diameter of 5.4 Å was estimated from permeation studies. Mutating D541, a residue involved in high-affinity Ca2+ binding, altered the apparent pore diameter, indicating that this residue lines the narrowest part of the pore and is part of the selectivity filter (407). Pore lining amino acids were determined by cysteine scanning mutagenesis (SCAM). Cysteines introduced in a region preceding D542 for TRPV5 and D541 for TRPV6 displayed a cyclic pattern of reactivity to cysteine reacting agents indicative of a pore helix. The location of the cation-selective filter was identified at the outer part of the pore helix. The pattern of covalent modification of cysteines supports a KcsA homology-based three-dimensional model (96). The external vestibule in TRPV5 and TRPV6 may build up the three structural domains consisting of a coiled structure that is connected to a 15-amino acid pore helix followed by the selectivity filter (probably a coiled structure with D542 and D541 as the narrowest part) and another coiled structure before the beginning of TM6. This is the first structural model of a TRP channel pore.
The pore region of TRPV5/6 contains additional negatively charged amino acids (Fig. 10A) that only have minor effects on the Ca2+ permeation properties. In another study, the role of D542 for blockage of monovalent currents through TRPV5 by Ca2+ and Mg2+ was supported, however, not for determining Ca2+ permeability (196). However, all of our data including recent findings from concatemers and a detailed study of TRPV6 pore properties clearly demonstrated the role of D542/D541 for Ca2+ selectivity (182, 286, 408, 409, 412). Another surprising result was the triple pore mutant F534/E535/L536 (84). This mutant has been reported to induce nonfunctional channels. However, any mutation of E535 was without effect for TRPV5 or TRPV6 permeation (196, 286, 408).
3. Mechanism of high Ca2+ selectivity
Our understanding of the pore properties of TRPV5/6 refers to the following mechanism of high Ca2+ selectivity, which is similar to the mechanisms of permeation by binding as proposed earlier for L-type Ca2+ channels (165, 381). Under Ca2+-free conditions, Mg2+ will bind at a site within the channels, which is mainly determined by D542/D541. At low driving forces, e.g., less negative membrane potentials, the driving force is not sufficient to move Mg2+ out of this binding well and the channel is blocked. At hyperpolarization, Mg2+ will be moved from this high-affinity site, thereby allowing monovalent cations to permeate the channels. Vice versa, at very positive membrane potentials, Mg2+ will move through the pore towards the extracellular space. Because of the higher affinity of the binding site for Ca2+ versus Mg2+, Ca2+ will outnumber Mg2+ at this site and is now blocking the pore for a movement of monovalent cations. If the inward driving force is high enough to remove Ca2+ from the high-affinity site, a fast inward permeation occurs supported by the low-affinity sites, which flank the high-affinity Ca2+ binding site. All our data so far support a crucial role of D542/D541 for establishing this functionally important site in the channel pore. The same mechanism has been reevaluated for tetrameric channels (182). Importantly, mutation of only one aspartate residue in the tetrameric channels already results in loss of Ca2+ selectivity and voltage dependence. Obviously, D542/D541 reflects a dominant effect on the unique pore properties of these Ca2+ channels.
A nearly complete inward rectification is a further hallmark of the TRPV5/6 channels in addition to the high Ca2+ selectivity. This inward rectification is only partially due to blockage by Mg2+, which is removed by hyperpolarization. Thus the most prominent part of rectification remains in the absence of intracellular Mg2+ and is not due to blockage by endogenous polyamines. This intrinsic rectification of the channels is, however, dramatically reduced by neutralization of D542/D541, indicating that this site is also involved in rectification (409).
In the generally used overexpression systems, TRPV5/6 are constitutively open at a low intracellular Ca2+ concentration and negative voltage. The above-described mechanism of removing channel blockage by Ca2+ or Mg2+ is necessarily part of the gating mechanism. We cannot exclude other mechanisms that might influence gating of TRPV5/6. However, essential prerequisites of channel opening are 1) low intracellular Ca2+ to remove Ca2+-dependent inactivation; 2) hyperpolarization, e.g., an increased driving force for the permeating cation which must be sufficient to move Ca2+ from the high affinity site; and 3) removal of the open pore blockage by Ca2+ and Mg2+.
C. Modulation of Channel Activity
1. Intracellular Ca2+
TRPV5 and TRPV6 are subject to Ca2+-dependent feedback inhibition (282, 404). Both channels rapidly inactivate during hyperpolarizing voltage steps, and this inactivation is reduced when Ba2+ or Sr2+ was used as charge carriers. This inhibition was dependent on the extracellular Ca2+ concentration and occurred also in cells buffered intracellularly with 10 mM BAPTA. Currents also disappeared during repetitive activation by short hyperpolarizing pulses. This decay of the current was significantly diminished when Ca2+ was replaced by Ba2+ as charge carrier and abolished when extracellular Ca2+ was lowered to 1 nM, again indicating that a Ca2+-operated process inhibits TRPV5/6 activity. These regulatory processes were strongly influenced by the surrounding Ca2+ concentrations. Elevation of the extracellular Ca2+ concentration significantly increased the rate of current decay. Ca2+ influx is a prerequisite for this phenomenon because the Ca2+-impermeable D542A mutant lacks a monovalent current decay in response to repetitive stimulation (282). These data suggest that the TRPV5/6 channels are downregulated by Ca2+ influx through the channel and thus likely by increasing the Ca2+ concentration in a microdomain near the pore region, thereby inducing feedback inhibition of the channel. This could be a crucial mechanism for regulation of TRPV5/6 under physiological conditions. It was shown that this inhibition is highly Ca2+ sensitive with calculated affinity values down to 100 nM. When measured directly in inside-out patches, half-maximal inhibition of TRPV5 currents occurred at ∼200 nM (282). With the consideration of the high-affinity mechanism of Ca2+-dependent TRPV5/6 inhibition, the presence of intracellular Ca2+ buffer proteins such as calbindins will play an important role in the regulation of channel activity (173, 175, 181).
Recovery from inhibition upon washout of extracellular Ca2+ (whole cell configuration) or removal of Ca2+ from the inner side of the channel (inside-out patches) is slow in both conditions. Half-maximal recovery was reached after 100–135 s (282). The slow recovery of TRPV5 (and to a lesser extent TRPV6) from their Ca2+-induced inhibited state is an intriguing feature of both TRPV5/6. Surprisingly, recovery is much slower than the inhibition, although this putative microdomain is accessible for intracellular Ca2+, and it does not correlate with the removal of intracellular Ca2+ either, since full recovery occurs much later than restoration of the basal Ca2+ level in non-Ca2+-buffered cells, or after removing Ca2+ from the inner side of excised membrane patches (282). The similar rates of recovery in whole cell and excised patch experiments might indicate that it is controlled by a complex consisting of pore and microdomain. Furthermore, these data suggest that other processes than rapid binding and slow dissociation of Ca2+ could be involved and might hint to interaction with other regulatory proteins.
The molecular mechanism of this feedback inhibition remains, however, unclear at the moment. The first mechanistic insight into Ca2+-dependent inactivation of TRPV6 included a Ca2+-dependent binding of CaM to the carboxy terminus (Fig. 11, A and B). (276). A CaM binding site was identified which seemed to be responsible for Ca2+-dependent inactivation in TRPV6. The minimal structure of this site comprises an arginine-rich motif between the positions 691 and 711, NWERLRQGTLRRDLRGIINR, which includes a conserved PKC phosphorylation site RQGTLRR. Phosphorylation of this site inhibited Ca2+-dependent inactivation of TRPV6. This sequence, however, is only partially conserved in TRPV6. Truncation of TRPV6, N696X, in which most of the CaM binding motif is removed, also exhibited a decreased inhibition of TRPV6 by the intracellular Ca2+ concentration (229, 281).
Other carboxy-terminal truncations and mutants modulated Ca2+-dependent inactivation of TRPV5 (Fig. 11C). Deletion of the last 30 amino acids of the carboxy terminus of TRPV5, G701X, decreased significantly the Ca2+ sensitivity. This carboxy-terminal part of TRPV5 does not comprise the Ca2+-CaM binding site of TRPV6. Another critical stretch for Ca2+-dependent inactivation of TRPV5 was found upstream in the carboxy terminus. Analysis of truncations at amino acid 635, 639, 646, 649, and 653 disclosed a critical stretch involved in Ca2+-dependent inactivation between position 649 and 653 (Fig. 11D). Detailed mutation analysis revealed that mutation of A650 and F651 decreased already the Ca2+ sensitivity of TRPV5. C653X showed a decreased Ca2+ sensitivity, comparable to G701X, while E649X lacked Ca2+-dependent inactivation (Fig. 11C). Likely, cells expressing truncations shorter than E649 do not survive and could only be restored in the presence of the high-affinity blocker ruthenium red, suggesting that these truncations exhibited a deleterious Ca2+ influx (288).
It is well known that acidification of the apical medium inhibits transcellular Ca2+ absorption across primary cultures of rabbit CNT and CCD cells (37). It was, therefore, interesting to evaluate modulation of TRPV5 by pH. 45Ca2+ uptake in TRPV5 expressing Xenopus laevis oocytes is inhibited by acidification of the incubation medium (179). Indeed, extracellular acidification reduced currents through TRPV5 carried by either monovalent or divalent cations, which was confirmed by Peng et al. (303). Additionally, extracellular pH also affected current kinetics, extracellular Mg2+ blockage, and Ca2+ affinity. The gating component of monovalent cation currents through TRPV5 was delayed at alkaline pH and as well as blockage by extracellular Mg2+. Mg2+ blockage of monovalent currents was shifted from an IC50 of ∼62 at pH 7.4 to ∼300 μM at pH 6.0 and ∼40 μM at pH 8.5, indicating that blockage of TRPV5 by extracellular Mg2+ is reduced under more alkaline conditions. Blockage of monovalent cation currents through TRPV5 by Ca2+ is less efficient at low pH (in the presence of 100 μM extracellular Ca2+ ∼160 nM at pH 7.4 and ∼5 μM at pH 6.0) (405). Both results indicate that Ca2+ and Mg2+ binding in the channel are weakened at higher proton concentrations. For L-type Ca2+ channels and cyclic nucleotide-gated channels, it was suggested that protonation of a single glutamate residue in the pore region is responsible for the dramatic changes in divalent cation affinity for the channel with respect to variations in the extracellular pH (316, 328, 376). This could be analogous for TRPV5 in which a single aspartate residue determines the Ca2+ permeation and Mg2+ blockage of the channel. Extrapolating the pH influence to the in vivo situation, this effect could at least in part provide the molecular basis of acidosis-induced calciuresis. Interestingly, the molecular mechanism of TRPV5 blockage by protons is mechanistically understood taking into account the pore structure described above. Mutation of the glutamate at position 522 to glutamine (E522Q) preceding the pore helix decreases inhibition of the channel by extracellular protons. Thus pH sensitivity is mainly mediated by glutamate at position 522 and may act as the “pH sensor” of TRPV5 (438).
Little is known about effective pharmacological tools to modulate TRPV5/6. Ruthenium red and econazole appeared to be the most effective inhibitors. TRPV5/6 are efficiently blocked by inorganic cations. The profile of TRPV5 blockage is Pb2+ = Cu2+ = Gd3+ > Cd2+ > Zn2+ > La3+ > Co2+ > Fe2+ >> Fe3+, with IC50 values between 1 and ∼10 μM (283, 303–305, 406). The inorganic polycationic dye ruthenium red, which binds to phospholipids, inhibits TRPV5 and TRPV6 in a voltage-dependent manner. Blockage is attenuated at depolarizing potentials. Currents of monovalent cations through TRPV5 are blocked with an IC50 of ∼100 nM at −100 mV. Transcellular Ca2+ transport in primary cultures of immunodissected rabbit CNT and CCD is inhibited in the same concentration range (283). Antimycotic imidazoles such as econazole and miconazole are highly effective inhibitors with IC50 values of ∼1–2 μM. These compounds block, however, voltage independent (283). Blockers of store-operated Ca2+ entry channels (SOC), e.g., SKF96365 and 2-aminoethoxydiphenylborate (2-APB), are nearly ineffective for TRPV5 and TRPV6 (283, 336, 410). Another SOC blocker, the adenyl cyclase inhibitor MDL 12330A (399), exerts a half-maximal inhibition of TRPV5 at ∼20 μM (283). Xestospongin, a noncompetitive inositol 1,4,5-trisphosphate receptor antagonist, seems, however, to block TRPV6 (402). TRPV5 is not or very weakly affected by capsazepine, a selective blocker of capsaicin receptor TRPV1. Arginine-rich peptides, such as dynorphins, inhibit currents through TRPV1 (311). TRPV5 is, however, resistant to both dynorphin A and the fragment 10 consisting of the first 10 amino acid residues of dynorphin A. Also, blockers of protein tyrosine kinases such as tyrphostin B46 and genistein, CaM antagonists calmidazolium R24571 and trifluoperazine, and channel blockers such as mibefradil, quinidine, nifedipine, niflumic acid, and verapamil, have only small or no effects on TRPV5 activity (283). Interestingly, some pharmacological tools differentiate between TRPV5 and TRPV6. Blockage by ruthenium red and Cd2+ is much more sensitive for TRPV5 than TRPV6 (181, 182). Blockage of TRPV6 by capsaicin has been described (84) that is ineffective for TRPV5 (Nilius, unpublished data).
D. CRAC and TRPV6
One of the still unsolved problems in channel physiology is the nature and the mechanism of activation of SOC. Twenty-four years ago Casteels and Droogmans (68) showed that depletion of agonist-sensitive intracellular stores stimulates the rate of Ca2+ uptake from the extracellular solution in vascular smooth muscle cells. The first electrical measurement of SOC was achieved in mast cells, and this current was referred to as “calcium release-activated calcium current” (CRAC) (189). CRAC is still the best-characterized SOC. It is a highly Ca2+-selective, inwardly rectifying channel; permeable to Ca2+, Sr2+, Ba2+, but not Mg2+; exhibits anomalous mole fraction behavior; becomes permeable for monovalent cations in the absence of extracellular divalent cations; and is inactivated by an increase in the intracellular Ca2+ concentration (300). These properties are reminiscent to the above-described biophysical properties of TRPV5/6 except the activation by depletion of intracellular Ca2+ stores. Recently, Yue et al. (444) reported that TRPV6 was activated by store depletion and manifests the pore properties of CRAC in RBL cells and might, therefore, be CRAC or at least part of the CRAC pore (444). This store-operated mode could, however, only be observed in COS cells in a restricted posttransfection time window when the current density was lower than in other studies. A strong argument for TRPV6 being CRAC was at that time apparent identical single-channel conductance for monovalent cations (208). It has also been reported in several forthcoming studies that a correlation between CRAC and TRPV6 exists: 1) an upregulation of CRAC channel was observed in RBL cells by transfection with TRPV6 (336); 2) in prostate cancer cells, which express TRPV6, antisense oligonucleotides downregulated TRPV6 and CRAC and antiandrogens unregulated both TRPV6 and CRAC (395); and 3) expression of a “dominant negative” pore mutant of TRPV6 in Jurkat cells, the already mentioned triple mutant F534A/E535A/L536A, attenuated endogenous CRAC (84). However, several features are incompatible with the proposed equality of the TRPV6 and CRAC or CRAC pores. Open pore blockage by intracellular Mg2+, which is a hallmark of TRPV5/6 gating (409, 410, 412), is completely absent in CRAC. In fact, CRAC behaves like TRPV5/6 in the absence of intracellular Mg2+ even when a high free intracellular Mg2+ concentration is present (also Fig. 4). As explained, this Mg2+ blockage is caused by binding of Mg2+ at the aspartate residue D541/542. Mutation of this residue attenuated intracellular Mg2+ blockage and Ca2+ selectivity of TRPV5/6 and decreased the inward rectification. Interestingly, Mg2+ blockage and high Ca2+ selectivity was also abolished if a single aspartate is mutated in a concatemeric homotetramer of TRPV6, indicating that this site is dominant for regulation of these important pore properties (182, 408–410). Recently, TRPV6 antisense and siRNA knockdown approaches inhibited TRPV6-derived currents in mast cells, but failed to inhibit CRAC currents. These results render it improbable that TRPV6 is a component of native CRAC channels in mast cells (203). It seems, therefore, unlikely that even heteromultimers containing TRPV6 form the highly Ca2+-selective CRAC pore.
More striking evidence against TRPV5/6 being CRAC comes now surprisingly from single-channel measurements. It has been recently shown that in extracellular solutions that are free of divalent cations and under conditions of reduced intracellular Mg2+, the dominant cation current in Jurkat and RBL cells is through a channel first termed MagNuM (magnesium-nucleotide-regulated metal cation current), which is conducted by the LTRPC7 channel (162) or also called MIC (Mg2+ inhibited cation channels) (314). These channels have a single-channel conductance for monovalent cations of ∼45 pS as reported for TRPV6 by Yue et al. (444). Because most of the studies on CRAC have used divalent-free solutions on either side of the membrane to study selectivity, the single-channel conductance of CRAC was misinterpreted by the one of MagNuM/MIC/TRPM7 which are activated by a decrease in intracellular Mg2+/Mg-ATP (162). CRAC has indeed a single-channel conductance of ∼0.2 pS (209, 220) rather than ∼45 pS as TRPV6 (see also Refs. 16, 73, 313). Interestingly, estimation of the pore diameter also resulted in rather different values. TRPV6 has a diameter of ∼5.2 Å (408), whereas for CRAC a diameter of ∼3.8 Å was estimated (313).
However, other differences also exist including the striking permeation difference of CRAC and TRPV5/6 for Cs+ (Cs+ is much less permeable through CRAC than through TRPV5/6), the different time course of the monovalent currents through CRAC (inactivating) and TRPV5/6 (sustained), the striking differences in rectification (much less rectification for monovalent currents in CRAC than TRPV5/6) and effects of extracellular 2-APB (blockage of CRAC but even light potentiation of TRPV6) (for a review see Refs. 280, 410). In addition, we have never observed any store-operated activation for TRPV5 or TRPV6. However, TRPV6 expression in HEK293 cells does result in a constitutive open channel, is not activated by ionomycin-induced store depletion, and is inhibited by a rise in cytosolic Ca2+ concentration independently of the posttransfection time (410). Furthermore, TRPV6 seems not to be expressed in Jurkat cells, one of the preferred cells to study CRAC, and expression of TRPV6 in these cells does not reveal store-dependent properties (47). CRAC is obviously not TRPV5/6. A modulator role for store-operated Ca2+ influx cannot be excluded (75). As a bottom line, all Ca2+-permeable TRP channels contribute to changes in the intracellular Ca2+ concentration and may, therefore, also affect store-operated processes. There is so far no TRP channels for which a consensus for being CRAC has been settled and at least TRPV5 or TRPV6 are very unlikely CRAC/SOC (74, 75).
E. Molecular Structure of TRPV5 and TRPV6
The oligomerization of TRPV5 and TRPV6 channels has recently been unraveled (182). Cross-linking studies, coimmunoprecipitations, and molecular mass determination of TRPV5/6 complexes using sucrose gradient sedimentation showed that TRPV5 and TRPV6 form homo- and heterotetrameric channel complexes. As described in section v, TRPV5 and TRPV6 are coexpressed in some tissues, which allows oligomerization of these channels in vivo. For instance, immunohistochemical data in kidney clearly demonstrated coexpression of TRPV5 and TRPV6 in the DCT (182). However, coexpression of TRPV5 and TRPV6 in these tissues is until now only quantified at the mRNA level, indicating that TRPV6 is 100–10,000 more expressed than TRPV5. Quantification at the protein level of both channels is certainly important to address the stoichiometry in vivo. Heteromeric complex formation can modify the activity of members of the TRP family. The Drosophila TRP and TRPL members were identified first, and it has been shown that these proteins form heteromultimeric channels associated in a supramolecular signaling complex with receptors and regulators including PKC, CaM, and the scaffolding PDZ domain containing protein InaD (13, 233). Moreover, it has been shown for TRPC1 and TRPC3 that hetero-oligomers of these channels possess more distinctive properties than that of either channel alone (237). In addition, Strubing et al. (370) demonstrated that TRPC1 and TRPC5 are subunits of a heteromeric neuronal channel. Both TRPC proteins have overlapping distributions in the hippocampus. Coexpression of TRPC1 and TRPC5 in HEK293 cells resulted in a novel nonselective cation channel with a voltage dependence similar to N-methyl-d-aspartate (NMDA) receptor channels, but unlike that of any reported TRPC channel. Other TRPCs exclusively assemble into homo- or heterotetramers within the confines of TRPC subfamilies, e.g., TRPC4/5 and TRPC3/6/7 (187). Based on the ability of the TRPC channels to form functional homo- and heteromultimeric complexes, Tsiokas et al. (382) provided evidence that PKD2, which is functionally related to the TRPC proteins, interacts with TRPC1, suggesting a possible role of this protein in modulating Ca2+ entry in response to G protein-coupled receptor activation and/or store depletion (382). Within the TRPV family, the oligomeric structure of TRPV1 was studied by biochemical cross-linking, and the predominant existence of tetramers was suggested (207). The principles of TRP channel formation provide the conceptual framework to address the physiological role of distinct TRP members. Likewise, hetero-oligomerization of TRPV5 and TRPV6 might influence the functional properties of the formed Ca2+ channel. As TRPV5 and TRPV6 exhibit different channel kinetics with respect to Ca2+-dependent inactivation, Ba2+ selectivity and sensitivity for inhibition by ruthenium red, the influence of the heterotetramer composition on channel properties was investigated (Fig. 7). Concatemers were constructed consisting of four TRPV5 and/or TRPV6 subunits configured in a head-to-tail fashion (182). A different ratio of TRPV5 and TRPV6 subunits in these concatemers showed that the phenotype resembles the mixed properties of TRPV5 and TRPV6. An increased number of TRPV5 subunits in such a concatemer displayed more TRPV5-like properties, indicating that the stoichiometry of TRPV5/6 heterotetramers influences the channel properties. Consequently, regulation of the relative expression levels of TRPV5 and TRPV6 may be a mechanism to fine-tune the Ca2+ transport kinetics in kidney or other TRPV5/6-coexpressing tissues. The tetrameric organization of TRPV5/6 resembles that of the Shaker potassium channel, which is composed of four tandemly associated homologous domains (221, 234). The clustering of four subunits is assumed to create an aqueous pore centered at the fourfold symmetry axis (Fig. 10B) (234). This tetrameric architecture of TRPV5/6 implies that four of the aspartic residues [D542 (286), D541 (408)] form a negatively charged ring that functions as a selectivity filter for Ca2+ in analogy with voltage-gated Ca2+ channels (182). Interestingly, Niemeyer and co-workers (102) identified the third ankyrin repeat being a stringent requirement for physical assembly of TRPV6 subunits. Deletion of this repeat or mutation of critical residues within this repeat renders nonfunctional channels that do not coimmunoprecipitate or form tetramers. It was proposed that the third ankyrin repeat initiates a molecular zippering process that proceeds past the fifth ankyrin repeat and creates an intracellular anchor that is necessary for functional subunit assembly (102). Protein crystallography of these channels will be the most challenging approach to determine the molecular structure in the near future.
V. SITES OF EPITHELIAL CALCIUM TRANSPORT
The exchange of Ca2+ between higher organisms and the environment takes place across epithelia, including the gastrointestinal tract, bone, kidney, and gills. Ca2+ transport occurring in these Ca2+-absorbing epithelial tissues is realized by paracellular and transcellular Ca2+ transport as outlined in sections ii and iii.
The renal handling of Ca2+ has been the subject of intensive investigation over the last years. The kidney plays an essential role in the maintenance of the Ca2+ balance by regulating the Ca2+ excretion of the body. On a daily basis, ∼8 g of Ca2+ is filtered at the glomeruli of which <2% is excreted into the urine. As a consequence, filtered Ca2+ is extensively absorbed as it passes through the individual nephron segments.
1. Proximal tubule
The proximal tubules, including proximal convoluted (PCT) and proximal straight (PST) tubules, are responsible for absorbing the majority of Ca2+. Micropuncture studies have demonstrated that ∼70% of the Ca2+ is absorbed in these segments (99, 122, 374, 385). Ca2+ transport along the proximal tubule proceeds essentially as an isosmotic process that is based on early micropuncture studies demonstrating that Ca2+, Na+, and water are absorbed in parallel (reviewed in Ref. 371). This means that the majority of Ca2+ reabsorption in these segments is energetically passive and follows the local Na+ reabsorption.
2. Limb of Henle
In the thin descending and ascending limbs of Henle, the permeability for Ca2+ is very low, and basically we can conclude that significant net Ca2+ transport does not occur in this segment (318, 331). This low Ca2+ permeability is particularly striking in view of the comparatively high permeability to Na+ and Cl− (331). Because these thin limbs of Henle do not transport Ca2+, the thick ascending limb of Henle (TALH) is responsible for the Ca2+ reabsorption between the bend of the loop and the start of the DCT (14, 53, 56, 92, 93, 120, 129, 192, 274, 331, 372, 373, 432). Approximately 20% of the Ca2+ filtered at the glomeruli is absorbed in Henle’s loop (371). In several studies Ca2+ reabsorption in cortical TALH was examined. Bourdeau and Burg (54) provided evidence that Ca2+ transport was driven by the electrochemical gradient, indicative of a passive absorption process. Similar findings were described by Shareghi and Agus (343), who concluded that Ca2+ reabsorption in the TALH is passive and driven by the large lumen-positive membrane potential. In contrast, Imai (192) postulated an active Ca2+ transport component in cortical TALH segments, whereas his studies in medullary TALH segments are more consistent with a passive mechanism. These findings were confirmed in mouse kidney by Friedman (120), whereas Wittner et al. (433) demonstrated that Ca2+ transport in the mouse cortical TALH is entirely passive. Interestingly, Ca2+ transport in cortical segments was stimulated by PTH without an increase in the transepithelial potential difference (120). At variance, Wittner et al. (433) showed that PTH-stimulated passive Ca2+ transport by increasing the electrical driving force and, therefore, the permeability for the paracellular pathway. Immunohistochemical studies on mouse and rat kidney sections did not provide evidence for the presence of the identified Ca2+ transport proteins including TRPV5, TRPV6, calbindins, NCX1, and PMCA1b in TALH segments (171, 240). Interestingly, a new protein, named paracellin 1 (PCLN-1), expressed in human TALH tight junctions, possibly plays a critical role in the control of passive Ca2+, and also Mg2+, reabsorption, since mutations of PCLN-1 are present in patients with the hypomagnesemia hypercalciuria syndrome (HHS) (43). In these patients, renal Ca2+ reabsorption is impaired as expected. This study was the first to demonstrate that homozygous mutations of PCLN-1 result in a selective defect in paracellular divalent cation reabsorption in the TALH, with intact sodium chloride reabsorption ability in this segment. Altogether, this segment definitely plays a significant role in the process of Ca2+ reabsorption, mainly due to paracellular Ca2+ transport. The contribution of active Ca2+ transport is, however, questionable (264).
3. DCT and CNT
The fine-tuning of Ca2+ excretion in the kidney occurs in the distal part of the nephron and amounts to 15% of the filtered load of Ca2+ (79). This section consists of DCT and CNT. The CNT lies just distal to the DCT, arising abruptly in rabbits and gradually in most other species (63, 80, 204, 240). The CNT contains, in contrast to DCT, in addition to principal cells also intercalated cells. The relative contribution of these two segments to active Ca2+ reabsorption appears to differ between the various species (38, 346–349, 371). In these nephron segments, Ca2+ reabsorption occurs against the existing electrochemical gradient. Together with the fact that the tight junctions are relatively impermeable for Ca2+, this substantiates that Ca2+ is reabsorbed in these segments through an active transcellular pathway. By comparing the fractional Ca2+ delivery at the beginning of the DCT and the final urine, early studies indicated Ca2+ reabsorption in these segments (385). Over the last years, several studies indicated that in most mammalian species the DCT is the primary nephron segment of active Ca2+ transport. However, the exact sites for transcellular Ca2+ reabsorption along the distal part of the nephron are still being questioned. Extensive studies revealed that the localization of the Ca2+ transport proteins, including TRPV5, TRPV6, calbindins, NCX, and PMCA, is restricted to the late distal part of the DCT (DCT2) and the CNT (171, 240). Micropuncture investigations of Ca2+ transport in the rat distal convolution (comprising DCT1, DCT2, CNT, and the initial CCD) aimed to distinguish between the first two (including DCT1, -2) and last two (including CNT and initial CCD) segments. Microperfusion studies by Costanzo and Windhager (78) showed similar Ca2+ transport rates in these segments, whereas experiments by Greger et al. (154) suggested that transcellular Ca2+ transport occurs predominately in CNT.
Transepithelial Ca2+ transport depends on the activity of Ca2+-transporting proteins in the apical and basolateral plasma membranes and the cytosol of the epithelial cells. Loffing et al. (240) and several studies by Hoenderop and Bindels and co-workers (171, 279) demonstrated in the mouse coexpression of the Ca2+ transport proteins including TRPV5, TRPV6, PMCA1b, NCX1, and calbindin-D28K in DCT2 and CNT, with the highest immunochemical abundance in DCT2, and a gradual decrease along CNT (Fig. 12A). The parallel axial reduction of apical TRPV5 and of basolateral Ca2+ extrusion machinery, i.e., PMCA1b and NCX1, indicates a progressive decrease of transcellular Ca2+ transport rates along CNT. TRPV5, PMCA1b, and NCX1 abruptly disappeared at the transition to the CCD, consistent with the notion that transcellular Ca2+ transport is negligible in CCD (240). A minority of cells along CNT lacked immunopositive staining for TRPV5 and the other Ca2+-transporting proteins. These negative cells were identified as intercalated cells (171, 240). Taken together, these findings strongly suggest that, in the mouse, the major sites of transcellular Ca2+ transport are DCT2 and, probably to a lesser extent, CNT. Whether Ca2+ transport occurs in DCT1 of mice, and if so at which rate is unknown. The occurrence of weak immunostaining for NCX1 and PMCA1b would be in line with active Ca2+ transport in this segment. At variance with the presence of these basolateral Ca2+-transporting proteins, TRPV5 and TRPV6 were not detectable in DCT1 (171, 240, 279). This raises the question whether other apical Ca2+ entry pathways might play a role in transcellular Ca2+ transport in DCT1 or confirms the absence of an active transport process. The existence of such pathways in DCT has been suggested by RT-PCR data obtained from isolated rat tubules (443) and an immortalized mouse DCT cell line (12). However, these cell preparations do not represent solely DCT1, and the corresponding findings should, therefore, be interpretated with great care. Combined micropuncture experiments, immunohistochemistry of the Ca2+ transport proteins, and corresponding nephron-specific mouse knockout models are needed to investigate whether active Ca2+ transport occurs in DCT1.
4. Collecting duct
The collecting duct is the last segment of the nephron and extends from the connecting tubule in the cortex through the outer and inner medulla to the tip of the papilla. It can be divided into at least three regions, based primarily on their localization in the kidney. These include CCD, the outer medullary collecting duct (OMCD), and the inner medullary collecting duct (IMCD). The collecting duct consists of principal cells that are responsible for salt and water reabsorption and intercalated cells that play a major role in H+ and HCO3− secretion. These latter cells constitute ∼30% of the cells in the CCD, OMCD, and IMCD and lack the typical Ca2+ transport proteins (Fig. 12A) (171). Based on differences between the fractional Ca2+ delivery at the last accessible segment in the distal part of the nephron on the surface of the cortex and in the final urine, it was suggested that ∼3% of the filtered Ca2+ is absorbed by the CCD (385). Because net transport occurs against the electrochemical gradient in the absence of water transport, the mechanism is thought to be active. In isolated perfused rabbit CCD, Shareghi et al. (344) and Bourdeau and Hellstrom-Stein (55) reported a low Ca2+ permeability and an insignificant net flux that is driven by the membrane potential and not affected by PTH and 1,25-(OH)2D3. However, other groups demonstrated a significant Ca2+ transport in this segment. Limited Ca2+ transport studies are performed in IMCD tubules. Lechene and co-workers (31) measured Ca2+ transport along the IMCD in rats by the microcatheterization technique. In this study, net Ca2+ reabsorption occurs along the IMCD and the fractional Ca2+ reabsorption is not altered by thyroparathyroidectomy (31). Ca2+ transport by the IMCD of normal rats was also studied using the in vitro microperfusion technique. Net fluxes of Ca2+ were measured using 45Ca as a tracer, and Ca2+ influx was independent of Na+ transport, was not blocked by verapamil, but was increased by Ca2+ transtubular gradient (244). In this respect, it is interesting to discuss a recent study that addresses the localization of TRPV6 in the mouse nephron. TRPV6 expression was also detected in OMCD and IMCD and displayed a distinct apical localization (279). However, the absence of supportive Ca2+ transport proteins, as confirmed by the lack of immunopositive staining for calbindins, questioned the involvement of these cells in Ca2+ reabsorption. Second, intercalated cells also expressed TRPV6. The consistent apical localization would imply that TRPV6 has a functional role as apical Ca2+ entry channel. Hypothetically, TRPV6-mediated Ca2+ influx could affect transport processes in these segments, for instance, as part of a hormonal signaling cascade. In particular, vasopressin was shown to induce Ca2+ influx across the apical membrane of collecting duct cells, and alterations in intracellular Ca2+ levels are known to affect Na+ reabsorption in these tubules (130, 364, 390). Taken together, the present data suggest that Ca2+ reabsorption in the collecting duct is primarily located in the CCD that accounts for maximally 3% of the total amount of Ca2+ filtered at the glomerulus.
B. Gastrointestinal Tract
The small intestine is the largest part of the gastrointestinal tract and is composed of the duodenum, jejunum, and ileum. The large intestine starts after the small intestine in the digestive track and consists of cecum, colon, and rectum. The mechanisms of Ca2+ absorption belong to the best studied processes, deficiencies of which cause significant health problems throughout the world. Intestinal Ca2+ absorption is a crucial control system in the regulation of Ca2+ homeostasis, because it facilitates the entry of dietary Ca2+ into the extracellular compartment. Ca2+ is absorbed by two distinct mechanisms including passive (paracellular) and active (transcellular) transport as outlined in sections ii and iii, and their relative magnitude of importance is set by the dietary Ca2+ content. Active transcellular Ca2+ absorption is located largely in the duodenum and upper jejunum, whereas paracellular Ca2+ absorption occurs throughout the entire length of the intestine (60). Chyme moves down the intestinal lumen in ∼3 h, spending only a short time in the duodenum, but over 2 h in the distal part of the small intestine. In the situation that dietary Ca2+ intake is low, transcellular Ca2+ transport accounts for a substantial fraction of the absorbed Ca2+, and vice versa, when Ca2+ intake is high (58, 59).
Transcellular Ca2+ absorption can be described in three sequential cellular steps: entry, intracellular diffusion, and extrusion (see sect. iii). The Ca2+-binding protein calbindin-D9K is responsible for intracellular diffusion of Ca2+ in the enterocyte, and its gastrointestinal expression has been studied in many species. Yamagishi et al. (436) examined the calbindin-D9K mRNA expression in the gastrointestinal tract of cattle by Northern blot analysis. Studies in several animal models have shown that calbindin-D9K and PMCA1b are both expressed in patterns that are compatible with roles in Ca2+ absorption, being found in villous cells of the proximal duodenum and gradually decreased distally, demonstrating vitamin D dependence and decreased expression with aging (Fig. 12B) (7, 117, 190, 379, 436, 445). The expression was highest at the most proximal region. In addition, NCX1 is expressed in the basolateral membrane of mammalian enterocytes (164). In fish enterocytes, NCX1 appears to be the main mechanism by which transcellular fluxes of Ca2+ are extruded from the cells at the basolateral surface (115). In contrast, in mammalian enterocytes, Ca2+ extrusion is predominantly dependent on PMCA1b activity (398).
Recent studies addressed the expression of TRPV5 and TRPV6 in the gastrointestinal tract (Fig. 12B). Initially, Northern blot analysis showed expression of rabbit TRPV5 in duodenum and jejunum, whereas ileum was negative. However, these hybridizations were performed before the identification of the TRPV6 member using a full-length cDNA as a probe that does not discriminate between the two homologous members (178). Subsequent studies using isoform-specific probes, quantitative PCR analysis, and immunohistochemical analysis determined the expression of both channels in the gut. These studies showed that TRPV6 is at the mRNA level at least three orders of magnitude higher expressed than TRPV5 (301, 387, 392). However, at the protein and functional level, it is not clear whether TRPV5 plays a significant role in the intestine, but in line with the observed quantitative mRNA data, a predominant role of TRPV6 is expected in the intestine. Initial characterization of TRPV5 knockout mice that exhibit Ca2+ hyperabsorption, probably mediated by the increased TRPV6 and calbindin-D9K expression levels, is in agreement with this notion (see sect. vii) (180). Additional immunohistochemical and functional studies are needed to address a possible role of TRPV5 in the small intestine. In the intestine, TRPV6 expression was found in duodenum, jejunum, and cecum where it colocalizes with calbindin-D9K and PMCA1b (Fig. 12B) (181, 304, 305). A comparable study by Hediger and co-workers (305) demonstrated expression of TRPV6 throughout the entire digestive tract from esophagus to colon. For unknown reasons, Flockerzi and co-workers (430) failed to detect TRPV6 (CaT-L) expression in the small intestine and colon.
Recently, TRPV6 and TRPV5 mRNA levels were quantified in mouse by quantitative PCR analysis and normalized for cDNA input. This study demonstrated that TRPV6 mRNA in order of decreasing expression is present in duodenum, cecum > colon >> ileum. Likewise, TRPV5 was expressed in kidney >>> duodenum, cecum, whereas ileum and colon were negative (279). Variation in TRPV5 and TRPV6 expression patterns found in different studies in intestine, esophagus, ileum, and colon might be caused by the low abundance of either TRPV5 or TRPV6 in these tissues or differences between species resulting in failure of detection. Immunocytochemical techniques have been used to examine the distribution of TRPV6, TRPV5, calbindin-D9K, and PMCA1b protein in enterocytes (171). TRPV6 and TRPV5 were localized along the brush-border membrane, whereas calbindin-D9K and PMCA1b were expressed cytosolic and at the basolateral membrane, respectively (394, 447). However, at the mRNA level, TRPV6 is 100 times more abundant in the intestine compared with TRPV5. Detailed immunohistochemical studies by Zhuang et al. (447) indicated expression of TRPV6 on the brush-border apical surface of intestinal villi in the entire small intestine and colon. Taken together, these findings are consistent with the postulated role of the epithelial Ca2+ channel TRPV6 as the major transcellular mediator of Ca2+ uptake from the intestinal lumen.
A contribution of the stomach in Ca2+ absorption from the gastrointestinal tract has been postulated from in vivo and in vitro studies in different ruminant species, although this organ is not generally considered to play a major role (339). Unidirectional flux rates of Ca2+ across rumen wall epithelia of sheep were measured in vitro by applying the Ussing-chamber technique in the absence of electrochemical gradients. Under these conditions, significant unidirectional Ca2+ flux rates suggest the presence of active mechanisms for Ca2+ transport (339). Interestingly, quantitative PCR analyses demonstrated a significant expression of TRPV6 in the stomach, whereas a lower expression of TRPV5 was observed (181, 279). These data were confirmed at the protein level in which strong labeling of TRPV6 was observed. TRPV6 expression was predominantly expressed in the upper segments of the gastric glands compared with the lower fragments (447). In this study, it was suggested that TRPV6 plays a role in these mucus-secreting cells in maintaining the intracellular Ca2+ balance after mucus secretion to refill the depleted cellular Ca2+ stores (444, 447). Functional studies showed that the voltage-operated Ca2+ channel blocker verapamil had no significant effect on Ca2+ transport in rumen that is in line with the insensitivity of the epithelial Ca2+ channels for this inhibitor (178, 303, 339). In this study a Ca2+/H+ exchange mechanism in the apical membrane of rumen epithelial cells was postulated that does not seem to be under the control of 1,25-(OH)2D3. Because vanadate did not affect Ca2+ absorption, basolateral Ca2+ extrusion occurs independently from the Ca2+ pump activity and may be accomplished via Na+/Ca2+ exchange (339). Future studies using isolated mucus-secreting cells from TRPV6 and TRPV5 knockout mice should address the physiological role of these channels in more detail.
During pregnancy Ca2+ absorption in the placenta is solely responsible for the nutrient supply to the developing fetus. The Ca2+ needs of the fetus increase progressively throughout the pregnancy. Ca2+ is actively transported across the placenta from the maternal to the fetal circulation in late gestation to meet the requirements of the rapidly mineralizing skeleton and to maintain an extracellular level of Ca2+ that is physiologically appropriate for the development of fetal tissues (27–29, 62, 132, 227, 258–261, 310). Ca2+ is transported by the syncytiotrophoblasts, cells that line the chorionic villi tissue and correspond to the epithelial layer separating the maternal and fetal circulation (105). In analogy to Ca2+ transfer across the intestinal and renal distal tubular cells (see sect. iii), it has been postulated that Ca2+ enters the Ca2+-transporting cells through maternal-facing basement membranes. Subsequently, Ca2+ diffuses across these cells by calbindin-D9K and calbindin-D28K and is actively extruded at the fetal-facing basement membranes by a Ca2+-ATPase (28, 29). With the use of animal models, it was demonstrated that the placental expression of calbindin-D9K increases more than 100-fold over the last 7 days of gestation, whereas the expression of the Ca2+-ATPase increases twofold over the same interval (146). Interestingly, initial Northern blot analysis indicated that TRPV5 and TRPV6 are both expressed in placenta (178, 304). Quantitative PCR analysis in human placenta revealed a robust mRNA expression of TRPV6, whereas TRPV5 was expressed at a very low level (301). Lafond and co-workers (260) demonstrated the expression of TRPV5 and TRPV6 by RT-PCR in cytotrophoblasts freshly isolated from human term placenta. Again, a higher mRNA expression of TRPV6 compared with TRPV5 was measured. This study indicated that the pattern of TRPV6 and TRPV5 expression correlates with the Ca2+ uptake potential along the human trophoblasts isolated from term placenta. These studies provided evidence for a role of the epithelial Ca2+ channels in basal Ca2+ influx during active Ca2+ transport by the syncytiotrophoblast.
Bone, the major Ca2+ store of the body, is an important tissue involved in Ca2+ homeostasis. Ca2+ removal from and redistribution in bone is mediated by specific bone cells, namely, osteoblasts and osteoclasts. The translocation of Ca2+ from the extracellular fluid compartment into the mineralizing matrix is not well understood at the molecular level. Knowledge on how bone directly contributes to serum Ca2+ homeostasis is virtually absent. It can be via bone resorption and formation, but these processes are relatively slow for rapid responses to changes in serum Ca2+. To understand Ca2+ movement in bone and Ca2+ homeostasis, it is crucial to identify and characterize the Ca2+ transport processes in bone. The osteoclast is a cell unique in its ability to resorb bone and becomes exposed to high Ca2+ concentrations in the millimolar range (42). It is generally accepted that, during resorption, osteoclasts can sense changes in their ambient Ca2+ concentration. An increased ambient Ca2+ concentration triggers a sharp cytosolic Ca2+ increase through both Ca2+ release and Ca2+ influx. The bone-forming osteoblasts form a layer that later calcifies. Several reports have described calbindin-D9K expression in rat and human osteoblast-like cells (17, 104, 420). In these studies, immunoreactive calbindin-D9K was localized in the cytoplasm of osteoblasts. Moreover, calbindin-D28K was expressed at low levels in several osteoblastic cell lines and at high levels in primary cultures of murine osteoblastic cells. Its localization in osteoblasts involved in bone formation and in their cell processes suggests a role in Ca2+ transport from these cells toward the sites of active bone mineralization. Interestingly, TRPV5 and TRPV6 mRNA have been detected in bone cells isolated from mouse femurs (279). To date, information on expression and localization of these proteins in bone cells is limited, and their functions in matrix mineralization are unclear.
3. Exocrine tissues
Exocrine tissues such as pancreas, testis, prostate, mammary gland, sweat gland, and salivary gland are not primarily implicated in transepithelial Ca2+ transport. Interestingly, several studies have demonstrated expression of Ca2+-transporting proteins in these exocrine tissues by Northern blot, PCR analysis, or immunohistochemistry. In pancreas, TRPV5 was observed in secretory granules of the β-cells where it colocalized with calbindin-D28K and insulin (194). This study suggested a role for TRPV5 in insulin secretion, although such a role could not be confirmed in studies on the pancreas of Zucker diabetic fatty (ZDF) rats, an animal model for type 2 diabetes mellitus. Interestingly, insulin secretion is impaired in vitamin D-deficient rats (202) but restored by vitamin D treatment (193). Sooy et al. (362) suggested that calbindin-D28K could control the rate of insulin release via regulation of the intracellular Ca2+ concentration. Like in kidney, in pancreatic β-cell NCX, PTH-related peptide (PTHrP), calbindin-D28K and receptors for 1,25-(OH)2D3 have been detected (76, 97, 198, 396). In contrast to TRPV5, TRPV6 mRNA is found in the exocrine acinar cell of the pancreas, but not in pancreatic duct and β-cells (430, 447). In the human pancreas, acinar cells showed a granular and apical membrane staining of TRPV6 (447). In addition, a weak signal was found along the basolateral membrane that could be involved in Ca2+ entry from the circulation. Ca2+ extrusion in the acinar pancreatic cell has been studied in great detail. With the use of confocal microscopy and Ca2+-sensitive fluorescent probes, it was shown that the secretory pole of the acinar cell is the major Ca2+ extrusion site following agonist stimulation (25, 26). It is well known that intracellular Ca2+ plays an essential role in exocytosis. Localization of Ca2+ influx through TRPV6 and Ca2+ extrusion at the luminal membrane would be important to direct physiological Ca2+ signals to the apical compartment of the acinar cell. Altogether, this indicates that in the pancreatic cell the Ca2+ homeostasis is a complex process in which many Ca2+ transport proteins might play a role and functional studies are needed to investigate the function of the Ca2+ transport proteins in the pancreas.
Several studies demonstrated a robust TRPV6 expression in prostate. Recently, quantitative PCR measurements and immunohistochemical studies indicated that the prostate contains the highest TRPV6 expression levels of all tested tissues (111, 176, 265, 279, 305, 306, 430, 447). Although the exact function in this organ remains to be elucidated, previous reports have suggested that TRPV6 expression correlates with prostate carcinoma tumor grade (111, 306). The first indication of TRPV6 as tumor progression marker came from expression studies in human prostate cancer cell lines (306). Expression of TRPV6 was elevated in prostate cancer samples compared with benign prostatic hyperplasia specimens and positively correlated with Gleason grade in prostate cancer. TRPV6 mRNA was downregulated by androgen and induced by a specific androgen receptor antagonist in LNCaP cells, suggesting that the expression of this channel is negatively regulated by androgen. Flockerzi and co-workers (430) identified a transcript from rat duodenum, named CaT-L, which was expressed in locally advanced prostate cancer and metastatic and androgen-insensitive prostatic lesions, but was undetectable in healthy prostate tissue (430). Detailed characterization of CaT-L demonstrated that this gene is identical to TRPV6. Recently, it was reported that CaT-L or TRPV6 is not expressed in benign prostate tissues including benign prostate hyperplasia, but is upregulated in prostate cancer. Other reports indicated that TRPV6 mRNA (279) and protein (447) are also expressed in normal prostate tissue. In prostatectomy specimens from 97 clinically organ-confined tumors, TRPV6 expression correlated significantly with the Gleason score, pathological stage, and extraprostatic extension (111). Together these findings indicate that TRPV6 expression is associated with prostate cancer progression and, therefore, represents a prognostic marker and a promising target for new therapeutic strategies to treat advanced prostate cancer.
Interestingly, several exocrine glands including sweat, salivary, and mammary glands do express one or both of the epithelial Ca2+ channels. Freeman and co-workers (447) demonstrated that TRPV6 is localized in the mammary gland along the luminal membrane of the epithelial cells in contrast to localization in the sweat gland where the channel was detected on both the luminal and basolateral membrane surfaces. Differences in the localization of TRPV6 may reflect diverse physiological functions in these exocrine glands. To date, TRPV5 expression in these exocrine glands remains to be established. In addition, functional studies are needed to examine the physiological function of the epithelial Ca2+ channels in these tissues.
VI. REGULATION OF EPITHELIAL CALCIUM TRANSPORT
A. Regulation by 1,25-(OH)2D3
It is commonly accepted that vitamin D3 is one of the main hormones controlling Ca2+ balance (322). There are two sources of vitamin D3 in the body. It is either ingested from the diet or synthesized in the skin from its precursor 7-dehydrocholesterol in the presence of sunlight (271). Vitamin D3 itself is physiologically inactive. It will undergo an activation process, involving 25-hydroxylation in the liver followed by 1α-hydroxylation in the kidney to synthesize the biologically active 1,25-(OH)2D3. The latter reaction step occurs in mitochondria of the renal proximal tubule (116). Whether these proximal tubular cells produce 1,25-(OH)2D3 depends on the Ca2+ status of the body. In the situation that Ca2+ is sufficient with adequate dietary Ca2+ intake and normal plasma Ca2+ concentration, 1α-OHase activity is low because there is no need for additional Ca2+. However, when Ca2+ is insufficient, with a low dietary Ca2+ intake and decreased plasma Ca2+ concentration, the activity of this enzyme increases to produce 1,25-(OH)2D3 to ensure that additional Ca2+ will be absorbed from the gastrointestinal tract.
From a historical point of view, the biological role of 1,25-(OH)2D3 in active intestinal Ca2+ absorption is most studied. Orr and co-workers (141, 350) discovered many decades ago that vitamin D3 is required for intestinal Ca2+ absorption. Ample studies confirmed this initial study that established the role of 1,25-(OH)2D3 in active Ca2+ absorption. More recently, a similar role in the distal part of the nephron was demonstrated as discussed in section v (38, 175). The effects underlying at least part of these processes are mediated by the interaction of 1,25-(OH)2D3 with the nuclear VDR in a ligand-dependent manner (159). This genomic mechanism of action is similar to that of other steroid hormones and is mediated by stereospecific interaction of 1,25-(OH)2D3 with the VDR which heterodimerizes with the retinoid X receptor (RXR) (211). After interaction with the vitamin D response element (VDRE) in the promoter of target genes, transcription proceeds through the interaction of VDR with coactivators and with the transcription machinery. Importantly, VDR is expressed in epithelia that play a role in Ca2+ (re)absorption. The functional significance of target proteins as well as the functional significance of proteins involved in the transport and metabolism of vitamin D is also of major importance. In general, transcellular Ca2+ transport in the small intestine and the distal part of the nephron is facilitated by the Ca2+ transport proteins, namely, TRPV5, TRPV6, the calbindins, NCX1 and PMCA1b, and stimulated by 1,25-(OH)2D3 primarily via a genomic action. The contribution of each individual transporter to the overall stimulatory action of 1,25-(OH)2D3 has been addressed in several studies as described in the following section.
Early studies demonstrated that 1,25-(OH)2D3 stimulates the expression of calbindin-D9K and calbindin-D28K in humans and many animal models. Promoter studies on these mammalian calbindin genes have demonstrated functional VDREs that interact with nuclear factors and may mediate, at least in part, the enhanced expression of these genes by 1,25-(OH)2D3 (88, 140). The appearance of calbindin-D9K protein in the gut and Ca2+ transport coincide as a function of time in response to 1,25-(OH)2D3 (200). However, situations are reported in which Ca2+ absorption diminishes while calbindin remains high in the small intestine (158, 363). Apparently, a tight correlation between calbindin expression and Ca2+ transport is not always present. It is not known whether the expression of TRPV6 is affected under this condition, but these observations suggest that calbindin-D9K is not the rate-limiting step in Ca2+ absorption. Previous studies have indicated that other proteins play an additional role in transcellular Ca2+ (re)absorption which led to an analysis of the basolateral efflux and apical influx Ca2+ transporters.
2. Basolateral extrusion mechanisms
The effect of 1,25-(OH)2D3 on the basolateral extrusion systems, NCX1 and PMCA1b, is less clear and remains controversial. Although NCX1 plays a dominant role in the extrusion process in renal cells (121, 184, 243, 427), many studies failed to establish a direct regulation by vitamin D. Exposure of 1,25-(OH)2D3 to primary cultures of rabbit DCT and CNT did not noticeably alter NCX1 expression (391). However, repletion studies with vitamin D-deficient animal models consistently demonstrated a vitamin D-dependent regulation of renal NCX1 and duodenal PMCA1b (see sect. vi). Importantly, these studies in mice indicated that renal PMCA1b expression and intestinal NCX1 expression were not significantly regulated by vitamin D. PMCA1b is the only isoform predominantly expressed in small intestine and kidney (8, 391). In kidney, it has been shown experimentally that during variable circumstances the extrusion capacity of PMCA1b is more than adequate, suggesting that the Ca2+ exit step is not necessarily a prime target for regulation by 1,25-(OH)2D3 (391). This could perhaps explain the difficulty to observe a consistent stimulatory effect of 1,25-(OH)2D3 on renal PMCA1b (169, 392). Recently, Kip and Strehler (214) demonstrated that 1,25-(OH)2D3 is a positive regulator of PMCA in MDCK cells. Interestingly, 1,25-(OH)2D3 caused a decrease of PMCA protein content in the apical membrane fraction and a concomitant increase of the pumps in the basolateral membrane (214). In addition, these authors could demonstrate a significant increase in the expression of PMCA upon stimulation by 1,25-(OH)2D3 that correlated with the magnitude of transcellular Ca2+ transport. In small intestine, many groups have shown that 1,25-(OH)2D3 upregulates PMCA1b protein expression (8, 9, 387, 392). Furthermore, 1,25-(OH)2D3 enhanced PMCA1b mRNA stability and activity (147, 148). However, other reports in which primary cultures of renal cells and animal models were used failed to show significant regulation of renal PMCA1b expression level by 1,25-(OH)2D3 (391, 392). Conversely, run-off reporter gene assays using 1.7 kb of the human PMCA1 promoter expressed in distal tubular cell lines demonstrated mRNA downregulation by 1,25-(OH)2D3 (149). Taken together, a consistent stimulatory effect of 1,25-(OH)2D3 on NCX1, but not on PMCA1b, in the kidney has been established. In the small intestine, however, PMCA1b is the vitamin D-regulated extrusion system.
3. Epithelial Ca2+ channels
Recent studies consistently indicated that the expression of TRPV5 and TRPV6 is tightly controlled by 1,25-(OH)2D3 (51, 61, 112–114, 169, 174, 279, 359, 387, 392, 435). The first evidence for this vitamin D sensitivity was obtained in in vivo studies in which vitamin D3-depleted rats were repleted (174). This was accompanied by normalization of the plasma Ca2+ concentration and an increase in the amount of TRPV5 mRNA and protein expression in the kidney. As a different approach to elicit vitamin D genomic responses, a single dose of 1,25-(OH)2D3 was administered to mice (392). Quantitative PCR data demonstrated an upregulation of TRPV5 in kidney (3-fold) and TRPV6 in duodenum (6-fold). Recently, Nijenhuis et al. (279) addressed the localization and 1,25-(OH)2D3-dependent regulation of TRPV6 in the kidney. Intraperitoneal injection of 100 ng 1,25-(OH)2D3 in mice resulted in a increase in TRPV6 mRNA (2-fold) and protein (3-fold) expression in the kidney. Although renal TRPV6 expression was upregulated by 1,25-(OH)2D3, the effect of this hormone on TRPV5 expression was more impressive. Analysis of putative promoter regions of human and murine TRPV5/6 genes revealed potential vitamin D response elements in line with the previously observed functional data (174, 279). Song et al. (359) also examined the expression of TRPV5 and TRPV6 mRNA in duodenum and kidney of mice. Following a single dose of 1,25-(OH)2D3, induction of duodenal TRPV6 mRNA occurred within 3–6 h and preceded the induction of intestinal Ca2+ absorption. In addition, this study described that the intestinal TRPV6 mRNA level increased 30-fold at weaning, coincident with the induction of calbindin-D9K expression. In contrast, renal TRPV6 and TRPV5 mRNA expression was equal until weaning when TRPV5 mRNA is induced and TRPV6 mRNA levels drop 70% (359). The observed vitamin D-dependent regulation of TRPV5 and TRPV6 was subsequently extensively studied using many different cell lines and animal (knockout) models as discussed in the next paragraph.
4. Vitamin D-dependent regulation of Ca2+ transport in animal models
Powerful tools to investigate the vitamin D dependency of target proteins are knockout mice. Within the past few years a wealth of new information has been obtained from studies using knockout and transgenic mice. Several genetically modified mouse models including VDR (392, 440) and 1α-OHase (86, 298) knockout mice, in which the vitamin D system has been inactivated, have been created to systematically dissect the genetic regulation of Ca2+ transport genes and their functional consequences on transcellular Ca2+ transport. St-Arnaud and co-workers (86) generated 1α-OHase knockout mice that represent a unique animal model for pseudovitamin D-deficiency rickets (PDDR), since these mice display undetectable 1,25-(OH)2D3 concentrations, hypocalcemia, secondary hyperparathyroidism, and failure to thrive (86). In addition, the 1α-OHase−/− mice developed distinct histological evidence of rickets and osteomalacia. Interestingly, there was a correlative relationship between the expression level of TRPV5, calbindin-D28K, and NCX1 proteins in kidney; TRPV6, calbindin-D9K, and PMCA1b in duodenum; and the serum Ca2+ concentration (169, 386, 387). Normalization of the plasma Ca2+ concentration by 1,25-(OH)2D3 was associated with a restoration of the expression level of the Ca2+ transporters, confirming the essential role of these proteins in active 1,25-(OH)2D3-mediated Ca2+ (re)absorption. The concerted regulation of TRPV5/6 and the other Ca2+-transporting proteins guarantees sufficient capacity during high transport rates. Calbindin-D regulates the Ca2+ influx across the apical membrane by buffering intracellular Ca2+ and thus controlling feedback inhibition of TRPV5/6 channel activity (404).
Analogous observations were made from experiments performed with VDR knockout mice (392, 423). In these hypocalcemic mice, urinary Ca2+ excretion is inappropriately high, suggesting renal Ca2+ wasting due to disturbed Ca2+ reabsorption. It has been demonstrated in this mouse model that duodenal TRPV6 levels are dramatically and consistently downregulated (392, 423). Intriguingly, the observed expression pattern indicated that, among the candidate Ca2+-transporting genes, mainly TRPV6 is severely impaired in VDR-knockout mice.
In addition, the correlation between vitamin D and the expression level of the Ca2+ transport proteins has been addressed in several cell models. Wood and co-workers (113, 435) studied in detail the vitamin D sensitivity of TRPV6, calbindin-D9K, and PMCA1b in Caco-2 cells. A clear correlation between the 1,25-(OH)2D3-induced expression of TRPV6, calbindin-D9K, and PMCA1b and transcellular Ca2+ transport was established in this intestinal cell line. In contrast, Barley et al. (20) could not confirm the generally observed vitamin D-dependent sensitivity of TRPV6 in duodenal biopsies from 20 normal volunteers. However, there was a 10-fold variation between the lowest and the highest level of TRPV6 expression. In addition, the subjects used were a mixed population of men and women of age 25–71 years, which made it hard to disclose a relationship between TRPV6 expression and vitamin D metabolites. Likewise, cell lines and primary cultures have been established from the distal part of the nephron including DCT and CNT (38, 91, 391). Measurements of calbindin-D28K expression in control and 1,25-(OH)2D3-treated renal epithelial cells indicated a direct relationship between 1,25-(OH)2D3-induced calbindin expression and transcellular Ca2+ transport (391). Taken together, vitamin D-deficient animal models and epithelial cell lines demonstrated a consistent 1,25-(OH)2D3 sensitivity of TRPV5, TRPV6, and the calbindins and to a lesser extent the basolateral extrusion systems NCX1 and PMCA1b.
B. Regulation by PTH
The parathyroid glands play a key role in maintaining the extracellular Ca2+ concentration through their capacity to sense even minute changes in the level of blood Ca2+ from its normal level. The Ca2+-sensing receptor (CaSR) is the mechanism through which the parathyroid chief cells sense variations in the Ca2+ concentration and release PTH. In response to low blood Ca2+ levels, PTH is secreted into the circulation and then acts primarily on kidney and bone, where it activates the PTH/PTHrP receptor. This receptor directly enhances the tubular Ca2+ reabsorption, and it stimulates the activity of 1α-OHase and, thereby, increases the 1,25-(OH)2D3-dependent absorption of Ca2+ from the intestine. Several groups localized PTH/PTHrP receptor mRNA in rat kidney to glomerular podocytes, PCT, PST, cortical thick ascending limb, and DCT, but the receptor was not detected in the thin limb of Henle’s loop or in CD (231, 437). PTH stimulates active Ca2+ reabsorption in the distal part of the nephron (154). As outlined in section iii, Bindels et al. (38) and Gesek and Friedman (124) have used immunodissected cell lines from rabbit DCT and CNT and mouse DCT, respectively, to investigate PTH-stimulated Ca2+ transport. These studies demonstrated equivocally that PTH increases transepithelial Ca2+ transport via a dual signaling mechanism involving PKA- and PKC-dependent processes (124, 170). Various mechanisms of PTH action have been postulated including membrane insertion of apical Ca2+ channels (12), opening of basolateral chloride channels resulting in cellular hyperpolarization (127), and modulation of PMCA activity (383). Friedman and colleagues (12) examined Ca2+ influx in single cultured cells from distal renal tubules sensitive to PTH by measuring intracellular Ca2+. Their results demonstrated that PTH activates dihydropyridine-sensitive channels responsible for Ca2+ entry. Once inserted or activated, these dihydropyridine-sensitive channels could mediate Ca2+ entry into these Ca2+-transporting epithelial cells (12). This finding is in contrast to the channel characteristics of TRPV5 that is dihydropyridine insensitive (178). It is well possible that these dihydropyridine-sensitive Ca2+ channels play a role in signal transduction processes to maintain the cellular Ca2+ homeostasis. Recently, it was reported that parathyroidectomy in rats resulted in decreased serum PTH levels and hypocalcemia, which was accompanied by decreased levels of TRPV5, calbindin-D28K, and NCX1 (388). Supplementation with PTH restored serum Ca2+ concentrations and abundance of these Ca2+ transporters in kidney. These data suggest that PTH affects renal Ca2+ handling through the regulation of the expression of the active renal Ca2+ transport proteins, including the epithelial Ca2+ channel TRPV5. Furthermore, it was demonstrated that PTH stimulates the PMCA activity by increasing the affinity for Ca2+ in the distal tubule, which is in contrast to 1,25-(OH)2D3 that did not directly affect the basolateral membrane PMCA activity (383).
Immunohistochemical analysis of rat duodenal sections showed localization of the PTH/PTHrP receptor in epithelial cells along the villus with intense staining of brush-border and basolateral membranes and cytoplasm (135). Interestingly, the receptor was absent in goblet cells. Direct effects of PTH have been reported on Ca2+ uptake by isolated rat duodenal cell preparations enriched in enterocytes. The first indication of a direct effect of PTH on the intestine was accomplished by perfusion experiments of isolated duodenal loops with PTH that increased Ca2+ transport (272, 273). These findings were confirmed by Picotto et al. (309) who demonstrated that PTH significantly stimulates enterocyte Ca2+ influx. This Ca2+ influx was blocked by the Ca2+ channel antagonists verapamil and nitrendipine.
In bone, PTH can induce a rapid release of Ca2+ from the bone matrix, but it also mediates long-term changes in Ca2+ metabolism by acting directly on the bone-forming osteoblasts and indirectly on bone-resorbing osteoclasts by increasing their number and activity. However, molecular mechanisms of PTH action in mediating Ca2+ transport in bone and intestine remain poorly understood.
C. Regulation by Calcitonin
Calcitonin, a 32-amino acid peptide hormone produced primarily by the thyroid, and its receptor are well known for their ability to regulate osteoclast-mediated bone resorption (186). Calcitonin has been suggested to be a renal Ca2+-conserving hormone and may share similar signaling mechanisms with PTH. Previously, this peptide was infused into groups of acutely thyroparathyroidectomized rats that had been treated with calcitonin for 12 days resulting in a marked inhibitory effect on renal Ca2+ excretion (67). The hypocalcemic and hypophosphatemic calcitonin is secreted by mammalian thyroid parafollicular cells. It has been reported that calcitonin increases Ca2+ reabsorption in mouse DCT cells, but the molecular mechanism is not known yet. It was postulated that calcitonin increases Cl− conductance in DCT cells, resulting in membrane hyperpolarization and activation of Ca2+ entry through Ca2+ channels (139). Interestingly, renal 1α-OHase gene expression is strictly upregulated at the transcriptional level through its gene promoter by PTH and calcitonin, whereas 1,25-(OH)2D3 itself has a negative effect on gene expression (205). Consequently, increased 1α-OHase gene activity results in elevated 1,25-(OH)2D3 levels and enhanced intestinal and renal Ca2+ (re)absorption. Previous experiments demonstrated that calcitonin inhibits bone resorption and decreases Ca2+ efflux from isolated cat tibiae that underlies its widespread clinical use for the treatment of bone disorders, including Paget’s disease, osteoporosis, and hypercalcemia of malignancy.
D. Regulation by Stanniocalcin
Stanniocalcin is another hypocalcemic hormone that is originally identified in fish (414). In these animals stanniocalcin exerts its antihypercalcemic effect by regulating Ca2+ and phosphate transporters in the gills, intestine, and kidney (136a). Interestingly, this hormone has also been recently identified in humans (415). In mammals, stanniocalcin is expressed in multiple organs including Ca2+-transporting epithelia like intestine, colon, kidney, and placenta (70, 71). It is released by specialized organs, the corpuscles of Stannius that are located adjacent to the kidney and scattered throughout the kidney (415). Immunoreactivity for stanniocalcin was detected in the limb of Henle, macular densa cells, DCT, and CCD (297). Stanniocalcin acts locally in kidney and gut to modulate Ca2+ and phosphate excretion, and its overexpression in mice results in high serum phosphate, dwarfism, and increased metabolic rate (109, 401). The main function of stanniocalcin, similar to that of calcitonin, appears to be the prevention of hypercalcemia. Importantly, stanniocalcin was upregulated by 1,25-(OH)2D3 treatment (188, 297). Nevertheless, the true physiological role for stanniocalcin in mammals is less clear, and future studies are needed to establish the functional importance of this hormone.
E. Regulation by Estrogens
Estrogen deficiency results in a negative Ca2+ balance and bone loss in postmenopausal women (289, 441, 442). Estrogen deficiency after menopause results in bone loss, which is associated with a rise in plasma and urinary Ca2+ (442). It has been generally described that the rise in plasma and urine Ca2+ are secondary to an increase in bone resorption. However, some studies have shown that the rise in urinary Ca2+ at menopause is not due to an increase in filtered load, suggesting that estrogen also has an effect on renal Ca2+ handling (1). In addition to bone, the intestine and kidney are also potential sites for estrogen action and involved in Ca2+ handling and regulation. There is increasing evidence that estrogen exerts a physiological role in the regulation of renal and intestinal Ca2+ (re)absorption. In vivo studies showed that estrogen deficiency is associated with increased renal Ca2+ loss, which can be corrected by estrogen replacement therapy (290, 315). Furthermore, estrogen receptors also reside in proximal and distal tubules of the kidney and in duodenum and colon. However, the underlying mechanism by which estrogen may act on Ca2+ (re)absorption is still poorly understood. In addition, it has not been conclusively established whether there is a direct effect of estrogen on Ca2+ transport or indirectly mediated by an effect on vitamin D metabolism. It was demonstrated that estrogen upregulates the expression of TRPV5 in kidney in a 1,25-(OH)2D3-independent manner (386). In ovariectomized 1α-OHase knockout mice, 17β-estradiol replacement therapy resulted in upregulation of renal TRPV5, but not the other Ca2+ transporters, mRNA, and protein levels, leading to normalization of plasma Ca2+ levels (386). Thus renal TRPV5 expression is, independent of vitamin D, transcriptionally controlled by estrogen. By upregulating TRPV5 expression, estrogen could be positively involved in Ca2+ reabsorption. Recent findings suggest that also TRPV6 expression is regulated by estrogen, as duodenal expression of TRPV6 mRNA of 1α-OHase knockout mice and ovariectomized rats is upregulated after 17β-estradiol administration (387). Van Cromphaut et al. (393) reported that renal TRPV5 and duodenal TRPV6 expression are reduced in estrogen receptor α (ERα) knockout mice and upregulated by estrogen treatment. In this study, TRPV6 expression was enhanced in both pregnant VDR knockout mice and wild-type littermates. Furthermore, in lactating mice, renal TRPV5 mRNA and duodenal TRPV6 expression levels increased 2 and 13 times, respectively. It is clear that the expression of both epithelial Ca2+ channels is influenced by the estrogen status. Estrogens, hormonal changes during pregnancy, and lactation have distinct, vitamin D-independent effects at the genomic level on active duodenal Ca2+ absorption mechanisms, mainly through a major upregulation of TRPV6. The estrogen effects seem to be mediated solely by ERα. Nevertheless, it remains to be clarified whether these changes have functional implications on Ca2+ (re)absorption. These data confirm that estrogens and vitamin D are independent potent regulators of the expression of TRPV6, which is involved in active intestinal Ca2+ absorption. Together, these data indicate that the function of estrogen in maintenance of the Ca2+ balance might be at least in part fulfilled by regulation of TRPV5 and TRPV6 levels, thereby controlling (re)absorption of the amount of Ca2+ that is needed for bone calcification.
The mechanism of estrogen-controlled upregulation of epithelial Ca2+ channel mRNA remains to be elucidated. Interestingly, Weber et al. (423) recently described an estrogen-responsive element in the promoter sequence of the mouse TRPV6 gene, which was absent in the mouse TRPV5 gene. Alternatively, transcriptional activation by the estrogen-liganded estrogen receptor can be mediated through other elements including activator protein 1 (AP-1) binding sites and GC-rich stimulatory protein (Sp1) binding sites (224, 334). Importantly, the human TRPV5 promoter contains several of these AP-1 and Sp1 sites (174). In the 5′-upstream region of the translational initiation site in the mouse TRPV5 gene, several of these AP-1 and Sp1 binding sites can be found that could be involved in the positive effect observed on TRPV5 mRNA expression by 17β-estradiol treatment. Detailed promoter analysis is necessary to identify the regulatory sites involved in this estrogen-mediated regulation of TRPV5 and TRPV6.
F. Regulation by Thyroid Hormone
There is ample evidence that thyroid dysfunction is associated with disturbances of Ca2+ and phosphate homeostasis (81–83, 222, 223, 263). Hypercalcemia is frequently observed in humans with thyrotoxicosis, and similar observations were made in animals (103, 160, 223, 333). It is interesting that long-term hyperthyroid state is associated with Ca2+ malabsorption and increased bone resorption. Several effects of thyroid hormone on the kidney, intestine, and bone resemble those of 1,25-(OH)2D3. Actions at the cellular level constituting their function as Ca2+-regulating hormones have not been addressed until recently. Kumar and Prasad (223) demonstrated that Ca2+ uptake into brush-border membrane vesicles and Ca2+ efflux from the basolateral membrane of enterocytes was significantly increased in hyperthyroid rats and decreased in hypothyroid animals. Comparable observations were made in the kidney by these investigators (222). It was postulated that thyroid hormone increases the affinity of TRPV6 for Ca2+ in the brush-border membrane of the enterocyte. However, experiments to confirm this hypothesis remain to be performed.
G. Regulation by Dietary Ca2+
Dietary Ca2+ is one nutrient that has been the focus of multiple studies in an effort to discover its preventive actions. It has been implicated in the reduction of risk in osteoporosis, whereas a low-Ca2+ diet avoids kidney stone formation. Obesity, hypertension, and even cancer are less well-known areas in which increasing dietary Ca2+ has a positive outcome (232). A significant link between Ca2+ intake and bone mass has been reported. Although recommended daily allowance of calcium is 600 mg/day for adults, ∼850 mg/day or more is recommended later in life.
The power of Ca2+ supplementation is best illustrated by the use of VDR and 1α-hydroxylase knockout models. The bone phenotype of VDR-ablated mice can be completely rescued by feeding the animals a high-Ca2+, high-phosphorus, high-lactose diet. In addition, the PDDR phenotype of mice deficient for the 1α-OHase gene has been rescued by feeding them with the high-Ca2+ diet. The rescue regimen consisted of feeding a diet containing 2% (wt/wt) Ca2+ from 3 wk of age until death at 8.5 wk of age. Serum analysis revealed that the rescue diet corrected the hypocalcemia and secondary hyperparathyroidism (87, 168, 169).
Subsequently, the expression level of the Ca2+ transport proteins was studied in the knockout mice models described above. Several aforementioned studies provided evidence that the Ca2+ transport proteins, including TRPV5 and TRPV6 channels, are regulated by 1,25-(OH)2D3. It is, however, difficult to distinguish the effects of hypocalcemia from those of vitamin D deficiency. Therefore, studies were performed in VDR and 1α-OHase knockout mice fed a normal and high-Ca2+ rescue diet (169, 392). Importantly, the reduced expression level of renal TRPV5, calbindin-D28K, and NCX1 in the 1α-OHase−/− mice was restored by high dietary Ca2+ intake and accompanied by normalization of the plasma Ca2+ concentration. In line with the 1,25(OH)2D3-dependent regulation, dietary Ca2+ controls also the other Ca2+ transport proteins. In contrast, this Ca2+-enriched rescue diet reduced the expression of renal TRPV5 and calbindin-D28K in 1α-OHase+/− mice that exhibit normal serum vitamin D and Ca2+ levels. It is known that under physiological conditions, plasma Ca2+ acts via a negative-feedback mechanism that eventually leads to suppression of the 1α-OHase-activity that decreases Ca2+ reabsorption and expression of Ca2+ transport proteins. Likewise, the expression of the intestinal Ca2+ transport proteins, TRPV6, calbindin-D9K, and PMCA1b, were normalized by this rescuing Ca2+ diet (387).
Comparable observations were made in VDR knockout mice where duodenal TRPV5 and TRPV6 mRNA levels were upregulated by dietary Ca2+ (392). Studies with VDR−/− and 1α-OHase−/− mice revealed that Ca2+ supplementation can upregulate gene transcription encoding for Ca2+ transporters in the absence of circulating 1,25-(OH)2D3, but the molecular mechanism of this vitamin D-independent Ca2+-sensitive pathway remains elusive. It is, however, likely that in addition to the identified 1,25-(OH)2D3 response elements in the promoter regions of TRPV5/6 and calbindin-D genes also Ca2+-responsive elements are present. Several elements have been proposed to function as Ca2+-sensitive transcriptional regulators including the serum responsive element and the cAMP/Ca2+-responsive element (133). Of interest is the identification of a Purkinje cell expression specific element in the calbindin-D28K gene that functions as a Ca2+-sensitive transcriptional regulatory mechanism (10). This mechanism may play a role in fine-tuning the Ca2+ buffer capacity of Purkinje cells. Detailed promoter analysis is needed to investigate if these domains are also present in the promoter of the Ca2+ transport genes.
H. Regulation by the Ca2+-Sensing Receptor
The CaSR plays a pivotal role in the regulation of Ca2+ homeostasis by sensing subtle changes in circulating Ca2+ concentration (185). In the parathyroid gland, the CaSR represents the molecular mechanism by which parathyroid cells detect changes in blood ionized Ca2+ concentration, modulate PTH secretion accordingly, and thus maintain serum Ca2+ levels within a narrow physiological range. Interestingly, in the kidney, the CaSR regulates renal Ca2+ excretion and influences the transepithelial movement of water and other electrolytes. The CaSR is expressed at numerous sites along the nephron including the apical membrane of the PT, the basolateral membrane of the medullary and cortical TAL and DCT, in some cells of CCD, and at the apical membrane of the IMCD. Thus CaSR can apparently be trafficked either to the apical or to the basolateral membrane depending on the tubule cell type. The receptor is located in the epithelial cells of the PT, where it is located in close proximity to the numerous apical transporters of various nutrients and electrolytes that reside within this portion of the nephron (329). There is evidence that the CaSR in the PT antagonizes or limits the effects of PTH as stimulator of 1,25-(OH)2D3 production and phosphate excretion (11). The water-impermeable TAL is responsible for the reabsorption of ∼20–25% of filtered Ca2+. The extracellular Ca2+ concentration at the basolateral side of these cells will increase as Ca2+ reabsorption occurs. Elevation of plasma Ca2+ or Mg2+ levels modulates mineral ion transport in the loop of Henle. The observations can be explained by the action of the CaSR, which permits the reabsorbed Ca2+ to feed back onto the cell decreasing Ca2+/Mg2+ reabsorption and, thereby, preventing hypercalcaemia (421). Motoyama and Friedman (264) demonstrated that CaSR activation in cortical TAL by the agonist NPS R-467 or Gd3+ indirectly inhibits passive Ca2+ transport and directly suppresses PTH-induced transcellular Ca2+ transport. Immunohistochemical investigation of rat kidney reveals CaSR immunostaining at the basolateral and occasionally apical membranes of DCT (329). However, the extent to which CaSR expression overlaps with the Ca2+ transport proteins including TRPV5 is not known. In MDCK cells, basolateral exposure to CaSR-stimulatory concentrations of extracellular Ca2+ or neomycin inhibits unidirectional Ca2+ reabsorption across these cells, possibly via the inhibition of PMCA (44). Studies in immortalized mDCT cells showed that 1,25-(OH)2D3-induced Mg2+ influx is also inhibited upon exposure to elevated extracellular Ca2+ concentrations, an effect that was blocked by an anti-CaSR monoclonal antibody or by transfection with antisense CaSR oligodeoxynucleotides (330). Furthermore, it has been demonstrated that a high luminal Ca2+ concentration activates the CaSR receptor in the apical membrane of the IMCD. As a consequence, the vasopressin-elicited water permeability is blunted by a PKC-mediated aquaporin-2 retrieval mechanism to reduce water reabsorption and prevent a further rise in urinary Ca2+ concentration (335). These data are the first analyses of hypercalcemia-induced alterations in arginine vasopressin-regulated water permeability and membrane transporters in IMCD. It is hypothesized that alterations in IMCD transport occur during hypercalcemia, allowing the body to dispose of excess Ca2+ without forming Ca2+-containing renal stones.
CaSR forms a unique molecular target for drugs that can directly alter the activity of the receptor, thereby manipulating the extracellular Ca2+ balance. Several calcimimetic compounds, like NPS R-467, have been described that activate CaSRs on parathyroid cells and suppresses serum levels of PTH and Ca2+. However, the involvement of intestinal and renal Ca2+ (re)absorption in the NPS R-467-induced hypocalcaemia is not clear. A recent study by van Abel et al. (388) investigated the effect of NPS R-467 on the expression of intestinal and renal Ca2+ transport proteins, including the epithelial Ca2+ channels TRPV5 and TRPV6. To this end, mice were infused with NPS R-467 via osmotic minipumps for 7 days. Treatment with NPS R-467 reduced serum PTH levels in a dose-dependent manner, which was accompanied by a significant decrease in serum Ca2+ concentration. Quantitative PCR and biochemical analysis of duodenal and renal samples demonstrated an overall downregulation of mRNA expression levels and proteins involved in active transcellular Ca2+ (re)absorption. The effects of the treatment with NPS R-467 on the expression of Ca2+ transporters could involve a direct action of CaSRs in kidney and/or intestine; however, previous studies have also indicated that the acute hypocalcemic response to calcimimetic compounds results from the inhibition of PTH secretion (388).
I. Regulation by Associated Proteins
To date, little information is available concerning the molecular players responsible for regulating the activity of TRPV5 and TRPV6. A number of regulatory proteins have recently been described that modify the biophysical, pharmacological, and expression properties of ion channels and transporters by direct interactions (233). These newly identified associated proteins have facilitated the elucidation of important molecular pathways modulating transport activity. Until now three regulatory proteins, i.e., calmodulin, S100A10-annexin 2, and 80K-H, have been described that associate with TRPV5 and/or TRPV6 (145, 276, 394).
CaM was identified as a TRPV6-interacting member and is a ubiquitous cytosolic protein known to regulate the activity of different ion channels, Ca2+ pumps, and other proteins in a Ca2+-dependent manner (276). CaM consists of four Ca2+-binding EF-hand structures, which are localized in the amino and carboxy terminus. Ca2+ binding to CaM is highly cooperative with Ca2+ binding first to the carboxy-terminal EF-hands, which have the highest affinity for Ca2+, followed by Ca2+ binding to lower affinity sites located in the amino terminus (418). At rest, when Ca2+ concentrations are low, Ca2+ entry through voltage-gated Ca2+ channels, cyclic nucleotide-gated channels, and NMDA receptors is enabled. Upon Ca2+ influx, activated Ca2+-CaM inactivates the above-mentioned influx pathways as well as TRPL-mediated currents in Drosophila photoreceptors (341). In addition, other members of the TRP family are possibly regulated by CaM. CaM binding to TRPV1 was restricted to a 35-amino acid segment in the carboxy terminus, and deletion of this CaM binding segment prevented TRPV1 desensitization (293). Members of the TRPC family have been shown to bind CaM (52, 225, 439). CaM acts as a Ca2+ sensor in the Ca2+-dependent feedback inhibition of TRPC1 as demonstrated using Ca2+-insensitive CaM mutants (356). Another study using CaM inhibitors described that TRPC6 is regulated by CaM (52).
There is a rapid Ca2+-dependent inactivation of TRPV6 channels that seems to be independent of CaM binding and could be due to a local effect of Ca2+ on the intracellular pore-forming region (276). This conclusion was based on the observation that the inactivation behavior of TRPV6 was altered in mutants that do not bind CaM. In addition, there was a slower inactivation that was found to be Ca2+/CaM dependent. Although CaM does not bind to TRPV6 at Ca2+ concentrations normally present in cells at rest, binding was observed at increased Ca2+ concentrations with maximal binding at a concentration of 60 μM. Ca2+/CaM binding inactivates TRPV6 channels, although additional Ca2+-dependent and -independent inactivation or rundown mechanisms must exist. Removal of CaM binding does not affect the initial rapid phase of inactivation, nor does it affect the slow inactivation.
Interestingly, binding of Ca2+/CaM, however, can be prevented by PKC-mediated phosphorylation of a threonine residue within the CaM binding site. Niemeyer et al. (276) concluded that PKC activity thus may act as a switch that can regulate the amount of Ca2+ influx through TRPV6 channels by altering their inactivation behavior. These results suggest a model in which TRPV6-expressing cells can have a substantial Ca2+ influx, where phosphorylation of TRPV6 can act as a positive-feedback system, delaying the inactivation process. The described mechanism of competitive regulation of TRPV6 by PKC and CaM is, however, restricted to human TRPV6, since this particular PKC site in the CaM-binding motif is not conserved in the other species (276). Recently, the corresponding region in mouse TRPV6 was shown to bind CaM (167). Detailed analysis of the binding region predicted a caseine kinase motif, but no significant phosphorylation could be detected in this particular domain. The regulation of TRPV6 by CaM was recently confirmed by Lambers et al. (229). By combination of pull-down assays and coimmunoprecipitations, it was demonstrated that CaM binds to both TRPV5 and TRPV6 in a Ca2+-dependent fashion. The binding of CaM to mouse TRPV6 was localized to the transmembrane domain and consensus CaM-binding motifs located in the amino [1–5-10 motif, TRPV6-(88—97)] and carboxy termini [1–8-14 motif, TRPV6-(643—656)], suggesting a mechanism of regulation involving multiple interaction sites (229). Electrophysiological experiments demonstrated that HEK293 cells heterologously expressing TRPV5 or TRPV6 and CaM mutants revealed that TRPV6, but not TRPV5, is negatively regulated by an inactive CaM mutant. This finding is remarkable given the high homology between both channels, similar Ca2+-dependent regulation of channel activity, and binding of CaM to both channels. Furthermore, the effect of CaM was mediated by the high Ca2+ affinity EF-hand structures 3 and 4 present in the carboxy terminus of CaM (229).
2. S100A10-annexin 2
Recently, an auxiliary protein of TRPV5 and TRPV6 was identified by screening a mouse kidney cDNA library using the yeast two-hybrid system. A bait was constructed with the cytoplasmic carboxy-terminal region of TRPV5. This study described the identification of the first auxiliary protein for TRPV5 and TRPV6, named S100A10, which specifically associates with the carboxy termini of these epithelial Ca2+ channels (394). S100A10 is a 97-amino acid protein member of the S100 superfamily that is present in a large number of organisms including vertebrates, insects, nematodes, and plants. S100A10 is predominantly present as a heterotetrameric complex with annexin 2, which has been implicated in numerous biological processes including endocytosis, exocytosis, and membrane-cytoskeleton interactions (136). Annexin 2 interacts with actin and is postulated to bind to the cytoplasmic face of membrane rafts to stabilize these domains, thereby providing a link to the actin cytoskeleton. Several members of the S100 protein family form heteromeric complexes with annexins: S100A11 with annexin 1 (247), S100A6 with annexin 2 (380), and S100A10 is often found tightly associated with annexin 2 to form a tetrameric complex (326). Van de Graaf et al. (394) provided the first evidence of a regulatory role for the S100A10-annexin 2 heterotetramer in vitamin D-mediated Ca2+ (re)absorption in general and in particular in TRPV5 and TRPV6 functioning. The association of S100A10 with TRPV5 and TRPV6 was restricted to a short peptide sequence VATTV located in the carboxy termini of these channels. This stretch is conserved among all identified species of TRPV5 and TRPV6 (394). Interestingly, the TTV sequence in the S100A10 binding motif resembles an internal type I PDZ consensus binding sequence, which is S/TXV (360). However, S100A10 does not contain PDZ domains, indicating that the TRPV5-S100A10 interaction has a different nature. The first threonine of the S100A10 interaction motif was identified as a crucial determinant for binding. Furthermore, the activity of TRPV5 and TRPV6 was abolished when this particular threonine was mutated, demonstrating that this motif is essential for channel function. Malfunctioning of these mutant channels was accompanied by a major disturbance in their subcellular localization, indicating that the S100A10-annexin 2 heterotetramer facilitates the translocation of TRPV5 and TRPV6 channels to the plasma membrane. The importance of annexin 2 in this process was demonstrated by a siRNA-based downregulation of annexin 2 that significantly inhibited the currents through TRPV5 and TRPV6, indicating that annexin 2 in conjunction with S100A10 is crucial for TRPV5 activity. In line with the cortical localization of annexin 2 and its postulated function in organizing certain plasma membrane domains, these findings provided the first functional evidence for a regulatory role of annexin 2 controlling Ca2+ channel trafficking. Interestingly, previous studies indicated that the background K+ channel (TASK1) is associated with S100A10 via its carboxy-terminal sequence SSV (142). The S100A10 interaction promoted the translocation of TASK1 to the plasma membrane producing functional K+ channels. This sequence resembles the binding motif in TRPV5 and TRPV6 identified in the present study, suggesting a shared structural S100A10 binding pocket. However, this motif is absent in the tetrodotoxin-insensitive voltage-gated Na+ channel (Nav1.8), which has been shown to bind S100A10 via its amino terminus and essential for plasma membrane trafficking (296). Interestingly, in line with the epithelial Ca2+ channels, S100A10 expression was found to be vitamin D sensitive (394). In addition, annexin 2 expression levels have been shown to increase upon prolonged incubation with 1,25-(OH)2D3 (252). Coregulation of TRPV5/6, S100A10, and annexin 2 could fulfill the essential control of trafficking of these channels to the plasma membrane. 1,25-(OH)2D3 has been shown to exert its effects by slow and rapid mechanisms. The slower genomic effects are mediated by interaction with the nuclear VDR. Recently, it was reported that annexin 2 serves as a membrane receptor for 1,25-(OH)2D3 and mediates the rapid effect of the hormone on intracellular Ca2+. It has been demonstrated that 1,25-(OH)2D3 is specifically bound to annexin 2 in the plasma membrane of rat osteoblast-like cells (18, 19). Partially purified plasma membrane proteins and purified annexin 2 exhibited specific and saturable binding for 1,25-[3H](OH)2D3. The results suggest that annexin 2 may serve as a receptor for rapid actions of 1,25-(OH)2D3. However, there is still a debate whether there is functional interaction between 1,25-(OH)2D3 and annexin 2 (253a). Taken together, these findings show that the S100A10-annexin 2 complex is a significant component for the trafficking of ion channels to the plasma membrane in general and in particular a major regulator of TRPV5 and TRPV6 function and, therefore, the Ca2+ homeostasis.
By the use of cDNA microarrays, Gkika et al. (145) identified 80K-H as a protein involved in the Ca2+-dependent control of TRPV5 (145). 80K-H was initially identified as a PKC substrate, but its biological function remains to be established (166). This recent study demonstrated a specific interaction between 80K-H and TRPV5, colocalization of both proteins in the kidney, and similar transcriptional regulation by 1,25-(OH)2D3 and dietary Ca2+ (145). Furthermore, 80K-H directly bound Ca2+, and inactivation of its two EF-hand structures totally abolished Ca2+ binding. Electrophysiological studies using 80K-H mutants showed that three domains of 80K-H (the two EF-hand structures, the highly acidic glutamic stretch, and the His-Asp-Glu-Leu sequence) are critical determinants for TRPV5 activity. Importantly, inactivation of the EF-hand pair reduced the TRPV5-mediated Ca2+ current and increased the TRPV5 sensitivity to intracellular Ca2+, accelerating the feedback inhibition of the channel. None of the 80K-H mutants altered the TRPV5 plasma membrane localization nor the association of 80K-H with TRPV5, suggesting that 80K-H has a direct effect on TRPV5 activity (145). Taken together, 80K-H acts as a novel Ca2+ sensor controlling TRPV5 channel activity.
Diuretics such as furosemide and thiazides are frequently used in the clinical practice and are known to alter Ca2+ metabolism. All of them are usually administered alone or in a combination of diuretics over a long period of time and are known to disturb Ca2+ reabsorption leading to various symptoms including alterations in structure and stability of bone (422). However, despite the widespread use of these diuretics, the molecular mechanism underlying their action on Ca2+ reabsorption remains incompletely understood.
Loop diuretics, such as furosemide, inhibit the Na+-K+-2Cl− (NKCC2) transporter present in the apical membrane of the TALH resulting in a reduction in NaCl reabsorption and K+ recycling across the apical membrane. This action diminishes the lumen-positive potential, which is the driving force for paracellular Ca2+ reabsorption in this particular nephron segment and explains the hypercalciuric effect of furosemide (123). The calciuric effect of furosemide enhances the delivery of Ca2+ to DCT and CNT, which are the primary sites of active Ca2+ reabsorption. It is at present unknown whether these latter nephron segments partly compensate the hypercalciuric effect of the loop diuretics. In this respect, it is interesting to study the effect of furosemide on the expression and activity of the Ca2+ transport proteins. This knowledge could also provide a rationale for TRPV5 or calbindin-D28K activating treatment during furosemide application to reduce the side effect caused by the hypercalciuric action of furosemide. Until now, data about the molecular regulation of these proteins during furosemide treatment were not available.
Thiazides are the most widely prescribed drugs today, particularly by being the mainstay of first-line therapy in hypertension. In addition, these diuretics have, in contrast to loop diuretics, the unique characteristic of decreasing Na+ reabsorption while increasing Ca2+ reabsorption (79). Their hypocalciuric effect provides therapeutic opportunities in, for instance, idiopathic hypercalciuria and nephrolithiasis. Furthermore, thiazides have been shown to increase bone mineral density and decrease fracture risk, spiking interest in the favorable long-term effects of these diuretics in counteracting osteoporosis (321). However, the exact molecular mechanism responsible for this thiazide-induced hypocalciuria has been a subject of discussion.
Thiazides increase renal Na+ excretion by inhibiting the Na+-Cl− cotransporter (NCC) present in the apical membrane of DCT cells (255). This inhibition of Na+ reabsorption results in increased renal salt and water loss and thereby decreases extracellular volume (255). Costanzo and Windhager (79) showed in pioneering micropuncture and microperfusion experiments that chlorothiazide can stimulate Ca2+ transport in DCT in situ. Their data are restricted to the acute effects of these diuretics on electrolyte transport and are difficult to translate to chronic thiazide treatment. Several theories, emphasizing an activation of transcellular Ca2+ transport processes in DCT, were subsequently advanced to explain the hypocalciuria during thiazide administration. Stimulation of apical Ca2+ entry has been suggested, mediated by hyperpolarization of the plasma membrane secondary to NCC inhibition (138). Alternatively, enhanced basolateral Na+/Ca2+ exchange by upregulation of NCX1 secondary to decreased intracellular Na+ concentration has also been suggested as the principal responsible mechanism (101, 121, 123). However, the significant downregulation of NCX1 transcripts during thiazide treatment shown in this latter study is hard to reconcile with the proposed stimulation of Na+/Ca2+ exchange activity. Furthermore, calbindin-D28K mRNA and protein abundance were also decreased during thiazide treatment (278). Calbindin-D28K facilitates diffusion of Ca2+ through the cytosol and simultaneously serves as an intracellular Ca2+ buffer to protect the cell from toxic Ca2+ levels (107). This substantial decrease of the Ca2+ diffusion and buffering capacity would be detrimental in the presence of increased apical Ca2+ entry and increased transcellular Ca2+ transport. The above-described hypotheses rely on substantial colocalization in DCT of NCC and the proteins involved in active Ca2+ transport. Extensive immunohistochemical studies demonstrated only minor overlap, whereas the Ca2+ transporters (i.e., TRPV5, calbindin-D28K, NCX1, and PMCA1b) completely colocalized (171, 240, 278). Interestingly, Loffing et al. (239) demonstrated that the DCT epithelium had lost the structural characteristics of electrolyte transporting epithelia after chronic thiazide treatment and the cells were in different stages of apoptosis. In apoptotic cells, calbindin-D28K and PMCA1b were strongly decreased, and the NCC protein was shifted from the luminal membrane to the basal membrane and was found additionally in small membrane vesicles in intercellular and peritubular spaces. The consistently decreased expression of the Ca2+ transporters during thiazide treatment strongly argues against stimulation of active Ca2+ transport processes in chronic thiazide treatment. Furthermore, transcripts of NCC were drastically reduced in homogenates of kidney cortex and almost absent in damaged DCT cells (278). All other tubular segments were unaffected by the treatment.
On the other hand, extracellular volume contraction leading to increased proximal Na+ reabsorption and thereby increasing the electrochemical gradient driving passive Ca2+ transport in proximal tubule segments has been suggested as a possible additional effect, further decreasing overall Ca2+ excretion. In addition to inhibition of NCC by thiazides, mutations in the gene encoding NCC have been shown to cause Gitelman’s syndrome (354). These patients suffer from hypovolemia, hypokalemic alkalosis, hypomagnesemia, and hypocalciuria (254). The mechanism causing the hypocalciuria in Gitelman’s syndrome has not been elucidated, but in general, similar hypotheses have been postulated as for the thiazide action (101). Early studies by Weinman and Eknoyan (424) already demonstrated that the escape from the chronic effects of chlorothiazide is due to a decrease in the glomerular filtration rate and to an increase in fractional reabsorption in the proximal tubule. Nijenhuis et al. (278) showed that the HCTZ-induced hypocalciuria was accompanied by a significant decrease in body weight compared with controls, illustrating that extracellular volume contraction occurred. Since Na+ depletion resulted in a similar hypocalciuria, it is likely that the extracellular fluid volume (ECV) contraction by itself is responsible for the thiazide-induced hypocalciuria. This is further supported by the finding that Na+ repletion during HCTZ treatment, thereby preventing the ECV contraction, normalized the calciuresis. In literature, volume contraction has merely been suggested as a contributing factor in thiazide-induced hypocalciuria (121, 123). The underlying mechanism is an enhancement of proximal Na+ reabsorption observed in volume contraction by diuretics and Na+ restriction (110, 319, 416), which leads to an increase in the electrochemical driving force for passive Ca2+ reabsorption (57). Indeed, thiazides are known to decrease lithium clearance, which is generally accepted as an inverse estimation of proximal Na+ reabsorption (215).
In conclusion, chronic thiazide treatment exerts two major effects. First, thiazides induce a hypovolemia that stimulates the proximal electrolyte reabsorption and explains the hypocalciuria. Second, thiazide exposure leads to structural degeneration of DCT resulting in downregulation of the particular ion transporters. These data inquire revision of the generally postulated hypotheses on hypocalciuria in chronic thiazide treatment and Gitelman’s disease, indicating a critical role of ECV contraction and the resulting enhancement of passive Ca2+ reabsorption.
Immunosuppressants like the calcineurin inhibitors tacrolimus (FK506) and cyclosporin A, next to glucocorticoids, such as dexamethasone, are widely prescribed drugs in various disorders and for organ transplant recipients. Although their immunosuppressive actions are accomplished by distinct mechanisms, FK506 and dexamethasone are both known to induce significant side effects on mineral homeostasis. These drugs are associated with an increased bone turnover, a negative Ca2+ balance, and hypercalciuria, perturbations that can ultimately result in osteoporosis (323, 332, 367). Furthermore, hypomagnesemia is a well-known additional consequence of FK506 treatment (6, 251). The kidney is essential to both Ca2+ and Mg2+ homeostasis by providing the main excretory route for these divalent ions. However, the exact mechanisms by which these immunosuppressants provoke renal divalent wasting are unknown. The hypercalciuria during treatment with these drugs has been attributed to increased bone resorption as well as decreased renal Ca2+ reabsorption (242, 323, 324). Until now, the cascade of cellular and molecular events leading to impaired renal Ca2+ handling during FK506 and glucocorticoid treatment is largely unknown. In theory, downregulation of Ca2+ transport proteins in the distal part of the nephron may be involved in the pathogenesis of hypercalciuria and hypermagnesuria during drug treatment. Previous reports showed reduced calbindin-D28K levels during FK506 treatment, suggesting that FK506 could affect active Ca2+ transport (2). Recently, it was demonstrated that FK506 treatment significantly increased urinary Ca2+ excretion, accompanied by a downregulation of the renal mRNA expression of TRPV5 and calbindin-D28K, and a specific reduction of the protein abundance of these Ca2+ transport proteins in DCT and CNT (277). The fact that serum Ca2+ concentrations and glomerular filtration rate did not differ from controls confirmed that impaired Ca2+ reabsorption rather than an increased filtered load caused the hypercalciuria. In addition, the expression of the major intestinal Ca2+ transport proteins was not increased by FK506 treatment, excluding that hypercalciuria is secondary to upregulation of active Ca2+ absorption. These data strongly supported the hypothesis that FK506 induces a primary defect of renal active Ca2+ reabsorption by specifically downregulating the proteins involved in active Ca2+ transport.
The molecular mechanism underlying the downregulation of the Ca2+ transport proteins by FK506 remains elusive. In previous studies, plasma 1,25-(OH)2D3 was either unaltered or moderately increased, while plasma PTH was not affected by similar doses of FK506, which excludes that the reduced Ca2+ transport protein expression levels are secondary to decreased circulating levels of these calciotropic hormones (2, 241). Interestingly, several groups demonstrated that the immunosuppressive action of FK506 depends on the inhibition of the Ca2+-dependent phosphatase calcineurin in T-lymfocyten (131, 238, 295). Calcineurin is not known to be involved in renal Ca2+ reabsorption, but the calcineurin inhibitor cyclosporin A increased urinary Ca2+ excretion and decreased calbindin-D28K protein levels, suggesting that calcineurin inhibition may play a role in the impairment of Ca2+ reabsorption by these drugs (2, 366). In addition, FK506 binds to intracellular immunophilins called FK506-binding proteins (FKBPs), which have been implicated as ion channel regulators (157, 338, 352). In particular, FKBP4 was shown to bind and regulate the Ca2+-permeable Drosophila TRPL channel, and this binding was disrupted by the addition of FK506 (150). Site-directed mutagenesis showed that mutations of P702Q or P709Q in the highly conserved TRPL sequence “701LPPPFNVLP709” eliminated interaction of the TRPL with the FK506-binding protein of Drosophila (dFKBP59). Detailed sequence alignment of TRPV5/6 species indicated that this domain is not conserved in the TRPV subfamily. Furthermore, several intracellular Ca2+ release channels were shown to be modulated by binding of FKBPs (250). Therefore, it is tempting to speculate that FKBPs are potential associated proteins regulating epithelial Ca2+ channel expression or activity.
VII. CHARACTERIZATION OF EPITHELIAL CALCIUM CHANNEL KNOCKOUT MICE
A. TRPV5 Knockout Mice
The characterization of TRPV5 and TRPV6 knockout mice should reveal the diseases that are associated with epithelial Ca2+ channel dysfunction. Recently, Hoenderop et al. (180) generated TRPV5 null (TRPV5−/−) mice by genetic ablation of TRPV5 to investigate the requirement of TRPV5 functioning in renal and intestinal Ca2+ (re)absorption. Interestingly, metabolic studies demonstrated that TRPV5−/− mice exhibit a robust calciuresis, since significantly more Ca2+ was excreted in the urine compared with wild-type (TRPV5+/+) littermates. The urinary Ca2+ concentration of the knockout mice reached values of 20 mM compared with 6 mM for TRPV5+/+ littermates. Serum analysis showed that TRPV5−/− mice have normal plasma Ca2+ concentrations, but significantly elevated 1,25-(OH)2D3 levels compared with TRPV5+/− and TRPV5+/+ mice. To pinpoint the defective site of the Ca2+ reabsorption along the nephron in vivo, micropuncture studies were performed in these transgenic mice that combine classical and new research tools in a way that promises to yield important new insights into single-nephron function. Quantitative free-flow collections of tubular fluid revealed unaffected Ca2+ reabsorption in TRPV5−/− mice up to the last surface loop of the late proximal tubule (LPT) (Fig. 13A). In contrast, mean Ca2+ delivery to puncturing sites within distal convolution (DC; DCT and CNT), was significantly enhanced in TRPV5−/− mice. Because K+ secretion occurs along the distal nephron sites accessible to micropuncture (together with water reabsorption in CNT and CCD), the distal luminal potassium concentration was used as an indicator of the distal collection site (Fig. 13B). Based on the shape of the relationship between distal luminal K+ concentration and fractional Ca2+ delivery, it is evident that in contrast to TRPV5+/+ mice, fractional Ca2+ delivery increases with a higher K+ concentration, indicating a defect in Ca2+ reabsorption along the DCT and CNT which is consistent with the localization of TRPV5.
Interestingly, polyuria and polydipsia were consistently observed in TRPV5−/− mice compared with TRPV5+/− and TRPV5+/+ littermates. Polyuria facilitates the excretion of large quantities of Ca2+ by reducing the potential risk of Ca2+ precipitations. It is known that a high luminal Ca2+ concentration activates the CaSR in the apical membrane of the IMCD. As a consequence, the arginine vasopressin-elicited water permeability is blunted by a PKC-mediated aquaporin-2 retrieval mechanism to reduce water reabsorption and prevent a further rise in urinary Ca2+ concentration and possibly stone formation (335). The hypercalciuria-induced polyuria has been observed in humans (253) and animal models (119, 317). Furthermore, TRPV5−/− mice produced urine that was significantly more acidic compared with TRPV5+/+ and TRPV5+/− littermates. Acidification of the urine is also known to prevent renal stone formation in hypercalciuria, since Ca2+ precipitates will not be formed at pH 5–6 (24).
In general, Ca2+ hyperabsorption by the small intestine is favored as compensation for renal Ca2+ wasting. Ca2+ absorption was assessed in preliminary experiments by measuring serum 45Ca2+ at early time points after oral gavage. A significant increase in the rate of 45Ca2+ absorption was observed in TRPV5−/− mice compared with wild-type littermates, indicating a compensatory role of the small intestine (Fig. 13C). TRPV6 and calbindin-D9K expression levels were significantly upregulated in TRPV5−/− mice consistent with this increased Ca2+ absorption (Fig. 13E) (180).
Surprisingly, inactivation of the TRPV5 gene was accompanied by a decrease in the renal calbindin-D28K and NCX1 mRNA expression (Fig. 13D). Because transcellular Ca2+ reabsorption in DCT and CNT was abolished in TRPV5−/− mice, the simultaneous decrease in calbindin-D28K and NCX1 mRNA levels, in the presence of elevated 1,25-(OH)2D3 levels, suggests a regulatory mechanism primarily controlled by TRPV5. This means that TRPV5 or the Ca2+ influx through TRPV5 possibly controls the transcription of the other Ca2+ transport genes including calbindin-D28K and NCX1. Although downregulation of renal calbindin-D28K is secondary to TRPV5 ablation in the TRPV5−/− mice, the reduced calbindin-D28K level may further augment the severity of the hypercalciuria. In this respect, it is interesting to compare the hypercalciuria in calbindin-D28K and TRPV5 knockout mice. Calbindin-D28K knockout mice fed a regular Ca2+ diet displayed an approximately twofold increase in the urinary Ca2+ excretion compared with wild-type littermates, whereas TRPV5−/− mice fed the same Ca2+ diet excreted approximately six times more Ca2+ in the urine compared with their littermates (180, 361). These findings underscore the gatekeeper function of TRPV5 in the process of Ca2+ reabsorption.
Furthermore, microcomputed tomography analyses of the femur demonstrated that trabecular thickness in the femoral head of TRPV5−/− mice was significantly reduced compared with TRPV5+/+ mice (Fig. 14). Trabecular bone volume, tissue volume, and bone fraction were not different between the genotypes (180). This could not be explained by a difference in trabecular number. Alternatively, it is possible that the trabeculae are longer, i.e., protrude further into the bone marrow cavity and thereby compensate for reduced trabecular thickness. Analyses of the diaphysis showed that cortical bone volume, cortical volume fraction, and cortical bone thickness were reduced in TRPV5−/− versus TRPV5+/+ mice. These initial data from TRPV5−/− mice demonstrated that TRPV5 is the gatekeeper in active Ca2+ reabsorption. Ablation of the TRPV5 gene seriously disturbs renal Ca2+ handling, causing increased 1,25-(OH)2D3 plasma levels, Ca2+ hyperabsorption, and reduced bone formation. All of these deficiencies have been reported frequently in patients with idiopathic hypercalciuria, although the molecular basis for this disorder remains unknown.
B. TRPV6 Knockout Mice
Hediger and co-workers (34) addressed the functional role of TRPV6 in Ca2+ absorption by inactivation of the mouse TRPV6 gene. These TRPV6 null (TRPV6−/−) mice were placed on a Ca2+-deficient diet and subsequently challenged in a 45Ca2+ absorption assay. TRPV6−/− mice showed a consistent decrease in Ca2+ absorption over time. From these initial data it was concluded that TRPV6−/− mice show a significant Ca2+ malabsorption, suggesting that TRPV6 is indeed the rate-limiting step in 1,25-OH2D3-dependent Ca2+ absorption. However, future studies are needed to address the functions of TRPV6 in detail.
Ca2+ play a fundamental role in many cellular processes, and its extracellular concentration is kept under strict control to allow proper physiological functions. The transepithelial Ca2+ (re)absorption determines the influx and efflux of Ca2+ to the extracellular Ca2+ pool and is controlled by several hormones. This review has focused on the identification, function, and regulation of the Ca2+ transport proteins. Over the last years significant advances were achieved in the Ca2+ homeostasis field. The picture that emerges from the recent data pointed to a cellular model for transepithelial Ca2+ transport in which distinct players execute an essential role. The identification of the new epithelial Ca2+ channels, TRPV5 and TRPV6, gave insights in a new molecular concept of Ca2+ influx. Unique physiological functions important for body Ca2+ homeostasis have been attributed to these gatekeeper channels, and many key questions in the study of transepithelial Ca2+ transport can now be addressed. What is the three-dimensional structure of these Ca2+ channels? Valuable information can be deduced from a structural models of TRPV5/6 channels. Future work should develop computational approaches to construct and refine three-dimensional models of these gatekeeper channels by incorporating and integrating available and new experimental data. How do intracellular Ca2+ control epithelial Ca2+ channel activity to allow sufficient Ca2+ uptake and subsequent transcellular Ca2+ movement? How is this Ca2+ movement organized in the absorbing epithelial cell, and which signaling pathways are involved? What is the underlying mechanism of the stimulatory effect of the calciotropic hormones? In this respect, it will be important to identify regulatory domains in the epithelial Ca2+ channels, binding partners of TRPV5/6 that participate in the channel complex and the signaling pathways controlling channel activity. Furthermore, what is the function of TRPV5/6 in unexplored tissues as testis, brain, exocrine tissues, stomach, and bone? Tissue-specific knockout mice models will help us to address these key functions of the Ca2+-transporting proteins. We have seen that ablation of the TRPV5 gene seriously disturbs renal Ca2+ handling resulting in compensatory intestinal hyperabsorption and bone abnormalities. These deficiencies in Ca2+ handling have been reported frequently in patients with idiopathic hypercalciuria, although the molecular basis is unknown and genetic defects in TRPV5, and/or TRPV6 remain to be identified. It is clear that despite considerable advances, much remains to be learned, and the field of Ca2+ homeostasis will remain active for many more years.
We thank the members of our laboratories for valuable discussion, advice, and critical reading of this manuscript.
This work was supported by Dutch Organization of Scientific Research Grants Zon-Mw 016.006.001, Zon-Mw 902.18.298, NWO-ALW 810.38.004, NWO-ALW 805–09.042, NWO-ALW 814–02.001, and NWO 812–08.002; Stomach Liver Intestine Foundation Grant MWO 03–19; Human Frontiers Science Program Grant RGP32/2004; Dutch Kidney Foundation Grants C10.1881 and C03.6017; the Belgian Federal Government; the Flemish Government; Onderzoeksraad Katholieke Universiteit Leuven (Leuven, Belgium) Grants GOA 99/07, FWO G.0237.95, FWO G.0214.99, FWO G.0136.00, and FWO G.0172.03; and the Belgian Prime Minister’s Office Interuniversity Poles of Attraction Program.
Address for reprint requests and other correspondence: R. J. M. Bindels, 160 Cell Physiology, University Medical Center Nijmegen, PO Box 9101, NL-6500 HB Nijmegen, The Netherlands (E-mail:)
- Copyright © 2005 the American Physiological Society