The fine tuning of Ca2+ excretion in the kidney takes place in the distal nephron, which consists of the distal convoluted tubule, connecting tubule, and initial portion of the cortical collecting duct. In these segments, Ca2+ is reabsorbed through an active transcellular pathway. The apical influx of Ca2+ into the distal renal cell is presumably the rate-limiting step in this process, and its molecular identity has remained obscure so far. The recently discovered epithelial Ca2+ channel (ECaC) exhibits the expected properties for being the gatekeeper in transcellular Ca2+ reabsorption. The characteristics and potential physiological role of ECaC will be discussed in this review. Our knowledge of the mechanisms involved in the regulation of transcellular Ca2+ transport has advanced rapidly since the development of cell models originating from distal tubular cells. Studies using these models indicate that hormones including arginine vasopressin, PGE2, adenosine, ATP, and atrial natriuretic peptide should be considered as calciotropic hormones controlling renal Ca2+ handling. Evidence is now beginning to emerge that the stimulating calciotropic hormones utilize new cAMP-independent pathways to stimulate Ca2+ reabsorption. These new findings allow the development of a comprehensive and detailed model of the process of transcellular calcium transport in the kidney whereby the individual contribution of the participating transporters can now be fully appreciated.
- calcium channel
- parathyroid hormone
- vitamin D3
- atrial natriuretic peptide
- arginine vasopressin
- guanosine 3′5′-cyclic monophosphate
- adenosine 3′5′-cyclic monophosphate
the kidney is critically important for the maintenance of overall Ca2+ homeostasis in mammals by determining the excretion of Ca2+ from the body and by conversion of vitamin D to its active metabolite, 1,25-dihydroxyvitamin D3 [1,25(OH)2D3] (52). To maintain a net Ca2+ balance, >98% of the Ca2+ filtered at the glomerulus must be reabsorbed along the nephron (21, 52). The major part of the Ca2+reabsorption takes place along the proximal tubule and thick ascending limb of Henle's loop through a paracellular, and therefore passive, pathway. The remaining Ca2+ reabsorption occurs in the distal part of the nephron and amounts to 15% of the filtered load of Ca2+. This section consists of distal convoluted tubules (DCT), connecting tubules (CNT), and the initial portion of the cortical collecting duct (CCD). The relative contribution of these individual segments to active Ca2+ reabsorption appears to differ among the various species (4). 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+ ions, this substantiates that Ca2+ is reabsorbed in these segments through an active transcellular pathway (21).
Transcellular Ca2+ transport is generally envisaged as a three-step process consisting of passive entry of Ca2+across the apical membrane, cytosolic diffusion of Ca2+bound to vitamin D3-sensitive calcium-binding proteins (calbindin-D28K), and active extrusion of Ca2+across the opposite basolateral membrane (Fig.1) (4). From an energetic perspective it is attractive to consider the apical influx of Ca2+ the rate-limiting step in this process and, therefore, presumably the final regulatory target for stimulatory and inhibitory hormones (21, 44, 66). The molecular nature of this apical influx mechanism has remained obscure so far, but the recently cloned epithelial Ca2+channel (ECaC) exhibits the defining characteristics of transepithelial Ca2+ transport (28). The potential role of ECaC in Ca2+ transporting distal nephron cells will be discussed in this review.
The rate of active Ca2+ transport is tightly controlled by the calciotropic hormones, and classical representatives of this group are 1,25(OH)2D3, parathyroid hormone (PTH), and calcitonin (4, 21, 52). 1,25(OH)2D3 stimulates Ca2+ reabsorption in a genomic fashion analogous to classic steroid hormones, whereas PTH and calcitonin have been postulated to act in a cAMP-dependent manner (4, 21, 42, 58-60). Recent studies have implicated other hormones as potential regulators of active Ca2+ reabsorption in renal tubular cells (25-27, 32,58, 59). These investigations also provided evidence for new signaling pathways that are independent of cAMP. On the basis of these emerging new concepts on the regulation of transepithelial Ca2+transport, we will expand the molecular model for hormone-regulated active Ca2+ reabsorption.
Molecular Identity of the Apical Ca2+ Entry Channel
In the past, the mechanism responsible for the apical entry of Ca2+ has been extensively characterized in DCT and CNT, but these studies have not led to the unambiguous identification of this Ca2+ transporter. From the available evidence it is likely that several distinct Ca2+ channels, which resemble the classified voltage-dependent Ca2+ channels in part, are present in the apical membrane of the distal nephron (19, 21, 42, 53,54, 68). However, the physiological significance of these channels with respect to Ca2+ reabsorption remains unclear, because it was previously shown that apical administration of a variety of Ca2+ channel antagonists failed to affect Ca2+reabsorption (5).
To determine the molecular structure of this apical Ca2+channel, several experimental strategies have been employed. The classic approach of purification of a transporter followed by amino acid sequencing of the isolated peptide has been hampered by the lack of a rich source of channel protein. In another cloning strategy, a calbindin-D28K-affinity column has been used to purify associated proteins potentially involved in transepithelial Ca2+ reabsorption (29). Alternatively, investigators have focused on homology-based cloning strategies by using sequences of previously described voltage-operated Ca2+ channels (3, 61,67). So far these attempts did not result in the identification of the apical Ca2+ entry transporter. As a new initiative, functional expression cloning in Xenopus laevis oocytes was applied by our group. This technique has previously been successful for the cloning of several transporters (12). Our rationale to use the latter technique was threefold: 1) mRNA isolated from primary cultures of rabbit kidney CNT and CCD, when injected in oocytes, induced a 45Ca2+ uptake two to three times above background (43); 2) transcellular Ca2+transport in primary cultures of rabbit CNT and CCD is 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 oocytes.
On the basis of the above-mentioned characteristics, an expression cloning strategy was established to identify the apical Ca2+ entry channel. A cDNA library from polyA+RNA isolated from primary cultures of rabbit kidney CNT and CCD was generated and screened for Ca2+ uptake activity in oocytes in the presence of a cocktail of Ca2+ channel antagonists, including nifedipine, verapamil, and Ba2+ (28). After an extensive screening procedure we were able to isolate a single transcript encoding for a novel epithelial Ca2+ channel, named ECaC in analogy to the epithelial Na+ channel (ENaC) (12). ECaC consists of six transmembrane-spanning domains and a short hydrophobic stretch predicted as the pore-forming region (Fig.2, A and B). Northern blot analysis revealed that ECaC is exclusively expressed in 1,25(OH)2D3-responsive epithelia, namely, kidney, small intestine, and placenta. Immunocytochemical analysis in kidney sections revealed that ECaC is localized in the apical domain of distal tubular segments (28). Colocalization studies showed that ECaC is expressed only in calbindin-D28K-positive segments, which in the rabbit mainly consist of CNT. These findings underline the putative function of ECaC as the apical Ca2+ entry mechanism mediating transepithelial Ca2+ reabsorption.
The primary structure of ECaC does not resemble that of the already cloned Ca2+ channels, such as voltage-gated, ligand-operated, and transient receptor potential (Trp) Ca2+ channels (40, 71, 72). Detailed analysis of the pore-forming region and the flanking transmembrane segment 6showed that there is low but significant homology with the two recently cloned capsaicin receptors (13, 14) and the family of Trp channels (72). The highest homology (30%) was found with the capsaicin receptors, which have been implicated in the sensory pathway of pain transmission (57). The predicted topology of ECaC is similar to that described for the capsaicin receptors, the α-subunit of the voltage-gated Ca2+ channels, the Trp channels, and the Shaker K+ channel (9, 34, 40, 66). These latter two channels have been shown to form homotetramers to constitute a functional ion channel. On the basis of these findings we speculate that ECaC is also a tetrameric ion channel (Fig. 2 C). Analysis of the phylogenetic relationship between ECaC and previously described voltage-gated Ca2+ channels underlines that there is a significant evolutionary distance with these channels, indicating that ECaC belongs to another family (Fig. 2 D). Of the Trp-related sequences reported, those that resemble ECaC most closely are VR1, VRL, and OSM-9, but the homology to these channels is low (<30%), indicating that ECaC may form a distant subgroup within this growing family of proteins.
ECaC contains potential regulatory sites including putative phosphorylation sites for protein kinase C (PKC), cAMP-dependent protein kinase (cAK) and cGMP-dependent protein kinase (cGK), and structural domains such as N-linked glycosylation sites and ankyrin repeats (Fig. 2 A). These ankyrin repeats are present in many membrane proteins including Trp channels and capsaicin receptors and interact with the cytoskeleton to assemble and stabilize proteins in the plasma membrane (56, 70). Generally, the functions of the ankyrin repeats are compatible with a role in protein-protein interactions (10).
The S4 segment of voltage-gated Ca2+ channels, which constitutes the voltage sensor that promotes channel opening in response to depolarization, contains four to six positively charged amino acid residues (35, 66). The corresponding segment in ECaC lacks the positively charged lysine and arginine residues, which is consistent with our finding that ECaC is insensitive to depolarization (26). Similarly, Ba2+ and Sr2+, which are highly permeant in voltage-gated Ca2+ channels (22), do not interfere with ECaC function. It is interesting to highlight two additional features of ECaC functioning. First of all, Mn2+, a potent inhibitor of Ca2+ entry in ECaC-expressing oocytes, competes with Ca2+ for apical entry into the renal cell over a wide range of transepithelial Ca2+ transport rates (44). In addition, the previously shown inhibition of the apical Ca2+ influx pathway by external H+ ions could be mimicked in ECaC-expressing oocytes (5).
Taken together, the above-described pharmacological and functional properties of ECaC are largely identical to those described for transcellular Ca2+ transport across monolayers of primary cultures of rabbit CNT and CCD (28). Thus ECaC exhibits the defining characteristics for being the apical Ca2+ transporter in transepithelial Ca2+ reabsorption.
Cytosolic Ca2+ Transport and Signaling: An Apparent Contradiction
Renal cells involved in transcellular Ca2+ transport are continuously challenged by substantial Ca2+ traffic through the cytosol, while simultaneously maintaining low levels of cytosolic Ca2+. Calbindin-D28K, which is a 1,25(OH)2D3-dependent calcium-binding protein that is present in the cytosol of cells lining the distal part of the nephron (8, 33, 60), could fulfill this paradoxical function. Theoretical considerations and experimental evidence suggest that calbindin-D28K facilitates the cytosolic diffusion of Ca2+ from the apical influx to the basolateral efflux sites and acts as a cytosolic Ca2+ buffer to maintain low cytosolic Ca2+ levels during changes in transcellular Ca2+ transport (4, 18). Because of the slow binding kinetics of calbindin-D28K, Ca2+signaling can occur independently of transcellular Ca2+movement mediated by calbindin-D28K (31). Depending on the vitamin D3 state, the cytosolic calbindin-D28Kconcentration reaches values in the submillimolar range, which is indeed sufficient to fulfill the above-mentioned functions (6, 17). Recently, a homozygous mutant of calbindin-D28Kgene-knockout mice was generated by gene targeting that developed normally without apparent signs of disturbed Ca2+homeostasis (1). These animals exhibited a three to four times higher urinary Ca2+ excretion compared with wild-type mice, but their plasma Ca2+ levels remained normal (30). This suggests that the hypercalciuria induced by calbindin-D28K deficiency is easily compensated for by, for instance, increased intestinal absorption of Ca2+.
Calcium Extrusion Mechanisms
The efflux of Ca2+ occurs against a considerable electrochemical gradient, and two Ca2+ transporters are present in the basolateral membrane to extrude Ca2+: a Na+/Ca2+ exchange mechanism (NCX) and a Ca2+-ATPase (PMCA) (4, 21, 60).
To date, three genes for NCX, designated NCX1, NCX2, and NCX3, have been identified in mammals (41), but in kidney only NCX1 isoforms were reported (36, 19, 63). The expression of NCX is restricted to the distal nephron, where it is predominantly localized along the basolateral membrane (45, 60). It has been shown that the operation of this exchanger is a prerequisite for transcellular Ca2+transport (7), which suggests that Ca2+ extrusion across the basolateral membrane is mediated primarily by NCX. Unfortunately, specific inhibitors of NCX are not available to substantiate the relative importance of this exchanger for overall Ca2+reabsorption.
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 (4, 37, 60). PCR analysis showed the expression of all four presently known isoforms in the kidney (37). On the basis of the fact that PMCA1 and PMCA4 are widespread, whereas 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 (51). 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 (60). Furthermore, PMCA1b is the predominant isoform expressed in rabbit small intestine, a site also involved in active Ca2+ transport. These data suggest that PMCA1 isoforms take part in the process of transcellular Ca2+ transport.
CALCIOTROPIC HORMONES: AN EVER-EXPANDING FAMILY?
The use of primary cultures and immortalized cell lines originating from distal tubule cells greatly facilitated the search for hormones other than 1,25(OH)2D3, PTH, and calcitonin that may play a role in the regulation of active Ca2+reabsorption and provided exciting new insights into the mechanisms of action of calciotropic hormones (6, 21, 42). At present, hormones identified as being potentially calciotropic include arginine vasopressin (AVP), PGE2, adenosine, ATP, nitric oxide (NO), and atrial natriuretic peptide (ANP) (Fig.3).
AVP is the key regulator of water reabsorption in the distal nephron. However, studies employing primary cultures of immunodissected rabbit CNT and CCD cells show that AVP can also stimulate Ca2+transport after interaction with basolateral V2 receptors (59). Importantly, recent investigations provide evidence that AVP may indeed have a Ca2+-sparing effect in vivo (46). Regarding the physiological relevance of these findings, it is tempting to speculate that increased Ca2+ reabsorption during AVP-stimulated water reabsorption may provide a means of decreasing the risk of urolithiasis.
Presently, three potential pathways for auto- or paracrine regulation of active Ca2+ reabsorption have been identified. The first pathway involves the production and release of a broad spectrum of prostanoids, of which only PGE2 has a dual effect on Ca2+ transport (58). On the one hand, PGE2interacts with apical and basolateral EP2 and/or EP4 receptors to stimulate Ca2+ transport, whereas on the other hand it interacts with basolateral EP3receptors to inhibit the stimulatory action of other calciotropic hormones. PGE2 is the predominant autacoid in the CCD where, thus far, it is mainly implicated as a negative regulator of AVP-stimulated water and Na+ transport (11). A second pathway involves the release of purine nucleosides and nucleotides. Of these, adenosine increases transcellular Ca2+ transport to the same extent as do PTH, AVP, PGE2, and ANP, and subsequent pharmacological characterization revealed the involvement of apical A1 receptors (25). Until now, effects of adenosine have only been reported in amphibian kidney A6 cells, where an A1 receptor agonist was shown to increase Na+transport (23). Conversely, ATP inhibits the action of stimulatory calciotropic hormones, and this effect is mediated via both apical and basolateral P2y receptors (32). A similar inhibitory effect of ATP has been observed on AVP-stimulated water transport in perfused tubules of rabbit CCD (47). These findings suggest that extracellular ATP may exert a general inhibitory effect on ion and water transport in the distal part of the nephron. Finally, a third auto- or paracrine pathway involves the local production of NO, leading to stimulation of active Ca2+ reabsorption (27). It is well known that NO activates guanylyl cyclase, suggesting the involvement of cGMP. At present, it is not established under which conditions the kidney produces the para- or autocrine hormones and whether they are of physiological relevance for active Ca2+ reabsorption.
It has been suggested that ANP exerts its natriuretic effect, at least in part, via receptors expressed in the collecting duct (38). These observations have recently been extended by Hoenderop and co-workers (27), who discovered that ANP and cGMP itself induce a Ca2+influx in freshly immunodissected CNT and CCD, which was further substantiated in primary cultures of these segments. ANP and NO stimulation of Ca2+ transport in these cultures was specifically mediated by membrane-targeted cGK II. This detailed molecular characterization was made possible by (re)introducing different cGK gene products and mutants in cells that had become deficient in cGK II on culturing.
The glycoprotein stanniocalcin, although not yet examined for its action on Ca2+ reabsorption, is a promising candidate for the function of calciotropic hormone. The hormone was originally identified in bony fish, where it regulates Ca2+ and phosphate homeostasis (62). Recently, the human and mouse homologs of stanniocalcin have been characterized (15, 16, 64) and demonstrated to be abundantly present in many segments of the nephron, including DCT and collecting duct (64). Thus far, only a stimulatory action on phosphate reabsorption has been demonstrated (39).
Signaling Pathways in Active Ca2+ Reabsorption: The Question of the Role of cAMP
With the exception of the slow-acting 1,25(OH)2D3, which acts primarily at the genomic level, all other stimulatory hormones are fast acting and, in all but a few cases (NO and ANP), their effect on Ca2+transport was accompanied by a marked increase in adenylyl cyclase activity. Together with the finding that membrane-permeable analogs of cAMP readily increased Ca2+ reabsorption, this is in agreement with the classic concept that cAMP mediates the stimulatory action of the fast-acting calciotropic hormones. However, recent observations seriously questioned such a role of cAMP. The adenosine A1 receptor agonistN 6-cyclopentyladenosine readily increased Ca2+ transport at concentrations at which it did not affect the rate of cAMP accumulation measured in the presence of an inhibitor of cyclic nucleotide phosphodiesterase activity (25). Conversely, the A2 agonist CGS-21680 readily stimulated the accumulation of cAMP but only marginally affected Ca2+ transport. Furthermore, determination of the effect of the calciotropic hormones on the actual cellular cAMP content rather than the rate of cAMP accumulation showed no significant increase after stimulation with adenosine or PGE2 (26). Moreover, although PTH and AVP did increase the cellular cAMP content, abolishment of this increase by the potent adenylyl cyclase inhibitor 2′,5′-dideoxyadenosine did not interfere with their capability to stimulate Ca2+transport. On the other hand, the potent PKC inhibitor chelerythrine markedly inhibited the stimulatory effect of all hormones tested and, because downregulation of phorbol ester-sensitive PKC isoforms did not interfere with the stimulatory action of these hormones, it was concluded that they act via a novel pathway involving a phorbol ester-insensitive PKC isoform. Intriguingly, experiments performed with amphibian kidney A6 cells have led to a similar conclusion with respect to the signaling pathway involved in adenosine- and AVP-stimulated Na+ reabsorption (23, 24). A major role of PKC in PTH-stimulated Ca2+ reabsorption was also deduced from experiments using primary cultures of mixed mouse DCT and cortical thick ascending limb cells (20). Similarly, phospolipase C activation has been implicated in calcitonin stimulation of Ca2+uptake (42). Although the experiments performed with primary cultures of rabbit CCD and CNT exclude a role for cAMP in the hormonal regulation of Ca2+ transport, addition of membrane-permeable cAMP analogs, activators of adenlyl cyclase, or inhibitors of cyclic nucleotide phosphodiesterase did have a stimulatory effect (27, 59). At present, the only explanation is that these drugs, in contrast to the hormones, evoke a generalized rather than a localized increase in cAMP and thus reach compartments or molecules that are involved in transcellular Ca2+transport. The question on the physiological relevance of the PTH- and AVP-induced increase in cellular cAMP content remains. An attractive possibility is that cAMP may mediate genomic effects of these hormones. Indeed, a recent study shows that cAMP potentiates the 1,25(OH)2D3-induced increase in 25-hydroxyvitamin D3-24-hydroxylase mRNA content (65).
Hormones that control transcellular Ca2+ transport may act at the genomic level to change the rate of transcription of a transporter molecule or at the level of the transport protein itself to alter its activity or to promote its insertion or removal from the plasma membrane. How a calciotropic hormone exerts its regulatory effect can be deduced from its speed of action. Thus the relatively slow onset of the action of 1,25(OH)2D3 can be taken as evidence for an effect at the genomic level, whereas, conversely, the virtually instantaneous action of PTH and the other stimulatory hormones is in agreement with an effect on an already existing Ca2+ transport protein. Indeed, 1,25(OH)2D3 has been demonstrated to increase the mRNA levels of both calbindin-D28K (60) and the Na+/Ca2+ exchanger (36), suggesting that the transcription of these transport proteins is upregulated. The slow-acting nature of 1,25(OH)2D3 was revealed by the fact that a significant increase in calbindin-D28Kprotein was observed only after 2 days of incubation in the presence of the steroid. Other studies indicate that the steroid may also influence Ca2+ reabsorption by altering the expression levels of the PTH receptor (50) and 25-hydroxyvitamin D3-24-hydroxylase (65). The steroid did not noticeably alter the mRNA and protein content of the plasma membrane Ca2+-ATPase in primary cultures of rabbit CNT and CCD (60). Finally, future studies will show whether the steroid has an effect on the level of expression of ECaC.
In contrast to 1,25(OH)2D3, PTH and the other stimulatory hormones including ANP and the NO liberator sodium nitroprusside (SNP) virtually instantaneously increase the rate of Ca2+ transport (27). But whereas PTH, AVP, PGE2, and adenosine readily increased Ca2+transport, the presence of cGK II was required for the stimulatory action of ANP and SNP. This unequivocally demonstrates that ANP and SNP act solely via cGMP, whereas, as discussed above, PTH and the other stimulatory hormones act via a phorbol ester-insensitive PKC isoform. At present, the exact target of these two kinases is unknown. It is, however, reasonable to believe that the regulation of transcellular Ca2+ transport occurs most efficiently at the level of the Ca2+ influx pathway. Recent elucidation of the primary structure of the potential Ca2+ influx channel, ECaC, reveals the presence of several PKC and cGK/cAK consensus sites (28). Future studies will show whether the phosphorylation state of ECaC is indeed affected by stimulatory and possibly also inhibitory calciotropic hormones.
POSSIBLE IMPLICATIONS FOR HUMAN DISEASES
Familial disturbances in renal handling of Ca2+ have been frequently reported (2). For some disorders the underlying pathophysiological mechanism has been defined at the molecular level. For instance, the identification of the extracellular calcium-sensing receptor (CaSR) and subsequent mutation analysis have explained familial hypocalciuric hypercalcemia by inactivating mutations whereas activating mutations in this receptor result in hypocalcaemia with hypercalciuria (2, 55). The molecular basis of other inherited Ca2+ disorders, such as idiopathic hypercalciuria, remains to be elucidated. This latter disorder is usually associated with primary Ca2+ hyperabsorption or primary renal tubule leak of Ca2+ (2). In the past, the vitamin D receptor and 1α-hydroxylase were excluded as candidate genes associated with idiopathic hypercalciuria (49, 69). It will be particularly important to analyze the human ECaC locus for linkage with this disorder. Gain of function mutations in ECaC may well be the cause of absorptive hypercalciuria, whereas loss of function mutations could result in renal hypercalciuria. Furthermore, Gitelman's syndrome, an inherited form of hypokalemic alkalosis, is accompanied by hypocalciuria possibly resulting from an indirect activation of the apical Ca2+channel (48). This disorder results primarily from the loss of function of the thiazide-sensitive Na-Cl cotransporter. The proposed explanation for the diminished Ca2+ excretion is that impairment of apical NaCl entry hyperpolarizes the cell and lowers the intracellular Na+ concentration, stimulating the entry of Ca2+ through the apical Ca2+ channel and facilitating the exchange of Na+ for Ca2+across the basolateral membrane, respectively (19, 48). The recent elucidation of ECaC makes it now possible to study the above-mentioned possibilities with molecular genetic and physiological approaches.
There are many questions that need to be addressed. For example, what is the physiological and clinical relevance of ECaC for the overall Ca2+ homeostasis? Which regulatory proteins do interact with ECaC and control its function? Can drugs be developed that directly regulate the activity of ECaC to effectively prevent hypercalciuria or age-related malabsorption of Ca2+? What is the physiological relevance of the newly discovered calciotropic hormones? Future investigations, particularly performed at the whole-animal level, will certainly address these questions adequately and will advance our understanding of Ca2+ homeostasis in health and disease.
The authors gratefully acknowledge the valuable help of Dr. Ole Madsen with the phylogeny analysis.
Address for reprint requests and other correspondence: R. J. M. Bindels, 162 Cell Physiology, Univ. of Nijmegen, PO Box 9101, NL-6500 HB Nijmegen, The Netherlands (E-mail:).
This work was supported in part by a grant from the Dutch Organization of Scientific Research (NWO-ALW 805–09.042).
- Copyright © 2000 the American Physiological Society