3. Transcellular Calcium Transport in the Distal Tubule: The Role of Calbindin-D28k

Transcription

1 3. Transcellular Calcium Transport in the Distal Tubule: The Role of Calbindin-D28k 3.1. Physiology of calcium reabsorption in the nephron The following short review places the emphasis on the transcellular calcium transport in the distal tubule. Paracellular and transcellular calcium transport The major load of filtered calcium is being reabsorbed by passive, paracellular pathways in the proximal tubule. This passive calcium transport is proceeded as an interplay with the countercurrent Na π /Cl ª movement. By this instance, the driving force of the calcium flux will be the transepithelial voltage, the magnitude of which will be set by the rate of sodium absorption. This voltage is oriented positive in the lumen, and therefore, when the sodium reabsorption increases, the transepithelial voltage also increases and thereby increases paracellular calcium reabsorption (Bourdeau & Burg 1979, Friedman & Gesek 1995). The transcellular calcium reabsorption in renal epithelial cells is a three-step process involving entry of the calcium ion across the luminal membrane, transport through the cytoplasm, and finally extrusion across the basolateral membrane. This calcium transport is dependent upon three crucial fractors: The differences in calcium concentrations intra- and extracellularly, the electrical gradient across the cell membranes, and the permeability of the cell membranes for calcium. The intracellular calcium concentration is in the range of 10,000-fold lower than extracellular calcium concentration and as the cell interior is electrically negative with respect to extracellular fluid, this combined electrochemical gradient for calcium favors entry into and opposes efflux from the cell. However, as the calcium ion is highly positive charged and as the cell membrane consists of a non-polar lipid bilayer, passive diffusion across the membrane is prevented (Friedman & Gesek 1995). Hence, the influx of calcium into cells requires the participation of a facilitating protein or other mechanism. Luminal membrane calcium channels have been described in the DCT (Bacskai & Friedman 1990) and at present, two different groups of channels are known: One is activated by camp, is voltage insensitive and has a conductance of ps (Poncet et al 1992). The other is activated by PTH, thiazides and calcitonin, is voltage sensitive and has a conductance of 3 to 10 ps. Little is presently known about the molecular identity of calcium channels that mediate apical entry in tubular epithelial cells (Friedman & Gesek 1995). The precise location of calcium channels in the nephron and the role they play in overall renal tubule cell regulation of calcium await further investigation (Suki & Rouse 1996). The next step in the transcellular calcium transport is the flux of calcium through the cytoplasm of the distal tubular cells. The only described mechanism for this calcium flux is the binding of calcium to calbindin-d28k (Johnson & Kumar 1994, Friedman & Gesek 1995). This will be discussed in section 3.2 and 3.3 concerning localization and functional considerations of renal calbindin-d28k. The final extrusion of calcium across the basolateral membrane and into the interstitial space takes place against a large electrochemical potential difference, and it must therefore be active and energy demanding. Its regulation is essential for the maintenance of cytosolic calcium concentrations in the submicromolar range. Two possible mechanisms have been identified: The high affinity plasma membrane, calmodulin stimulated, calcium-dependent adenosine triphosphatase (Ca ππ -ATPase) generally referred to as PMCA, and the Na π /Ca ππ exchanger. At least five major PMCA isoforms exist, which exhibit approximately 85% homology, whereas the same isoforms, cloned from different species, exhibit up to 99% homology (Doucet & Katz 1982, Suki & Rouse 1996). PMCA activity is present in all nephron segments, and indeed in many other tissues not involved in active transcellular calcium transport. PMCA1 isoforms have been found throughout the kidney and has been designated a housekeeping pump function. The PMCA1b and PMCA2 isoforms only has been detected in the cortical thick ascending limbs and in DCT and are presumed to be involved in basolateral calcium extrusion (Borke et al. 1989, Van Baal et al. 1996). However, the functional role of any of these isoforms in cellular calcium transport remains to be determined. The other mechanism for calcium extrusion is the Na π / Ca ππ exchanger, which has been described in several tissues including the kidney, where it initially was thought to be present in both proximal tubule (Taylor & Windhager 1979), DCT (Ramachandran & Brunette 1989), connecting tubule (Shimuzu et al. 1990), and collecting tubule (Taniguchi et al. 1989). By using Na π /Ca ππ exchanger nucleotid sequence obtained from cardiac sarcolemma, the corresponding mrna to this enzyme was examined in the kidney demonstrating the existance of a similar mrna in both cortex and in much lower concentrations in medulla (Reilly & Shugrue 1992). Parallel studies by the use of polymerase chain reaction revealed the localization of the Na π / Ca ππ exchanger m-rna in DCT and only low or no concentrations in the other nephron segments (Yu et al. 1992). Immunoblotting by the use of monoclonal antibodies generated against the isolated canine cardiac sarcolemmal Na π / Ca ππ exchanger reported the renal Na π /Ca ππ exchanger to be exclusively located in the connecting tubules (Bourdeau et al. 1993, Reilly et al. 1993). More recently, three

2 REGULATION OF RENAL CALBINDIN-D28k 11 isoforms of the Na π /Ca ππ -exchanger were isolated using homology-based reverse transcription-polymerase chain reaction with RNA from a DCT cell line (White et al. 1996). The molecular structure of the renal Na π /Ca ππ exchanger has not been identified and its ion binding sites have so far not been determined, and at the present state, further studies are necessary to localize and characterize the Na π / Ca ππ exchanger in the kidney (Suki & Rouse 1996, Friedman & Gesek 1995). Calcium transport along different nephron segments Proximal tubule. Of the total amount of calcium in plasma only 60 65% is filtered at the glomerulus, as the remaining is bound to plasma proteins, especially albumin. The movement of ionized calcium and low molecular-weight calcium complexes is not restrained by the filtration membrane of the corpusculum, the calcium concentration in the glomerular filtrate and ultrafiltrates of plasma being equal (Friedman & Gesek 1995, Suki & Rouse 1996). More than 70% of the filtered calcium load is reabsorbed in the proximal tubule principally by paracellular transport consistent with a passive, voltage dependent Ca ππ flux. In this part of the nephron, a lumen-positive transepithelial potential difference in combination with a luminal calcium concentration above the plasma calcium concentration is regarded as being sufficient to drive the calcium reabsorption (Suki & Rouse 1996). The recent description of calcium channels in the luminal membranes of the proximal tubule opens the possibility of a potential active, transcellular calcium transport in this segment (Zhang & O Neil 1996 aπb). This active calcium flux probably takes place in the most proximal and most distal segments of the proximal tubule, where the electrochemical driving forces are unfavorable for calcium reabsorption (Suki & Rouse 1996). Loop of Henle. In the loop of Henle, approximately 20% of the filtered calcium is reabsorbed, and as the permeability for calcium in both thin descending and thin ascending limb is minimal, the calcium movement is believed to take place exclusively in the ascending thick limb. In this part of the nephron, both a passive, paracellular pathway and an active, hormone-regulated transcellular transport contribute to the net calcium reabsorption. In the thick ascending loop of Henle, hormonal and phamacological affection of sodium transport results in proportional changes of paracellular calcium transport: The stimulation of sodium reabsorption by antidiuretic hormone or glucagon increases calcium reabsorption, whereas inhibition of sodium reabsorption by loop diuretics impedes the passive calcium reabsorption and causes increased urinary calcium excretion (Friedman 1988, De Rouffignac et al. 1991). Several studies have described an increased transcellular calcium reabsorption both in vivo and in vitro after stimulation with PTH and with calcitonin in this part of the nephron, presumably by mechanisms similar to those of the distal tubule (Bourdeau & Burg 1980, Di Stefano et al 1990). Distal convoluted tubule (DCT). In the DCT approximately 10% of the glomerular filtered calcium is reabsorbed by active hormone regulated transcellular transport. As more than 85% of the calcium has already been reabsorbed in the preceding segments, up to 90% of the calcium load delivered to the distal tubule may be reabsorbed here (Constanzo et al aπb). The DCT measures approximately 1 mm in length in both the rabbit and in the rat and it consists mainly of one cell type, the DCT cell, though in its terminal portion a second cell type appears, the intercalated cell (IC cell) which is more typical in the collecting duct system. Electronmicrographically, the DCT cell presents with extensive microprojections of both luminal surface and basal membrane, many vesicles in the apical region of the cytoplasma, numerous elongate mitochondria, and short microvilli on the luminal surface (Dørup 1985, Madsen & Tisher 1986, Kriz & Bankir 1988). Passive paracellular calcium transport in the DCT is minimal because of the high transepithelial resistance and low calcium permeability and further because the luminal calcium concentration in the distal nephron is below that of plasma. Calcium ions in the DCT are therefore actively reabsorbed against both chemical concentration and electrical potential differences (Constanzo et al aπb). The mechanism of entry at the luminal membrane has been examined in vesicle preparations and found to be inhibited by Na π, stimulated by K π and an alkaline ph, but unaffected by changes in potential (Brunette et al. 1992). In this nephron segment the urinary calcium excretion is finely regulated by the effect of the calciotropic hormones PTH, calcitonin and 1,25-(OH) 2 vitamin D on the transcellular calcium reabsorption (Friedman & Gesek 1995). Calbindin-D28k is found in very high concentrations in the cytosol of the DCT cells as will be discussed in section 3.2. The connecting tubule (CNT). The morphological appearance of the transition from the distal tubule to the collecting duct varies considerably between species; however in both rat and human kidney, the CNT shows a gradual transition due to intermingling of cells from DCT and collecting duct together with the specific CNT cells and the IC cells. The CNT cells have fewer microprojections and fewer vesicles than the cells of the DCT, whereas IC cells exhibit a darker cytoplasm with numerous tubulovesical membrane structures in the apical region. IC cells are always separated from one another by at least one CNT cell (Dørup 1985, Madsen & Tisher 1986, Kriz & Bankir 1988). This segment may perform a net Ca ππ -efflux of approximately 2 pmol/min/mm which is 10-fold higher than that found in neighbouring nephron segments (Morel et al 1976, Imai 1981). The mechanism of calcium transport is like in DCT principally transcellular and mainly regulated by PTH consistent with the presence of PTH-sensitive adenylate cyc-

3 12 CLAUS HEMMINGSEN lase in both segments, but also thiazides exert their action in this segment (Morel et al. 1976). Calbindin-D28k is present in the DCT cells and to a lesser extent in the CNT cells of this segment and its functional role will be discussed in section 3.2. The collecting duct (CD). The main cell type of this segment is the CD cell or principal cell which is characterized by a light cytoplasm with only few mitochondria and vesicles. The apical plasma membrane is relatively smooth with few microvilli and a single centrally placed cilium. The basal plasma membrane have infoldings, whereas lateral interdigitations between adjacent cells are lacking. The IC cell constitutes approximately 40% of the cells in the cortical area of the collecting duct in the rat (Madsen & Tisher 1986). The transepithelial voltage in CD is lumen-negative and the luminal fluid calcium concentration is below that of the peritubular fluid. Therefore, an electrical potential difference favors net calcium secretion via paracellular pathways. As the permeability for calcium is low in this segment, loss of calcium in this segment does not take place under normal conditions (Bourdeau & Hellstrom-Stein 1982). The cortical CD of the rat is believed to transport only small amounts of calcium and mainly by paracellular pathway, but a recent study performed in rabbits described a transcellular calcium transport comparable of that in DCT involving calbindin-d28k, the Na π /Ca ππ exchanger and Ca ππ -ATPase (Van Baal et al. 1996). In the medullary CD, a reabsorption of 3 7% of the total filtered load of calcium has been implied by indirect assessment and is believed to be active and transcellular, though details of this transport are not known. A functional relationship of calcium transport to the structural differences in the distal nephron is not well documented (Friedman & Gesek 1995, Suki & Rouse 1996). Calbindin-D28k is expressed in low concentrations in the CD cells (See section 3.2). Regulation of renal calcium handling in the distal tubule. PTH. Urinary calcium excretion is in part regulated by the action of parathyroid hormone (PTH) on calcium reabsorption in the distal nephron and connecting tubules. The mechanism by which PTH increases calcium reabsorption is incompletely defined (Lau & Bourdeau 1995). PTH stimulates calcium uptake through the apical plasma membrane (Bourdeau & Lau 1989), where binding of PTH to the basolateral membrane receptor induces activation of the dihydropyridine-sensitive Ca ππ -channels (Bacskai & Friedman 1990). This activation requires activation of both protein kinase A and protein kinase C (Friedman et al. 1996). At the basolateral membrane PTH enhance the activity of the Na π /Ca ππ -exchanger but not of the Ca ππ -ATPase (Bouhtiauy et al. 1991). The conductance of the Cl ª - channels in the basolateral membrane is increased by PTH stimulation, resulting in an increased Cl ª diffusion potential and increased intracellular hyperpolarization (Shimuzu et al. 1990, Gesek & Friedman 1992, Friedman & Gesek 1993). Normal levels of 1,25-(OH) 2 D are necessary for optimal expression of PTH-stimulated renal tubular calcium reabsorption (Fraher et al. 1992). Vitamin D. The research concerning the effects of vitamin D on the kidney presents a vast amount of conflicting data which are presented in major textbooks as controversial (Bourdeau & Attie 1994, Suki & Rouse 1996). Vitamin D is considered a pro-hormone and its biological activities are in a great majority due to its hormonally active derivative 1,25(OH) 2 D (Haussler et al. 1968, Holick et al. 1971, Cancela et al. 1988). Vitamin D depletion, both with and without intact parathyroid glands decreases calcium reabsorption, whereas administration of 1,25(OH) 2 D enhances calcium reabsorption (Ney et al. 1968, Constanzo et al. 1974). Examination of isolated membranes from proximal and distal tubules derived from vitamin D depleted rats suggested, that 1,25(OH) 2 D exerts its effect only in distal tubule, and that calcium transport is affected both at the luminal and at the basolateral membrane (Bouhtiay et al. 1993). In transformed mouse DCT cells it has been shown that 1,25(OH) 2 D modulates PTH-dependent calcium entry at a step between the PTH receptor and the activation of Cl ª channels, which involves message transcription (Friedman & Gesek 1993 (b)). Immunocytochemical analysis of rabbit connecting and cortical collecting tubules suggested that 1,25(OH) 2 D stimulation of transcellular calcium transport involves an increase in the gene expression of calbindin-d28k but not of the Na π /Ca ππ -exchanger and Ca ππ -ATPase (Van Baal et al. 1996). Calcitonin. Infusion of calcitonin lowers circulating calcium concentration by inhibition of bone resorption and by increasing the overall urinary calcium excretion. In the distal convoluted tubule of the kidney, however, pharmacological doses of calcitonin increases the active calcium reabsorption by a stimulation of the adenylate cyclase activity (Quamme 1980, Elalouf et al. 1983, Shimutzu et al. 1990) and by increasing membrane hyperpolarization (Gesek & Friedman 1993) which opens calcium channels. A recent study on purified luminal and basolateral DCT cell membranes (Zuo et al. 1997) concluded that calcitonin increases calcium transport in DCT through two mechanisms: the opening of low affinity calcium channels in the luminal membrane and the stimulation of the Na π /Ca ª exchanger in the basolateral membrane. Both actions depend on the activation of the adenylate cyclase. Calcitonin receptors have been described on osteoclasts and tubular cells, but not in the intestine, where administration of calcitonin has no effect on the calcium transport (Bourdeau & Attie 1994). A direct action of calcitonin on the stimulation of 1,25(OH) 2 D production was suggested by studies in thyroparathyroidectomized rats (Jaeger et al. 1986) and in vitro (Kawashima et al. 1981). Calcium. The way isolated changes of plasma calcium concentrations influence tubular calcium handling is difficult to

4 REGULATION OF RENAL CALBINDIN-D28k 13 study in vivo, as compensatory changes of the other parameters of the calcium metabolism will affect the results. Increased plasma concentrations of calcium increases the filtered load of calcium entering the nephron. This effect is opposed by a decline in GFR secondary to a decline in the ultrafiltration coefficient of the glomerular membrane dependent upon PTH (Humes et al. 1978, Suki & Rouse 1996). Hypocalcemia, when associated with hypoparathyroidism, therefore, may result in a decrease in plasma ultrafiltrable calcium, which results in a decline in the filtered calcium load. In the distal tubule, increased peritubular calcium concentrations decrease the reabsorption of calcium as shown by micropuncture studies (Sutton et al. 1983), but PTH independent effects of abnormal plasma calcium per se has not been examined. The application of in vivo microperfusion with increased calcium load to the distal tubule did show a load dependence, i.e. as the amount of calcium delivered to the distal tubule increases, its reabsorption increases proportionally (Constanzo et al (b)). Remarkably, no saturation of this mechanism was observed over a seven-fold range. The authors suggested that the increased calcium load would increase the passive calcium entry across apical plasma membranes causing increased intracellular calcium activity followed by proportionally elevated extrusion of calcium by the basolateral membrane. No bearing point on this assumption is presently available (Suki & Rouse 1996). The possible direct effect on the expression of calbindin-d28k of different calcium loads to the distal nephron has never been examined. Phosphorus. Systemic phosphorus loading increases calcium reabsorption in the distal tubule propably secondary to the stimulation of PTH secretion (Wong et al. 1985). Phosphorus depletion is associated with hypercalcuria, which is only partially corrected by administration of PTH, suggesting a direct effect of phosphorus depletion on renal tubule calcium transport in addition to the depression of circulating PTH (Grabie et al. 1978). Several studies have suggested the effect of low plasma phosphate to be located in the loop of Henle (Imai 1981, Guruprakash et al. 1985), however, the exact nephron location and mechanism of this effect of phosphate is still disputed. Magnesium. A close relation between calcium and magnesium metabolism has been known since 1909, when it was first shown in animals that infusion of calcium produced an increase in urinary excretion of magnesium and vice versa (Mendel & Benedict 1909) though the interplay between magnesium and calcium is generally thought to be of significance at extreme concentrations of magnesium only (Quamme 1993). The renal magnesium excretion is regulated by reabsorption in both the loop of Henle and distal tubule of the nephron. Of the ultrafiltrable magnesium, % is reabsorbed by the proximal tubule and % is reabsorbed in the thick ascending limb predominantly by the paracellular pathway. Pharmacological doses of various hormones including parathyroid hormone and calcitonin influence the magnesium reabsorption in these segments (De Rouffignac et al. 1993) though they have never been shown to physiologically control renal magnesium handling. In the distal tubule, parathyroid hormone has been demonstrated to increase magnesium reabsorption, while significance of other hormones remains to be established (De Rouffignac et al. 1993). In the rat, the distal tubule reabsorbs magnesium by 5 10% of the filtered load, i.e % the amount delivered to this segment and probably representing the fine adjustment of the plasma concentration (Quamme 1993). Profound hypermagnesemia competetively restrains the reabsorption of calcium in the distal tubules and vice versa. This suggests that a related saturable mechanism may exist for the renal reabsorption of magnesium and calcium (Quamme 1993) Localization of calbindin-d28k. In the kidney, calbindin-d28k is exclusively localized in the distal nephron, where it is found i abundance in DCT cells and to a lesser extent in CNT and CD cells (Fig. 1) (Roth et al. 1981, Taylor et al. 1982, Van Baal et al. 1996). The intercalated, mitochondria-rich cells in these tubular segments are negative for this calcium binding protein. Interestingly, calbindin-d28k, which is present in most epithelia that actively absorb calcium, is absent in the thick ascending loop of Henle (Bourdeau & Burg 1980, De Rouffignac et al. 1991), where transcellular calcium reabsorption takes place stimulated by PTH and calcitonin. Renal calbindin-d28k is an intracellular protein, which is evenly distributed through the cytosol and in the nucleus without anatomical association with membranes or filamentous elements (Roth et al. 1981, Bindels et al. 1991, Van Baal et al. 1996, Liu et al. 1996). The exact functional role of calbindin-d28 in the kidney has not been established (Christakos et al. 1997), but the co-localization of the protein with other effectors of the calcium transport along the nephron and within the renal epithelial cells has been examined in recent papers: The VDR is believed to be localized in all segments of the nephron as RT-PCR has revealed the expression of mrna for VDR i all segments (Iida et al. 1993, Liu et al. 1996). In the distal tubule, this co-localization with calbindin-d28k was found consistent with the hypothesis that 1,25(OH) 2 D acts via the VDR to stimulate calbindin-d28k synthesis, whereas the presence of VDR mrna in other parts of the nephron was interpreted as an indication for other genomically mediated actions of 1,25(OH) 2 D within the kidney in addition to stimulation of calbindin-d28k synthesis (Liu et al. 1996).. The calcium channels in the luminal membrane are believed to be exclusively located in the DCT. They are activated by camp or PTH and calcitonin, but their relation to calbindin-d28 await further investigation (Bacskai & Friedman 1990, Poncet et al. 1992, Friedman & Gesek 1995, Suki & Rouse 1996). Plasma membrane Ca ππ -ATP-

5 14 CLAUS HEMMINGSEN ases are present in all nephron segments, but the PMCA1b and PMCA2 isoforms have only been detected in CTAL and DCT and are presumed to be involved in basolateral calcium extrusion (Doucet & Katz 1982, Borke et al. 1989, Van Baal et al. 1996). The Na π /Ca ππ exchanger is primarily localized in DCT (Ramachandran & Brunette 1989, Yu et al. 1992) and the connecting tubules (Bourdeau et al. 1993, Reilly & Shugrue 1993) and in only low or no concentrations in the other segments of the nephron. The regulation of Ca ππ -ATPases and Na π /Ca ππ exchangers is, however, not vitamin D dependent (Van Baal et al. 1996) though any dependency upon calbindin-d28k remains to be established Functional considerations. Over the years, several theories on the functional role of renal calbindin-d28k have been proposed. In the following, the carrier theory, the buffer theory, the basolateral membrane calcium-pump activator theory, and the Ca ππ - Mg ππ -ATPase activator theory will briefly be discussed. The carrier theory is based upon the calbindin-based facilitated diffusion model of vitamin D-dependent movement of calcium through the cell which is also described for calbindin-d9k in the intestinal enterocyte (Bronner & Stein 1988). Based on a series of assumptions, Bronner and Stein calculated that the presence of calbindin-d28k in DCT raised the possible total calcium content by three orders of magnitude, thereby allowing the transcellular calcium movement to be more than 70 times greater than what could be predicted from the diffusion rate alone (Bronner & Stein 1988). Though these theory-based calculations remains to be verified, the general belief is that a major function of calbindin-d28k is to act as a calcium shuttle to transport calcium across the renal epithelia (Friedman & Gesek 1995, Suki & Rouse 1996, Christakos et al. 1997). The buffer theory is based upon the fact that intracellular calcium concentrations are 10 4 lower than extracellular calcium concentrations and calcium transporting cells therefore need buffer proteins to prevent intracellular Ca ππ concentrations from reaching toxic levels. Calbindin-D28k is present in abundance in a variety of cells involved in calcium transport and might therefore represent such a buffer protein. In the distal tubule, the concentration of calbindin- D28k is estimated to approximately 60 mm, whereas the cytosolic free Ca ππ concentration without calbindin-d28k is less than 0.6 mm (Bronner & Stein 1988, Johnson & Kumar 1994). The presence of calbindin-d28k effectively lowers the cytosolic free Ca ππ concentrations during Ca ππ transport and consequently more Ca ππ may enter the cell without causing a significant rise in the Ca ππ concentration (Johnson & Kumar 1994). At the basolateral membrane, Ca ππ becomes dissociated from calbindin-d28k and associates with plasma PMCA, which has a greater calcium-binding affinity than does calbindin-d28k. The basolateral membrane calcium-pump activator theory was described by Bouhtiauy et al. (1994), who showed that calbindin-d28k selectively increased calcium uptake into luminal cell membrane vesicles obtained from distal nephron cells and suggested that calbindin-d28k might be connected to calcium entry through the calciumchannels of the apical plasma membrane. The Ca ππ -Mg ππ -ATPase activator theory was presented on the assumption that calbindin-d28k co-localize with this calcium-pump in the kidney (Borke et al. 1989). The theory was based upon reports of a direct effect of rat kidney calbindin-d28k on human erythrocyte Ca ππ -Mg ππ -ATPase activity (Morgan et al. 1986, Borke et al. 1988,). The physiological significance of these results remains to be established. In conclusion, no data exist to define which theory on the functional role of the renal calbindin-d28k is true. In contrast to intestinal calbindin-d9k the 28k protein is expressed in measurable concentrations even without vitamin D stimulation. This may suggest a constitutive role of calbindin-d28k for the cell. Calbindin-D28k is a phylogenetic well-preserved protein found in a variety of tissues and may possess different functional capabilities. In fact, elements of all the described functional theories may be true.