- Email: [email protected]
γ-L-glutamyl-L-DOPA inhibits Na+-phosphate cotransport across renal brush border membranes and increases renal excretion of phosphate
Kidney International, Vol. 55 (1999), pp. 1832–1842
g-L-glutamyl-L-DOPA inhibits Na1-phosphate cotransport across renal brush border membranes and increases renal excretion of phosphate FREDERICO G.S. DE TOLEDO, MICHAEL A. THOMPSON, CHAD BOLLIGER, GERTRUDE M. TYCE, and THOMAS P. DOUSA Renal Pathophysiology Laboratory, Department of Physiology and Biophysics, Mayo Clinic and Foundation, Mayo Medical School, Rochester, Minnesota, USA
g-L-glutamyl-L-DOPA inhibits Na1-phosphate cotransport across renal brush border membranes and increases renal excretion of phosphate. Background. For treatment of phosphate (Pi) overload in various pathophysiological states, an agent that selectively increases renal Pi excretion would be of major value. Previously, we have shown that dopamine (DA) inhibits Na1-Pi cotransport in renal epithelia. However, the administration of DA or its immediate precursor L-DOPA increases DA in multiple tissues. Synthetic dipeptide g-L-glutamyl-L-DOPA (gludopa) can serve as an inactive precursor (pro-pro-drug) of DA. This study tested the hypothesis that, because of the unique colocalization of g-glutamyltransferase (g-GT), aromatic amino acid decarboxylase, Na1-Pi cotransporter, and Na1-L-DOPA cotransporter in brush border membrane (BBM) of proximal tubular cells, gludopa may elicit phosphaturia by action of DA generated within the kidney. Methods. Thyroparathyrectomized rats were given placebo, or gludopa, or gludopa 1 g-GT inhibitor acivicin. Urinary excretion of Pi, Ca21, Na1, K1, DA, cAMP, and cGMP was determined, and Na1-Pi cotransport was measured in BBM prepared from kidneys of rats at the end of the experiment. Results. The administration of gludopa resulted in: (a) an inhibition of Na1-Pi cotransport, but not cotransport of Na1proline and Na1-alanine in BBM; (b) an increase (1300%) of fractional excretion (FE) of Pi and a drop (235%) of plasma Pi, whereas the plasma levels and FEs of Ca21, Na1, and K1 were unchanged; (c) an increase in urinary excretion of cAMP, but not cGMP; (d) a 1000-fold increase of urinary excretion of DA, without a change in excretion of norepinephrine; and (e) an incubation of gludopa with BBM in vitro, which caused a release of L-DOPA, and the in vivo administration of acivicin, which blocked actions of gludopa to inhibit Na1-Pi cotransport and to increase urinary excretions of Pi and DA. Conclusions. We conclude that colocalization of enzymes of biotransformation, BBM transporters, and the autocrine/ Key words: dopamine, gludopa, phosphaturia, acivicin, brush border membrane, renal excretion. Received for publication August 4, 1998 and in revised form December 16, 1998 Accepted for publication December 16, 1998
1999 by the International Society of Nephrology
paracrine DA system in cells of proximal tubules constitutes a cellular basis for the potent and specific phosphaturic action of gludopa.
Regulation of renal excretion of inorganic phosphate (Pi) is a key process in maintaining body Pi homeostasis [1]. Most of the Pi that is ultrafiltered in glomeruli is reabsorbed in proximal tubules. Sodium-phosphate (Na1-Pi) cotransport across luminal brush border membranes (BBM) is a rate-limiting [2] and regulated [3] step in the translocation of Pi from lumen of proximal tubules to peritubular capillaries [1]. Diminished renal excretion of Pi in a variety of pathophysiological states, such as hypoparathyroidism, pseudohypoparathyroidism, tumoral calcinosis, and, most frequently, renal insufficiency of various origins, results in chronic accumulation and retention of Pi and in hyperphosphatemia [4–6]. A surfeit of Pi in renal insufficiency, in turn, leads to secondary hyperparathyroidism and results in renal osteodystrophy in its various manifestations [4, 5, 7]. In principle, Pi overload can be ameliorated by either dietary or pharmacotherapeutic maneuvers that reduce the quantum of Pi that is absorbed in the gastrointestinal tract [4, 5, 7], an increase of renal Pi excretion [8–11], or a combination of both. The decrease of Pi quantum absorbed from small intestine can be achieved by dietary restriction of the amount of ingested Pi, that is, feeding with low-Pi diet and/or with the use of “Pi binders,” cationic compounds that bind dietary Pi in intestinal lumen and prevent its absorption across the wall of the small intestine [4, 5, 7]. However, therapeutic measures to correct Pi overload, based on these principles, are still unsatisfactory. A low Pi diet is difficult to design and maintain; moreover, a diet with low Pi should be used with caution in children with renal insufficiency who are in a growth anabolic phase [7]. Treatment with both
1832
de Toledo et al: Gludopa inhibits Na1-Pi cotransport
1833
Fig. 1. Scheme of biotransformation and action of gludopa in proximal tubular cell, resulting in increased phosphate (Pi) excretion. Gludopa (g-glutamyl-L-DOPA) from glomerular ultrafiltrate—tubular fluid binds as a substrate to g-glutamyltransferase (1. g-GT) on the surface of brush border membrane (BBM). g-GT catalyzes release of L-DOPA, and L-DOPA is translocated into cell interior by the [Na1o] . [Na1i] gradient energized cotransporter (s) across BBM [27]. In cytoplasm, L-DOPA is converted by L-aromatic amino acid decarboxylase (2. L-AADC) to dopamine that exits to interstitial fluid and also into tubular lumen. Dopamine may act also on the adjacent cells. Dopamine binds to a dopamine receptor (DA1 subtype) that is coupled to adenylate cyclase (AC) and increases synthesis of cAMP; increased cAMP activates protein kinase A (PKA) and thereby inhibits Na1-Pi cotransporter (d) in BBM. This inhibition is the basis for decreased Pi reabsorption in proximal tubules and results in phosphaturia (discussed earlier in this article).
aluminum- and calcium-based Pi binders has serious side effects [4, 5, 7], and this treatment modality also has a low degree of patient compliance. Thus, in patients with mild or moderate renal insufficiency with partially preserved renal function or in patients with normal renal function who retain Pi, such as hypoparathyroidism or pseudohypopathyroidism [4], the availability of a novel pharmacotherapy that selectively increases the renal excretion of Pi would be of major benefit. To date, however, attempts to increase renal Pi excretion by pharmacological inhibition of Pi reabsorption in proximal tubules has met with little success [12]. The administration of parathyroid hormone (PTH), an archetypal hormone that elicits phosphaturia [2, 3], also primarily increases tubular reabsorption of Ca21, mobilizes bone Ca21, and can lead to hypercalcemia. Moreover, PTH should be given by the parenteral route. The administration of phosphonoformic acid (PFA), a specific competitive inhibitor of Na1-Pi cotransport in the luminal BBM of proximal tubules [8, 9], markedly increases urinary excretion of Pi in rats fed either normal- or low-Pi diet [8, 9], and also in rats with experimental renal insufficiency because of 5/6 nephrectomy [10, 11]. However, the relatively low affinity of PFA (Ki > 1 mm) for Na1-Pi cotransporter [8, 13] and concerns that have been raised about potential toxic effects of larger doses of PFA when administered to severely compromised patients [12, 14] currently precluded its further preclinical testing and use
in clinical practice [12, 14]. Thus, a suitable pharmacotherapy that would selectively increase the renal excretion of Pi is not available. Via stimulating dopamine (DA) receptors (DA1 subtype) and by activating the cAMP signaling pathway, dopamine elicits an inhibition of Na1-Pi cotransport in cultured renal epithelial opossum kidney (OK) cells [15, 16], an established in vitro model of Na1-Pi cotransport in renal proximal tubules [15, 16]. Intravenous infusion of DA to thyroparathyrectomized rats selectively decreased the Na1-Pi cotransport across renal BBM of proximal tubules [6] and resulted in an increased fractional excretion (FE) of Pi [17, 18]. Based on these observations, we hypothesized that DA generated intrarenally, via the inhibition of Na1-Pi cotransport across BBM, can decrease proximal tubular reabsorption of Pi and elicit phosphaturia [6, 15, 17, 18]. DA should be administered by intravenous (i.v.) infusion and thence has multiple effects in nonrenal organs, other than the increase of Pi excretion. To specifically modulate renal tubular function, DA should be generated in situ in renal parenchyma. Not only DA itself, but also its immediate precursor, 3,4-dihydroxyphenylalanine (L-DOPA), when added to OK cells in vitro inhibits Na1-Pi cotransport [15, 16], and in vivo administration of L-DOPA elicits phosphaturia [18]. Aromatic L-amino acid decarboxylase (AADC), a cytosolic enzyme [19] that catalyzes decarboxylation of L-DOPA to DA, is present exclusively in proximal tubules
1834
de Toledo et al: Gludopa inhibits Na1-Pi cotransport
Fig. 2. Plasma levels (A) and urinary excretions (B) of Pi, Ca21, and Na1 and K1 in control rats (C) and in rats injected with gludopa (GD). Plasma levels are in mm, and urinary excretion of electrolytes is expressed as a fractional excretion (% on ordinate) relative to urinary excretion and plasma level of creatinine (Methods section). Urinary excretion of Pi was significantly (P , 0.02) increased (D , 1300%), whereas plasma level of Pi was significantly (P , 0.05) decreased (D , 235%). Plasma levels as well as urinary excretions (FE) of Ca21, Na1, and K1 were not different between controls and GD-treated rats. Each bar denotes mean 6 sem (N 5 5 to 6).
[20], and it stands to reason that DA generated from L-DOPA locally in proximal tubules can act as an autocrine/paracrine agent primarily on proximal tubule cells. When administered parenterally or orally, L-DOPA is decarboxylated to DA not only in the kidney but also in numerous extrarenal tissues [21], and as a result, generated DA has many other effects, including hemodynamic effects, in many nonrenal tissues. Synthetic dipeptide g-L-glutamyl-L-dihydroxyphenylalanine (g-L-glutamyl-LDOPA, or gludopa) can serve as a pro-drug [22] of L-DOPA, and consequently of DA [21–24], in a manner that is sharply targeted to the kidney [21]. Gludopa is enzymatically converted to L-DOPA by g-glutamyltranspeptidase (g-GT), a very highly active enzyme present in the BBM of proximal tubules [25, 26], and the L-DOPA is transported across BBM [27] and then decarboxylated in proximal tubules by cytosolic AADC to DA [19, 20]. (g-GT catalyzes the enzymatic release of L-DOPA from gludopa either by transpeptidation, that is, glutamic acid forms a g-peptidic bond with another amino acid, if available, or hydrolyzes gludopa to L-DOPA and free glutamic acid [24]. Because it is not known whether one or both types of reaction occurs in vivo, we refer to action of g-GT to gludopa as the “release” or “generation” of L-DOPA from gludopa.) Thus,
colocalization of g-GT and AADC in proximal tubules, but not in other nephron segments [20, 26], should allow the conversion of L-DOPA to DA in proximal tubules by a two-step enzymatic reaction. Then in situ generated DA can modulate the functions of proximal tubular cells, including the rate of Na1-Pi cotransport (Fig. 1). It is well known that the administration of gludopa to humans [23, 24] and to experimental animals [28, 29] causes an increase in renal excretion of Na1 that is sometimes accompanied with increases of glomerular filtration rate, urine flow, and renal blood flow, but are without prohibitive side effects [23, 24, 30]. To date, the effects of gludopa on the renal excretion of Pi and homeostasis of Pi have not been considered or investigated. We experimentally tested the hypothesis that gludopa is converted to DA in proximal tubule cells, which inhibits Na1-Pi cotransport across BBM and elicits phosphaturia by the mechanism outlined in Figure 1. METHODS Surgically thyroparathyroidectomized male rats weighing 250 to 280 g were used for the in vivo experiments. Rats were kept in metabolic cages at least three days prior to the experiment while receiving free access to
de Toledo et al: Gludopa inhibits Na1-Pi cotransport
Fig. 3. Urinary excretions of cAMP and cGMP, dopamine, and norepinephrine in control rats (C) and in rats injected with gludopa (GD). Excretions of cAMP and cGMP are expressed as a percentage of change relative to the paired control, taken as 100%. Excretions of cAMP (P , 0.05) and of dopamine (P , 0.001) were (*) significantly increased (by t-test). Excretions of dopamine and norepinephrine are expressed in nmol/mg creatinine (ordinates). Each bar denotes mean 6 sem (N 5 7).
standard rat chow and drinking water with the addition of 0.35% Ca21 glutamate. As reported [31], gludopa injected intraperitoneally (i.p.) was rapidly absorbed (t1/2ab 5 6 min), and maximum plasma concentrations were maintained for approximately two hours postinjection, followed by a monophasic elimination decline [31]. Therefore, the maximum effect of gludopa can be anticipated within a three hour time span after i.p. injection. In our experiments, rats were injected i.p. with either gludopa (20 mg/kg body wt) or controls with glutamic acid (5.24 mg/kg body wt), both dissolved in 5% glucose. The dose of 20 mg/kg body wt gludopa was chosen based on known pharmacokinetics of gludopa in rat [31] and was 2.5 times lower than a dose of 50 mg/kg, which elicits modest natriuresis [32]. Rats were then kept in metabolic cages, and urine was collected for the three-hour postinjection period. At the end of three hours, rats were anesthetized by an i.m. injection of 0.1 mg/100 g body wt of a solution of equal volumes of 20 mg/ml xylazine and 100 mg/ml ketamine hydrochloride [33]. Blood was collected. Kidneys were removed, and the dissected renal cortical tissue was used for isolation of BBM [8, 13]. In experiments examining the effects of g-GT inhibitor acivicin in vivo, one group of rats was injected first with
1835
i.p. acivicin (10 mg/kg body wt) one hour prior to the injection of gludopa, and a second dose of acivicin (10 mg/kg body wt) was administered simultaneously with gludopa (20 mg/kg body wt). In the second group, rats received gludopa (20 mg/kg body wt) only, and in the third group, rats received only an i.p. injection of the vehicle (5% glucose). To allow the comparison between groups, experiments were conducted using a paired design [6, 9, 10]. That is, control rats and rats treated with test agents were taken into the experiment at the same date, were followed simultaneously, and were sacrificed at the same day. Also, preparation of BBM, transport and enzyme measurements, and other analyses were conducted in the same experimental session using identical reagents and procedures [6, 8–10, 17]. Brush border membrane vesicles were prepared with the Mg21-precipitation method, as described in our previous studies [8, 9, 13]. The fraction of basolateral membranes (BLM) was prepared by differential centrifugation and by the Percoll gradient centrifugation procedure [34]. Activities of leucine aminopeptidase (LAP) and g-GT in BBM and BLM were determined using chromogenic substrates, as described in our previous studies [8, 13, 25], and also Na1,K1-ATPase activity was determined as an oubain-sensitive ATPase, also described in our previous communications [8, 13]. Protein concentration in fractions was determined using Lowry’s methods [35]. The fraction of BBM showed significantly higher activities of g-GT and LAP, and significantly lower activity of Na1,K1-ATPase as compared with BLM (Fig. 7). The BLM had a significantly higher activity of Na1,K1ATPase and, conversely, rather lower activities of g-GT and LAP than BBM (Fig. 7). To determine the rate of generation of L-DOPA from gludopa in vitro, fractions of BBM and BLM prepared from normal rats were incubated with a substrate of 0.1 mm gludopa in a medium containing 40 mm Tris-HCl, 5 mm glycine-glycine, and 0.01% (wt/vol) sodium metabisulfite (pH 5 7.4). The incubation was terminated by the addition of ice-cold perchloric acid at a final concentration of 5%. The precipitate was removed by centrifugation at 3800 3 g, and the content of L-DOPA in the supernatant was determined by high-performance liquid chromatography (HPLC) with electrochemical detection, as previously described [16, 36]. The content of DA and norepinephrine in the urine was also determined by HPLC [36]. Transport studies in BBM vesicles The Na1-gradient–dependent uptakes of 32Pi, [3H]-Lalanine, and [3H]-L-proline by BBM vesicles were measured by the rapid filtration technique, as described in our previous communications [8, 9, 13, 25]. The initial concentrative “uphill” uptake was measured at 5 and 10 seconds, and uptake at the equilibrium period of 90 minutes [8, 13]. The Na1-independent uptakes of 32Pi
1836
de Toledo et al: Gludopa inhibits Na1-Pi cotransport
Fig. 4. Capacity for Na1-gradient–dependent Na1-Pi cotransport (A), Na1-[3H]-proline cotransport (B), and Na1-[3H]-alanine cotransport across BBM isolated from kidneys of control rats (C) or rats injected with gludopa (GD). All three transport systems were determined in aliquots of the same preparation of BBM vesicles and are expressed as an uptake of solute pmol/mg protein (ordinates) at indicated time periods. (A) Na1-dependent uptake of Pi was significantly (t-test) lower in GD-treated rats than in controls in the uphill concentrative phase, at 5 (–31%) and 10 seconds (–28%), but not at the 90-minute equilibrium point. In contrast, the capacity for Na1-dependent cotransports of [3H]proline (B) or of [3H]-alanine (C) was not different between controls and GD-treated rats. Each bar denotes mean 6 sem (N 5 4 to 7).
and of [3H]-L-proline and [3H]-L-alanine were negligible (less than 8% of the uptake in the presence of Na1 gradient), and therefore, the values without Na1 were not routinely measured and subtracted [9, 25]. Urine and plasma concentrations of Ca21, Na1, and K1 were determined by atomic absorption spectrophotometry. Pi in plasma and urine was measured by Chen’s colorimetric methods [37]. Creatinine was determined by the colorimetric method, and the content of cAMP and cGMP was determined by radioimmunoassay [38]. Gludopa was synthesized in the Mayo Clinic Protein Core Facility. L-DOPA, glutamic acid, DA, and acivicin were all purchased from Sigma (St. Louis, MO, USA). 32 Pi, [3H]-L-proline, [3H]-L-alanine, and radioimmunoassay kits for cAMP and cGMP were purchased from NEN (Boston, MA, USA). Other chemicals, all of the highest purity grades available, were purchased from Sigma. The results were evaluated statistically, unless stated other-
wise, with the use of Student’s t-test for group or pair comparisons, as appropriate. Values of P less than 0.05 were considered statistically significant.
RESULTS Intraperitoneal administration of gludopa resulted in at least a fourfold increase in FE of Pi (FEPi), but the FEs of Ca21, Na1, and K1 were not changed (Fig. 2). After three hours of gludopa injection, the plasma concentration of Pi significantly decreased (235%), but no significant changes were found in the plasma levels of Ca21, Na1, and K1 (Fig. 2). In response to gludopa, the urinary excretion of cAMP showed a significant increase (D > 180%), but the excretion of cGMP did not change significantly (Fig. 3). Finally, in response to gludopa, the urinary excretion of DA increased more than 1000-fold,
de Toledo et al: Gludopa inhibits Na1-Pi cotransport
Fig. 5. Kinetics of Na1-Pi cotransport inhibition in response to gludopa as evaluated by a double-reciprocal plot. The Vmax for Na1-Pi cotransport in BBM from (s) (221 6 49 nmol Pi/second/mg protein, N 5 3) was significantly (P , 0.05, paired t-test) lower than Vmax in BBM from corresponding controls (d; 288 6 38 pmol/second/mg protein; N 5 3). On the other hand, KmPi (242 6 34 mm Pi; N 5 3) did not significantly differ from KmPi for controls (232 6 26 mm Pi; N 5 3). Data are mean 6 sem.
whereas urinary excretion of norepinephrine did not change at all (Fig. 3). The Na1-gradient–dependent uptake of 32Pi by BBM vesicles prepared from renal cortex of gludopa-injected rats was markedly decreased (D > 230%) in the concentrative phase (Fig. 4). As measured in the same BBM membranes, the Na1-gradient–dependent uptake of [3H]-L-proline or [3H]-L-alanine was not altered in response to the administration of gludopa (Fig. 4). Neither uptake of [3H]-L-proline, [3H]-L-alanine, nor 32Pi at equilibrium period (90 min) was decreased (Fig. 4). The decrease in Na1-gradient–dependent uptake of 32Pi was due to a lower Vmax without a change in KmPi (Fig. 5). Because plasma for Pi determination was collected and kidneys for the preparation of BBM were harvested at the same time, the relationship of these two parameters was examined. In response to gludopa administration, the D% decrease in rate of Na1-Pi cotransport across BBM showed positive correlation with D% decrease of plasma Pi in the same rats (Fig. 6). When gludopa was added directly in vitro to isolated BBM vesicles, the Na1-dependent uptake of 32Pi was not changed (data not shown). Incubation of gludopa with BBM in vitro resulted in a time-dependent release of L-DOPA that was linear with time at least up to 10 minutes (Fig. 7). In contrast, the rate of L-DOPA release after incubation of gludopa with BLM was five times lower than with BBM (Fig. 7). The generation of L-DOPA from gludopa catalyzed both by
1837
Fig. 6. Correlation between the extent of decrease of Na1-Pi cotransport capacity of BBM and decrease of plasma levels of Pi in the same animals. Both parameters are expressed as -D% relative to paired controls that received solvent only, with the solid line representing the regression line, and dashed lines the 95% confidence limits. Ordinate, -D% of plasma [Pi]; abscissa, -D% of Na1-Pi cotransport, determined as 32Pi uptake by BBM at five-second intervals (Methods and Fig. 4). The correlation was significant: r 5 0.88 6 0.11, P , 0.05, t-test.
BBM and BLM was almost completely inhibited by acivicin [38], a specific irreversible inhibitor of g-GT (Fig. 7). Finally, we investigated whether the release of L-DOPA from gludopa by the action of g-GT was indeed an essential first step in the biotransformation pathway that produces phosphaturia (Fig. 1). The in vivo administration of acivicin [39] prior to and simultaneously with gludopa (discussed in the Methods section) completely blocked the decrease in Na1-gradient–dependent uptake of 32Pi across BBM, as well as abolished the increase in urinary excretion of Pi and increase of urinary excretion of DA (Fig. 8). The addition of acivicin in vitro to BBM vesicles had no effect on Na1-dependent 32Pi uptake by BBM (data not shown). DISCUSSION The experimental results presented herein report that gludopa administered in vivo acts as a selective phosphaturic agent and elucidate the mechanism of its action. Administration of gludopa decreased the capacity of Na1-gradient–dependent transport of 32Pi across renal BBM, but Na1-dependent cotransport of [3H]-L-proline or [3H]-L-alanine was not altered (Fig. 4). This finding is similar to the effect of infused DA on BBM [6], and suggests that administered gludopa did not cause a generalized inhibition of multiple Na1-gradient–dependent
1838
de Toledo et al: Gludopa inhibits Na1-Pi cotransport
Fig. 7. Activities of membrane-bound enzymes in fraction of brush border membranes (BBM) and fraction of basolateral membranes (BLM) prepared from rat kidney cortex, and generation of L-DOPA from gludopa by incubation with BBM and BLM. BBMs are enriched in g-glutamyltransferase (g-GT; A) and leucine aminopeptidase (LAP; B); BLMs are enriched in Na1, K1-ATPase (C). (D) Generation of L-DOPA from gludopa after incubation with either BLM or BBM fractions from rat kidney cortex. Gludopa (0.1 mm) was incubated with membrane fractions for 10 minutes, and L-DOPA and gludopa were determined by high-performance liquid chromatography (Methods section). The addition of acivicin, a g-glutamyltransferase inhibitor [38], suppressed generation of L-DOPA from gludopa both in BBM and BLM; see black portions at the bottom of the bars. The rate of enzymatic release of L-DOPA from gludopa was significantly (P , 0.01, t-test) higher in BBM (D 1394%) than in BLM. Each bar denotes mean 6 sem (N 5 3). (E) Under these experimental conditions, L-DOPA generation from gludopa was linearly proportional to a time of at least up to 10 minutes (D), both by BBM (d) and BLM (s). Data are the means 6 sem from three experiments. Ordinates and the specific activities of enzymes: g-GT (mmol/min/mg protein); LAP (nmol/min/mg protein); Na1,K1ATPase (mmol Pi/min/mg protein 3100). Each bar denotes mean 6 sem (N 5 3). *Denotes BBM values significantly different between BLM.
transporters within BBM of proximal tubules. This observation also indicates that the decrease of Na1-Pi uptake by BBM vesicles prepared from gludopa-treated rats was not due to increased conductance of BBM for Na1, a change that would cause a faster dissipation of the [Na1o] . [Na1i] gradient, as it was observed in BBM prepared from uremic kidneys [40], because it should similarly affect all Na1-solute cotransports. The lack of change in uptake of [3H]-L-proline and [3H]-alanine by BBM at the equilibrium period (90 min) also indicates that the volume of BBM vesicles was not altered in response to gludopa [8, 13, 25]. As reported by other investigators, infusion of gludopa decreases the activity of Na1,K1-ATPase in proximal convoluted tubules [41]. Decreased Na1,K1-ATPase activity may lead to a decrease of Na1 extrusion from the interior of proximal tubular cells across the BLM and may increase intracellular Na1 concentration, and hence, it could decrease of [Na1o] . [Na1i] gradient across the BBM. In principle, lower [Na1o] . [Na1i] gradient across BBM in situ may contribute to decrease of Na1-Pi reabsorption in intact cells. However, the nature of dependence of Na1-Pi cotransport across BBM [2, 8, 13] on the external concentration of Na1 (KmNa1 > 70 mm) indicates that a relatively modest decline in Na1,K1ATPase activity in response to gludopa [40] is unlikely to decrease the Na1 gradient [Na1o] . [Na1i] across
BBM sufficiently to noticeably influence the rate of Na1-Pi cotransport across BBM in situ [2, 8, 13]; moreover, the decrease in [Na1o] . [Na1i] gradient would also affect other Na1-solute cotransports in BBM. Furthermore, DA decreases Na1-H1 antiport across BBM [42, 43] and thus the Na1 entry from tubular lumen. This factor would counteract the effect of decreased Na1,K1-ATPase on intracellular Na1 and on the [Na1o] . [Na1i] gradient. The kinetics of Na1-Pi cotransport across BBM in response to gludopa, that is, the decreased Vmax without a change in KmPi (Fig. 5), resemble the changes of Na1-Pi cotransport across BBM in response to phosphaturic hormones that act via cAMP [2, 3], and differ from the competitive inhibition of Na1-Pi cotransporter by phosphaturic agents such as PFA that interact with BBM [8, 9]. This finding is consistent with the notion that Na1-Pi cotransport is not decreased because of the direct action of gludopa Na1-Pi cotransport. The addition of gludopa directly to BBM gludopa did not inhibit Na1-Pi cotransport in vitro, a finding also consistent with the notion that gludopa must be first converted to L-DOPA and DA [44]. Our finding that L-DOPA is generated from gludopa at a much higher rate by BBM than by BLM (Fig. 7) is consistent with the predominant [25], but not exclusive [45], localization of g-GT activity in BBM [8, 13, 25]. Suppression of L-DOPA formation from gludopa
de Toledo et al: Gludopa inhibits Na1-Pi cotransport
1839
Fig. 8. Effect of g-GT inhibitor acivicin (AVCN) on renal response to gludopa in vivo. Fractional excretion of phosphate (FEPi; A), urinary excretion of dopamine (B) and Na1Pi cotransport (C; measured as Na1-dependent 32 Pi-uptake by BBM at five-seconds; values of the control group is taken as 100% for paired comparison) were compared using controls (C), gludopa-treated (GD), and rats treated with gludopa plus acivicin (GD 1 AFCN). See Methods section for details. In A and B, each bar denotes mean 6 sem (N 5 4). In C each bar denotes mean 6 sem (N 5 3).
in vitro by the specific inhibitor acivicin (Fig. 7) affirms that the reaction is catalyzed by g-GT. Although the injection of L-DOPA resulted in an increase of DA both in kidney and in various other organs [21], injection of gludopa increases DA content preferentially in the kidney, in which the DA content increased approximately 10 times or more than in other studied tissues [21]. This huge 1000-fold increase of urinary DA in response to gludopa observed in our study (Fig. 3) suggests that, within the kidney parenchyma, enzymatic conversion of gludopa to DA occurs specifically in proximal tubules from where it egresses both to tubular and
peritubular fluid [46]. Thus, DA has ready access to DA receptors in membranes of proximal tubular cells, which are likely to be the first structures that come into contact with DA generated de novo from gludopa. The striking increase in urinary DA without significant change in excretion of norepinephrine (Figs. 3 and 8) shows that gludopa is converted to DA in non-neural tissues that lack DA-b-hydroxylase. To date, virtually all of the studies on gludopa have been focused on investigating its antihypertensive and natriuretic effects [21–24, 28–30]. It is amply documented that gludopa administration causes natriuresis by inhib-
1840
de Toledo et al: Gludopa inhibits Na1-Pi cotransport
iting Na1 reabsorption in tubules [24, 27, 28, 40]. DA inhibits Na1 reabsorption in proximal tubules [40, 47] as well as in distal nephron segments, especially in collecting ducts [48]. Natriuresis elicited by gludopa is likely due to an inhibition of Na1 reabsorption both in proximal and distal tubular segments. Inhibition of Na1 reabsorption in collecting ducts probably determines the overall rate of Na1 excretion; DA inhibits Na1,K1-ATPase in collecting ducts to an even greater degree than in proximal tubules [48]. Dopamine is generated from gludopa only in proximal tubules, and it ought to be delivered to cells of distal segments, namely collecting ducts, either via tubular fluid or perhaps also via peritubular interstitial fluid. Therefore, we suggest that gludopa given in lower doses can generate a quantum of DA in situ that is adequate to inhibit Na1-Pi cotransport and to decrease Pi reabsorption locally in proximal tubules, but such an amount of DA is not sufficient to reach, and act on, distal tubular segments. As a consequence, at lower doses, gludopa can inhibit Na1-Pi cotransport (Fig. 4) and Pi reabsorption (Fig. 2) in proximal tubules without causing natriuresis. A recent study showed that in lower doses, injected gludopa increases urinary excretion of DA without raising Na1 excretion [28]. This feature should be an important factor in designing the use of gludopa as a selective phosphaturic agent. The increase in urinary excretion of cAMP but not of cGMP (Fig. 3) is consonant with the notion that DA, generated in proximal tubules from gludopa, inhibits Na1-Pi cotransport via activation of cAMP signaling pathway (Fig. 1), as it was documented in our previous studies of OK cells [15, 16]. DA receptors were found both in luminal BBM and antiluminal BLM of proximal tubular cells [49]. However, adenylate cyclase activity in BLM is many-fold higher than in BBM [50], and therefore, DA generated from gludopa acts on DA receptors (probably DA1 subtypes) that are coupled positively to adenylate cyclase in BLM [51]. Dipeptide gludopa is a paradigm of an organ-targeted pro-drug, where the specific site of action is determined by cell phenotype, that is, expression of enzymes and transporters in the target cells. The initial and determining step in the sequence leading to the conversion of gludopa to L-DOPA is the action of g-GT, documented by the finding that acivicin blocks the release of L-DOPA in vitro (Fig. 7) and all effects of gludopa in vivo (Fig. 8). This initial step as well as all other steps in gludopa biotransformation, dynamics, and action are colocalized in proximal tubule cells (Fig. 1). A decrease of plasma Pi in response to gludopa (Fig. 2), which correlates inversely with the inhibition of Na1-Pi cotransport in BBM (Fig. 6), suggests that gludopainduced phosphaturia may have a significant impact on the overall Pi homeostasis. Furthermore, the action of gludopa is sharply targeted to proximal tubules. Appar-
ently, gludopa has little or no side effects [22–24, 27, 30], and the dipeptide is metabolized to natural molecules, that is, L-DOPA, DA, and glutamate. It should be stressed, however, that in this study we report on the selective phosphaturic effect and basic mechanisms of gludopa action. Whether or not the phosphaturic effect of gludopa is dose dependent, whether it persists after prolonged periods of administration, and how extensive it is in rats with intact parathyroid glands remains to be investigated in future studies. In this initial study, the effect of gludopa was studied in thyroparathyrectomized rats to eliminate the effects of PTH and calcitonin on Na1-Pi cotransport. This pathophysiological state resembles hypoparathyroid patients devoid of PTH, who retain Pi and in whom gludopa might be of therapeutic value. According to preliminary results, gludopa increases FEPi in intact rats and in partially nephrectomized rats with reduced renal function (abstract; Berndt et al, J Am Soc Nephrol 8:558A, 1997). The effect of gludopa in the syndrome of hyperparathyroidism remains to be examined. However, in this context, it is of interest that DA was shown to sensitize the kidney to the phosphaturic action of PTH [17]. The bioavailability of gludopa administered orally was reported to be low [30]. However, for reasons discussed earlier in this study, the quantum of gludopa that is adsorbed from the intestine may not be sufficient to produce natriuresis [28], but may be adequate for eliciting the phosphaturic effect. Moreover, various maneuvers potentially may allow gludopa to be effective even when administered orally. For example, oral coadministration of gludopa with monoaminooxidase type B (MAO-B) inhibitor, which slows down DA catabolism, increased the urinary excretion of Pi (T.P. Dousa, preliminary results). In conclusion, this study reports a selective phosphaturic effect of gludopa and elucidates basic features of the mechanism by which gludopa, given in vivo, inhibits Na1-Pi cotransport in the BBM of proximal tubules. Our results support the hypothesis that this phosphaturic effect of gludopa is due to the paracrine/autocrine action of DA generated in situ from the dipeptide in cells of proximal tubules (Fig. 1). We surmise that the key aspects of gludopa’s phosphaturic effect make this compound, in principle, potentially suited for novel, kidneytargeted pharmacotherapy of Pi overload. ACKNOWLEDGMENTS This study was supported by National Institutes of Health grants DK-30759 and DK-197215 and by the Mayo Foundation. We are grateful to Dr. D. McCormick, Department of Biochemistry and Molecular Biology, for the synthesis of gludopa. Ms. C.A. Davidson provided secretarial assistance. Reprint requests to Thomas P. Dousa, M.D., Ph.D., Mayo Clinic and Foundation, Renal Pathophysiology Laboratory, Guggenheim 9, Rochester, Minnesota 55905, USA. E-mail: [email protected]
de Toledo et al: Gludopa inhibits Na1-Pi cotransport
APPENDIX Abbreviations used in this article are: AADC, aromatic L-amino acid decarboxylase; BBM, brush border membrane; BLM, basolateral membrane; DA, dopamine; FE, fractional excretion; g-GT, g-glutamyltransferase; gludopa, dipeptide g-glutamyl-L-dihydroxyphenylaline; LAP, leucine aminopeptidase; L-DOPA, 3,4-dihydroxyphenylaline; OK, opossum kidney; PFA, phosphonoformic acid; Pi, phosphate.
20. 21. 22.
REFERENCES 1. Suki WN, Rouse D: Renal transport of calcium, magnesium, and phosphate, in Brenner and Rector’s The Kidney (5th ed), edited by Brenner BM, Philadelphia, W.B. Saunders, 1996, pp 472–515 2. Gmaj P, Murer H: Cellular mechanisms of inorganic phosphate transport in kidney. Physiol Rev 66:36–70, 1986 3. Kempson SA, Dousa TP: Current concepts of regulation of phosphate transport in renal proximal tubules. Biochem Pharmacol 35:721–726, 1986 4. Knochel JP, Agarwal R: Hypophosphatemia and hyperphosphatemia, in Brenner and Rector’s The Kidney (5th ed), edited by Brenner BM, Philadelphia, W.B. Saunders, 1996, pp 1086–1133 5. Delmez JA, Slatopolsky E: Hyperphosphatemia: Its consequence and treatment in patients with chronic renal disease. Am J Kidney Dis 19:303–317, 1992 6. Issac J, Glahn RP, Appel MA, Onsgard M, Dousa TP, Knox FG: Mechanism of dopamine inhibition of renal phosphate transport. J Am Soc Nephrol 2:1601–1607, 1992 7. Ghazali A, Ben Hamida F, Bouzernidj M, el Esper N, Westeel PF, Fournier A: Management of hyperphosphatemia in patients with renal failure. Curr Opin Nephrol Hypertens 2:566–579, 1993 8. Szczepanska-Konkel M, Yusufi ANK, Vanscoy M, Webster SK, Dousa TP: Phosphonocarboxylic acids as specific inhibitors of Na1-dependent transport of phosphate across renal brush border membrane. J Biol Chem 261:6375–6383, 1986 9. Vanscoy M, Loghman-Adham M, Onsgard M, SzczepanskaKonkel M, Homma S, Knox FG, Dousa TP: Mechanism of phosphaturia elicited by administration of phosphonoformate in vivo. Am J Physiol 255:F984–F994, 1988 10. Loghman-Adham M, Szczepanska-Konkel M, Dousa TP: Phosphate transport in brush border membranes from uremic rats: Response to phosphonoformic acid. J Am Soc Nephrol 3:1253–1259, 1992 11. Brooks DP, Ali SM, Contino LC, Stack E, Fredrickson TA, Feild J, Edwards RM: Phosphate excretion and phosphate transporter messenger RNA in uremic rats treated with phosphonoformic acid. J Pharmacol Exp Ther 281:1440–1445, 1997 12. Loghman-Adham M: Adaptation to changes in dietary phosphorus intake in health and in renal failure. J Lab Clin Med 129:176–188, 1987 13. Szczepanska-Konkel M, Yusufi ANK, Dousa TP: Interactions of [14C]-phosphonoformic acid with renal cortical brush-border membranes: Relationship to the Na1-phosphate cotransporter. J Biol Chem 262:8000–8010, 1987 14. Loghman-Adham M: Use of phosphonocarboxylic acids as inhibitors of sodium-phosphate cotransport. Gen Pharmacol 27:305–312, 1996 15. Glahn RP, Onsgard MJ, Tyce GM, Chinnow SL, Knox FG, Dousa TP: Autocrine/paracrine regulation of renal Na1-phosphate cotransport by dopamine. Am J Physiol 264:F618–F622, 1993 16. de Toledo FGS, Beers KW, Berndt TJ, Thompson MA, Tyce GM, Knox FG, Dousa TP: Opposite paracrine effects of 5-HT and dopamine on Na1-Pi cotransport in opossum kidney cells. Kidney Int 52:152–156, 1997 17. Isaac J, Berndt TJ, Chinnow SL, Tyce GM, Dousa TP, Knox FG: Dopamine enhances the phosphaturic response to parathyroid hormone in phosphate-deprived rats. J Am Soc Nephrol 2:1423– 1429, 1992 18. Cuche JL, Marchand GR, Greger RF, Lang RC, Knox FG: Phosphaturic effect of dopamine in dogs: Possible role of intrarenally produced dopamine in phosphate regulation. J Clin Invest 58:71–76, 1976 19. Yamazaki N, Sudo J: Activities of aromatic L-amino acid decar-
23. 24. 25.
26. 27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37. 38.
39.
40.
41.
42.
1841
boxylase with L-dopa as substrate in brush-border and basolateral membranes and cytoplasm obtained from rat renal cortex. Jpn J Pharmacol 46:193–196, 1988 Hayashi M, Yamaji Y, Kitajima W, Saruta T: Aromatic L-amino acid decarboxylase activity along the rat nephron. Am J Physiol 258:F28–F33, 1990 Wilk S, Mizoguchi H, Orlowski M: A kidney-specific dopamine precursor. J Pharmacol Exp Ther 206:227–232, 1978 Casagrande C, Merlo L, Ferrini R, Miragoli G, Semeraro C: Cardiovascular and renal action of dopaminergic prodrugs. J Cardiovasc Pharmacol 14:S40–S59, 1989 Lee RM: Five years’ experience with g-L-glutamyl-L-dopa: A relatively renally specific dopaminergic prodrug in man. J Auton Pharmacol 10:S103–S108, 1990 Lee MR: Dopamine and the kidney, in Advances in Renal Physiology (chapt 7), New York, Liss, 1986, pp 218–247 Yusufi ANK, Murayama N, Gapstur SM, Szczepanska-Konkel M, Dousa TP: Differential properties of brush-border membrane vesicles from early and late proximal tubules of rat kidney. Biochim Biophys Acta 1191:117–132, 1994 Shimada H, Endou H, Sakai F: Distribution of g-glutamyl transpeptidase and glutaminase isozymes in the rabbit nephron. Jpn J Pharmacol 32:121–129, 1982 Armando I, Nowicki S, Aguirre J, Barontini M: A decreased tubular uptake of dopa results in defective renal dopamine production in aged rats. Am J Physiol 268:F1087–F1092, 1995 Wang Z-Q, Siragy HM, Felder RA, Carey RM: Intrarenal dopamine production and distribution in the rat: Physiological control of sodium excretion. Hypertension 29:228–234, 1997 Casson IF, Clayden DA, Cope GF, Lee MR: The protective effect of g-glutamyl-L-dopa on the glycerol treated rat model of acute renal failure. Clin Sci 65:159–164, 1983 Freestone S, Ellis R, MacDonald TM, Lee MR: Effect of prolonged intravenous infusion of gludopa in humans and a preliminary study of its oral bioavailability. Am J Hypertens 3:126S–129S, 1990 Cummings J, Matheson LM, Maurice L, Smyth JF: Pharmacokinetics, bioavailability, metabolism, tissue distribution and urinary excretion of g-L-glutamyl-L-dopa in the rat. J Pharm Pharmacol 42:242–246, 1990 Casson IF, Clayden DA, Cope GF, Lee MR: The protective effect of g-glutamyl-L-dopa on the glycerol treated rat model of acute renal failure. Clin Sci 65:159–164, 1983 Khraibi AA, Knox FG: Renal interstitial hydrostatic pressure during natriuresis in hypertension. Am J Physiol 255:R756–R759, 1988 Sacktor B, Rosenbloom IL, Liang CT, Cheng L: Sodium gradientand sodium plus potassium gradient-dependent L-glutamate uptake in renal basolateral membrane vesicles. J Membr Biol 60:63– 71, 1981 Lowry OH, Rosebrough NJ, Farr AL, Randall RJ: Protein measurements with the Folin phenol reagent. J Biol Chem 193:265– 275, 1951 Dousa MK, Tyce GM: Free and conjugated plasma catecholamines, DOPA and 3-O-methyldopa in humans and in various animal species. Proc Soc Exp Biol Med 188:427–434, 1988 Chen PS, Toribara TY, Warner H: Microdetermination of phosphorus. Anal Chem 28:1756–1758, 1956 Shah SV, Northrup TE, Hui YSF, Dousa TP: Action of serotonin (5-hydroxytryptamine) on cyclic nucleotides in glomeruli of rat renal cortex. Kidney Int 15:463–472, 1979 Stole E, Smith TK, Manning JM, Meister A: Interaction of g-glutamyl transpeptidase with acivicin. J Biol Chem 269:21435– 21439, 1994 Hruska KA, Klahr S, Hammerman MR: Decreased luminal membrane transport of phosphate in chronic renal failure. Am J Physiol 242:F17–F22, 1982 Eklo¨f A-C, Holtba¨ck U, Sundelo¨f M, Chen S, Aperia A: Inhibition of COMT induces dopamine-dependent natriuresis and inhibition of proximal tubular Na1,K1-ATPase. Kidney Int 52:742–747, 1997 Felder RA, Campbell T, Albrecht F, Jose PA: Dopamine inhibits
1842
43.
44.
45.
46.
de Toledo et al: Gludopa inhibits Na1-Pi cotransport
Na1-H1 exchanger activity in renal BBMV by stimulation of adenylate cyclase. Am J Physiol 259:F297–F303, 1990 Winaver J, Burnett JC Jr, Tyce GM, Dousa TP: ANF inhibits Na1-H1 antiport in proximal tubular BBM: The role of dopamine. Kidney Int 38:1133–1140, 1990 Jeffrey RF, MacDonald TM, Marwick K, Lee MR: The effect of carbidopa and indomethacin on the renal response to g-L-glutamyl-L-dopa in normal man. Br J Pharmacol 25:195–201, 1988 Spater HW, Poruchynsky MS, Quintana N, Inoue M, Novikoff AB: Immunocytochemical localization of g-glutamyltransferase in rat kidney with protein A-horseradish peroxidase. Proc Natl Acad Sci USA 79:3547–3550, 1982 Soares-Da-Silva P, Serrao MP, Vieira-Coelho MA: Apical and basolateral uptake and intracellular fate of dopamine precursor L-dopa in LLC-PK1 cells. Am J Physiol 274:F243–F251, 1998
47. Satoh T, Cohen HT, Katz AI: Different mechanisms of renal NaK-ATPase regulation by protein kinases in proximal and distal nephron. Am J Physiol 265:F399–F405, 1993 48. Takemoto F, Cohen HT, Satoh T, Katz AI: Dopamine inhibits Na-K-ATPase in single tubules and cultured cells from distal nephron. Pflu¨gers Arch Eur J Pharmacol 421:302–306, 1992 49. Felder CC, McKelvey AM, Gitler MS, Eisner GM, Jose PA: Dopamine receptor subtypes in renal brush border and basolateral membranes. Kidney Int 36:183–193, 1989 50. Schwartz IL, Shalatz LJ, Kinne-Saffran E, Kinne R: Target cell polarity and membrane phosphorylation in relation to the mechanism of action of antidiuretic hormone. Proc Natl Acad Sci USA 71:2595–2599, 1974 51. Lokhandwala MF, Amenta F: Anatomical distribution and function of dopamine receptors in the kidney. FASEB J 5:3023–3030, 1991
g-L-glutamyl-L-DOPA inhibits Na1-phosphate cotransport across renal brush border membranes and increases renal excretion of phosphate FREDERICO G.S. DE TOLEDO, MICHAEL A. THOMPSON, CHAD BOLLIGER, GERTRUDE M. TYCE, and THOMAS P. DOUSA Renal Pathophysiology Laboratory, Department of Physiology and Biophysics, Mayo Clinic and Foundation, Mayo Medical School, Rochester, Minnesota, USA
g-L-glutamyl-L-DOPA inhibits Na1-phosphate cotransport across renal brush border membranes and increases renal excretion of phosphate. Background. For treatment of phosphate (Pi) overload in various pathophysiological states, an agent that selectively increases renal Pi excretion would be of major value. Previously, we have shown that dopamine (DA) inhibits Na1-Pi cotransport in renal epithelia. However, the administration of DA or its immediate precursor L-DOPA increases DA in multiple tissues. Synthetic dipeptide g-L-glutamyl-L-DOPA (gludopa) can serve as an inactive precursor (pro-pro-drug) of DA. This study tested the hypothesis that, because of the unique colocalization of g-glutamyltransferase (g-GT), aromatic amino acid decarboxylase, Na1-Pi cotransporter, and Na1-L-DOPA cotransporter in brush border membrane (BBM) of proximal tubular cells, gludopa may elicit phosphaturia by action of DA generated within the kidney. Methods. Thyroparathyrectomized rats were given placebo, or gludopa, or gludopa 1 g-GT inhibitor acivicin. Urinary excretion of Pi, Ca21, Na1, K1, DA, cAMP, and cGMP was determined, and Na1-Pi cotransport was measured in BBM prepared from kidneys of rats at the end of the experiment. Results. The administration of gludopa resulted in: (a) an inhibition of Na1-Pi cotransport, but not cotransport of Na1proline and Na1-alanine in BBM; (b) an increase (1300%) of fractional excretion (FE) of Pi and a drop (235%) of plasma Pi, whereas the plasma levels and FEs of Ca21, Na1, and K1 were unchanged; (c) an increase in urinary excretion of cAMP, but not cGMP; (d) a 1000-fold increase of urinary excretion of DA, without a change in excretion of norepinephrine; and (e) an incubation of gludopa with BBM in vitro, which caused a release of L-DOPA, and the in vivo administration of acivicin, which blocked actions of gludopa to inhibit Na1-Pi cotransport and to increase urinary excretions of Pi and DA. Conclusions. We conclude that colocalization of enzymes of biotransformation, BBM transporters, and the autocrine/ Key words: dopamine, gludopa, phosphaturia, acivicin, brush border membrane, renal excretion. Received for publication August 4, 1998 and in revised form December 16, 1998 Accepted for publication December 16, 1998
1999 by the International Society of Nephrology
paracrine DA system in cells of proximal tubules constitutes a cellular basis for the potent and specific phosphaturic action of gludopa.
Regulation of renal excretion of inorganic phosphate (Pi) is a key process in maintaining body Pi homeostasis [1]. Most of the Pi that is ultrafiltered in glomeruli is reabsorbed in proximal tubules. Sodium-phosphate (Na1-Pi) cotransport across luminal brush border membranes (BBM) is a rate-limiting [2] and regulated [3] step in the translocation of Pi from lumen of proximal tubules to peritubular capillaries [1]. Diminished renal excretion of Pi in a variety of pathophysiological states, such as hypoparathyroidism, pseudohypoparathyroidism, tumoral calcinosis, and, most frequently, renal insufficiency of various origins, results in chronic accumulation and retention of Pi and in hyperphosphatemia [4–6]. A surfeit of Pi in renal insufficiency, in turn, leads to secondary hyperparathyroidism and results in renal osteodystrophy in its various manifestations [4, 5, 7]. In principle, Pi overload can be ameliorated by either dietary or pharmacotherapeutic maneuvers that reduce the quantum of Pi that is absorbed in the gastrointestinal tract [4, 5, 7], an increase of renal Pi excretion [8–11], or a combination of both. The decrease of Pi quantum absorbed from small intestine can be achieved by dietary restriction of the amount of ingested Pi, that is, feeding with low-Pi diet and/or with the use of “Pi binders,” cationic compounds that bind dietary Pi in intestinal lumen and prevent its absorption across the wall of the small intestine [4, 5, 7]. However, therapeutic measures to correct Pi overload, based on these principles, are still unsatisfactory. A low Pi diet is difficult to design and maintain; moreover, a diet with low Pi should be used with caution in children with renal insufficiency who are in a growth anabolic phase [7]. Treatment with both
1832
de Toledo et al: Gludopa inhibits Na1-Pi cotransport
1833
Fig. 1. Scheme of biotransformation and action of gludopa in proximal tubular cell, resulting in increased phosphate (Pi) excretion. Gludopa (g-glutamyl-L-DOPA) from glomerular ultrafiltrate—tubular fluid binds as a substrate to g-glutamyltransferase (1. g-GT) on the surface of brush border membrane (BBM). g-GT catalyzes release of L-DOPA, and L-DOPA is translocated into cell interior by the [Na1o] . [Na1i] gradient energized cotransporter (s) across BBM [27]. In cytoplasm, L-DOPA is converted by L-aromatic amino acid decarboxylase (2. L-AADC) to dopamine that exits to interstitial fluid and also into tubular lumen. Dopamine may act also on the adjacent cells. Dopamine binds to a dopamine receptor (DA1 subtype) that is coupled to adenylate cyclase (AC) and increases synthesis of cAMP; increased cAMP activates protein kinase A (PKA) and thereby inhibits Na1-Pi cotransporter (d) in BBM. This inhibition is the basis for decreased Pi reabsorption in proximal tubules and results in phosphaturia (discussed earlier in this article).
aluminum- and calcium-based Pi binders has serious side effects [4, 5, 7], and this treatment modality also has a low degree of patient compliance. Thus, in patients with mild or moderate renal insufficiency with partially preserved renal function or in patients with normal renal function who retain Pi, such as hypoparathyroidism or pseudohypopathyroidism [4], the availability of a novel pharmacotherapy that selectively increases the renal excretion of Pi would be of major benefit. To date, however, attempts to increase renal Pi excretion by pharmacological inhibition of Pi reabsorption in proximal tubules has met with little success [12]. The administration of parathyroid hormone (PTH), an archetypal hormone that elicits phosphaturia [2, 3], also primarily increases tubular reabsorption of Ca21, mobilizes bone Ca21, and can lead to hypercalcemia. Moreover, PTH should be given by the parenteral route. The administration of phosphonoformic acid (PFA), a specific competitive inhibitor of Na1-Pi cotransport in the luminal BBM of proximal tubules [8, 9], markedly increases urinary excretion of Pi in rats fed either normal- or low-Pi diet [8, 9], and also in rats with experimental renal insufficiency because of 5/6 nephrectomy [10, 11]. However, the relatively low affinity of PFA (Ki > 1 mm) for Na1-Pi cotransporter [8, 13] and concerns that have been raised about potential toxic effects of larger doses of PFA when administered to severely compromised patients [12, 14] currently precluded its further preclinical testing and use
in clinical practice [12, 14]. Thus, a suitable pharmacotherapy that would selectively increase the renal excretion of Pi is not available. Via stimulating dopamine (DA) receptors (DA1 subtype) and by activating the cAMP signaling pathway, dopamine elicits an inhibition of Na1-Pi cotransport in cultured renal epithelial opossum kidney (OK) cells [15, 16], an established in vitro model of Na1-Pi cotransport in renal proximal tubules [15, 16]. Intravenous infusion of DA to thyroparathyrectomized rats selectively decreased the Na1-Pi cotransport across renal BBM of proximal tubules [6] and resulted in an increased fractional excretion (FE) of Pi [17, 18]. Based on these observations, we hypothesized that DA generated intrarenally, via the inhibition of Na1-Pi cotransport across BBM, can decrease proximal tubular reabsorption of Pi and elicit phosphaturia [6, 15, 17, 18]. DA should be administered by intravenous (i.v.) infusion and thence has multiple effects in nonrenal organs, other than the increase of Pi excretion. To specifically modulate renal tubular function, DA should be generated in situ in renal parenchyma. Not only DA itself, but also its immediate precursor, 3,4-dihydroxyphenylalanine (L-DOPA), when added to OK cells in vitro inhibits Na1-Pi cotransport [15, 16], and in vivo administration of L-DOPA elicits phosphaturia [18]. Aromatic L-amino acid decarboxylase (AADC), a cytosolic enzyme [19] that catalyzes decarboxylation of L-DOPA to DA, is present exclusively in proximal tubules
1834
de Toledo et al: Gludopa inhibits Na1-Pi cotransport
Fig. 2. Plasma levels (A) and urinary excretions (B) of Pi, Ca21, and Na1 and K1 in control rats (C) and in rats injected with gludopa (GD). Plasma levels are in mm, and urinary excretion of electrolytes is expressed as a fractional excretion (% on ordinate) relative to urinary excretion and plasma level of creatinine (Methods section). Urinary excretion of Pi was significantly (P , 0.02) increased (D , 1300%), whereas plasma level of Pi was significantly (P , 0.05) decreased (D , 235%). Plasma levels as well as urinary excretions (FE) of Ca21, Na1, and K1 were not different between controls and GD-treated rats. Each bar denotes mean 6 sem (N 5 5 to 6).
[20], and it stands to reason that DA generated from L-DOPA locally in proximal tubules can act as an autocrine/paracrine agent primarily on proximal tubule cells. When administered parenterally or orally, L-DOPA is decarboxylated to DA not only in the kidney but also in numerous extrarenal tissues [21], and as a result, generated DA has many other effects, including hemodynamic effects, in many nonrenal tissues. Synthetic dipeptide g-L-glutamyl-L-dihydroxyphenylalanine (g-L-glutamyl-LDOPA, or gludopa) can serve as a pro-drug [22] of L-DOPA, and consequently of DA [21–24], in a manner that is sharply targeted to the kidney [21]. Gludopa is enzymatically converted to L-DOPA by g-glutamyltranspeptidase (g-GT), a very highly active enzyme present in the BBM of proximal tubules [25, 26], and the L-DOPA is transported across BBM [27] and then decarboxylated in proximal tubules by cytosolic AADC to DA [19, 20]. (g-GT catalyzes the enzymatic release of L-DOPA from gludopa either by transpeptidation, that is, glutamic acid forms a g-peptidic bond with another amino acid, if available, or hydrolyzes gludopa to L-DOPA and free glutamic acid [24]. Because it is not known whether one or both types of reaction occurs in vivo, we refer to action of g-GT to gludopa as the “release” or “generation” of L-DOPA from gludopa.) Thus,
colocalization of g-GT and AADC in proximal tubules, but not in other nephron segments [20, 26], should allow the conversion of L-DOPA to DA in proximal tubules by a two-step enzymatic reaction. Then in situ generated DA can modulate the functions of proximal tubular cells, including the rate of Na1-Pi cotransport (Fig. 1). It is well known that the administration of gludopa to humans [23, 24] and to experimental animals [28, 29] causes an increase in renal excretion of Na1 that is sometimes accompanied with increases of glomerular filtration rate, urine flow, and renal blood flow, but are without prohibitive side effects [23, 24, 30]. To date, the effects of gludopa on the renal excretion of Pi and homeostasis of Pi have not been considered or investigated. We experimentally tested the hypothesis that gludopa is converted to DA in proximal tubule cells, which inhibits Na1-Pi cotransport across BBM and elicits phosphaturia by the mechanism outlined in Figure 1. METHODS Surgically thyroparathyroidectomized male rats weighing 250 to 280 g were used for the in vivo experiments. Rats were kept in metabolic cages at least three days prior to the experiment while receiving free access to
de Toledo et al: Gludopa inhibits Na1-Pi cotransport
Fig. 3. Urinary excretions of cAMP and cGMP, dopamine, and norepinephrine in control rats (C) and in rats injected with gludopa (GD). Excretions of cAMP and cGMP are expressed as a percentage of change relative to the paired control, taken as 100%. Excretions of cAMP (P , 0.05) and of dopamine (P , 0.001) were (*) significantly increased (by t-test). Excretions of dopamine and norepinephrine are expressed in nmol/mg creatinine (ordinates). Each bar denotes mean 6 sem (N 5 7).
standard rat chow and drinking water with the addition of 0.35% Ca21 glutamate. As reported [31], gludopa injected intraperitoneally (i.p.) was rapidly absorbed (t1/2ab 5 6 min), and maximum plasma concentrations were maintained for approximately two hours postinjection, followed by a monophasic elimination decline [31]. Therefore, the maximum effect of gludopa can be anticipated within a three hour time span after i.p. injection. In our experiments, rats were injected i.p. with either gludopa (20 mg/kg body wt) or controls with glutamic acid (5.24 mg/kg body wt), both dissolved in 5% glucose. The dose of 20 mg/kg body wt gludopa was chosen based on known pharmacokinetics of gludopa in rat [31] and was 2.5 times lower than a dose of 50 mg/kg, which elicits modest natriuresis [32]. Rats were then kept in metabolic cages, and urine was collected for the three-hour postinjection period. At the end of three hours, rats were anesthetized by an i.m. injection of 0.1 mg/100 g body wt of a solution of equal volumes of 20 mg/ml xylazine and 100 mg/ml ketamine hydrochloride [33]. Blood was collected. Kidneys were removed, and the dissected renal cortical tissue was used for isolation of BBM [8, 13]. In experiments examining the effects of g-GT inhibitor acivicin in vivo, one group of rats was injected first with
1835
i.p. acivicin (10 mg/kg body wt) one hour prior to the injection of gludopa, and a second dose of acivicin (10 mg/kg body wt) was administered simultaneously with gludopa (20 mg/kg body wt). In the second group, rats received gludopa (20 mg/kg body wt) only, and in the third group, rats received only an i.p. injection of the vehicle (5% glucose). To allow the comparison between groups, experiments were conducted using a paired design [6, 9, 10]. That is, control rats and rats treated with test agents were taken into the experiment at the same date, were followed simultaneously, and were sacrificed at the same day. Also, preparation of BBM, transport and enzyme measurements, and other analyses were conducted in the same experimental session using identical reagents and procedures [6, 8–10, 17]. Brush border membrane vesicles were prepared with the Mg21-precipitation method, as described in our previous studies [8, 9, 13]. The fraction of basolateral membranes (BLM) was prepared by differential centrifugation and by the Percoll gradient centrifugation procedure [34]. Activities of leucine aminopeptidase (LAP) and g-GT in BBM and BLM were determined using chromogenic substrates, as described in our previous studies [8, 13, 25], and also Na1,K1-ATPase activity was determined as an oubain-sensitive ATPase, also described in our previous communications [8, 13]. Protein concentration in fractions was determined using Lowry’s methods [35]. The fraction of BBM showed significantly higher activities of g-GT and LAP, and significantly lower activity of Na1,K1-ATPase as compared with BLM (Fig. 7). The BLM had a significantly higher activity of Na1,K1ATPase and, conversely, rather lower activities of g-GT and LAP than BBM (Fig. 7). To determine the rate of generation of L-DOPA from gludopa in vitro, fractions of BBM and BLM prepared from normal rats were incubated with a substrate of 0.1 mm gludopa in a medium containing 40 mm Tris-HCl, 5 mm glycine-glycine, and 0.01% (wt/vol) sodium metabisulfite (pH 5 7.4). The incubation was terminated by the addition of ice-cold perchloric acid at a final concentration of 5%. The precipitate was removed by centrifugation at 3800 3 g, and the content of L-DOPA in the supernatant was determined by high-performance liquid chromatography (HPLC) with electrochemical detection, as previously described [16, 36]. The content of DA and norepinephrine in the urine was also determined by HPLC [36]. Transport studies in BBM vesicles The Na1-gradient–dependent uptakes of 32Pi, [3H]-Lalanine, and [3H]-L-proline by BBM vesicles were measured by the rapid filtration technique, as described in our previous communications [8, 9, 13, 25]. The initial concentrative “uphill” uptake was measured at 5 and 10 seconds, and uptake at the equilibrium period of 90 minutes [8, 13]. The Na1-independent uptakes of 32Pi
1836
de Toledo et al: Gludopa inhibits Na1-Pi cotransport
Fig. 4. Capacity for Na1-gradient–dependent Na1-Pi cotransport (A), Na1-[3H]-proline cotransport (B), and Na1-[3H]-alanine cotransport across BBM isolated from kidneys of control rats (C) or rats injected with gludopa (GD). All three transport systems were determined in aliquots of the same preparation of BBM vesicles and are expressed as an uptake of solute pmol/mg protein (ordinates) at indicated time periods. (A) Na1-dependent uptake of Pi was significantly (t-test) lower in GD-treated rats than in controls in the uphill concentrative phase, at 5 (–31%) and 10 seconds (–28%), but not at the 90-minute equilibrium point. In contrast, the capacity for Na1-dependent cotransports of [3H]proline (B) or of [3H]-alanine (C) was not different between controls and GD-treated rats. Each bar denotes mean 6 sem (N 5 4 to 7).
and of [3H]-L-proline and [3H]-L-alanine were negligible (less than 8% of the uptake in the presence of Na1 gradient), and therefore, the values without Na1 were not routinely measured and subtracted [9, 25]. Urine and plasma concentrations of Ca21, Na1, and K1 were determined by atomic absorption spectrophotometry. Pi in plasma and urine was measured by Chen’s colorimetric methods [37]. Creatinine was determined by the colorimetric method, and the content of cAMP and cGMP was determined by radioimmunoassay [38]. Gludopa was synthesized in the Mayo Clinic Protein Core Facility. L-DOPA, glutamic acid, DA, and acivicin were all purchased from Sigma (St. Louis, MO, USA). 32 Pi, [3H]-L-proline, [3H]-L-alanine, and radioimmunoassay kits for cAMP and cGMP were purchased from NEN (Boston, MA, USA). Other chemicals, all of the highest purity grades available, were purchased from Sigma. The results were evaluated statistically, unless stated other-
wise, with the use of Student’s t-test for group or pair comparisons, as appropriate. Values of P less than 0.05 were considered statistically significant.
RESULTS Intraperitoneal administration of gludopa resulted in at least a fourfold increase in FE of Pi (FEPi), but the FEs of Ca21, Na1, and K1 were not changed (Fig. 2). After three hours of gludopa injection, the plasma concentration of Pi significantly decreased (235%), but no significant changes were found in the plasma levels of Ca21, Na1, and K1 (Fig. 2). In response to gludopa, the urinary excretion of cAMP showed a significant increase (D > 180%), but the excretion of cGMP did not change significantly (Fig. 3). Finally, in response to gludopa, the urinary excretion of DA increased more than 1000-fold,
de Toledo et al: Gludopa inhibits Na1-Pi cotransport
Fig. 5. Kinetics of Na1-Pi cotransport inhibition in response to gludopa as evaluated by a double-reciprocal plot. The Vmax for Na1-Pi cotransport in BBM from (s) (221 6 49 nmol Pi/second/mg protein, N 5 3) was significantly (P , 0.05, paired t-test) lower than Vmax in BBM from corresponding controls (d; 288 6 38 pmol/second/mg protein; N 5 3). On the other hand, KmPi (242 6 34 mm Pi; N 5 3) did not significantly differ from KmPi for controls (232 6 26 mm Pi; N 5 3). Data are mean 6 sem.
whereas urinary excretion of norepinephrine did not change at all (Fig. 3). The Na1-gradient–dependent uptake of 32Pi by BBM vesicles prepared from renal cortex of gludopa-injected rats was markedly decreased (D > 230%) in the concentrative phase (Fig. 4). As measured in the same BBM membranes, the Na1-gradient–dependent uptake of [3H]-L-proline or [3H]-L-alanine was not altered in response to the administration of gludopa (Fig. 4). Neither uptake of [3H]-L-proline, [3H]-L-alanine, nor 32Pi at equilibrium period (90 min) was decreased (Fig. 4). The decrease in Na1-gradient–dependent uptake of 32Pi was due to a lower Vmax without a change in KmPi (Fig. 5). Because plasma for Pi determination was collected and kidneys for the preparation of BBM were harvested at the same time, the relationship of these two parameters was examined. In response to gludopa administration, the D% decrease in rate of Na1-Pi cotransport across BBM showed positive correlation with D% decrease of plasma Pi in the same rats (Fig. 6). When gludopa was added directly in vitro to isolated BBM vesicles, the Na1-dependent uptake of 32Pi was not changed (data not shown). Incubation of gludopa with BBM in vitro resulted in a time-dependent release of L-DOPA that was linear with time at least up to 10 minutes (Fig. 7). In contrast, the rate of L-DOPA release after incubation of gludopa with BLM was five times lower than with BBM (Fig. 7). The generation of L-DOPA from gludopa catalyzed both by
1837
Fig. 6. Correlation between the extent of decrease of Na1-Pi cotransport capacity of BBM and decrease of plasma levels of Pi in the same animals. Both parameters are expressed as -D% relative to paired controls that received solvent only, with the solid line representing the regression line, and dashed lines the 95% confidence limits. Ordinate, -D% of plasma [Pi]; abscissa, -D% of Na1-Pi cotransport, determined as 32Pi uptake by BBM at five-second intervals (Methods and Fig. 4). The correlation was significant: r 5 0.88 6 0.11, P , 0.05, t-test.
BBM and BLM was almost completely inhibited by acivicin [38], a specific irreversible inhibitor of g-GT (Fig. 7). Finally, we investigated whether the release of L-DOPA from gludopa by the action of g-GT was indeed an essential first step in the biotransformation pathway that produces phosphaturia (Fig. 1). The in vivo administration of acivicin [39] prior to and simultaneously with gludopa (discussed in the Methods section) completely blocked the decrease in Na1-gradient–dependent uptake of 32Pi across BBM, as well as abolished the increase in urinary excretion of Pi and increase of urinary excretion of DA (Fig. 8). The addition of acivicin in vitro to BBM vesicles had no effect on Na1-dependent 32Pi uptake by BBM (data not shown). DISCUSSION The experimental results presented herein report that gludopa administered in vivo acts as a selective phosphaturic agent and elucidate the mechanism of its action. Administration of gludopa decreased the capacity of Na1-gradient–dependent transport of 32Pi across renal BBM, but Na1-dependent cotransport of [3H]-L-proline or [3H]-L-alanine was not altered (Fig. 4). This finding is similar to the effect of infused DA on BBM [6], and suggests that administered gludopa did not cause a generalized inhibition of multiple Na1-gradient–dependent
1838
de Toledo et al: Gludopa inhibits Na1-Pi cotransport
Fig. 7. Activities of membrane-bound enzymes in fraction of brush border membranes (BBM) and fraction of basolateral membranes (BLM) prepared from rat kidney cortex, and generation of L-DOPA from gludopa by incubation with BBM and BLM. BBMs are enriched in g-glutamyltransferase (g-GT; A) and leucine aminopeptidase (LAP; B); BLMs are enriched in Na1, K1-ATPase (C). (D) Generation of L-DOPA from gludopa after incubation with either BLM or BBM fractions from rat kidney cortex. Gludopa (0.1 mm) was incubated with membrane fractions for 10 minutes, and L-DOPA and gludopa were determined by high-performance liquid chromatography (Methods section). The addition of acivicin, a g-glutamyltransferase inhibitor [38], suppressed generation of L-DOPA from gludopa both in BBM and BLM; see black portions at the bottom of the bars. The rate of enzymatic release of L-DOPA from gludopa was significantly (P , 0.01, t-test) higher in BBM (D 1394%) than in BLM. Each bar denotes mean 6 sem (N 5 3). (E) Under these experimental conditions, L-DOPA generation from gludopa was linearly proportional to a time of at least up to 10 minutes (D), both by BBM (d) and BLM (s). Data are the means 6 sem from three experiments. Ordinates and the specific activities of enzymes: g-GT (mmol/min/mg protein); LAP (nmol/min/mg protein); Na1,K1ATPase (mmol Pi/min/mg protein 3100). Each bar denotes mean 6 sem (N 5 3). *Denotes BBM values significantly different between BLM.
transporters within BBM of proximal tubules. This observation also indicates that the decrease of Na1-Pi uptake by BBM vesicles prepared from gludopa-treated rats was not due to increased conductance of BBM for Na1, a change that would cause a faster dissipation of the [Na1o] . [Na1i] gradient, as it was observed in BBM prepared from uremic kidneys [40], because it should similarly affect all Na1-solute cotransports. The lack of change in uptake of [3H]-L-proline and [3H]-alanine by BBM at the equilibrium period (90 min) also indicates that the volume of BBM vesicles was not altered in response to gludopa [8, 13, 25]. As reported by other investigators, infusion of gludopa decreases the activity of Na1,K1-ATPase in proximal convoluted tubules [41]. Decreased Na1,K1-ATPase activity may lead to a decrease of Na1 extrusion from the interior of proximal tubular cells across the BLM and may increase intracellular Na1 concentration, and hence, it could decrease of [Na1o] . [Na1i] gradient across the BBM. In principle, lower [Na1o] . [Na1i] gradient across BBM in situ may contribute to decrease of Na1-Pi reabsorption in intact cells. However, the nature of dependence of Na1-Pi cotransport across BBM [2, 8, 13] on the external concentration of Na1 (KmNa1 > 70 mm) indicates that a relatively modest decline in Na1,K1ATPase activity in response to gludopa [40] is unlikely to decrease the Na1 gradient [Na1o] . [Na1i] across
BBM sufficiently to noticeably influence the rate of Na1-Pi cotransport across BBM in situ [2, 8, 13]; moreover, the decrease in [Na1o] . [Na1i] gradient would also affect other Na1-solute cotransports in BBM. Furthermore, DA decreases Na1-H1 antiport across BBM [42, 43] and thus the Na1 entry from tubular lumen. This factor would counteract the effect of decreased Na1,K1-ATPase on intracellular Na1 and on the [Na1o] . [Na1i] gradient. The kinetics of Na1-Pi cotransport across BBM in response to gludopa, that is, the decreased Vmax without a change in KmPi (Fig. 5), resemble the changes of Na1-Pi cotransport across BBM in response to phosphaturic hormones that act via cAMP [2, 3], and differ from the competitive inhibition of Na1-Pi cotransporter by phosphaturic agents such as PFA that interact with BBM [8, 9]. This finding is consistent with the notion that Na1-Pi cotransport is not decreased because of the direct action of gludopa Na1-Pi cotransport. The addition of gludopa directly to BBM gludopa did not inhibit Na1-Pi cotransport in vitro, a finding also consistent with the notion that gludopa must be first converted to L-DOPA and DA [44]. Our finding that L-DOPA is generated from gludopa at a much higher rate by BBM than by BLM (Fig. 7) is consistent with the predominant [25], but not exclusive [45], localization of g-GT activity in BBM [8, 13, 25]. Suppression of L-DOPA formation from gludopa
de Toledo et al: Gludopa inhibits Na1-Pi cotransport
1839
Fig. 8. Effect of g-GT inhibitor acivicin (AVCN) on renal response to gludopa in vivo. Fractional excretion of phosphate (FEPi; A), urinary excretion of dopamine (B) and Na1Pi cotransport (C; measured as Na1-dependent 32 Pi-uptake by BBM at five-seconds; values of the control group is taken as 100% for paired comparison) were compared using controls (C), gludopa-treated (GD), and rats treated with gludopa plus acivicin (GD 1 AFCN). See Methods section for details. In A and B, each bar denotes mean 6 sem (N 5 4). In C each bar denotes mean 6 sem (N 5 3).
in vitro by the specific inhibitor acivicin (Fig. 7) affirms that the reaction is catalyzed by g-GT. Although the injection of L-DOPA resulted in an increase of DA both in kidney and in various other organs [21], injection of gludopa increases DA content preferentially in the kidney, in which the DA content increased approximately 10 times or more than in other studied tissues [21]. This huge 1000-fold increase of urinary DA in response to gludopa observed in our study (Fig. 3) suggests that, within the kidney parenchyma, enzymatic conversion of gludopa to DA occurs specifically in proximal tubules from where it egresses both to tubular and
peritubular fluid [46]. Thus, DA has ready access to DA receptors in membranes of proximal tubular cells, which are likely to be the first structures that come into contact with DA generated de novo from gludopa. The striking increase in urinary DA without significant change in excretion of norepinephrine (Figs. 3 and 8) shows that gludopa is converted to DA in non-neural tissues that lack DA-b-hydroxylase. To date, virtually all of the studies on gludopa have been focused on investigating its antihypertensive and natriuretic effects [21–24, 28–30]. It is amply documented that gludopa administration causes natriuresis by inhib-
1840
de Toledo et al: Gludopa inhibits Na1-Pi cotransport
iting Na1 reabsorption in tubules [24, 27, 28, 40]. DA inhibits Na1 reabsorption in proximal tubules [40, 47] as well as in distal nephron segments, especially in collecting ducts [48]. Natriuresis elicited by gludopa is likely due to an inhibition of Na1 reabsorption both in proximal and distal tubular segments. Inhibition of Na1 reabsorption in collecting ducts probably determines the overall rate of Na1 excretion; DA inhibits Na1,K1-ATPase in collecting ducts to an even greater degree than in proximal tubules [48]. Dopamine is generated from gludopa only in proximal tubules, and it ought to be delivered to cells of distal segments, namely collecting ducts, either via tubular fluid or perhaps also via peritubular interstitial fluid. Therefore, we suggest that gludopa given in lower doses can generate a quantum of DA in situ that is adequate to inhibit Na1-Pi cotransport and to decrease Pi reabsorption locally in proximal tubules, but such an amount of DA is not sufficient to reach, and act on, distal tubular segments. As a consequence, at lower doses, gludopa can inhibit Na1-Pi cotransport (Fig. 4) and Pi reabsorption (Fig. 2) in proximal tubules without causing natriuresis. A recent study showed that in lower doses, injected gludopa increases urinary excretion of DA without raising Na1 excretion [28]. This feature should be an important factor in designing the use of gludopa as a selective phosphaturic agent. The increase in urinary excretion of cAMP but not of cGMP (Fig. 3) is consonant with the notion that DA, generated in proximal tubules from gludopa, inhibits Na1-Pi cotransport via activation of cAMP signaling pathway (Fig. 1), as it was documented in our previous studies of OK cells [15, 16]. DA receptors were found both in luminal BBM and antiluminal BLM of proximal tubular cells [49]. However, adenylate cyclase activity in BLM is many-fold higher than in BBM [50], and therefore, DA generated from gludopa acts on DA receptors (probably DA1 subtypes) that are coupled positively to adenylate cyclase in BLM [51]. Dipeptide gludopa is a paradigm of an organ-targeted pro-drug, where the specific site of action is determined by cell phenotype, that is, expression of enzymes and transporters in the target cells. The initial and determining step in the sequence leading to the conversion of gludopa to L-DOPA is the action of g-GT, documented by the finding that acivicin blocks the release of L-DOPA in vitro (Fig. 7) and all effects of gludopa in vivo (Fig. 8). This initial step as well as all other steps in gludopa biotransformation, dynamics, and action are colocalized in proximal tubule cells (Fig. 1). A decrease of plasma Pi in response to gludopa (Fig. 2), which correlates inversely with the inhibition of Na1-Pi cotransport in BBM (Fig. 6), suggests that gludopainduced phosphaturia may have a significant impact on the overall Pi homeostasis. Furthermore, the action of gludopa is sharply targeted to proximal tubules. Appar-
ently, gludopa has little or no side effects [22–24, 27, 30], and the dipeptide is metabolized to natural molecules, that is, L-DOPA, DA, and glutamate. It should be stressed, however, that in this study we report on the selective phosphaturic effect and basic mechanisms of gludopa action. Whether or not the phosphaturic effect of gludopa is dose dependent, whether it persists after prolonged periods of administration, and how extensive it is in rats with intact parathyroid glands remains to be investigated in future studies. In this initial study, the effect of gludopa was studied in thyroparathyrectomized rats to eliminate the effects of PTH and calcitonin on Na1-Pi cotransport. This pathophysiological state resembles hypoparathyroid patients devoid of PTH, who retain Pi and in whom gludopa might be of therapeutic value. According to preliminary results, gludopa increases FEPi in intact rats and in partially nephrectomized rats with reduced renal function (abstract; Berndt et al, J Am Soc Nephrol 8:558A, 1997). The effect of gludopa in the syndrome of hyperparathyroidism remains to be examined. However, in this context, it is of interest that DA was shown to sensitize the kidney to the phosphaturic action of PTH [17]. The bioavailability of gludopa administered orally was reported to be low [30]. However, for reasons discussed earlier in this study, the quantum of gludopa that is adsorbed from the intestine may not be sufficient to produce natriuresis [28], but may be adequate for eliciting the phosphaturic effect. Moreover, various maneuvers potentially may allow gludopa to be effective even when administered orally. For example, oral coadministration of gludopa with monoaminooxidase type B (MAO-B) inhibitor, which slows down DA catabolism, increased the urinary excretion of Pi (T.P. Dousa, preliminary results). In conclusion, this study reports a selective phosphaturic effect of gludopa and elucidates basic features of the mechanism by which gludopa, given in vivo, inhibits Na1-Pi cotransport in the BBM of proximal tubules. Our results support the hypothesis that this phosphaturic effect of gludopa is due to the paracrine/autocrine action of DA generated in situ from the dipeptide in cells of proximal tubules (Fig. 1). We surmise that the key aspects of gludopa’s phosphaturic effect make this compound, in principle, potentially suited for novel, kidneytargeted pharmacotherapy of Pi overload. ACKNOWLEDGMENTS This study was supported by National Institutes of Health grants DK-30759 and DK-197215 and by the Mayo Foundation. We are grateful to Dr. D. McCormick, Department of Biochemistry and Molecular Biology, for the synthesis of gludopa. Ms. C.A. Davidson provided secretarial assistance. Reprint requests to Thomas P. Dousa, M.D., Ph.D., Mayo Clinic and Foundation, Renal Pathophysiology Laboratory, Guggenheim 9, Rochester, Minnesota 55905, USA. E-mail: [email protected]
de Toledo et al: Gludopa inhibits Na1-Pi cotransport
APPENDIX Abbreviations used in this article are: AADC, aromatic L-amino acid decarboxylase; BBM, brush border membrane; BLM, basolateral membrane; DA, dopamine; FE, fractional excretion; g-GT, g-glutamyltransferase; gludopa, dipeptide g-glutamyl-L-dihydroxyphenylaline; LAP, leucine aminopeptidase; L-DOPA, 3,4-dihydroxyphenylaline; OK, opossum kidney; PFA, phosphonoformic acid; Pi, phosphate.
20. 21. 22.
REFERENCES 1. Suki WN, Rouse D: Renal transport of calcium, magnesium, and phosphate, in Brenner and Rector’s The Kidney (5th ed), edited by Brenner BM, Philadelphia, W.B. Saunders, 1996, pp 472–515 2. Gmaj P, Murer H: Cellular mechanisms of inorganic phosphate transport in kidney. Physiol Rev 66:36–70, 1986 3. Kempson SA, Dousa TP: Current concepts of regulation of phosphate transport in renal proximal tubules. Biochem Pharmacol 35:721–726, 1986 4. Knochel JP, Agarwal R: Hypophosphatemia and hyperphosphatemia, in Brenner and Rector’s The Kidney (5th ed), edited by Brenner BM, Philadelphia, W.B. Saunders, 1996, pp 1086–1133 5. Delmez JA, Slatopolsky E: Hyperphosphatemia: Its consequence and treatment in patients with chronic renal disease. Am J Kidney Dis 19:303–317, 1992 6. Issac J, Glahn RP, Appel MA, Onsgard M, Dousa TP, Knox FG: Mechanism of dopamine inhibition of renal phosphate transport. J Am Soc Nephrol 2:1601–1607, 1992 7. Ghazali A, Ben Hamida F, Bouzernidj M, el Esper N, Westeel PF, Fournier A: Management of hyperphosphatemia in patients with renal failure. Curr Opin Nephrol Hypertens 2:566–579, 1993 8. Szczepanska-Konkel M, Yusufi ANK, Vanscoy M, Webster SK, Dousa TP: Phosphonocarboxylic acids as specific inhibitors of Na1-dependent transport of phosphate across renal brush border membrane. J Biol Chem 261:6375–6383, 1986 9. Vanscoy M, Loghman-Adham M, Onsgard M, SzczepanskaKonkel M, Homma S, Knox FG, Dousa TP: Mechanism of phosphaturia elicited by administration of phosphonoformate in vivo. Am J Physiol 255:F984–F994, 1988 10. Loghman-Adham M, Szczepanska-Konkel M, Dousa TP: Phosphate transport in brush border membranes from uremic rats: Response to phosphonoformic acid. J Am Soc Nephrol 3:1253–1259, 1992 11. Brooks DP, Ali SM, Contino LC, Stack E, Fredrickson TA, Feild J, Edwards RM: Phosphate excretion and phosphate transporter messenger RNA in uremic rats treated with phosphonoformic acid. J Pharmacol Exp Ther 281:1440–1445, 1997 12. Loghman-Adham M: Adaptation to changes in dietary phosphorus intake in health and in renal failure. J Lab Clin Med 129:176–188, 1987 13. Szczepanska-Konkel M, Yusufi ANK, Dousa TP: Interactions of [14C]-phosphonoformic acid with renal cortical brush-border membranes: Relationship to the Na1-phosphate cotransporter. J Biol Chem 262:8000–8010, 1987 14. Loghman-Adham M: Use of phosphonocarboxylic acids as inhibitors of sodium-phosphate cotransport. Gen Pharmacol 27:305–312, 1996 15. Glahn RP, Onsgard MJ, Tyce GM, Chinnow SL, Knox FG, Dousa TP: Autocrine/paracrine regulation of renal Na1-phosphate cotransport by dopamine. Am J Physiol 264:F618–F622, 1993 16. de Toledo FGS, Beers KW, Berndt TJ, Thompson MA, Tyce GM, Knox FG, Dousa TP: Opposite paracrine effects of 5-HT and dopamine on Na1-Pi cotransport in opossum kidney cells. Kidney Int 52:152–156, 1997 17. Isaac J, Berndt TJ, Chinnow SL, Tyce GM, Dousa TP, Knox FG: Dopamine enhances the phosphaturic response to parathyroid hormone in phosphate-deprived rats. J Am Soc Nephrol 2:1423– 1429, 1992 18. Cuche JL, Marchand GR, Greger RF, Lang RC, Knox FG: Phosphaturic effect of dopamine in dogs: Possible role of intrarenally produced dopamine in phosphate regulation. J Clin Invest 58:71–76, 1976 19. Yamazaki N, Sudo J: Activities of aromatic L-amino acid decar-
23. 24. 25.
26. 27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37. 38.
39.
40.
41.
42.
1841
boxylase with L-dopa as substrate in brush-border and basolateral membranes and cytoplasm obtained from rat renal cortex. Jpn J Pharmacol 46:193–196, 1988 Hayashi M, Yamaji Y, Kitajima W, Saruta T: Aromatic L-amino acid decarboxylase activity along the rat nephron. Am J Physiol 258:F28–F33, 1990 Wilk S, Mizoguchi H, Orlowski M: A kidney-specific dopamine precursor. J Pharmacol Exp Ther 206:227–232, 1978 Casagrande C, Merlo L, Ferrini R, Miragoli G, Semeraro C: Cardiovascular and renal action of dopaminergic prodrugs. J Cardiovasc Pharmacol 14:S40–S59, 1989 Lee RM: Five years’ experience with g-L-glutamyl-L-dopa: A relatively renally specific dopaminergic prodrug in man. J Auton Pharmacol 10:S103–S108, 1990 Lee MR: Dopamine and the kidney, in Advances in Renal Physiology (chapt 7), New York, Liss, 1986, pp 218–247 Yusufi ANK, Murayama N, Gapstur SM, Szczepanska-Konkel M, Dousa TP: Differential properties of brush-border membrane vesicles from early and late proximal tubules of rat kidney. Biochim Biophys Acta 1191:117–132, 1994 Shimada H, Endou H, Sakai F: Distribution of g-glutamyl transpeptidase and glutaminase isozymes in the rabbit nephron. Jpn J Pharmacol 32:121–129, 1982 Armando I, Nowicki S, Aguirre J, Barontini M: A decreased tubular uptake of dopa results in defective renal dopamine production in aged rats. Am J Physiol 268:F1087–F1092, 1995 Wang Z-Q, Siragy HM, Felder RA, Carey RM: Intrarenal dopamine production and distribution in the rat: Physiological control of sodium excretion. Hypertension 29:228–234, 1997 Casson IF, Clayden DA, Cope GF, Lee MR: The protective effect of g-glutamyl-L-dopa on the glycerol treated rat model of acute renal failure. Clin Sci 65:159–164, 1983 Freestone S, Ellis R, MacDonald TM, Lee MR: Effect of prolonged intravenous infusion of gludopa in humans and a preliminary study of its oral bioavailability. Am J Hypertens 3:126S–129S, 1990 Cummings J, Matheson LM, Maurice L, Smyth JF: Pharmacokinetics, bioavailability, metabolism, tissue distribution and urinary excretion of g-L-glutamyl-L-dopa in the rat. J Pharm Pharmacol 42:242–246, 1990 Casson IF, Clayden DA, Cope GF, Lee MR: The protective effect of g-glutamyl-L-dopa on the glycerol treated rat model of acute renal failure. Clin Sci 65:159–164, 1983 Khraibi AA, Knox FG: Renal interstitial hydrostatic pressure during natriuresis in hypertension. Am J Physiol 255:R756–R759, 1988 Sacktor B, Rosenbloom IL, Liang CT, Cheng L: Sodium gradientand sodium plus potassium gradient-dependent L-glutamate uptake in renal basolateral membrane vesicles. J Membr Biol 60:63– 71, 1981 Lowry OH, Rosebrough NJ, Farr AL, Randall RJ: Protein measurements with the Folin phenol reagent. J Biol Chem 193:265– 275, 1951 Dousa MK, Tyce GM: Free and conjugated plasma catecholamines, DOPA and 3-O-methyldopa in humans and in various animal species. Proc Soc Exp Biol Med 188:427–434, 1988 Chen PS, Toribara TY, Warner H: Microdetermination of phosphorus. Anal Chem 28:1756–1758, 1956 Shah SV, Northrup TE, Hui YSF, Dousa TP: Action of serotonin (5-hydroxytryptamine) on cyclic nucleotides in glomeruli of rat renal cortex. Kidney Int 15:463–472, 1979 Stole E, Smith TK, Manning JM, Meister A: Interaction of g-glutamyl transpeptidase with acivicin. J Biol Chem 269:21435– 21439, 1994 Hruska KA, Klahr S, Hammerman MR: Decreased luminal membrane transport of phosphate in chronic renal failure. Am J Physiol 242:F17–F22, 1982 Eklo¨f A-C, Holtba¨ck U, Sundelo¨f M, Chen S, Aperia A: Inhibition of COMT induces dopamine-dependent natriuresis and inhibition of proximal tubular Na1,K1-ATPase. Kidney Int 52:742–747, 1997 Felder RA, Campbell T, Albrecht F, Jose PA: Dopamine inhibits
1842
43.
44.
45.
46.
de Toledo et al: Gludopa inhibits Na1-Pi cotransport
Na1-H1 exchanger activity in renal BBMV by stimulation of adenylate cyclase. Am J Physiol 259:F297–F303, 1990 Winaver J, Burnett JC Jr, Tyce GM, Dousa TP: ANF inhibits Na1-H1 antiport in proximal tubular BBM: The role of dopamine. Kidney Int 38:1133–1140, 1990 Jeffrey RF, MacDonald TM, Marwick K, Lee MR: The effect of carbidopa and indomethacin on the renal response to g-L-glutamyl-L-dopa in normal man. Br J Pharmacol 25:195–201, 1988 Spater HW, Poruchynsky MS, Quintana N, Inoue M, Novikoff AB: Immunocytochemical localization of g-glutamyltransferase in rat kidney with protein A-horseradish peroxidase. Proc Natl Acad Sci USA 79:3547–3550, 1982 Soares-Da-Silva P, Serrao MP, Vieira-Coelho MA: Apical and basolateral uptake and intracellular fate of dopamine precursor L-dopa in LLC-PK1 cells. Am J Physiol 274:F243–F251, 1998
47. Satoh T, Cohen HT, Katz AI: Different mechanisms of renal NaK-ATPase regulation by protein kinases in proximal and distal nephron. Am J Physiol 265:F399–F405, 1993 48. Takemoto F, Cohen HT, Satoh T, Katz AI: Dopamine inhibits Na-K-ATPase in single tubules and cultured cells from distal nephron. Pflu¨gers Arch Eur J Pharmacol 421:302–306, 1992 49. Felder CC, McKelvey AM, Gitler MS, Eisner GM, Jose PA: Dopamine receptor subtypes in renal brush border and basolateral membranes. Kidney Int 36:183–193, 1989 50. Schwartz IL, Shalatz LJ, Kinne-Saffran E, Kinne R: Target cell polarity and membrane phosphorylation in relation to the mechanism of action of antidiuretic hormone. Proc Natl Acad Sci USA 71:2595–2599, 1974 51. Lokhandwala MF, Amenta F: Anatomical distribution and function of dopamine receptors in the kidney. FASEB J 5:3023–3030, 1991