Introduction

Abstract Introduction Materials & Methods Results Discussion References

Phosphonates serve different functions. Arylphosphonates inhibit protein-tyrosine and serine/threonine phosphatases (Kole et al., 1995; Segal et al., 1996). Phosphonocarboxylates are antiviral (Chrisp and Clissold, 1991; Eriksson and Öberg, 1984) and inhibit intestinal and renal phosphate transport (Loghman-Adham et al., 1987; Szczepanska-Konkel et al., 1986). Diphosphonates inhibit bone reabsorption (Fleisch, 1983), are used as ^99mTc complexers in scintigraphic bone scan (for literature, see Palmer et al., 1992) and have anti-inflammatory properties (Dunn et al., 1993). The phosphonocarboxylate compounds, phosphonoformate (foscarnet) and phosphonoacetate, are reabsorbed from the gut, freely filtered in the glomerulus and act on the luminal side of the proximal tubule to inhibit competitively sodium-dependent phosphate reabsorption (Van Scoy et al., 1988). They are transported in cotransport with Na⁺ by the phosphate transporter in intestinal brush-border membranes (Tsuji and Tamai, 1989). They bind in a Na⁺-dependent fashion to the phosphate transporter in renal brush-border membranes, but as far as phosphonoformate (foscarnet) is concerned, are not translocated, although some degree of uptake cannot be excluded (Szczepanska-Konkel et al., 1987). Diphosphonates are secreted by the kidney (Troehler et al., 1975; Lin et al., 1992) by an as yet uncharacterized renal transport system. After the multispecificity of the organic anion transporters in the proximal renal tubule (PAH transporter, dicarboxylate transporters, sulfate transporters) was amply documented and after the molecular features of substrates which interact with the different transporters were investigated extensively (for review, see Ullrich 1997), it was a challenge to find out whether pharmacologically applied and related alkyl/arylphosphonates and diphosphonates also fit in the predicted scheme of interaction. Since the crucial molecular parameters were known, we tested from a pool of available compounds those which were necessary to check our hypothesis. Furthermore it was not known to what extent these compounds interact and inhibit the luminal Na⁺/P_i reabsorption in the proximal renal tubule in situ under standard conditions. So far all kinetic Na⁺/P_i studies had been performed with brush-border membrane vesicles. It was found that in the submillimolar range only phosphonoformate (foscarnet), etidronate and clodronate inhibit luminal phosphate transport. As predictable from previous structure-activity studies for the contraluminal PAH, sulfate and dicarboxylate transporter, the phosphonocarboxylates interacted with these transporters according to their hydrophobicity and charge distribution. Among the diphosphonates tested, only etidronate and clodronate have a moderate affinity to the sulfate transporter, whereas the aminodiphosphonates have no (or only a low) affinity to the contraluminal anion transporters.

	Materials and Methods

Abstract Introduction Materials & Methods Results Discussion References

The experiments were performed on male Wistar rats (Wistar Kirchborchen, raised germ-free in the Max Planck Institut für Hirnforschung, Frankfurt, Germany), with 180 to 200 g body weight, which were fed Altromin standard diet with free access to water. For the phosphate transport experiments the animals fasted over night. They were anesthetized by injection of thiobutabarbital (Inactin, Byk-Gulden, Konstanz, Germany), 120 to 150 mg/kg b.wt. i.p., and placed on a heated operating table (thermostat control at 37°C). An incision was made on the left flank. The kidney was separated from the surrounding fascia, and after the capsule had been removed, the kidney was immobilized in a plastic cup, resting on cotton wool covered by paraffin oil heated to 37°C.

The luminal apparent K_i values against P_i were evaluated by the luminal stop-flow tubular microperfusion method as described in previous studies from this laboratory (Sheridan et al., 1983). The tubular lumen was punctured with three different sharpened glass pipettes (tip diameter, 5-8 µm): one containing blue castor oil to block the flow of ultrafiltrate in the proximal tubule, a second pipette in close proximity to the first containing HCO₃ steady state solution with [³²P_i], 0.01 mmol/l, and [³H]inulin as volume marker splitting the oil blockade column when injected and a third pipette some distance from the others, but in the same tubule, for collecting the steady state solution after 3 s contact time (acoustic signal). The procedure for evaluating the contraluminal apparent K_i values for PAH, sulfate and succinate has been described previously (Ullrich and Rumrich, 1990; Ullrich et al., 1988). The renal artery and vein were clamped for each measurement of the disappearance rate of radiolabeled substances. The proximal convoluted tubules then collapsed as the luminal fluid was reabsorbed, whereas glomerular filtration ceased. Immediately thereafter, a thick superficial capillary was impaled by an oil-filled pipette (tip diameter, ~6 µm) for sample collection. At a distance of 100 to 140 µm from this glass pipette another blood vessel was punctured with a filling pipette (tip diameter, ~8 µm). A rapid injection of isotonic solution was administered through the filling pipette, which contained 0.1 mmol/l [³H]PAH and [¹⁴C]inulin or 0.01 mmol/l [³⁵S]sulfate and [¹⁴C]inulin or 0.15 mmol/l [¹⁴C]succinate and [³H]inulin as extracellular space marker. After 2, 4 or 1 s as checked by acoustic signal, the test solution was withdrawn into the sampling pipette. The radioactivity was measured in a Betamatik 1 Contron scintillation counter with Picofluor 15 (Packard, Frankfurt, Germany) as scintillation fluid. Because the control values of different animals varied, the experimental values were normalized to the respective controls by multiplying the values of each animal x by a factor: control x/control n, where n is the mean of more than 50 experiments. In the [³²P_i] experiments the HCO₃ steady state solution for the luminal perfusion contained (in mmol/l): NaCl, 105; NaHCO₃, 25; KCl, 4.0; CaCl₂, 1.5; MgCl₂, 1.0; raffinose, 31; and was gassed with 95% O₂/5% CO₂. The capillary perfusate for the [³H]PAH and [³⁵S]sulfate experiments contained (in mmol/l): Na gluconate, 154; KCl, 4; and was gassed with pure oxygen. The capillary perfusate for the [¹⁴C]succinate experiments contained (in mmol/l): NaCl, 130; NaHCO₃, 25; KCl, 4.0; CaCl₂, 1.5; MgCl₂, 1.0; gassed with 95% N₂/5% CO₂. The pH was set to 7.4. All substances added replaced an equivalent amount of Cl or gluconate so that the osmolarity remained constant. Apparent K_i values were calculated as described for luminal transporters by Sheridan et al. (1983) and for contraluminal transporters by Fritzsch et al. (1984) with a computer program and applying one or two concetrations of inhibitory substrates. Apparent K_i values were used as operational values, because competitive inhibition was assumed but not explicitly proven (for discussion, see Ullrich et al., 1991). All results are reported as the mean ± S.E.M. Statistical analysis was performed by analysis of variance followed by unpaired t test.

The source and specific activities of the radioisotopes were as follows: [³H]PAH, 6.8 Ci/mmol; [³⁵S]sulfate, 1400 Ci/mmol; [¹⁴C]inulin, 15 mCi/g; [³H]inulin, 164 mCi/g, and [³²P_i] carrier-free were obtained from Du Pont (NEN), Dreieich, Germany. [¹⁴C]Succinate with a specific activity of 111 mCi/mmol as obtained from Amersham Buchler (Frankfurt/Main, Germany). The sources of the tested phosphonates are listed in the legend to the table.

	Results

Abstract Introduction Materials & Methods Results Discussion References

Luminal P_i transporter. In 18 rats, which fasted overnight, time-dependent disappearance of luminal ³²P_i at 0.01, 0.1 and 1.0 mmol/l starting concentration and 1, 2, 3 and 5 s contact time was measured. A quasilinear decrease of luminal P_i concentration was seen for the first 3 s. An Eadie-Hofstee plot of the 2- and 3-s values gave a K_m,l,Pi of 0.41 mmol/l and a V_max value of 0.54 pmol cm¹ s¹.

Contraluminal PAH transporter. Among the three phosphonate groups tested only those compounds interacted with the PAH transporter which have one electronegative ionic charge and a satisfactory hydrophobic back bone, i.e., phenylphosphonate app. 5.4 mmol/l; 2-hydroxy-5-nitrobenzylphosphonate app. 0.9 mmol/l; 2-naphthylmethylphosphonate app. 0.8 mmol/l; 2-naphthylhydroxymethylphosphonate app. 1.3 mmol/l; and 2-naphthyldifluoromethylphosphonate app. 4.4 mmol/l. As expected, 4-phosphonobutyrate, in which three hydrophobic methylene groups are interposed between the negative ionic charges, also interacted with the PAH transporter. In this regard, the phosphonocarboxylate series resembles the dicarboxylate series in which the affinity to the PAH transporter increases 27-fold from succinate to glutarate (Ullrich et al., 1987a).

Contraluminal sulfate transporter. The prerequisite of a substrate for the interaction with the contraluminal sulfate transporter is an electronegative charge accumulation (Ullrich et al., 1985c; Fritzsch et al., 1989). This condition is evidently fulfilled by the following compounds: chlorofluoromethylphosphonate 5.6 mmol/l; 2-hydroxy-5-nitrobenzylphosphonate 2.9 mmol/l; 2-naphthylmethylphosphonate 0.5 mmol/l, 2-naphthylhydroxymethylphosphonate 3.2 mmol/l; 2-naphthyldifluoromethylphosphonate 0.35 mmol/l; and phosphonoacetate 3.7 mmol/l. It is also fulfilled by those diphosphonates which carry an hydroxy or two chloro groups on the interposed methylene group, i.e., etidronate 3.1 mmol/l; clodronate 1.5 mmol/l; pamidronate 7.2 mmol/l; and alendronate 9.9 mmol/l. The two compounds that have an amino group in their molecule, i.e., pamidronate and alendronate, have a lower affinity to the sulfate transporter than the two other compounds, i.e., etidronate and clodronate. The reason that naphthyl groups per se, as in 2-naphthylmethylphosphonate, promote interaction with the sulfate transporter will be discussed below.

Contraluminal dicarboxylate transporter. Among the 14 compounds tested against the contraluminal dicarboxylate transporter only 3-phosphonopropionate and 4-phosphonobutyrate inhibited contraluminal succinate uptake (app. 4.2 and 7.0 mmol/l, respectively). These affinities are 40 times lower than the affinities of the corresponding dicarboxylates, i.e., succinate ( 0.11 mmol/l) and glutarate ( 0.16 mmol/l) (Ullrich et al., 1984a). With 2-phosphonobutyrate and phosphonoacetate the distance between the negative ionic charges of the carboxylic and the phosphono group is apparently too small for interaction with the contraluminal dicarboxylate transporter. The fact that the distance of the negative ionic charges was too short might also be the reason that the tested diphosphonates did not interact with the contraluminal dicarboxylate transporter. However, because of their bulkiness two phosphono groups might also not interact sufficiently with the dicarboxylate transporter even when their distances are all right.

	Discussion

Abstract Introduction Materials & Methods Results Discussion References

Phosphonoformate (foscarnet) inhibits phosphate reabsorption in the mammalian kidney (Szczepanska-Konkel et al., 1986; Van Scoy et al., 1988). Inhibition of phosphate transport was also seen in rat brush-border membrane vesicles (app.K_i, 0.46 mmol/l) (Szczepanska-Konkel et al., 1986), and in agreement with this, in our preparation of microperfused rat proximal tubules ( 0.42 mmol/l) (table 1). This K_i value is in the same range as the K_m value of phosphate transport (0.41 mmol/l) in our experimental animals which fasted overnight. It would be very interesting to see, whether phosphonoformate (foscarnet) is transported in our setup or whether it behaves as in brush-border membrane vesicles where only binding, but no transport was observed (Szczepanska-Konkel et al., 1987). Unfortunately radiolabeled phosphonoformate was not at our disposal to test this. We could also confirm two other findings of Dousa and his group (Szczepanska-Konkel et al., 1986), namely that the simple arylphosphonates (dihalogen phosphonate and phenylphosphonate as examples) have no or only a low inhibitory potency against phosphate transport and that the affinity of the phosphonocarboxylates to the phosphate transporter quickly vanishes when the distance between the phosphonate and the carboxy group becomes larger.

The diphosphonates inhibit phosphate transport in brush-border membrane vesicles as well (Szczepanska-Konkel et al., 1986; Tenenhouse et al., 1980). Szczepanska-Konkel et al. (1986) found an inhibitory potency of ethanehydroxydiphosphonate (etidronate) that was between the inhibitory potency of phosphonoformate and phosphonoacetate. This agrees with our findings (table 1). An acute effect of diphosphonates on renal P_i excretion was not observed (Walton et al., 1974). Furthermore, a different effect of chronic ethanehydroxydiphosphonate application was seen in man, in which it augments the maximal transport capacity of renal phosphate reabsorption (Walton et al., 1975), and in rat, in which it diminishes it (Stoll et al., 1980). It remains to be explained why phosphonoformate and ethanehydroxydiphosphonate have a similar inhibitory potency against phosphate transport in brush-border membrane vesicles, whereas phosphate transport is acutely inhibited only by phosphonoformate and not by ethanehydroxydiphosphonate in the intact kidney. The reason for this may lie in the fact that the two compounds themselves are handled differently by the different transport mechanisms in the proximal renal tubules.

The interaction of the tested alkyl- and arylphosphonates with both the contraluminal PAH and sulfate transporters could be predicted from our previous studies on the specificity of these transport systems (Ullrich et al., 1985a, 1987b, 1988). Chlorofluoromethylphosphonate and ethylphosphonate do not interact with the contraluminal PAH transporter because they are not hydrophobic enough. Chlorofluoromethylphosphonate, however, interacts with the contraluminal sulfate transporter because of negative charge accumulation. The compounds that carry a benzene or naphthalene group are fairly hydrophobic and therefore fulfill a prerequisite for interaction with the PAH transporter. Why these compounds interact also with the contraluminal sulfate transporter is not so easy to interpret because the benzene and naphthalene ring structures can act as hydrophobic moieties as well as -electron negatively charged entities (Mecozzi et al., 1996). Nevertheless, we have seen similar interactions with correspondent benzoate, benzaldehyde and naphthylcarboxylate analogs. Thus, phenylphosphonate behaves similarly to benzoate (Ullrich et al., 1988), 2-hydroxy-5-nitrobenzylphosphonate behaves similarly to 2-hydroxy-5-nitro-substituted benzoate, benzaldehyde and benzenesulfonate (Ullrich et al., 1985a, 1988) and the 2-naphthylphosphonates behave similarly to 2-naphthylcarboxylate (Ullrich et al., 1987b).

The interaction of the tested phosphonocarboxylates and methylenediphosphonates with the contraluminal PAH transporter can be readily interpreted. All tested substances of these two classes of compounds have their two anionic charges too close together to interact with the PAH transporter except 4-phosphonobutyrate, which reaches the 4 Å hydrophobic distance that is necessary for interaction (Fritzsch et al., 1989).

The interaction of the tested phosphonocarboxylates and methylenediphosphonates with the contraluminal sulfate transporter is more complex to interpret. First, how is the interaction of similarly structured aliphatic dicarboxylates and disulfonates with the contraluminal sulfate transporter (Ullrich et al., 1985b, 1987a)? Among the aliphatic dicarboxylates only oxalate with no interposed methylene group interacts with the contraluminal sulfate transporter but not the longer dicarboxylates, in which one or more methylene groups are interposed (Ullrich et al., 1987a). On the contrary, aliphatic disulfonates tolerate one or two methylene groups for interaction with the sulfate transporter (methanedisulfonate, ethanedisulfonate; Ullrich et al., 1985b). In the present study phosphonoformate (foscarnet) surprisingly did not interact, but phosphonoacetate, in which one methylene group is interposed, did interact. Two interposed methylene groups as in 3-phosphonopropionate again prevent interaction. Methanediphosphonate in which one methylene group is interposed between the two phosphonate groups did not interact, whereas all tested compounds that had a hydroxy group on the interposed methylene group interacted. A direct analogy to this finding does not exist in our previous studies. But one may suppose that the partial electronegative charge or the hydrogen bond-forming capability of the hydroxy group might promote the observed interaction.

Second, by what as yet unidentified renal transport system(s) are the diphosphonates secreted? Certainly not by the contraluminal PAH transport system (table 1), which was in clearance and kidney slice studies already shown for etidronate, clodronate and pamidronate (Troehler et al., 1975, 1985; Lin et al., 1992) and not by the contraluminal dicarboxylate transport system (table 1). The most likely transport system for renal secretion of these diphosphonates is the contraluminal sulfate transport system (table 1). This suggestion, however, must be corroborated by transport studies with labeled diphosphonates and more or less specific inhibitors of the sulfate transport system (Ullrich et al., 1992). Finally, one should keep in mind that not only the renal contraluminal sulfate transporter interacts with sulfate, phosphate and bi-anionic compounds (Ullrich et al., 1984b, 1985b, 1987a), but also a structurally unrelated dicarboxylate transporter in yeast mitochondria (Kakhniashvili et al., 1997).

The contraluminal sulfate transporter by which diphosphonates seem to be secreted in the kidney was recently identified as SAT 1 (Markovich et al., 1994), originally cloned from hepatocytes (Bissig et al., 1994). By linkage disequilibrium mapping in patients with osteochondrodysplasia, Hästbacka et al. (1994) found mutations in the gene of the diastrophic dysplasia sulfate transporter which has strong amino acid similarities with SAT 1. Diminished transport of sulfate into diastrophic dysplasia chondrocytes may be responsible for the formation of undersulfated proteoglycans, with consecutively defective organization of collagen fibrils and bone formation. It is possible that the therapeutic active diphosphonates tested in this study are transported into osteoclasts by the diastrophic dysplasia sulfate transporter, in which they inhibit bone turnover. This must be verified by future experiments, however.