Introduction

Top Abstract Introduction Materials & Methods Results Discussion References

Inhibition of COX with nonsteroidal anti-inflammatory drugs usually results in vasoconstriction by eliminating vasodilator prostanoids. However, vasodilation can occur if production of TxA₂ and PGH₂ predominates as in angiotensin II-salt induced hypertension (Mistry and Nasjletti, 1988). We have reported that the response to COX inhibition with indomethacin in the rat isolated kidney can be influenced by the prevailing chloride concentration (Yin et al., 1995). The isolated kidney allows selective changes in chloride concentration to be examined for their influence on renal mechanisms while maintaining a normal sodium concentration and osmolality (Quilley et al., 1993). When the kidney was exposed to a high chloride concentration (117 mM), GFR and sodium excretion were depressed, and indomethacin produced an increase in GFR and natriuresis (Yin et al., 1995). Exposure to a low chloride concentration (87 mM) increased GFR and sodium excretion, both of which were reduced by indomethacin administration. Indomethacin was without effect when kidneys were exposed to a normal chloride concentration of 102 mM (Hilchey and Bell-Quilley, 1995).

These findings suggest that vasodilator-diuretic prostanoids predominate when chloride concentration is low and that vasoconstrictor-antidiuretic prostanoids predominate when chloride concentration is high. Thus, the response of the kidney to COX inhibition was a graded phenomenon conditioned by the chloride concentration. However, TxA₂, measured as immunoreactive TxB₂, did not increase in venous and urinary effluents of the rat kidney in response to hyperchloremia (Yin et al., 1995). We had predicted that TxA₂ would increase on the basis of the report of Bullivant et al. (1989). Because changes in TxA₂ concentrations at critical sites intrarenally, such as the glomerular afferent arteriole and the basolateral side of the proximal tubules, may not be reflected by changes in renal efflux of TxA₂, we could not eliminate TxA₂ as a mediator of the renal functional response to hyperchloremia. With regard to the other candidate mediator, PGH₂, we could not address its role in mediating the high-chloride effects because the highly labile PGH₂ rapidly degrades to stable products such as PGE₂ and PGF₂. To answer questions concerning the possible participation of TxA₂ and/or PGH₂ in mediating the renal functional effects of high chloride, pharmacological agents of demonstrated efficacy and selectivity were used.

We developed an experimental design that addressed the possible roles of both TxA₂ and PGH₂ in the renal response to hyperchloremia. We used the renal functional response to inhibition of COX with indomethacinnamely, increased GFR and natriuresisas the reference responses to hyperchloremia (Yin et al., 1995) to which corresponding responses produced by TxA₂ synthase blockade and TxA₂/PGH₂ receptor antagonists were compared. The first step was to determine whether TxA₂ mediates the renal functional effects of hyperchloremia. We eliminated TxA₂ because neither of two inhibitors of TxA₂ synthase modified the depressed GFR and antinatriuresis produced by high chloride. The next step was to address the contribution of PGH₂ to the hyperchloremic effect on renal function by blocking PGH₂ receptors. BMS 180291, the more selective antagonist of the TxA₂/PGH₂ receptor (Ogletree et al., 1993), did not increase GFR and sodium excretion produced by hyperchloremia, although the dose of BMS 180291 that we administered abolished the renal vasoconstrictor action of the TxA₂/PGH₂ receptor agonist, U 46619. These results excluded PGH₂ as a potential mediator of the high-chloride effect on renal function and provided additional evidence for excluding TxA₂ in this capacity. However, SQ 29548, the less selective antagonist of the TxA₂/PGH₂ receptor, had effects similar to those of indomethacin; it prevented the renal functional effects of hyperchloremia. The disparity between the two TxA₂/PGH₂ antagonists in reversing the depression of GFR and sodium excretion produced by hyperchloremia may be a function of one or both of two factors: 1) The antagonism produced by SQ 29548 is relatively nonselective; it attenuates the vascular actions of PGF₂, 8-epi-PGF₂ and 20-HETE in addition to antagonizing TxA₂ and PGH₂ (Ogletree et al., 1985; Takahashi et al., 1992; Laniado-Schwartzman et al., 1989). 2) Subtypes of TxA₂/PGH₂ receptors mediate different functional responses in a given cell type or tissue (Hirata et al., 1996; Meagher and FitzGerald, 1993). These subtypes differ in susceptibility to blockade by TxA₂/PGH₂ receptor antagonists and may account, in part, for the differential effects of SQ 29548 and BMS 180291 on the renal functional response to hyperchloremia.

	Materials and Methods

Top Abstract Introduction Materials & Methods Results Discussion References

Male Sprague-Dawley rats (300-400 g) were anesthetized with sodium pentobarbital (65 mg/kg, Arpo Pharmaceutical, Arcadia, CA). Heparin (1000 U/ml, Elkin Sinn Inc., Cherry Hill, NJ) and 0.5 ml mannitol (10% w/v, Sigma Corporation, St. Louis, MO) were injected i.v. as a 1:4 mixture for anticoagulation and to expand the ureter for cannulation by promoting a diuresis.

Animals were laparotomized, and the right kidney was isolated as described (Yin et al., 1995). The right ureter was cannulated with PE 10 coupled to PE 50 tubing to allow free exit of urine. The abdominal aorta and vena cava were ligated above the femoral bifurcation. The left renal artery was ligated. The right renal artery was then cannulated with a blunt 19-gauge needle via the mesenteric artery such that flow to the kidney was not interrupted. The vena cava above the kidney was then quickly ligated. The first 50 ml of blood-contaminated venous effluent was discarded, and the venous outflow was then directed to a reservoir from which it was pumped (Watson Marlow, model 5025, Wilmington, MA) through an 8-µm filter, oxygenated and heated to 37C°. The perfusate flow rate was adjusted to obtain a basal perfusion pressure of 90 mm Hg.

The values for RVR were 3.1 to 3.2 mm Hg/ml/min initially and 2.7 to 3.0 mm Hg/ml/min after 1 hr, i.e., at the conclusion of the experiments (P > .05). RVR in control groups did not differ from treatment groups. All RVR values decreased as a function of time, which we reported for the rat isolated perfused kidney under hyperchloremic (117 mM) conditions (Yin et al., 1995).

The perfusate was a modified Krebs-Henseleit buffer to which bovine serum albumin and Ficoll 70 were added as oncotic agents, as well as a mixture of amino acids to improve the functional indices of the experimental preparation, as described (Yin et al., 1995). Chloride concentration was 117 mM, which has been demonstrated to depress GFR and sodium excretion (Yin et al., 1995). Vitamin B₁₂ (50 µg/ml) was included for measurements of GFR (C.P. Quilley and McGiff, 1990).

Experimental Protocol

After starting renal perfusion, a 15 to 20-min equilibration period was observed, and renal function was examined during six 10-min clearance periods (figs. 1 and 2). Perfusate flow rate was determined from collections of venous effluent over 30 sec at the start and the end of each period. Urine was collected for the entire 10-min period, and the volume was determined gravimetrically. Pharmacological agents were added directly to the recirculating perfusate reservoir before initiation of arterial cannulation. Samples of urine and perfusate were taken for measurement of electrolytes and vitamin B₁₂. All samples were stored at 20C° pending colorimetric and electrolytic assays.

View larger version (K): [in this window] [in a new window]   Fig. 1.   Changes in GFR (panels A and D); fractional water excretion (% FE H2O, panels B and E) and fractional sodium excretion (% FE Na+, panels C and F) in control, rat isolated kidneys () compared with responses in the presence of 1 µM SQ 29548 () panels A, B and C), 1 µM BMS 180291 ( panels A, B and C), 5 µM OKY 046 ( panels D, E and F) and 10 µM indomethacin ( panels D, E and F).

View larger version (K): [in this window] [in a new window]   Fig. 2.   Changes in urine flow (panels A and D); sodium excretion rate (UNaV, panels B and E) and potassium excretion rate (UKV, panels C and F) in control, rat isolated kidneys () compared with responses in the presence of 1 µM SQ 29548 ( panels A, B and C), 1 µM BMS 180291 ( panels A, B and C), 5 µM OKY 046 ( panels D, E and F) and 10 µM indomethacin ( D, E and F).

Experimental groups of seven to eight kidneys, perfused with high chloride, were randomly assigned to five treatment groups: time vehicle control, BMS 180291 (1 µM), SQ 29548 (1 µM), OKY 046 (5 µM) and indomethacin (10 µM). Groups of six kidneys each were randomly used for either time- control or treatment with a second Tx synthase inhibitor, CGS 13080 (1 µM). Rats used for the Tx synthase inhibitor groups were pretreated 2 hr before surgery with an i.p. injection of either 20 mg/kg of OKY 046 or 2 mg/kg of CGS 13080.

The selection of the dose of BMS 180291 and SQ 29548, the TxA₂/PGH₂ receptor antagonists, was made on the basis of studies that examined inhibition of renal vasoconstrictor responses to graded bolus injections of the TxA₂/PGH₂ receptor agonist, U46619. Concentrations of 1 µM of BMS 180291 and SQ 29548 either abolished or greatly reduced the vasoconstrictor responses of kidneys to 0.03 to 1 µg U 46619, which was tested at the end of each experiment.

The ability of BMS 180291 to gain access to the tubular compartment was verified by Drs. Ogletree and Jemel of Bristol-Myers Squibb, Princeton, NJ, using HPLC quantitation of renal effluents. Levels in the urine were found to be approximately 25% of those observed in the perfusate.

The TxA₂ synthase inhibitor, OKY 046, was used in accordance with the reported inhibitory activity and the protocol described by Bullivant et al. (1989). The dosing schedule used for CGS 13080, which was based on previous experience (J. Quilley and McGiff, 1990), produced more than 80% inhibition of TxB₂ levels. The effectiveness of indomethacin inhibition of COX in our preparation has been described (Yin et al., 1995).

20-HETE Measurements and Renal Vascular Responses to 20-HETE

For measurements of 20-HETE, the right kidney was isolated as described previously (J. Quilley and McGiff, 1990) and perfused by means of a Watson-Marlow pump (model 502S) at a rate to maintain a basal perfusion pressure of 80 mm Hg. The perfusate was warmed (37°C), oxygenated (95% O₂/5% CO₂) Krebs-Henseleit buffer of the following composition (mM): NaCl 63.1, MgSO₄ 1.2, CaCl₂ 2.5, NaHPO₄ 1.2, KCl 4.4, dextrose 5.5 and NaHCO₃ 25. NaCl and Na acetate were then added to maintain a constant sodium concentration of 145 mM and chloride concentrations of 87 mM and 117 mM for the low- and high-chloride groups, respectively. The kidney was excised and the ureter cannulated in order to collect separate ureteral and venous effluents. This experimental preparation was used to obtain samples for 20-HETE analysis (fig. 3) and to define possible conversion of 20-HETE via COX to prostaglandin analogs (fig. 4).

View larger version (K): [in this window] [in a new window]   Fig. 3.   Renal release of 20-HETE during perfusion with high chloride (117 mM) and low chloride (87 mM). Venous effluents were collected at 5-, 15- and 30-min time-points. Solid columns: high chloride; open columns: low chloride. *P < .05.

View larger version (K): [in this window] [in a new window]   Fig. 4.   Effects of the inhibition of cyclooxygenase with indomethacin (10 µM) and blockade of TxA2/PGH2 receptors with either SQ (1 µM) or BMS (1 µM) on vasoconstrictor responses, expressed as increases in perfusion pressure (PP), to bolus doses of angiotensin II (10 ng), U 46619 (100 ng) and 20-HETE (10 µg) in isolated perfused kidneys. Vehicle control: solid column; Indo: open column; BMS 180291: diagonal column; SQ 29548: cross-hatched column. *P < .05.

20-HETE measurements. An equilibration period of 15 min was observed, followed by 2-min collections of venous and ureteral effluents obtained at intervals of 5, 15 and 30 min. Effluents were collected into ice-cold ethyl acetate, and a deuterium-labeled internal standard of 20-HETE (1 ng/ml) was added. Samples were then evaporated and washed three times with methanol. They were then dried, resuspended in methanol and subjected to reverse-phase HPLC (Hewlett-Packard 1050A). Samples were injected into an ultrasphere C18 column (250 × 4.6 mm, 5 µm, Beckman Instruments) and eluted with a linear gradient of 50% acetonitrile/water/acetic acid, 50% acetonitrile to pure acetonitrile in 10 min at a flow rate of 1 ml/min and a total run time of 20 min. The fraction containing 20-HETE (6.3 min-7.2 min) was collected, dried and resuspended in 100 µl of acetonitrile before derivitization. HETEs were esterified with pentafluorobenzyl bromide-nin-diisopropylethylamine (PFB 30 µl, DIPEA 30 µl, all from Sigma, St. Louis, MO) at room temperature for 30 min, followed by derivitization of hydroxyls with bis-trimethylsilyl-trifluoroacetamide and pyridine for 1 hr. Derivitized fractions were resuspended in iso-octane and injected into samples that were analyzed by HPLC and GC/MS for absolute concentrations of 20-HETE in the venous effluent.

Analysis of 20-HETE-induced renal vasoconstriction. Renal vascular responses to 20-HETE (10 µg) were obtained in the presence and absence of indomethacin (10 µM), BMS 180291 (1 µM) or SQ 29548 (1 µM) (fig. 4). Renal vascular responses to U 46619 (100 ng) were determined to verify the effectiveness of the PGH₂/TxA₂ receptor antagonists, whereas angiotensin II served as a negative control to exclude nonspecific effects of the antagonists on vasoconstrictor mechanisms.

Assays and Calculations

Perfusate and urinary electrolytes were measured using an automatic analyzer (Model 644, Ciba Corning, Medfield, MA). Electrolyte excretion rate was the product of urine flow and electrolyte concentration. GFR was calculated from colorimetric measurements at 550 nm of the concentration of vitamin B₁₂ in perfusate and urine (Model EL 309, Biotek Instruments, Windoshi, VT). Background readings at 630 nm were automatically subtracted. Fractional excretion of water and electrolytes was calculated by dividing the absolute excretion by the filtered load.

Statistical Analysis

Differences between groups were assessed by two-way analysis of variance, using the STATISTICA computer program (StatSoft, Tulsa, OK). Specific comparison between points was determined by using the Newman-Keuls post-hoc analysis for multiple comparisons with P < .05 taken to indicate statistical significance (Rosner, 1990).

	Results

Top Abstract Introduction Materials & Methods Results Discussion References

Kidneys were perfused throughout the experiments with high chloride concentrations (117 mM), which we have shown to depress GFR and sodium excretion (Yin et al., 1995); these values are similar to those previously reported by us for the rat isolated kidney under hyperchloremic conditions (Yin et al., 1995). The efficacy of the TxA₂/PGH₂ receptor antagonists was verified by measuring attenuation of the vasoconstrictor activity (registered as changes in renal perfusion pressure) of the TxA₂/PGH₂ mimic, U 46619. The renal pressor effect of U 46619 (1 µg) was reduced by SQ 29548 (1 µM) from a control value of 62 ± 9 to 26 ± 8 mm Hg (P < .05) and was abolished by BMS 180291 (1 µM). The doses of each TxA₂/PGH₂ antagonist, therefore, were sufficient either to attenuate greatly or to abolish a TxA₂/PGH₂-mediated effect on renal function.

Modification of hyperchloremic-induced changes in GFR. The initial GFRs were similar to those reported previously by us in the rat kidney perfused with high chloride (117 mM) and were depressed relative to those obtained under normochloremic conditions (Yin et al., 1995). GFR was increased, although not significantly, as a function of time in the control, hyperchloremic group, from an initial value of 0.6 ± 0.1 ml/min to a maximum of 0.9 ± 0.1 ml/min by periods 3 and 4 (fig. 1). Treatment with either indomethacin or SQ 29548 increased the GFR in periods 2 to 4 (fig. 1, A and D). In contrast to the positive effects of indomethacin and SQ 29548 on GFR (fig. 1, A and D), neither TxA₂ synthase inhibition with OKY 046 (fig. 1D) nor antagonism of the TxA₂/PGH₂ receptor with BMS 180291 (fig. 1A) affected the GFR, which remained at levels similar to those of the control group during the 1-hr period of the experiment.

Modification of hyperchloremic-induced changes in water and electrolyte excretion. In addition to increasing GFR, either inhibition of COX with indomethacin or blockade of the TxA₂/PGH₂ receptor with SQ 29548 also increased U_NaV and UV. Indomethacin doubled urine flow and potassium excretion rate and increased sodium excretion rate 3-fold as compared with those values for the time control group (fig. 2, D-F). The TxA₂/PGH₂ receptor antagonist, SQ 29548, produced an increase in salt and water excretion indistinguishable from that produced by indomethacin (fig. 2, A-C). In contrast to the effects of either indomethacin (fig. 2, D-F) or SQ 29548 (fig. 2, A-C), the more selective TxA₂/PGH₂ receptor antagonist, BMS 180291 (fig. 2, A-C), like the TxA₂ synthase inhibitor, OKY 046, (fig. 2, D-F), did not alter excretion rates of salt and water.

20-HETE concentrations in hyperchloremic vs. hypochloremic kidney. Renal release of 20-HETE increased in a time-dependent manner in kidneys perfused with high chloride (117 mM) as compared with kidneys perfused with low chloride (87 mM) (fig. 3). Renal release of 20-HETE was exclusively into the venous effluent; none was detected in the urinary effluent. After 15 and 30 min, rates of 20-HETE release into the venous effluent as quantified by GC/MS were 4.2 ± 1.0 ng/min and 5.2 ± 1.3 ng/min, respectively, in the high-chloride group as compared with 1.5 ± 0.4 ng/min and 2.5 ± 0.9 ng/min in the low-chloride group.

Renovascular actions of 20-HETE: Relationship to COX activity and TxA₂/PGH₂ receptors. The renal vasoconstrictor action of 20-HETE was analyzed in terms of possible transformation of 20-HETE by COX to prostaglandin endoperoxide and thromboxane analogs of 20-HETE (fig. 4). The renal vasoconstrictor action of 20-HETE (10 µg) was abolished by inhibition of COX with indomethacin (10 µM) as well as by blockade of the TxA₂/PGH₂ receptors with either BMS 180291 (1 µM) or SQ 29548 (1 µM). The renal vasoconstrictor effect of angiotensin II was unaffected by either inhibition of COX or antagonism of TxA₂/PGH₂ receptors, which demonstrates the absence of nonspecific effects of the inhibitors on vasoconstrictor mechanisms. The TxA₂/PGH₂ receptor antagonists, BMS 180291 and SQ 29548, prevented the renal vasoconstrictor response to U 46619, the TxA₂/PGH₂ mimic.

	Discussion

Top Abstract Introduction Materials & Methods Results Discussion References

The present study was occasioned by our published findings on the role of one or more COX products as mediators of the renal functional effects produced by changes in chloride concentration (Yin et al., 1995). In the earlier study, we found "a striking dependence of changes in renal function" on the prevailing chloride concentration. High chloride depressed GFR and U_NaV, which could be restored to levels found in normochloremic kidneys by inhibiting COX. These findings confirmed those of Wilcox et al. (1985) that hyperchloremia depressed renal function consequent to production of a COX-dependent mediator(s) in the canine kidney. The variable response of the kidney to inhibition of COX, therefore, could be predicted from the chloride concentration.

We designed this study to address potential mediators of the renal response to high chloride. Because we were primarily interested in intrinsic renal mechanisms evoked by hyperchloremia, we elected to continue our studies in the rat isolated kidney in order to exclude confounding systemic neural and hormonal influences, as well as blood-borne factors, and thereby to focus on intrarenal mechanisms. The rat isolated kidney was perfused at constant pressure with oncotic agents in a closed-circuit system, conditions that promoted reabsorption of greater than 99% of the filtered load of sodium while maintaining a GFR similar to that of the in situ blood-perfused kidney. An essential criterion for using the closed-circuit isolated kidney is that its response to hyperchloremia be identical to that of the in situ blood-perfused kidney (Wilcox et al., 1985) which is the case because GFR and sodium excretion are reduced on exposure to high chloride. We had regarded PGH₂ and/or TxA₂ as the probable mediators, although our initial study (Yin et al., 1995) failed to show a relationship between the renal efflux of either prostaglandin or TxA₂ and chloride-induced changes in renal function. TxA₂ mediation of the effects of high chloride on renal function was excluded because both inhibitors of TxA₂ synthase and a selective antagonist of the TxA₂/PGH₂ receptor, BMS 180291, failed to modify the renal response to high chloride. Thus, we did not confirm the findings of Wilcox et al. (1985) and Bullivant et al., (1989) that increased renal production of TxA₂ 1) accompanied the hyperchloremia and 2) mediated the renal functional effects of the latter. However, we did confirm the ability of SQ 29548 as well as indomethacin to attenuate the renal response to hyperchloremia. These differences between our studies and those of Wilcox et al. (1985) probably reflect different experimental conditions and species: rat isolated kidneys perfused with an artificial solution vs. in situ blood-perfused canine kidneys. We were also able to exclude a PGH₂-dependent mechanism on the basis of the failure of the specific antagonist of the TxA₂/PGH₂ receptor, BMS 180291, to modify the hyperchloremic-induced depression of GFR and sodium excretion. BMS 180291 was administered in a concentration that blocked the renal vasoconstrictor response to the TxA₂/PGH₂ mimic, U 46619. This lack of effect of BMS 180291 was surprising in view of the ability of SQ 29548, also an antagonist of the TxA₂/PGH₂ receptor, to prevent the effect of hyperchloremia on renal function. Indeed, the response of GFR and sodium excretion to SQ 29548 was virtually superimposable on the response to indomethacin.

Because the major candidates for mediating the renal functional effects of hyperchloremia, TxA₂ and PGH₂, were eliminated by the present study, a vasoconstrictor-antinatriuretic COX product remained to be identified. To address this issue, we considered that the ability of SQ 29548 to inhibit the renal functional effects of high chloride may reside in its relative non-selectivity as an antagonist of the TxA₂/PGH₂ receptor (Ogletree et al., 1985). First, however, we must consider the possibility that the efficacy of SQ 29548 derives either from its greater potency when compared with BMS 180291 or from a pharmacokinetic property of SQ 29548 that facilitates access to renal receptors in contrast to limited access of BMS 180291. The potency of BMS 180291 was evident in experiments designed to determine the doses of BMS 180291 and SQ 29548 required to inhibit the renal functional effects of hyperchloremia. During exposure to high chloride, the renal vasculature became progressively more sensitive to U 46619, the TxA₂/PGH₂ mimic, as we have described for other vasoconstrictor agonists (Quilley et al., 1993). By the end of each experiment, the heightened vasoconstrictor response to U 46619 was completely abolished by BMS 180291, whereas SQ 29548 attenuated the renal vascular action of U 46619 by more than 80%. Consequently, BMS 180291 is more effective as an antagonist of the TxA₂/PGH₂ receptor, because SQ 29548 did not produce a full blockade of the TxA₂/PGH₂ receptor.

The second question regarding the observed differences between SQ 29548 and BMS 180291 is based on a pharmacokinetic consideration: BMS 180291, unlike SQ 29548, did not affect high-chloride-induced antinatriuresis/antidiuresis because it was denied access to critical tubular sites. However, analysis of urine samples from control and BMS-treated kidneys indicated that BMS 180291 was present in relatively large amounts in the urine (personal communication, M.L. Ogletree). Thus, it is unlikely that the failure of BMS 180291 to produce similar increases in water and electrolyte excretion to SQ 29548 can be attributed to pharmacokinetic differences.

The ability of SQ 29548 to attenuate the renal functional response to hyperchloremia, as compared with the failure of BMS 180291, is in agreement with studies that have identified two subtypes of the TxA₂/PGH₂ receptor (Halushka et al., 1987; Furci et al., 1991; Hirata et al., 1996). These subtypes can be differentiated in platelets by their binding properties (reversible vs. irreversible) regarding TxA₂/PGH₂ receptor antagonists (Furci et al., 1991). The reversible site in platelets binds BMS 180291 and SQ 29548, whereas the irreversible site binds only SQ 29548 (Meagher and FitzGerald, 1993). These findings are complemented by the cloning of two isoforms of the human TxA₂ receptor, a placental (Hirata et al., 1996) and an endothelial (Raychowdhury et al., 1994 isoform), both of which are present in platelets. Moreover, subtypes of the TxA₂/PGH₂ receptor with distinguishing binding characteristics have also been described in the kidney (Folger et al., 1992). Recent studies support the presence in the kidney of at least two receptor subtypes that differ in their intrarenal localization: a renal vascular subtype (Abe et al., 1995) and a renal tubular subtype (Bresnahan et al., 1996). A provisional basis for interpreting the differences in the renal functional effects of SQ 29548 and BMS 180291, on the basis of the above studies, rests on the different affinities of these receptor antagonists for the renal tubular receptor subtype. Thus, to explain the present results, we suggest that both SQ 29548 and BMS 180291 bind to the TxA₂/PGH₂ renal vascular receptor subtype, because both antagonists inhibited the renal vasoconstrictor action of the TxA₂/PGH₂ agonist, U46619 (fig. 4), whereas only SQ 29548 either binds or gains access to the renal tubular receptor, because it, not BMS 180291, prevented the antidiuretic-antinatriuretic effects of high chloride (figs. 1 and 2). This interpretation must remain speculative until studies, both functional and molecular biological, become available that provide information about the basis of the postulated differential binding of TxA₂/PGH₂ receptor antagonists to subtypes of this receptor. For example, Abe et al. (1995) showed that cells transfected with the rat TxA₂/PGH₂ vascular receptor bound SQ 29548; however, these investigators did not include comparisons with BMS 180291. The ability of SQ 29548 vs. that of BMS 180291 to prevent hyperchloremia-induced depression of GFR can also be explained by the different distribution of TxA₂/PGH₂ receptor subtypes in the glomerulus. Thus, Bresnahan et al. (1996) found the renal tubular subtype to be most prominent in glomerular capillary loops, whereas Abe et al. (1995) found the renal arteriolar TxA₂/PGH₂ subtype to be prominent in glomerular mesangial cells. SQ 29548, which binds to both subtypes of the TxA₂/PGH₂ receptor, is more likely to influence GFR than BMS 180291 that binds to the mesangial subtype of Abe et al. (1995) rather than the glomerular capillary TxA₂/PGH₂ receptor subtype. The present study offers tentative conclusions regarding "the functional significance of the differences in the pattern of distribution" (Bresnahan et al., 1996) of the renal TxA₂/PGH₂ receptor subtypes: a TxA₂/PGH₂ receptor antagonist binding to both receptor subtypes (SQ 29548) would affect renal tubular as well as vascular function.

An essential characteristic for identifying the unknown mediator(s) is that inhibition of COX prevented its renal functional effects (figs. 1 and 2). Because several cytochrome P450-dependent AA products can be further metabolized by COX, a product of this pathway should be considered a potential candidate for mediating the high-chloride effects, particularly 20-HETE, which can be metabolized by COX to vasoconstrictor analogs of TxA₂ and PGH₂ (Laniado-Schwartzman et al., 1989; Hill et al., 1992). Indeed, the ability of 20-HETE to constrict the rat renal vasculature is COX-dependent, as seen in figure 4. Further, the renal vasoconstrictor response to the 20-HETE metabolite of COX is inhibited by both BMS 180291 and SQ 29548. This demonstration supports the candidacy of a 20-HETE metabolite of COX as a mediator of the renal functional effects of hyperchloremia and confirms the findings of Escalante et al. (1989) that the vasocontractile action of 20-HETE can be blocked either by inhibition of COX or by antagonism of the TxA₂/PGH₂ receptor. In addition, 20-HETE is a major product of AA metabolism in the kidney (Schwartzman et al., 1986), being produced by both renal tubules (Escalante et al., 1991) and blood vessels (Harder et al., 1995; McGiff, 1991). Finally, high chloride in the renal perfusate caused a 2-fold increase in efflux of 20-HETE from the kidney as compared with low chloride conditions of renal perfusion (fig. 3). This demonstration is essential to the proposed role of a COX-dependent 20-HETE metabolite as a mediator of the renal functional changes produced by hyperchloremia.

Given the information available, we conclude that a prostaglandin endoperoxide analog of 20-HETE is a leading candidate for mediating the renal functional effects of hyperchloremia and that this possibility has implications for mechanisms that regulate glomerular filtration. For example, the negative effects of hyperchloremia on renal blood flow and GFR may be related to activation of tubulo-glomerular feedback by Cl acting on the mascula densa (Schnermann et al., 1976). A significant component of the renal vascular mechanism, which is the effector limb of tubulo-glomerular feedback, has been proposed to be mediated by 20-HETE production in the preglomerular microcirculation (Harder et al., 1995). Zou et al. (1994) have shown that 20-HETE is a principal product of preglomerular microvessels, at which site the P450 arachidonate metabolite contributes critically to the myogenic response of renal arterioles and, thereby, to the autoregulation of RBF and possibly to tubulo-glomerular feedback.