Introduction

Top Abstract Introduction Materials & Methods Results Discussion References

NDPS¹ (fig. 1) was initially developed as an agricultural fungicide (Fujinami et al., 1972). In spite of promising activity, subsequent testing revealed that NDPS was a nephrotoxicant in rats after either acute or chronic administration (Sugihara et al., 1975; Rankin, 1982; Barrett et al., 1983). After acute exposure, NDPS produces renal proximal tubular necrosis that is characterized by increased BUN levels, elevated kidney weights, diuresis, proteinuria, glucosuria, and decreased uptake of organic ions by renal cortical slices (Rankin, 1982). The onset of toxicity is rapid, and renal functional changes can be observed within 3 hr after administration of NDPS (Rankin et al., 1984). Although concerns regarding its toxicity have precluded the use of NDPS as an agricultural chemical, it was proposed as a useful model nephrotoxicant (Sugihara et al., 1975). Therefore, a careful study of its in vivo disposition could contribute to our understanding of chemically induced nephrotoxicity.

View larger version (17K): [in this window] [in a new window]   Fig. 1.   Metabolic scheme for NDPS. Routes A and B represent two possible pathways for the production of 2-NDHSA and 3-NDHSA from NDPS (see Discussion for details). NDHPS, N-(3,5-dichloro-4-hydroxyphenyl)succinimide.

The mechanism of NDPS-mediated nephrotoxicity remains unknown; however, hepatic biotransformation is required to produce kidney damage (Rankin et al., 1986, 1987, 1988; Nyarko and Harvison, 1995; Nyarko et al., 1997). After oral dosing, NDPS (0.4 mmol/kg) was converted to multiple urinary metabolites in rats (Ohkawa et al., 1974). Less than 1% of the excreted radioactivity was present in urine as the parent compound. Five metabolites were separated by TLC, but approximately 20% of the total urinary radioactivity remained at the origin of the TLC plates. Using GC/MS, three of the resolved urinary metabolites were identified as NDPSA, DMA, and 2-NDHSA. The remaining two urinary metabolites, comprising approximately 34% of the resolved radioactivity, could not be fully characterized. Blood, tissues, and feces were not evaluated for the presence of NDPS metabolites. Thus, the in vivo metabolic profile of NDPS has not been fully elucidated. Furthermore, nothing is known about NDPS metabolism at early time points or with different doses.

The studies described in this report were therefore designed to reevaluate NDPS metabolism in vivo. [¹⁴C]NDPS was administered to male Fischer 344 rats at three different doses, i.e. 0.2 mmol/kg (nontoxic dose), 0.4 mmol/kg (threshold toxic dose), and 0.6 mmol/kg (toxic dose). The metabolic profile in urine, feces, blood, and several organ homogenates was determined by HPLC. Nephrotoxicity was assessed by measuring BUN levels, urine volume, and urine contents. Covalent binding of NDPS-derived radiolabel to liver and kidney proteins was also measured. NDPS was rapidly metabolized, which could help explain the early onset of nephrotoxicity. Furthermore, we found that kidney damage was correlated with a shift toward increased oxidative metabolism at higher doses of NDPS.

	Materials and Methods

Top Abstract Introduction Materials & Methods Results Discussion References

Materials and Animals. NDPS, [¹⁴C]NDPS (specific activity, 0.65 mCi/mmol; radiochemical purity, 99%), and NDPS metabolites were synthesized and purified as previously described (Fujinami et al., 1972; Shih and Rankin, 1989; Griffin et al., 1996). -Glucuronidase, glucose kit 510-A, urea nitrogen kit 640-A, and Ponceau S were obtained from Sigma Chemical Co. (St. Louis, MO). Sep-Pak cartridges were purchased from Waters (Milford, MA). Male Fischer 344 rats (150-175 g) were purchased from Charles River Laboratories (Wilmington, MA). Animals were housed in stainless steel hanging cages, with a 12-hr light/dark cycle, at approximately 22°C and 45-50% relative humidity. Food (laboratory rodent diet 5001; PMI Foods, St. Louis, MO) and water were freely available unless otherwise noted. Rats were allowed a 1-week acclimation period before use in any experiments.

Equipment. HPLC analyses were conducted using a Beckman system with a model 421 controller, two model 114M pumps, and a model 165 variable-wavelength detector set at 254 nm. The detector sensitivity was set at 0.1 absorbance units full scale. Chromatograms were printed on a Hewlett Packard model HP 3396A integrator. Fractions were collected with an ISCO Retriever III fraction collector. ¹H NMR analyses were conducted with a General Electric (Fremont, CA) GN-500 NMR spectrometer. Chemical shifts are reported in ppm downfield from tetramethylsilane.

Animal Treatments. Male Fischer 344 rats were randomly assigned to treatment groups (three or four rats/group). The rats were administered [¹⁴C]NDPS (approximately 10 µCi, ip, in corn oil) at 0.2, 0.4, or 0.6 mmol/kg. Control animals received corn oil only (3 ml/kg). After administration of NDPS, the rats were immediately placed in plastic metabolism cages. The animals were kept in the cages for 3 hr (0.6 mmol/kg only) or for 24 hr (0.2, 0.4, and 0.6 mmol/kg), during which time they were fasted but were allowed free access to water. During this period, urine and feces were collected. A final blood sample was collected by cardiac puncture under methoxyflurane anesthesia, and the animals were then sacrificed by cervical dislocation while still anesthetized. To measure tissue distribution and metabolites, the liver, kidneys, lungs, heart, testes, adrenal glands, and brain were removed from the carcasses of the 3-hr treatment groups. Because of the low levels of radioactivity, only livers and kidneys were removed from the carcasses of the 24-hr treatment groups.

Sample Preparation. Urine samples from each animal were divided into 0.5-ml aliquots for preparation. One aliquot of urine from each rat was precipitated with 1.5 ml of ice-cold acetonitrile and was then centrifuged at 2000 rpm. The supernatant was removed, passed through a C₁₈ Sep-Pak cartridge, and frozen until HPLC analysis (see below). A second aliquot of urine from each rat was incubated with -glucuronidase (15,000 units/ml) for 24 hr at 37°C and was then treated as described above. Feces (24-hr time point only) were weighed and then homogenized in approximately 30 ml of 25 mM Tris buffer (pH 7.4). A sample was removed for liquid scintillation counting, and the remainder was frozen for later analysis. After thawing, a 2.0-ml aliquot of each fecal sample was precipitated with 6.0 ml of cold acetonitrile and then treated as the urine samples. The fecal pellet from the centrifugation was washed twice with 1.0 ml of acetone and repelleted. Finally, the pellet was digested in 2.0 ml of 1.0 M NaOH and 1.0 ml of dimethylsulfoxide, and the digest was counted to determine nonextractable radioactivity.

HPLC Analysis. All samples (see above) were analyzed by the HPLC assay described previously (Griffin et al., 1996). The mobile phase (flow rate, 1.7 ml/min) consisted of a gradient between acetonitrile containing 3% tetrahydrofuran (solvent A) and water containing 2% acetic acid (solvent B). Analyses were conducted with a Beckman 5-µm C₁₈ column (4.6 mm × 25 cm) protected by a Waters Guard-Pak precolumn module containing µBondapak C₁₈ precolumn inserts. Using these HPLC conditions, NDPS and its metabolites showed the following retention times: NDHPSA, 2.9 min; N-(3,5dichloro-4-hydroxyphenyl)succinimide, 3.8 min; 2-NDHSA, 8.4 min; NDHS, 9.7 min; 3-NDHSA, 11.9 min; DMA, 13.7 min; NDPS, 16.1 min; NDPSA, 17.6 min (Griffin et al., 1996).

Nephrotoxicity Studies. Nephrotoxicity in the rats was assessed 24 hr after dosing. A blood sample was obtained by cardiac puncture, under methoxyflurane anesthesia, immediately before sacrifice. BUN values were assessed spectrophotometrically (kit 640-A; Sigma). The total volume of urine excreted over 24 hr was measured. Urinary protein content was measured by the method of Pesce and Strande (1973). The concentration of glucose in the urine was measured using a commercially available kit (kit 510-A; Sigma).

Protein Binding. After thawing, 2.0-ml aliquots of the 24-hr liver and kidney homogenates (see above) were added to 6.0 ml of cold acetonitrile, the resulting suspension was centrifuged at 2000 rpm for 20 min, and the supernatant was discarded. Protein binding was determined using a variation of the method used by Newton et al. (1983a). Briefly, the pellet was thoroughly broken up, washed in a series of solvents, and repelleted after each wash. In order, the solvents were 80% aqueous methanol (three washes), 0.4 M trichloroacetic acid, 3:1 ethanol/ethyl ether (two washes), and 80% aqueous methanol (two washes). The series of solvent washes and centrifugations was repeated until the activity of 1.0 ml of the supernatant was less than twice the background level (i.e. <60 dpm). The pellet was then digested in 250 µl of 1.0 N NaOH, and a 150-µl aliquot of the solution was counted. Protein concentrations were measured by the method of Lowry et al. (1951).

Metabolite Isolation and Identification. To isolate and identify the urinary metabolites, a single male Fischer 344 rat was dosed with NDPS (0.8 mmol/kg) and placed in a metabolism cage, and urine was collected for 48 hr. This dosage and time point were chosen to maximize recovery of the metabolites. The urine sample was centrifuged for 15 min at 500 rpm, diluted with methanol (1:1, v/v), vortex-mixed, and recentrifuged. Methanol was removed using a SpeedVac apparatus (Savant Instruments, Farmingdale, NY), and the remaining liquid was passed through a C₁₈ Sep-Pak cartridge (Waters). NDPS metabolites were separated on a Rainin Microsorb C₁₈ column, using a mobile phase (flow rate, 1.2 ml/min) consisting of water containing 1% trifluoroacetic acid (solvent A) and acetonitrile (solvent B). The following linear gradient program was used: starting conditions, 90% solvent A/10% solvent B; 0-15 min, 25% solvent A/75% solvent B; 15-17 min, 0% solvent A/100% solvent B; 17-20 min, 90% solvent A/10% solvent B. Peaks were collected, and the mobile phase was removed using the SpeedVac concentrator. ¹H NMR (500 MHz) analyses were performed with a General Electric (Fremont, CA) GN-500 spectrometer, in acetone-d₆.

Statistics. The data were analyzed by the Student t test, one-way analysis of variance, or the corresponding nonparametric tests. If significance was obtained in the analysis of variance, differences among the individual groups were determined by using the Student-Newman-Keuls test. A 5% level of significance was used for all statistical tests.

	Results

Top Abstract Introduction Materials & Methods Results Discussion References

Three known NDPS metabolites (NDPSA, 2-NDHSA, and DMA) and a previously unreported in vivo metabolite (3-NDHSA) were present in the 3-hr urine sample from a rat that had received 0.6 mmol/kg [¹⁴C]NDPS (fig. 2). NDPS was not detected in urine. At this early time point, the total recovered urinary activity was 5.7 ± 1.8% of the administered dose. Collectively, the nephrotoxic metabolites 2-NDHSA and 3-NDHSA (Rankin et al., 1988, 1989) accounted for a significantly greater amount (58.7 ± 0.8%) of the urinary radioactivity than did NDPSA (34.9 ± 1.1%) (fig. 3). A small amount of radioactivity (approximately 4%) (fig. 3, Unk) eluted in the solvent front (2-6 min) and did not correspond to any distinct peaks on the UV chromatograms. To hydrolyze possible glucuronide conjugates, the urine samples were treated overnight with -glucuronidase. However, there were no significant differences in the HPLC chromatograms (results not shown) or profiles of urinary metabolites before and after -glucuronidase treatment (fig. 3).

View larger version (21K): [in this window] [in a new window]   Fig. 2.   Typical HPLC chromatogram of rat urine after administration of [14C]NDPS. Male Fischer 344 rats were administered [14C]NDPS (approximately 10 µCi, 0.6 mmol/kg, ip in corn oil). Urine was collected for 3 hr and analyzed for the presence of metabolites by HPLC, as described in Materials and Methods. Separation was achieved with a Beckman C18 column, using a gradient of acetonitrile and water. Peaks were detected by UV absorbance at 254 nm. The column effluent was collected at 1-min intervals, and metabolites were quantitated by liquid scintillation counting. The absorbance scale was 0.1 absorbance units full scale.

View larger version (16K): [in this window] [in a new window]   Fig. 3.   Urinary metabolic profile of [14C]NDPS. Male Fischer 344 rats were administered [14C]NDPS (approximately 10 µCi, 0.6 mmol/kg, ip in corn oil). Urine was collected for 3 hr and analyzed for the presence of metabolites by HPLC, as described in Materials and Methods. Unk, an unknown metabolite that eluted in the solvent front (2-6 min). , control urine; , urine incubated with -glucuronidase overnight at 37°C. Results are expressed as means ± SE (N = 4).

NDPSA, 2-NDHSA, 3-NDHSA, and DMA were detected in rat plasma 3 hr after administration of 0.6 mmol/kg NDPS (chromatogram not shown). The estimated total radioactivity in blood accounted for 3.2 ± 0.2% of the administered dose. NDPSA was the major circulating NDPS metabolite (246.7 ± 25.1 nmol/ml of plasma) and contributed approximately 75% of the total radioactivity in blood. 2-NDHSA, 3-NDHSA, and DMA were present in plasma at concentrations of 33.7 ± 2.5, 16.6 ± 1.7, and 9.7 ± 2.6 nmol/ml, respectively. Although no distinct UV-absorbing peak for NDPS was visible in the HPLC chromatograms, we were able to quantitate a small amount of the parent compound (12.0 ± 8.1 nmol/ml) in the plasma by liquid scintillation counting of the column effluent. Additional metabolites that eluted in the solvent front (2-6 min) or late in the gradient run (>20 min) were detected in several plasma samples. These accounted for <4% of the plasma radioactivity and did not correspond to any of the synthetic standards.

NDPS (0.6 mmol/kg) tissue distribution was determined by measuring the amount of radioactivity in homogenates prepared from the major organs at 3 hr. Distribution (in nanomoles per gram of tissue) into the kidneys was significantly greater than that into the liver, testes, adrenal glands, lungs, heart, or brain (table 1). At this time point, we could account for 5.3 ± 0.9% of the total administered dose in the tissues that were examined, and the majority of this amount was present in the liver and kidneys (table 1). In fact, the levels of radioactivity in the hepatic and renal homogenates were high enough that we were able to quantitate the metabolites present (table 2). NDPSA was the major metabolite found in both tissues (approximately 40% of the total); however, a considerable amount of unmetabolized NDPS was also detected. In addition to the compounds found in plasma, we were able to detect two novel in vivo NDPS metabolites in the hepatic homogenates by liquid scintillation counting of the HPLC column effluent (table 2). Based on comparison of their retention times with those of synthetic standards, these metabolites were tentatively identified as NDHS and NDHPSA. Distribution of radioactivity (in nanomoles per gram) in the hepatic and renal homogenates (table 1) was comparable to (i.e. not significantly different from, p > 0.05) the total amount of NDPS metabolites present in each tissue as determined by HPLC (table 2).

Excretion and metabolism of [¹⁴C]NDPS (0.2, 0.4, and 0.6 mmol/kg) were also monitored for 24 hr after administration. The total recovery of radioactivity in urine, feces, liver, and kidneys at this time point was <100% at every dose (table 3) and was significantly lower at 0.6 mmol/kg NDPS than at 0.2 or 0.4 mmol/kg NDPS. Fecal excretion increased slightly with increasing dose, whereas total urinary recovery decreased sharply at the highest dose (table 3). Recovery of radioactivity in liver homogenates was significantly greater at 0.6 mmol/kg than at the lower doses. A similar increase in tissue radioactivity at the highest dose was also observed in the kidney homogenates, although the differences were not statistically significant (table 3). Distribution into liver and kidneys (177.7 ± 33.5 and 561.4 ± 176.1 nmol/g, respectively) was significantly greater at the highest dose than at the two lower doses (results not shown). In contrast to the results obtained at 3 hr, the amounts of radioactivity present in hepatic and renal homogenates 24 hr after administration of NDPS (0.6 mmol/kg) were too low for identification and quantitation of individual metabolites by HPLC.

The same four urinary metabolites that were detected in 3-hr urine samples were found in 24-hr urine samples (chromatograms not shown); however, the metabolic profile for NDPS was dose-dependent (fig. 4). For example, 2-NDHSA excretion was significantly increased at 0.6 mmol/kg, whereas the amount of 3-NDHSA remained constant at all three doses. Excretion of NDPSA was significantly less at the highest dose, compared with the lowest dose. Two unknown metabolites (fig. 4, Unk 1 and Unk 2) that eluted in the solvent front (2-6 min) or late in the chromatograms (>20 min) were also found in the 24-hr urine samples. The urinary elimination of these metabolites was unaffected by dose.

View larger version (25K): [in this window] [in a new window]   Fig. 4.   Dose-dependent urinary metabolic profile of [14C]NDPS. Male Fischer 344 rats were administered [14C]NDPS (approximately 10 µCi, ip in corn oil) at three different doses (0.2, 0.4, and 0.6 mmol/kg). Urine was collected for 24 hr and analyzed for the presence of metabolites by HPLC, as described in Materials and Methods. Unk 1 and Unk 2, unknown metabolites that eluted in the solvent front (2-6 min) or late in the gradient run (>20 min), respectively. , 0.2 mmol/kg; , 0.4 mmol/kg; , 0.6 mmol/kg. Results are expressed as means ± SE (N = 4). *, values that are significantly different (p < 0.05) from the corresponding results at a dose of 0.2 mmol/kg.

Analysis of NDPS metabolites in the feces was difficult. Approximately 25% of the total fecal radioactivity could not be recovered for HPLC analysis. The major fecal metabolite was 2-NDHSA; however, the amount excreted (10-30% of the total) was independent of the dose administered (results not shown). A significant amount of the parent compound (30-35% of the total) was present in the feces at all three doses.

Covalent binding of [¹⁴C]NDPS-derived radioactivity to hepatic and renal protein was also assessed at 24 hr. There was no detectable difference in binding between liver and kidney at any of the doses (fig. 5). Protein binding was dose-related in both tissues, with binding at the highest dose being significantly elevated, compared with that at the lowest dose.

View larger version (26K): [in this window] [in a new window]   Fig. 5.   Dose-dependent covalent binding of [14C]NDPS-derived radioactivity to hepatic and renal proteins. Male Fischer 344 rats were administered [14C]NDPS (approximately 10 µCi, ip in corn oil) at three different doses (0.2, 0.4, and 0.6 mmol/kg). Binding to tissue proteins at 24 hr was assessed as described in Materials and Methods. , 0.2 mmol/kg; , 0.4 mmol/kg; , 0.6 mmol/kg. Results are expressed as means ± SE (N = 4). *, values that are significantly different (p < 0.05) from the corresponding results at a dose of 0.2 mmol/kg. , values that are significantly different (p < 0.05) from the corresponding results at a dose of 0.4 mmol/kg.

¹H NMR (500 MHz) spectra for the synthetic standards and known urinary metabolites were very similar (results not shown). Peak assignments for synthetic 3-NDHSA were as follows (values represent chemical shifts, in ppm downfield from tetramethylsilane): 2.8 (dd, 2H, CH₂), 4.58 (m, 1H, CH), 7.1 (s, 1H, Ar-H⁴), and 7.9 (s, 2H, Ar-H^2,6). The following peaks were identified in the NMR spectrum of 3-NDHSA isolated from rat urine: 4.55 (m, 1H, CH), 7.1 (s, 1H, Ar-H⁴), and 7.7 (s, 2H, Ar-H^2,6). The methylene resonance in the isolated 3-NDHSA sample was obscured by a residual water peak.

Compared with controls and the two lower doses of NDPS, BUN levels were significantly elevated 24 hr after administration of 0.6 mmol/kg NDPS (table 4). Urine protein contents were also increased in a dose-dependent manner. NDPS (0.6 mmol/kg) produced marked diuresis in the rats. Kidney weights were significantly elevated at the highest dose of NDPS (table 4). Liver weights, body weights, and urine glucose contents were not altered by NDPS at any of the doses used in these experiments (results not shown).

	Discussion

Top Abstract Introduction Materials & Methods Results Discussion References

NDPS was previously shown to be extensively metabolized when administered orally to rats; however, a substantial proportion (approximately 34%) of the urinary metabolites could not be identified (Ohkawa et al., 1974). Because NDPS must undergo initial cytochrome P450-mediated biotransformation in the succinimide ring to exhibit nephrotoxicity (Rankin et al., 1986, 1987, 1988; Nyarko et al., 1997), it is important to more fully characterize the metabolic fate of this compound in vivo. Furthermore, kidney damage can be observed within 3 hr after dosing (Rankin et al., 1984), although nothing is known about the metabolism of NDPS at early time points. In addition, NDPS nephrotoxicity exhibits a steep dose-response relationship (Rankin et al., 1985), whereas in vivo biotransformation of this compound has been evaluated only at a single dose (0.4 mmol/kg) (Ohkawa et al., 1974). Therefore, the purpose of the experiments described in this report was to evaluate NDPS metabolism at an early time point and at several different doses.

Although only a small fraction (approximately 6%) of the dose was excreted 3 hr after administration of a nephrotoxic dose of [¹⁴C]NDPS (0.6 mmol/kg), we were able to detect several metabolites in rat urine (fig. 2). In addition to the known in vivo metabolites NDPSA, 2-NDHSA, and DMA (Ohkawa et al., 1974), we found 3-NDHSA (a structural isomer of 2-NDHSA) (fig. 1) in the urine samples. Together, these four metabolites accounted for approximately 96% of the urinary radioactivity. 3-NDHSA was previously detected only in vitro (Nyarko and Harvison, 1995; Griffin et al., 1996; Nyarko et al., 1997), but it was probably one of the metabolites that was not fully characterized by Ohkawa et al. (1974). Assignment of this metabolite as 3-NDHSA was confirmed by NMR. The remaining radioactivity (approximately 4%) eluted from the HPLC column in the solvent front and was not clearly visible as a peak in the UV chromatograms. We are uncertain regarding the composition of this peak, although it may contain polar metabolites such as succinic acid or malic acid (Ohkawa et al., 1974; Nyarko and Harvison, 1995; Griffin et al., 1996). Overall, these results are important because they confirm that nephrotoxic metabolites, such as 2-NDHSA and 3-NDHSA (Rankin et al., 1988, 1989), can be generated from NDPS at an early time point, when toxicity initially occurs (Rankin et al., 1984). Thus, rapid NDPS biotransformation correlates with the early onset of renal damage.

Treatment of urine samples with -glucuronidase had no effect on the urinary metabolic profile of NDPS (fig. 3), which suggests that glucuronide conjugates are not produced from this compound. These results are consistent with those of Ohkawa et al. (1974) and our previous in vitro experiments (Nyarko and Harvison, 1995). However, we cannot exclude the possibility that labile glucuronide conjugates may be produced from NDPS or one of its metabolites. For example, acyl glucuronides, which could be formed from any of the acidic NDPS metabolites, are prone to degradation and isomerization (Faed, 1984; Spahn-Langguth and Benet, 1992). Furthermore, Rankin et al. (1995, 1996, 1997) obtained indirect evidence that glucuronide or sulfate conjugates of NDHS or 2-NDHSA may contribute to the nephrotoxic effects of NDPS. The formation of putative conjugative metabolites from NDPS and their potential roles in the toxicity of this compound require further study.

NDPS-derived radioactivity levels were highest in the kidneys and liver, which suggests that these organs may be important sites for distribution and metabolism at 3 hr (table 1). Furthermore, distribution into the kidneys (in nanomoles per gram) was greater than that into the other organs. This implies that the kidneys are able to accumulate NDPS or its metabolites, which could be an important factor in the selective nephrotoxicity of this compound. The results also suggest that the nephrotoxic species are very potent, because only a small percentage (approximately 1.5%) of the total NDPS dose could be accounted for in the kidneys at this time point.

We were able to detect metabolites in both liver and kidney homogenates 3 hr after NDPS dosing (table 2). Because metabolites were found in plasma and there was no evidence for NDPS biotransformation by in vitro renal preparations (Henesey and Harvison, 1995; Griffin et al., 1996), we believe that the metabolites were generated in the liver and were transported to the kidneys through the circulation. In addition to NDPSA, 2-NDHSA, 3-NDHSA, and DMA, we were able to detect small amounts of NDHS and NDHPSA in the hepatic homogenates. These two metabolites were previously detected only when NDPS was incubated with isolated rat hepatocytes (Nyarko and Harvison, 1995; Nyarko et al., 1997). Formation of 2-NDHSA and 3-NDHSA from NDPS could occur by two different routes (fig. 1), i.e. hydroxylation of the succinimide ring to NDHS, followed by hydrolysis (route A), or hydrolysis of the succinimide ring to NDPSA, followed by hydroxylation (route B). Although we cannot exclude the participation of route B, the detection of NDHS in the liver homogenates indicates that route A definitely contributes to NDPS metabolism. Because NDHS is rapidly hydrolyzed to 2-NDHSA and 3-NDHSA (Shih and Rankin, 1989; Nyarko and Harvison, 1995; Griffin et al., 1996), our inability to detect NDHS in plasma, kidney homogenates, and urine is not surprising. The presence of NDHPSA in the renal homogenates, but not the urine, could indicate that this metabolite is accumulated in the kidneys. However, this may not be toxicologically relevant, because NDHPSA did not produce any adverse renal effects when administered to rats (Harvison et al., 1992).

We could account for only 14% of the administered NDPS at the 3-hr time point. One possible explanation for this finding is that NDPS was not yet completely absorbed from the peritoneal cavity. It is also conceivable that NDPS underwent distribution into other sites in the body that were not examined. Nevertheless, these results indicate that sufficient absorption, metabolism, and distribution occurred within 3 hr to explain the rapid onset of NDPS-induced nephrotoxicity (Rankin et al., 1984).

NDPS metabolism was also evaluated at a later time point (24 hr), when nephrotoxicity is fully established (Rankin et al., 1984). In these experiments, NDPS was administered at three different doses, i.e. 0.2 mmol/kg (nontoxic dose), 0.4 mmol/kg (threshold toxic dose), and 0.6 mmol/kg (toxic dose). The induction of nephrotoxicity in the rats was confirmed by marked diuresis, as well as elevations in BUN levels, urine protein contents, and kidney weights (table 4). Similar changes in kidney function were previously reported for NDPS (Rankin, 1982; Rankin et al., 1984, 1985; Kellner-Weibel et al., 1995; Nyarko et al., 1997). The recovery of administered radioactivity in the urine was significantly lower at 0.6 mmol/kg NDPS than at the two lower doses (table 3). Comparable results were reported for acetaminophen (Newton et al., 1983b; Tarloff et al., 1989) and are probably indicative of renal failure. The 24-hr urinary metabolic profile was similar to that obtained in the 3-hr study (figs. 3 and 4), and we were able to account for 85-90% of the urinary radioactivity as NDPSA, 2-NDHSA, 3-NDHSA, and DMA at all three doses. However, at a dose of 0.6 mmol/kg there appeared to be more oxidative metabolism (i.e. 2-NDHSA production) than hydrolytic metabolism (i.e. NDPSA production). These findings are important because 2-NDHSA is more toxic than NDPS (Rankin et al., 1988), whereas NDPSA is less toxic (Yang et al., 1985). Therefore, a shift toward increased formation of a nephrotoxic metabolite such as 2-NDHSA with increasing dose could account for the dose-response relationship reported by Rankin et al. (1985) and may indicate that route A (fig. 1) predominates over route B at higher doses. We could account for approximately 70% of the radioactivity at 0.6 mmol/kg NDPS (table 3). The remainder may be sequestered in tissues (e.g. fat) that were not examined.

Covalent binding to proteins can indicate that a reactive, potentially toxic metabolite was generated in tissues (Pumford and Halmes, 1997). In fact, selective covalent binding to renal proteins was previously reported for the potent nephrotoxicant para-aminophenol (Crowe et al., 1979). We found that NDPS-derived covalent binding increased with dose in both the liver and kidney; however, binding levels were not significantly different between the two tissues at any dose. In addition, binding paralleled the total tissue burden of radioactivity. This lack of organ selectivity may indicate that covalent binding does not contribute to NDPS-induced cytotoxicity in the kidney. However, it is possible that the critical molecular targets for covalent binding in the two organs are different (Cohen et al., 1997). A reassessment of covalent binding of NDPS to specific subcellular targets (e.g. mitochondria), rather than total protein binding, may be worthwhile.

In conclusion, we have demonstrated that NDPS metabolism occurs rapidly and may contribute to the early onset of nephrotoxicity that is observed with this compound. In addition, the extent of oxidative metabolism appears to be correlated with dose and increasing nephrotoxicity. These experiments have provided additional information on the correlation between NDPS-induced kidney damage and the in vivo biotransformation of this compound.