Relative roles of metabolism and renal excretory mechanisms in xenobiotic elimination by fish.

Renal clearance techniques were used to examine the relative contributions of metabolism and renal tubular transport in determining the rates of excretion of benzo(a)pyrene (BaP) and several of its phase I metabolites by southern flounder, Paralichthys lethostigma. Each compound (3H-labeled) was injected at a dose of 2.5 mumole/kg, producing plasma concentrations of 1 to 5 microM. Despite extensive plasma binding (greater than 95%), the uncorrected renal clearance of BaP-7,8-dihydrodiol exceeded the glomerular filtration rate (GFR) by more than 20-fold. Phenolic BaP metabolites also showed net secretion (1.5- to 3-fold). At times prior to 3 hr, BaP itself showed an average clearance of only 0.2 times the GFR. After 3 hr, BaP clearance increased to three times the GFR. Decreasing the dose of BaP injected also dramatically increased its clearance. Clearances of all four compounds studied were reduced by probenecid and other organic anion, including the herbicide 2,4-dichlorophenoxyacetic acid. HPLC analysis demonstrated that the bulk of the material excreted in the urine was not the parent compound, but sulfate or glucuronide conjugates of its phenolic or dihydrodiol metabolites. Excretion of sulfate conjugates predominated over the first 24 hr, whereas the glucuronide conjugates were the primary excretory products in succeeding days. In vitro, isolated renal tubules transported both glucuronide and sulfate conjugates, but sulfates were the preferred substrates. Isolated tubules were shown to be capable of catalyzing conjugation reactions, producing predominantly glucuronide conjugates. Liver slices produced both types of conjugates. Thus, the rapid excretion of BaP-7,8-dihydrodiol reflected a combination of two processes.(ABSTRACT TRUNCATED AT 250 WORDS)


Introduction
Aquatic organisms are exposed to a wide variety of foreign chemicals. Many of these chemicals are potentially toxic to the organisms themselves and to other animals, including man, which may feed on them. Furthermore, once within the aquatic organism, a number of these chemicals, e.g., the polycyclic aromatic hydrocarbons (PAHs), may be converted to still more toxic forms through metabolic oxidation, usually via the cytochrome P-450 monooxygenase system (1). Therefore, the primary focus of much of the work in this area has been on mechanisms ofmetabolic activation and on the enzyme systems, e.g., epoxide hydrolase and the glutathione transferases, which mediate detoxication of the activated, chemically reactive metabolites (2,3).
Much less attention has been given to the second means of reducing tissue concentrations ofpotentially toxic compounds, elimination from the body via urinary and biliary excretion. Certainly, metabolic studies in fish, as in mammals, amply demonstrate conversion oflipophilic xenobiotics to more polar forms. Conversion to lipophilic xenobiotics alone should accelerate elimination of toxic compounds (4,5). Indeed, the appearance of numerous polar xenobiotic metabolites in urine and bile has been widely documented (6)(7)(8). However, these studies have not addressed the possibility that different phase I metabolites might be excreted differently. For example, although the model PAH, benzo(a)pyrene (BaP), is known to be metabolized through a number of increasingly oxidized intermediates to its putative carcinogenic metabolite, BaP 7,8-t-dihydrodiol-9,10-epoxide, neither the extent ofexcretion nor the factors that regulate excretory rate are known for BaP or any of its metabolites. Clearly, such factors may play an important role in determining the retention and toxicity of PAHs to the marine organism itself and to those who consume such contaminated seafood.
In the studies reported in this paper, we have examined the urinary excretion of BaP by the southern flounder (Paralichthys lethostigma) and compared its excretion to that of several of its more polar metabolites, with particular attention to BaP-7,8-tdihydrodiol (BaP-7,8-diol), an immediate precursor to the putative ultimate carcinogenic BaP metabolite, the 7-8-diol-9,10-epoxide. We also examined several of the processes that contribute to the pattern and rate of BaP-7,8-diol excretion. These studies were conducted using a marine teleost for two reasons. cClearance ratio is independent of urine flow rate. bPlasma values are those determined for the midpoint of the collection period. dUrine flow was erratic during periods 0-2, estimated values based on average over first 90 min were used to calculated data for periods 0, 1, and 2. n = 7 for this mean (I and 2 excluded).
First, results from such studies should shed some light on the dynamics ofBaP and metabolites in an important class ofaquatic food organisms. Second, the marine teleost presents several distinct experimental advantages for exaning thie role ofrenal excretory processes in xenobiotic elimination (9). For example, because the marine teleost has an extensive renal portal system, the tubules are bathed in a blood flow that approaches the entire cardiac output. This anatomical arrangement allows maximal impact ofthe kidney on the excretion ofall solutes present in the circulation. It also permits direct experimental access to the blood perfusing the kidney, since the readily accessible caudal vein feeds directly into the renal portal system. Thus, application of in vivo renal clearance techniques to marine teleosts can provide a great deal of information about the renal handling of a given agent, particularly the contribution of secretory tubular transport toward its net elimination (9). Furthermore, the tubules themselves may be readily isolated for in vitro examination oftheir metabolic capacity and transport properties (9,10). As shown here, these studies demonstrate that BaP and its phase I metabolites are excreted into the urine at different rates. These differences may be ascribed in large part to the combined action of phase II metabolic events and the renal organic anion secretory transport system mediates their excretion.

Animals
Adult southern flounder (0.5 to 3.0 kg) ofboth sexes were netted in the wild in the vicinity of Matanzas Inlet, Florida. These fish were maintained in the laboratory in flowing, sand-filtered sea water for not less than 1 week following capture. Live shrimp orjuvenile fish were continuously available as food, approximating the diet of these fish in the wild.

Renal Clearance Determination
Clearance measurements were carried out as previously described (11). The glomerular filtration rate (GFR) marker, '4C-polyethyleneglycol (PEG) (molecular weight 4000; dose 250 mg/kg) and the 3H-labeled BaP or BaP metabolite (2.5 lsmole/kg) were injected together 1 hr prior to the first collection period. One-half of the dose was administered IV via the caudal vein as a priming dose and the remainder was given IM in the earlier work (11). Resulting plasma concentrations were 0.5 to 1.0 mg/mL for PEG and I to 5 ,M for BaP and metabolites and were stable over the period from I to 6 hr after injection. A typical experiment with BaP-7,8-diol is shown in Table 1.
In each experiment, clearances were calculated based on radioactivity in plasma and urine in the usual manner, i.e., the quantity oflabel appearing in the urine per unit time was divided by the plasma concentration of label. Calculated clearance values were not corrected for plasma binding ( > 95 % for BaP and each of its metabolites, as determined by ultrafiltration; data not shown). Thus, the values presented represent minimal estimates ofrenal clearance. This choice was based on several important considerations. First, because the concentrations of free compounds were so low, small errors in binding measurements would be grossly magnified in the final clearance calculations. Furthermore, it is not yet established how effectively tubular secretory transport may compete with plasma binding sites for access to these compounds. Thus, we have chosen the conservative course ofestimating tubular transport based on the assumption that all labeled BaP or metabolite was available for transport. Our second assumption was to use the clearance of label to estimate the clearance of each parent compound. We know that each compound was subject to extensive metabolism during the course of these in vivo studies. However, the important initial point was to focus on the relative rates ofexcretion for all compounds derived from a given parent molecule and to determine whether evidence could be obtained for the participation of tubular transport in the excretion process. As the data presented below clearly attest, differences in clearance rate and net tubular transport were readily apparent between BaP and its metabolites.
All later work was performed using an Altex ultrasphere ODS 5 um C-18 reverse phase column. Conditions used were a flow rate of 0.75 mL/min with 45% methanol:water (9 mL), 95% methanol:water (8.25 mL) and return to 45% methanol (27.75 mL). Standards eluted in the following order: BaP-3-glucuronide

Isolated Tubule and Slice Studies
Tubules were prepared from flounder kidney as originally described by Forster (10) with minor modifications (12). Approximately 10 mg of tissue was incubated in 2 mL ofoxygenated Tris Forster's saline (in mM; Na 140, K 2.5, Ca 1.5, Mg 1, Cl 147.5, and Tris 20 at pH 8.25) containing labeled BaP or BaP metabolite. After the desired incubation time, the tubules were blotted on filter paper, weighed, and homogenized in 0.5 mL ofdistilled water. Acetone (0.5 mL) was added, and the mixture was extracted three times with 4 mL of ethyl acetate. The ethyl acetate extract was evaporated to dryness under a stream ofdry nitrogen, and the residue was taken up in methanol for HPLC analysis as above. Liver slices (70.5 mm thick) were prepared using a Stadie-Riggs microtome. Incubation and processing procedures for slices were identical to those described above for renal tubules.
In experiments to compare the tubular transport ofglucuronide and sulfate conjugates, 4-methylumbelliferone (4-MU) and its conjugates were used as substrates. The isolated tubules were incubated as before with 10 ,M 4-MU or 4-MU conjugate. After incubation, the tissue was weighed and rapidly denatured by heating with 1 mL of 0.2 M acetate buffer (pH 5.0) in a boiling water bath for 10 min. (4-MU and its conjugates were stable for at least 60 min at 100°C; data not shown.) The tubules were then homogenized. To determine 4-MU itself, another 1 mL ofacetate buffer was added, and this mixture was extracted with 2.0 mL of ethyl acetate. Samples were dried and taken up in methanol for analysis by HPLC with fluorescence detection. 4-MU sulfate and glucuronide were enzymatically hydrolyzed prior to extraction with ethyl acetate and analysis as for 4-MU itself. Conditions used for hydrolysis were as described above for BaP conjugates. Correction for recovery of 4-MU and its conjugates was determined by analysis of known quantities of each compound performed in parallel with each group of unknowns. Recoveries ranged from 65 to 80%.

Chemicals
Both labeled and unlabeled BaP metabolites were obtained from the Cancer Research Program of the National Cancer Institute. 4-MU and its glucuronide and sulfate conjugates, (3glucuronidase and sulfatase, were purchased from Sigma Chemical Company (St. Louis, MO). All other chemicals were of the highest purity available from commercial sources.

Renal Clearances
In preliminary experiments, 3H-BaP and 14C-PEG (as GFR marker) were injected IM. As in previous studies (12), the more water soluble PEG was effectively absorbed into plasma from muscle and reached a stable plasma concentration within 1 hr. However, the more lipophilic BaP was not effectively absorbed. Therefore, in the experiments reported here, BaP was presented by IV injection into the caudal vein. The more water soluble BaP metabolites were presented in two doses, a priming dose of 1.25 jtmole/kg injected IV and a maintenance dose of 1.25. /&mole/kg injected IM. As shown in Tables 1 and 2, these protocols yielded relatively stable plasma concentrations of BaP and its metabolites. Note, however, that plasma concentrations following IV injection of 3H-BaP were somewhat higher than those obtained following IV plus IM injection of its phenolic and dihydrodiol metabolites, reflecting both the slower excretion rate for BaP and its route of administration.
Two major points should be noted when comparing the renal handling of BaP with that of its phenolic and dihydrodiol derivatives (Table 2). First, average clearance values over the first 3 hr after injection of labeled BaP were substantially lower than the GFR, demonstrating net reabsorption. However, upon examination ofthe final urine-to-plasma ratios for BaP and PEG, it was clear that by the end of the 4 to 6 hr experiment, label injected as BaP had begun to appear in the urine at a rate more rapid than the GFR marker, i.e., net secretion was taking place. In fact, BaP clearances measured at times exceeding 3 hr after injection averaged more than 10 times the earlier clearance values. Furthermore, as shown in Table 3, it was possible to dramatically increase the renal clearance of 3H-label injected as BaP by reducing the quantity ofBaP injected. Such changes in renal handling would be most readily explicable by metabolic conversion ofBaP to a form, or forms, subject to active tubular transport. In the first instance, increased metabolism followed increased contact time. In the second case, lowering the total amount of BaP injected could have increased the relative contribution ofmetabolites in the total mix of BaP-derived compounds. Indeed, as discussed below, this appears to be the case. aDose was 2.5 jAmole/kg for all compounds. The glomerular filtration rate was measured simultaneously using PEG (molecular weight 4000).
bThe clearance (C%) is calculated as (U. * V)/P. and is the volume ofplasma cleared of compoundx per unit time. Ux, concentration ofx in urine; P, concentration ofx in plasma; and V, rate of urine production.
cThe ratio of CX/GRF will be greater than 1 only for compounds added to the urine by the renal tubules (i.e., secreted). A ratio less than 1 means the compound was removed by the tubules (i.e., reabsorbed). GFR, glomerular filtration rate.
Despite evidence for tubular secretion ofBaP late (> 3 hr) in the experiments (Table 2) or after a low dose of BaP (Table 3), it was nevertheless clear that BaP phenols and dihydrodiols were cleared by the kidney far more rapidly than BaP itself immediately after injection ( Table 2). The phenolic products, l-OH-BaP and 7-OH-BaP, were intermediate in their clearance rates. In most fish, the mean clearance ratio was greater than unity, demonstrating net tubular secretion. By far, the most effectively eliminated derivative examined was BP-7,8-diol, which was cleared on the average 10 to 30 times the GFR. For this reason, subsequent experiments focused largely on BaP-7,8-diol, an important metabolite of BaP.
The influence ofaltered temperature and ofenzyme induction, both of which should alter metabolic rate, were assessed in fish given either BaP or BaP-7,8-diol. To examine the effects of temperature, fish were injected with 3H-BaP-7,8-diol in winter when the water temperature was 12 to 14WC, about -10°C lower than the water temperatures present during the initial clearance studies (in summer). Clearance ofthe diol fell to only one-tenth of the summer values. On the other hand, injection with 3-methylcholanthrene (10 mg/kg) 1 week prior to clearance determination, previously shown to increase metabolism via induction of cytochrome P450 in this species (13), enhanced the renal excretion of BaP but was without significant effect on BaP-7,8-diol clearance (Table 4). These results suggest that the excretion ofBaP requires oxidative metabolism but that excretion of the dihydrodiol does not. However, dihydrodiol excretion was temperature dependent, indicating that a subsequent metabolic step, e.g., conjugation, and/or the excretory process itself, was temperature dependent. Therefore, we examined the secretory mechanism and the sensitivity of excretion rate to manipulation ofthe secretory process. Table 1 and Figure 1 demonstrate the sensitivity of renal excretion to inhibition ofBaP-7,8-diol excretion to inhibition of the organic anion transport system by probenecid, the classical inhibitor of this system. Probenecid also reduced the clearances ofboth BaP 1-and 7-phenols (not shown). In addition, both the renal (Table S) and hepatic (not shown) accumulation of these phenols was markedly reduced by probenecid, as would be predicted iftheir accumulation and subsequent excretion were. mediated by the organic anion transport system known to be present in both organs.
Other organic anion, notably the anionic herbicide, 2,4-dichlorophenoxyacetic acid (2,4-D), were also effective inhibitors ofthe renal excretion ofBaP-7,8-diol. As shown in Figure 2A, administration ofa small dose (2.5 iamole/kg) of 2,4-D led to a prompt, but transient inhibition of BaP-7,8-diol clearance. As shown previously (6), 2,4-D is an excellent substrate for the organic anion system and is excreted very rapidly. Thus, the transient nature of the inhibition seen here apparently reflects elimination of the inhibitor. When a higher dose of 2,4-D (Amole/kg) was given prior to the administration of BaP-7,8-diol (Fig. 2B), inhibition was much more extensive  Net secretion of BaP-7,8-diol, expressed as the ratio of BaP-7,8-diol clearance over the glomerular filtration rate (GFR) was followed for 2.5 hr. Probenecid was administered IV at 155 min, and the clearance ratio was followed for 90 additional min. Over this interval, the clearance ratio was reduced dramatically, and net renal secretory transport ofBP-7,8-diol was nearly abolished.
(> 95 % over the first 2 hr), and clearance values were still substantially below control after more than 6 hr. Given the importance ofmetabolism in determining the efficacy ofelimination and transport, we examined the chemical nature ofthe labeled material appearing in the urine using HPLC techniques. As shown in Figure 3, the bulk of the radioactivity administered as BaP-7,8-diol appeared in the urine as a mixture of conjugated metabolites. On the first day after injection of 3H-7,8-diol, the largest fraction ofthe excreted label co-chromatographed with the sulfate conjugate standard. The glucuronide conjugate was also prominent; smaller portions were excreted as unchanged BaP-7,8-diol. In this flounder (Fig. 3), some label also chromatographed with the phenolic standard, but this peak was missing in the urine ofother fish. After the first day, the portion of the label excreted as the sulfate conjugate decreased markedly and was nearly undetectable on day 3. The identities of the several radioactive peaks were confirmed by enzymatic hydrolysis (Fig. 4). Treatment of the urine collected within the first day after 7,8-diol injection (high in putative sulfate conjugate) with sulfatase essentially eliminated the sulfate peak (Fig. 4). Similarly, treatment ofthe urine collected on days 2 and 3 with glucuronidase abolished the putative glucuronide peak. The rapid appearance and elimination ofthe sulfate conjugate of BaP-7,8-diol might reflect its preferential transport on the organic anion system. This possibility could not be assessed directly due to the unavailability of sufficient quantities of and sulfiae conjugates were commerially awailable and could be readily assayed fluorimetrically. Thus, we were able to expose isolated flounder tubules to 4-MU or its conjugates in vitro and determine the relative effectiveness oftubular transport for the three classes of substrate. As shown in Table 6, each of these xenobiotics was accumulated by the flounder renal tubules. The accumulation of4-MU itselfwas the most modest and reflected both entry ofthe parent molecule and subsequent conversion to conjugates. Therefore, accumulation of 4-MU is a maximal estimate of transport activity. Both 4-MU-glucuronide and 4-MU-sulfate were accumulated to a much greater extent than 4-MU itself. Furthermore, the accumulation ofconjugates was largely blocked by probenecid, which inhibits the organic anion transport system, and by cyanide, which should inhibit active trasport though blocking cellular respiration and ATPproduction. As suggested by the in vvo results with BaP-7,8-diol, the slfae conjugate yielded cellular accumulation ofapproximtely twice that of the glucuronide conjugate. Thus, it would appear that the pattern of excretion documented above does indeed rflect the secretory transport of both conjugates. The sulfate conjugate appeared in the greatest quantities initially and was eliminated first. The glucuronide continued to be htansported and eliminated over subsequent days. Finally, we mined the ability ofkidney and liver to produce these conjugated metabolites from BaP-7,8-diol. Isolated renal tubules or liver slices were incubated in Forster's saline with 10 FM 3H-BaP-7 8-diol for 4 (kidney) to 6 (liver) hr in vitro. As shown in Table 7, both tissues produced sulfate and glucuronide conjugates. However, there were several differences between the two tissues. Most notably, the kidney tubules were far more active in vitro, convting nearly 50% ofthe substrate to conjugates within 4 hr. In contrast, liver slices only produced about 20% conjugates after 6 hr of incubation. The pattern of metabolism was also quite diffent, with the kidney tubules producing more than 10 times as much glucuronide as sulfate conjugates. The liver also produced more glucuronide conjugate, but the ratio was less than 3    injection must have been produced elsewhere in the body and been transferred to the kidney via the circulation. This conclusion is also supported by the ability ofother organic anions (probenecid and 2,4-D) to inhibit both renal secretion ofBaP-7,8-diol conjugates ( Table 1, Figs. 1 and 2) and to reduce the accumulation of conjugates both in viwv (able 5) and in atro ( Table 6).

Discussion
In any atempt to assess the potential impct ofchemically contminated seafod on human consumers, one must be able to assess not only the degree ofcontamination, but also the retention of that contmination within the food organism for subsequent transfer up the food chain. The results reported above demonstate that several fiaors including ocidative me lism, conjugation reactions, and e ry mechans pLy importnt roles in determining both the retention and the potential toxicity of foreign chemicals. For the many lipophilic xenobiotics such as BaP, it has been recognized for years that oxidative metabolism via cytochrome P450 plays an important dual role (1). On one hand, oxidative metabolism increases water solubility ofthese xenobiotics, thus increasing their plasma concentrations and making them more readily available for urinary and/or biliary excretion. On the other hand, a significant number ofthese oxidative metabolites are biologically active in their own right, often exceeding their parent in reactivity and toxicity. Therefore, phase II reactions, which include epoxide hydrolase, glutathione transerase, and conjugation reactions with glucuronic acid, sulfate, or amino acids, normally assume critical roles in terminating the toxicity ofthese xenobiotics and their phase I metabolites by converting them to nontoxic products that are readily excreted. What has been demonstrated in these studies, as emphasized in Table  2, is that the different products ofoxidative metabolism are also handled in very specific and different ways by the available renal excretory mechanisms. Furthermore, as shown in Table 6, different conjugated metabolites (e.g., glucuronides versus sulfites) are also subject to differential handling y the renal transport system that mediates their elimination. Thus, the overall retention and the potential toxicity ofa given xenobiotic are determined by the specific combination ofthe actions ofphase I/phase I metabolic pathways and excretory mechanisms on that compound.
For BaP and the specific metabolites studied here, the factors discussed here sort out in the following manner. a) Increased oxidative metabolism accelerates the elimination ofBaP itself, as evidenced by the inreased rate ofexcretion ofBaP at later times after injection into the fish (Table 2), the accelerated loss ofBaPderived radioactivity when the injected dose was reduced (Table  3), and the increased excretion at higher environmental temperatures or after induction of the cytochrome P-450 monooxygenase system by 3-methylcholanthrene (Table 4). b) In contrast to the result for BaP itself, BaP-7,8-diol excretion was not significantly changed by P-450 induction (TIble 4), implying that subsequent steps, e. g., conjugation and renal secretory transport, were more critical for dihydrodiol excretion. c) The accelerate excetion ofall mur labeled BaP compounds depended upon their secretory transport on the organic anion system, as evidenced by the inhibition oftransport in vivo ( Table 1, Fig. 1) and in vitro (Tible 6) by probenecid and 2,4-D (Fig. 2). d) Sulfate-conjugated BaP metabolites were better substrates for renal organic anion transport dtan comparable glucuronide conjugates in vit (ble 6) and inviw (Figs. 3 and 4). e)The ability of the herbicide 2,4-D to reduce the rate of elimination of BaP-7,8-diol (Fig. 2) demonstrates the possibility that the simultaneous presence oftwo or more substrates for the organic anion system may retard the eliminaon oftoxic xenobiotics such as BaP-7,8-diol, increasing the likelihood of both systemic toxicity in the fish and enhanced ltranr ofxenobiotic to subsequent consumers, including man. f) Although both kidney and liver were shown to be capable ofconjugating BaP metabolites (TIble 7), the bulk ofthe excreted metabolites must have been produced outside the kidney, as evidenced by the reduced renal tissue levels and excretion following inhibition of the organic anion transport system by probenecid or 2,4-D (Tibles 1 and 5; Figs. 2 and 3).
In conclusion, these data emphasize that important interactions occur between metabolism and excretory mechanisms that togetherdeterminethe retentionofpotentially toxicchemicalsby fish. Withoutanunderstandingofboth processes, itis notpossibleto predicteidterheextentofaccumulation orthetoxicity ofthe retainedproducts. Thus, any attempttoaddress thepotential toxicity ofsuch chemicals intheexposedorganismor in subsequent consumers requires a thorough knowledge ofboth processes.