The thiopurines 6-mercaptopurine (6-MP¹; Fig. 1) and 6-thioguanine (6-TG; Fig. 1) are antimetabolites used for leukemia treatment (Elion, 1989). Thiopurine metabolism is complex and involves both activation and inactivation reactions. The primary activation pathway is the formation of the nucleotides 6-thioinosine monophosphate and 6-thioguanosine monophosphate from 6-MP and 6-TG, respectively; the former metabolite can subsequently be converted to 6-thioguanosine monophosphate (Fig. 2). These nucleotides interfere with cellular functions by inhibiting de novo biosynthesis and interconversion of normal purines (Martin, 1987). However, the further metabolism of 6-thioguanosine monophosphate and subsequent incorporation into nucleic acids are considered the crucial steps in thiopurine cytotoxicity (van Scoik et al., 1985; Martin, 1987). Another pathway in thiopurine metabolism is the S-methylation of thiopurine bases and nucleotides (Lennard, 1992; Krynetski et al., 1995; Weinshilboum, 2001). Methylated thiopurine bases have no direct biological activity, whereas methylated thiopurine nucleotides inhibit de novo purine biosynthesis (Martin, 1987). Inactivation of the thiopurines occurs primarily through oxidation of 6-MP to form thiouric acid (TU) and deamination of 6-TG to yield thioxanthine (TX) that can be further oxidized to TU (Fig. 2; Bronk et al., 1988).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 1.

Chemical structures of 6-MP, cis-AVTP, AZA, 6-TG, trans-AVTG, B.W. 57–323.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 2.

A simplified scheme of 6-MP and 6-TG metabolism.

Although the thiopurines are successful anti-leukemic agents, their bone marrow toxicity has limited their utility against solid tumors (Frank and Tornyos, 1962; Regelson et al., 1964; Moore et al., 1968). It may be possible to increase the use of the thiopurines against solid tumors by decreasing their systemic toxicity, for example by using prodrugs that are preferentially bioactivated in the tumor compared with the normal surrounding tissue. In this regard, we have been characterizing the biological activity of cis-6-(2-acetylvinylthio)purine (cis-AVTP; Fig. 1) and trans-6-(2-acetylvinylthio)guanine (trans-AVTG; Fig. 1) that undergo sequential addition-elimination reaction with GSH to yield 6-MP and 6-TG, respectively. We determined that the prodrugs have lower or comparable IC₅₀ values than 6-MP and 6-TG, respectively, when tested in the human renal cell carcinoma cell lines ACHN and A-498. Additionally, less bone marrow and intestinal toxicity was observed in mice treated with the prodrugs compared with 6-TG (Gunnarsdottir et al., 2002a,b). However, after multiple treatments with these compounds, a lower fraction [4 ± 2% (mean ± S.D.); Gunnarsdottir et al., 2002b] of administered trans-AVTG doses was recovered in mouse urine as 6-TG and its metabolites than was recovered after equimolar 6-TG treatments (27 ± 11%). Similarly, the recovery observed after multiple trans-AVTG treatments was lower than that observed after multiple cis-AVTP treatments (18 ± 6%). These data suggested that cis-AVTP, trans-AVTG, and 6-TG may have different tissue uptake and metabolism in the mouse.

The goals of this study were to characterize the tissue distribution of metabolites of cis-AVTP and trans-AVTG in the mouse and compare with distribution of metabolites of 6-TG to determine whether differences in tissue distribution or metabolite composition among these compounds may explain the differences in systemic toxicity and urinary metabolite excretion profiles observed after i.p. treatment of mice with single or multiple equivalent doses of cis-AVTP, trans-AVTG, and 6-TG (Gunnarsdottir et al., 2002b). Additionally, levels of reduced and oxidized GSH were measured in tissues to determine whether thiopurine prodrug administration was associated with changes in these parameters. Finally, the effect of glutathione S-transferases on thiopurine formation from the prodrugs was also assessed using rat liver cytosol as the enzyme source.

Materials and Methods

Chemicals. 6-MP, 6-TG, 6-MP riboside, 6-TG riboside, TX, GSH, the oxidized form of GSH (GSSG), sulfosalicylic acid (SSA), dithiothreitol (DTT), and trifluoroacetic acid (TFA) were purchased from Sigma-Aldrich Research (St. Louis, MO). HPLC-grade acetonitrile was purchased from EM Science (Gibbstown, NJ). cis-AVTP and trans-AVTG were synthesized as previously described (Gunnarsdottir et al., 2002a). Reference TU was synthesized by the enzymatic oxidation of TX by xanthine oxidase in buffer at 37°C according to the method of Bergmann and Ungar (1960). The TU formed in the reaction was purified on a Sephadex LH-20 column (Amersham Biosciences Inc. Amersham Pharmacia, Piscataway, NJ) using 5% acetonitrile adjusted to pH 2.5 with TFA. Fractions containing TU were combined and lyophilized. The purity of the final product was greater than 96% as assessed by HPLC. All other chemicals were of highest grade commercially available.

In Vivo Experiments and Metabolite Analysis. The tissue distribution of metabolites of cis-AVTP, trans-AVTG, and 6-TG was assessed in 6- to 10-week old CD-1 hetereozygous nude male mice (Charles River Laboratories, Inc., Wilmington, MA). All experiments were carried out in accordance with the Guide for the Care and Use of Laboratory Animals as recommended by the U.S. National Institutes of Health and the University of Wisconsin - Madison Animal Care Committee. The mice were kept on a 12 h light/dark schedule and allowed food and water ad libitum. The mice were injected (i.p.) with vehicle only or 42.5 μmol/kg of cis-AVTP, trans-AVTG, or 6-TG dissolved in buffer (0.78 mM phosphate, 137 mM NaCl, 5.4 mM KCl, and 4.2 mM NaHCO₃) with a few drops of dilute NaOH. This dose is higher than the 8.5 and 21.25 μmol/kg of doses used in a previous study examining the in vivo toxicity of multiple treatments of cis-AVTP, trans-AVTG, and 6-TG in order for the concentration of the compounds and their metabolites to reach and be above the limits of detection of the methods used in this study. The injection volume was between 0.4 and 0.6 ml. All treatments were carried out between 9 and 11 AM. The mice were sacrificed using a CO₂ chamber at 15, 30, or 45 min after injection. These time points were chosen based upon the results of preliminary experiments indicating that 6-MP and 6-TG have short half-lives and thus, the tissue concentration at time points longer than the ones chosen were expected to be close to or below the limits of detection of the methods used in this study. Each time course was repeated in triplicate, using one mouse per time point. Metabolites were analyzed in plasma, RBC, liver, intestine, and bile. Blood was collected via cardiac puncture. Liver and intestine were removed from all treated mice and kept in ice-cold buffer; the intestine was cut into 0.5- to 1-inch pieces and the luminal content rinsed out. When possible, the gall bladder was also harvested and the bile put into a clean microcentrifuge tube. Metabolites were analyzed in plasma to assess whether differences in plasma concentrations of the compounds and their metabolites might explain the differences in urinary excretion of metabolites previously observed after treatment of mice with these compounds (Gunnarsdottir et al., 2002b) because a positive correlation has been observed between plasma thiopurine area under the curve (AUC) and urinary excretion in children given 6-MP (Endresen et al., 1990) and in RBCs because thiopurine nucleotides accumulate in RBCs (Lennard, 1992). Metabolites were also assessed in liver and intestine because previous experiments indicated difference in liver and intestine toxicity among cis-AVTP, trans-AVTG, and 6-TG after treatment of mice with these compounds (Gunnarsdottir et al., 2002b), and in bile to assess whether the prodrugs and/or their metabolites are excreted through bile.

Our previous results showed that the prodrugs are rapidly converted to the thiopurines in vitro when incubated with equimolar concentrations of GSH (Gunnarsdottir et al., 2002a). Similar rapid conversion was observed in vivo because no unmetabolized prodrug was recovered in urine from mice treated with cis-AVTP or trans-AVTG (Gunnarsdottir et al., 2002b). Thus, we expected similar rapid conversion of the prodrugs to the thiopurines in tissues before and after the mice were sacrificed and tissues harvested and deproteinized by acid. Therefore, the prodrug and the thiopurine were considered to comprise a pool of thiopurine equivalents and were quantitated as thiopurines. Tissue samples were analyzed before and after acid hydrolysis because whereas the thiopurines and the corresponding ribosides were easily detected in unhydrolyzed samples, thiopurine nucleotides must be hydrolyzed to the corresponding thiopurines before analysis. Upon hydrolysis, thiopurine ribosides were also converted to the corresponding thiopurines. For that reason, the concentration of thiopurine nucleotides present in tissues was calculated as the difference between the concentration of thiopurines in hydrolyzed samples and the combined concentration of thiopurines and thiopurine riboside in unhydrolyzed samples. Thus, the compounds quantitated in unhydrolyzed samples were 6-MP or 6-TG, the thiopurine riboside, and the inactive metabolites TX and TU. The compounds quantitated in hydrolyzed samples were 6-MP or 6-TG, TX, and TU. Methylated thiopurines were not assessed in this study because the mouse does not readily methylate 6-MP or 6-TG (Elion et al., 1960b), and because the K_m value for the mouse thiopurine methyltransferase using 6-MP as substrate is above 0.75 mM (Otterness and Weinshilboum, 1987). Therefore, limited methylation was expected at the thiopurine tissue concentrations obtained in our studies. This agrees with our previous results where no methylated thiopurines were detected in urine from mice treated with cis-AVTP or trans-AVTG (Gunnarsdottir et al., 2002b).

Samples were prepared and analyzed based on the methods of Mawatari et al. (1998) and Dervieux and Boulieu (1998). Briefly, for analysis of metabolites in plasma and RBCs, approximately 650 μl of whole blood were added to a K-EDTA blood collection tube containing 0.5 mg DTT; DTT was added to the samples to protect the thiol group of the thiopurines and their metabolites from oxidation and to improve their recovery from the biological matrices (Dervieux and Boulieu, 1998; Mawatari et al., 1998). The blood harvested from the mice was centrifuged at 14,000g at 4°C for 5 min to separate the plasma from the RBCs. The plasma was transferred to a clean microcentrifuge tube into which were added 0.11 volumes of 50% (v/v) PCA to deproteinize the sample. A 200-μl volume of packed RBCs was put in a microcentrifuge tube containing 4 mg of DTT, diluted with 200 μl of ice-cold water, and deproteinized by the addition of 0.11 volumes of 50% (v/v) PCA. For analysis of metabolites in bile, the bile that was obtained was diluted with 300 μl of water into which were added 0.11 volumes of 50% (v/v) PCA. For analysis of metabolites in liver or intestine, approximately 500 to 600 mg of liver or 350 to 400 mg of small intestine were patted dry, and homogenized in three volumes of 6.67% (v/v) PCA in the presence of 20 μl of 1 M DTT. All samples were centrifuged at 20,000g for 5 min. Approximately half of the resulting supernatant was filtered through Acrodisc LC13 0.2-μm filters (Gelman Sciences, Ann Arbor, MI) and analyzed by HPLC as described below. The other half of the supernatant was transferred to a clean microcentrifuge tube and hydrolyzed by boiling for 45 min before the samples were filtered and analyzed by HPLC as described below. A portion of the liver and intestine homogenate was saved for protein measurements by the method of Lowry et al. (1951) using bovine serum albumin as standard. To ease the comparison of metabolite concentration among tissues, the amount of protein per gram of wet weight of liver or intestine was also measured.

Effect of Rat Liver Cytosol on the Formation of Thiopurines from the Prodrugs. To assess whether the prodrugs cis-AVTP and trans-AVTG are substrates for GSH S-transferases (GSTs), the prodrugs were incubated with GSH in the presence and absence of rat liver cytosol. The cytosol was prepared from male Sprague-Dawley rats (215–225 g; Sasco Laboratory, Omaha, NE). The rats were sacrificed by decapitation after sedation with CO₂, the liver removed and homogenized in 3 ml of buffer (0.1 M KH₂PO₄, 0.15 M KCl, 1.5 mM EDTA, pH 7.4) per gram tissue. The resulting homogenate was centrifuged at 48,000g for 30 min and the supernatant recentrifuged at 105,000g for 90 min. The resulting supernatant (cytosol fraction) was used as a source of GSTs in enzymatic assays. The assays were carried out as follows: GSH (5 mM final concentration) was preincubated in two test tubes for 2 min at 37°C in the presence of either buffer or diluted rat liver cytosol (1.0 mg/ml final protein concentration) in a shaking water bath. The reactions were initiated by the addition of cis-AVTP or trans-AVTG (100X solution in DMSO, 1 mM final concentration). The final reaction volume was 1 ml. After 0.5, 1, 2, 4, 6, 8, 10, 20, and 40 min, a 110-μl aliquot was removed from the test tubes and added to a microcentrifuge tube containing 11 μl of 10% TFA. The samples were filtered and kept on ice until analyzed by HPLC as described below.

HPLC Analyses. The HPLC system used for metabolite analysis consisted of two Gilson 306 pumps, a Gilson 119 UV/vis detector, and a Gilson 234 autoinjector (Gilson Medical Electronics, Middleton, WI). The column used was a Beckman ultrasphere ODS 5-μm reversed-phase C₁₈ (4.6 × 250 mm; Beckman Coulter, Inc., Fullerton, CA) with a Brownlee spheri-5 ODS 5 μm (4.6 × 30 mm) guard column (PerkinElmer, Norwalk, CT). Mobile phase for pump A consisted of water adjusted to pH 2.5 with TFA whereas the mobile phase for pump B was 1:1 acetonitrile/water mixture adjusted to pH 2.5 with TFA. For all analyses, the flow rate was 1 ml/min, and the detection wavelengths were 323 and 343 nm.

For analyses of metabolites generated in in vivo experiments, the injection volume was 50 μl. The gradient used for analyses was as follows: initially at 0% B for 3 min, increased to 12% B over 1 min, constant at 12% B for 6.5 min, increased to 100% B over 4.5 min, constant at 100% B for 4 min, decreased to 0% B over 5 min, and constant at 0% B for 8 min for a total run time of 32 min. Retention times obtained with this gradient were as follows: TU, 9.1 min; 6-TG, 9.4 min; 6-MP, 9.5 min; TX, 10.5 min; 6-MP, riboside 11.1 min; 6-TG riboside, 11.3 min; trans-AVTG, 17.3 min; cis-AVTP, 18.1 min. All metabolites were quantitated using standard curves that were generated by regression of peak area versus concentration of standard solutions spiked into the appropriate biological matrix. Metabolites in bile were not quantitated because of the small and variable amount of bile recovered from each mouse. The standard solutions were made up in dimethylsulfoxide and were 75-fold concentrated for the generation of standard curve in plasma and 80-fold concentrated for the standard curves in RBC, liver, or intestine. The standard solutions were diluted to their final concentration by the addition of 2-μl standard into a final volume of 150 μl for plasma or by the addition of 2.5-μl standard into a final volume of 200 μl for RBC, liver homogenate supernatant, or intestine homogenate supernatant. The standards were then processed identical to the samples with the exception of boiling. Limits of detection for the compounds were generally between 0.05 and 0.1 nmol/ml in the biological matrices examined, except the limits of detection for TU in plasma and RBC were 0.2 and 0.4 nmol/ml, respectively, and the limit of detection for 6-MP riboside in liver was 0.2 nmol/ml.

For analysis of metabolites generated in in vitro experiments, the injection volume was 20 μl. The gradient used for analyses was as follows: initially at 0% B for 2 min, increased to 15% B over 1 min, constant at 15% B for 3 min, increased to 75% B over 3 min, constant at 75% B for 3 min, decreased to 0% B over 3 min, and constant at 0% B for 5 min for a total run time of 20 min. Retention times obtained with this gradient were as follows: 6-TG, 7.9 min; 6-MP, 8.0 min; trans-AVTG, 12.7 min; cis-AVTP, 13.7 min. 6-TG and 6-MP were quantitated by using standard curves that were generated by regression of peak area versus concentration of standard solutions spiked into buffer. The limits of detection of 6-MP and 6-TG were 0.5 nmol/ml.

Assessment of GSH and GSSG Concentrations in Tissues. The effect of drug treatment on tissue GSH and GSSG levels was also examined. For assessment of GSH in plasma and RBCs, approximately 300 μl of whole blood was transferred to an EDTA tube and centrifuged at 14,000g at 4°C. The plasma was transferred to a clean microcentrifuge tube into which were added 0.11 volumes of 50% (w/v) SSA to deproteinize the sample. For analysis of GSH in RBCs, a 100-μl aliquot of packed RBCs was diluted with 200 μl of water and deproteinized by the addition of 0.11 volume of 50% (w/v) SSA. For analysis of GSH in liver or intestine, approximately 500 mg of liver and 350 to 400 mg of small intestine were patted dry, and homogenized in three volumes of 6.67% (w/v) SSA. The resulting homogenate was transferred to microcentifuge tubes except for a small portion that was saved for protein measurements as described above. All samples were centrifuged at 20,000g for 5 min. The resulting supernatant was transferred into clean microcentrifuge tubes and stored at -80°C until analysis. The method used for measurements of GSH and GSSG was based on the enzymatic recycling method of Tietze (1969), as described by Wild and Mulcahy (1999). In brief, for measurements of “total GSH” present in samples, the supernatant was diluted 1/3, 1/40, 1/125, and 1/40 for plasma, RBC, liver, and intestine, respectively, with 5% (w/v) SSA. Standards of GSH were also made up in 5% (w/v) SSA. Aliquots (10 μl) of the diluted supernatant and the standards were pipetted into a 96-well plate. A 100-μl volume of 143 mM phosphate buffer containing 6.3 mM EDTA, 1.05 mM 5,5′-dithio-bis(nitrobenzoic acid) and 0.35 mM NADPH at pH 7.5, was added to each sample, followed by the addition of 50 μl of 5 U/ml GSH reductase. The plate was read several times over 5 min at 412 nm. For the determination of GSSG in samples, the supernatant was diluted 1/2, 1/20, 1/5, and 1/5 for plasma, RBC, liver, and intestine, respectively. GSSG was measured using the assay described above after the free GSH had been derivatized by 2-vinylpyridine (Griffith, 1980). Briefly, 100 μl of the supernatant or GSSG standards were transferred to microcentrifuge tubes into which were added 2 μl of 2-vinylpyridine and 6 μl of triethanolamine. The tubes were vortexed and incubated for 1 h at room temperature in the dark. An aliquot of the derivatized samples or standards (10 μl) was then transferred to a 96-well plate, followed by the addition of phosphate buffer containing 5,5′-dithio-bis(nitrobenzoic acid) and NADPH and GSH reductase as described above. The reaction was monitored for 10 min at 412 nm. GSSG levels present intracellularly or in medium were calculated from the standard curve and expressed as nmol GSSG/ml or nmol GSSG/g tissue. The concentration of reduced GSH in a sample was determined by subtracting the molar amount of GSH equivalents coming from GSSG from the molar amount of total GSH calculated from the standard curve and expressed as nanomole of GSH per milliliter or as nanomole of GSH per gram of tissue.

Statistical Analysis and Calculations of Pharmacokinetic Parameters. All values are reported as the mean ± S.D. with the number (n) of experiments indicated in figure or table legends. One time course using three mice was treated as an independent experiment. AUC was calculated using the trapezoidal rule. Statistical analyses were carried out using Sigma Stat (SPSS Science Inc., Chicago, IL). Comparison of means was assessed using a t test or analysis of variance. Post hoc comparisons were carried out using the Student-Newman-Keul method or Dunnett's test for the comparisons of GSH or GSSG levels in drug treated mice versus mice receiving vehicle only, α was set at 0.05.

Results

Distinct metabolite profiles were observed in plasma, RBC, liver, intestine, and bile after treatment with cis-AVTP, trans-AVTG, and 6-TG. Whereas 6-MP or 6-TG were detected in all the tissues examined, including bile, the thiopurine ribosides, thiopurine nucleotides, and the biologically inactive compounds TX and TU had different tissue profiles and were not detected in all tissues (Figs. 3, 4, 5). After treatment with all the compounds, the concentration of 6-MP or 6-TG in plasma, RBC, and liver was highest at 15 min after which the concentrations fell significantly. Because the 15 min time point was the first time point assessed in this study, higher concentrations may have been achieved at earlier time points. In contrast, 6-MP or 6-TG concentrations in intestine were constant over the 15 to 45 min time period examined (Figs. 3, 4, 5). Low or undetectable levels of thiopurine ribosides and thiopurine nucleotides were detected in plasma and RBC after treatment with cis-AVTP, trans-AVTG, and 6-TG, whereas higher levels were measured in liver and intestine (Figs. 3, 4, 5). The highest concentrations of the biologically inactive metabolites TX and TU were detected in plasma, liver, and intestine after treatment with cis-AVTP, whereas 6-TG treatment yielded the highest concentration of TX and TU in RBCs. Similar to what was observed with the thiopurines, the highest concentration of thiopurine ribosides, thiopurine nucletides, TX and TU was often observed at 15 min in plasma, RBC, and liver whereas their concentration was relatively constant over the time period 15 to 45 min in intestine (Figs. 3, 4, 5).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 3.

The concentration of 6-MP (A), 6-MP riboside (B), 6-MP nucleotides (C), and the combined concentration of TX and TU in plasma, RBC, liver, and intestine after treatment of mice with 42.5 μmol/kg of cis-AVTP (D).

Values presented are the means ± S.D. from three independent experiments.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 4.

The concentration of 6-TG (A), 6-TG riboside (B), 6-TG nucleotides (C), and the combined concentration of TX and TU in plasma, RBC, liver and intestine after treatment of mice with 42.5 μmol/kg of trans-AVTG (D).

Values presented are the means ± S.D. from three-independent experiments.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 5.

The concentration of 6-TG (A), 6-TG riboside (B), 6-TG nucleotides (C), and the combined concentration of TX and TU in plasma, RBC, liver, and intestine after treatment of mice with 42.5 μmol/kg 6-TG (D).

Values presented are the means ± S.D. from three-independent experiments.

The AUC between 15 and 45 min can be used to assess the tissue burden of each metabolite during this time period (Table 1). After treatment with cis-AVTP, the AUC_(15–45) value obtained for 6-MP was significantly higher than the AUC_(15–45) value obtained for 6-TG after trans-AVTG treatment in plasma only, whereas treatment with 6-TG yielded AUC_(15–45) values for 6-TG that were significantly higher than those obtained after trans-AVTG treatment in all the tissues examined (Table 1). Similar findings were observed for the thiopurine ribosides where cis-AVTP treatment yielded significantly lower AUC_(15–45) value than trans-AVTG treatment in intestine only, whereas AUC_(15–45) values for 6-TG ribosides after 6-TG treatment were higher than those obtained after trans-AVTG treatment in all tissues (Table 1). No significant difference was observed between AUC_(15–45) for thiopurine nucleotides obtained after treatment with cis-AVTP and trans-AVTG in all the tissues examined. Interestingly, only in the intestine did 6-TG treatment yield higher AUC_(15–45) value for thiopurine nucleotides than trans-AVTG treatment. The AUC_(15–45) values for the combined concentration of TX and TU were generally lower after trans-AVTG treatment than after treatment with either cis-AVTP or 6-TG.

The ratio that is obtained when the AUC_(15–45) value for the combined concentrations of all metabolites in RBC, liver, or intestine is divided by the AUC_(15–45) value obtained for plasma, can be used to assess whether relative differences in tissue accumulation exist among metabolites of cis-AVTP, trans-AVTG, and 6-TG between 15 and 45 min after treatment (Table 2). cis-AVTP treatment yielded ratios that were significantly lower than the ratios obtained after trans-AVTG treatment for all tissues, whereas only the liver/plasma ratio obtained after 6-TG treatment was significantly lower than that obtained after trans-AVTG treatment (Table 2). When only the potentially cytotoxic metabolites thiopurine, thiopurine riboside, and thiopurine nucleotide were used to calculate the RBC/plasma, liver/plasma, and intestine/plasma ratios, similar results were obtained; treatment with cis-AVTP and 6-TG yielded significantly lower liver/plasma ratios than trans-AVTG treatment, and cis-AVTP treatment yielded lower intestine/plasma ratios than trans-AVTG treatment (data not shown).

To examine whether treatment of mice with 6-TG, trans-AVTG, or cis-AVTP affected the GSH or GSSG levels of plasma, RBC, liver, or intestine, the concentrations of reduced and oxidized GSH were measured in acidified supernatant of tissue homogenates. At 15 min, the concentrations of GSH in plasma and RBC of vehicle only treated mice were 0.06 (mean) ± 0.04 (S.D.) and 1.26 ± 0.51 μmol/ml, respectively, whereas the GSH concentrations in liver and intestine were 8.27 ± 0.73 and 2.48 ± 0.31 μmol/g tissue, respectively. The GSSG concentrations in plasma and RBC of vehicle only treated mice were 0.01 ± 0.01 and 0.37 ± 0.12 μmol/ml, respectively, and 0.33 ± 0.04 and 0.03 ± 0.01 μmol/g tissue for liver and intestine, respectively. The GSH and GSSG concentrations obtained at 30 and 45 min in vehicle only treated mice were not significantly different from those obtained at 15 min. Furthermore, no difference was observed in the concentrations of reduced or oxidized GSH in mice treated with cis-AVTP, trans-AVTG, and 6-TG compared with vehicle only treated mice at any of the time points examined.

To investigate whether GSTs can enhance the rate of formation of thiopurines from cis-AVTP and trans-AVTG, the prodrugs (1 mM) were incubated with GSH (5 mM) in the presence and absence of diluted rat liver cytosol. A small increase in the amount of 6-MP formed was observed when cis-AVTP and GSH were incubated in the presence of cytosol compared with incubations without rat liver cytosol (data not shown). The increase in the amount of 6-MP formed ranged from 10–30%, with the earliest time points exhibiting the largest increase. However, in incubations of 4 min or longer, no difference in 6-MP concentration was detected between incubations with and without rat liver cytosol. No or negligible increase was detected in the amount of 6-TG formed between incubations of trans-AVTG and GSH in the presence and absence of rat liver cytosol.

Discussion

In this work, we have provided for the first time a detailed analysis of metabolites found in tissues after treatment of mice with the novel GSH-activated thiopurine prodrugs cis-AVTP and trans-AVTG. Furthermore, our results demonstrate distinct patterns and distribution of metabolites depending on the compound administered.

The finding that the thiopurines and their metabolites were present in all the tissues assessed after treament with the prodrugs and 6-TG agrees well with previous studies on the tissue distribution of the thiopurines and their metabolites after 6-MP and 6-TG treatment. When marmoset monkeys were given radioactive 6-MP, radioactivity was detected in liver, bile, intestinal content, central nervous system, and bone marrow, whereas mice treated with radioactive 6-MP had radioactivity in blood, kidney, lung, liver, gut, spleen, sternum, and brain (Elion et al., 1954; Schouten et al., 1986). Similarly, radioactivity was detected in blood, kidney, lung, heart, liver, small intestine, spleen, and leg muscle of mice given radioactive 6-TG (Moore and LePage, 1958).

The metabolites detected in tissues after treatment with the prodrugs in this study are consistent with our previous findings where the thiopurines, the thiopurine riboside, TX and TU were detected in urine from mice treated with the prodrugs (Gunnarsdottir et al., 2002b), and where the thiopurines and their ribosides were detected in renal cell carcinoma cells after incubation with the prodrugs (Gunnarsdottir et al., 2002a). Similarly, 6-TG, 6-TG riboside, 6-TG nucleotides, and TU have previously been found in mouse tissues after treatment with 6-TG (Moore and LePage, 1958).

Because the time frame used in this study was relatively short, it is possible that the pattern of metabolite distribution and concentrations may have been different if other time frames had been examined. Although we acknowledge that the AUC_(15–45) values reported in this study only accurately describe the metabolite tissue burden between 15 and 45 min, the metabolite profiles and AUC_(15–45) values shown here agree with and support results obtained from previous experiments examining the tissue toxicity and urinary metabolite excretion in mice given single or multiple treatments with cis-AVTP, trans-AVTG, and 6-TG (Gunnarsdottir et al., 2002b).

In this study, we observed higher AUC_(15–45) values for thiopurines in plasma after treatment with cis-AVTP than after trans-AVTG (Table 1) treatment. Similarly, higher AUC_(15–45) values for 6-TG were observed after 6-TG treatment than after trans-AVTG treatment. Because a positive correlation exists between the plasma AUC value for 6-MP and the amount of 6-MP excreted in urine in children given 6-MP (Endresen et al., 1990), the current results may partly explain our previous findings where a higher fraction of the administered dose was recovered in urine from mice treated with cis-AVTP or 6-TG than from mice treated with trans-AVTG (Gunnarsdottir et al., 2002b).

After treatment with cis-AVTP and trans-AVTG, the AUC_(15–45) values for thiopurines and thiopurine ribosides were comparable except in plasma and intestine, respectively (Table 1). Similarly, the AUC_(15–45) values for cytotoxic thiopurine nucleotides were not significantly different in any tissue examined. Because trans-AVTG is a prodrug of the more toxic thiopurine 6-TG (Elion, 1989), these results may explain our previous results in which multiple treatment with trans-AVTG caused more severe intestinal toxicity than equivalent cis-AVTP treatments (Gunnarsdottir et al., 2002b). Additionally, significantly higher concentrations and AUC_(15–45) values were obtained for 6-TG, 6-TG riboside, and 6-TG nucleotides in the intestine after treatment with 6-TG than after treatment with trans-AVTG (Fig. 3, 4, 5; Table 1). These findings, suggesting higher tissue burden of toxic metabolites in intestine after treatment with 6-TG than after treatment with trans-AVTG, may also explain the more severe intestinal toxicity observed previously after multiple 6-TG treatments than after trans-AVTG treatments (Gunnarsdottir et al., 2002b).

Of interest is the finding that whereas the AUC_(15–45) values for 6-TG, 6-TG riboside, TX, and TU in all tissues and for 6-TG nucleotides in intestine were lower after trans-AVTG treatment than after 6-TG treatment, the AUC_(15–45) value for 6-TG nucleotides in liver was comparable after treatment with both compounds (Table 1). These findings suggest that between 15 and 45 min after treatment, a higher fraction of metabolites is present in the liver in the cytotoxic nucleotide form after trans-AVTG treatment than after 6-TG treatment. Similarly, our current results, showing a significantly higher liver/plasma ratio for trans-AVTG metabolites than for 6-TG metabolites, may suggest that the accumulation of metabolites in liver is relatively higher after trans-AVTG treatment than after 6-TG treatment. These results may partly explain why multiple treatments of mice with trans-AVTG but not equimolar concentrations of 6-TG caused a slight leakage of hepatocellular enzymes as was shown previously to occur after treatment of mice with these compounds (Gunnarsdottir et al., 2002b). These results also suggest that a difference in disposition may exist between trans-AVTG and 6-TG.

Whereas the concentration of total metabolites detected was higher after cis-AVTP treatment than after trans-AVTG treatment, a higher fraction of the cis-AVTP dose was inactivated compared with the trans-AVTG dose (Figs. 3 and 4; Table 1). The highest AUC_(15–45) value for inactive metabolites was obtained in the intestine that has high xanthine oxidase activity (Bronk et al., 1988). Because 6-MP is a substrate of xanthine oxidase, whereas 6-TG must be deaminated before it is oxidized by xanthine oxidase, the higher concentrations of inactive metabolites detected in the intestine after cis-AVTP treatment than after trans-AVTG or 6-TG treatments agree well with previously published results (Bronk et al., 1988).

We chose to use the intraperitoneal route of administration in our studies because it is the most commonly used route of administration in chemotherapy studies in mice. However, the route of administration can significantly affect the bioavailability, pharmacokinetic parameters, and antitumor efficacy of antineoplastic drugs. For example, in terms of 6-MP levels, the oral bioavailability of 6-[(1-methyl-4-nitro-5-imidazolyl)thio]purine (AZA; Fig. 1), a GSH-activated 6-MP prodrug, and a structural analog of cis-AVTP, was significantly higher than after 6-MP treatment, whereas the amount of 6-MP in the systemic circulation was less after oral dosing with AZA than 6-MP (Ding and Benet, 1979). Additionally, AZA had a better chemotherapeutic index against adenocarcinoma 755 when given p.o. than i.p. whereas the reverse was true for 6-MP (Elion et al., 1961). Similarly, the compound 2-amino-6-[(1-methyl-4-nitro-5-imidazolyl)thio]purine (B.W. 57–323) (Fig. 1), the 6-TG analog of AZA, had a superior chemotherapeutic index compared with 6-TG after oral administration, whereas it was less active against adenocarcinoma 755 when administered i.p. (Elion et al., 1960a). Therefore, it is of interest to determine the tissue distribution and disposition of cis-AVTP, trans-AVTG, and 6-TG after oral administration.

The finding that no change in tissue GSH status was detected after treatment of mice with cis-AVTP or trans-AVTG suggests that the prodrugs do not alter the thiol status at the doses used. However, it is conceivable that the prodrugs selectively altered the mitochondrial GSH pool that is smaller in size than the cytosolic pool, but because GSH and GSSG levels were only measured in whole tissue homogenate in our studies, the possible change in the GSH mitochondrial pool would not have been detected. Furthermore, even though no change was detected in GSH levels after treatment of mice with the prodrugs, it is possible that the GSH-dependent metabolism of cis-AVTP and trans-AVTG affects cellular functions and contributes to their antitumor activity. Previously, in studies characterizing the antitumor efficacy of B.W. 57–323, it was discovered that administration of sulfhydryl compounds prior to or simultaneously with B.W. 57–323 administration diminished the antitumor activity of B.W. 57–323. The authors concluded that something else than 6-TG release from B.W. 57–323 contributed to the antitumor activity of the compound (Elion et al., 1960a).

The observation that the amount of 6-MP formed from cis-AVTP was enhanced in the presence of rat liver cytosol suggests that the prodrug is a substrate for GSTs. However, because the increase in 6-MP formation was only observed during the first minutes of incubation, these findings suggest that the role of GSTs in metabolism of the prodrugs may be small because of the rapid rate of the nonenzymatic reaction.

Taken together, our results demonstrate that cis-AVTP, trans-AVTG, and 6-TG exhibit different tissue distribution and metabolite profiles in mice that may play a role in the mouse systemic toxicity and urinary metabolite excretion profile previously observed. The current and previous results (Gunnarsdottir et al., 2002b) show that prodrug-treated mice generally have lower metabolite tissue AUC_(15–45) levels and excrete lower fractions of the administered doses in urine than do 6-TG-treated mice. At present, the fate of the remainder of the administered prodrug dose is uncertain, but it is possible that biliary excretion of the prodrugs and their metabolites is more extensive than for 6-TG. The current results, demonstrating higher liver/plasma ratio for trans-AVTG metabolites than 6-TG metabolites, and previous results showing hepatocellular membrane damage after trans-AVTG treatment, lend credence to this hypothesis. Thus, further examination of the biliary excretion of the prodrugs compared with 6-MP and 6-TG and further investigations into the pharmacokinetic and tissue distribution properties for different routes of administration and at different doses of the prodrugs are warranted.