Introduction

Top Abstract Introduction Results Discussion References

Porfiromycin (PM, Fig. 1)¹, the aziridine N-methyl analog of mitomycin C (MC), is a bioreductive alkylating agent that has been shown to be preferentially cytotoxic to hypoxic tumor cells in vitro and in vivo (Iyer and Szybalski, 1964; Keyes et al., 1985; Fracasso and Sartorelli, 1986; Marshall and Rauth, 1988; Rockwell et al., 1993). Interestingly, PM generates a greater differential toxicity to hypoxic EMT6 and Chinese hamster ovary tumor cells versus their aerobic counterparts than that of MC (Rockwell et al., 1988; Belcourt et al., 1996). Furthermore, human phase I and phase II clinical trials have demonstrated that PM is tolerated at a dose 3-fold higher than that of MC (Foley et al., 1967; Loo et al., 1967; Izbicki et al., 1972; Grage et al., 1975; Baker et al., 1976). The findings indicate that PM is potentially superior to MC in clinical applications for the treatment of solid tumors. Currently, PM is under development for the treatment of head and neck cancers as an adjunct to radiation therapy in phase III clinical trials.

View larger version (13K): [in this window] [in a new window]   Fig. 1.   Structure of PM.

In vitro metabolism and bioreductive activation of MC under hypoxic conditions have been extensively studied (Schwartz, 1962; Tomasz and Lipman, 1981; Kennedy et al., 1982; Pan et al., 1984). The enzymes responsible for the reductive metabolism of MC are NADPH:cytochrome c (P-450) reductase (Keyes et al., 1984; Pan et al., 1986), xanthine oxidase (Pan et al., 1984, 1986), xanthine dehydrogenase (Gustafson and Pritsos, 1992), DT-diaphorase (Siegal et al., 1990; Pan et al., 1992, 1995; Robertson et al., 1992; Beall et al., 1994), and NADH:cytochrome b₅ reductase (Hodnick and Sartorelli, 1993). In contrast, the information on the metabolism and disposition of PM is limited. To date, three metabolites of PM resulting from biotransformation at the C-1 position in the presence of purified NADPH:cytochrome c reductase and xanthine oxidase have been identified (Pan and Iracki, 1988; Pan, 1990). Recently, we reported the identification of several C-1 and C-10 metabolites of PM arising from a rat liver preparation under aerobic conditions (Lang et al., 2000). However, the in vivo metabolism and disposition of PM under physiological conditions are still unclear.

To understand the metabolic fate of PM, we conducted the studies on the metabolism and disposition of PM in dogs and humans. In this article, we present: 1) urinary and fecal excretion of radioactivity derived from [methyl-³H]PM in dogs; 2) isolation and identification of urinary metabolites of PM in dogs and humans; 3) comparison of urinary metabolite profiles between the two species; 4) chemical syntheses of the major urinary metabolites; and 5) postulation of metabolic pathways of PM on the basis of the metabolite profiles.

Experimental Procedures

Materials. Porfiromycin injection, 15 mg, (Promycin) was manufactured at Chesapeake Biological Lab., Inc. (Owings Mills, MD). [Methyl-³H]PM (1.4 mg/ml of ethanol), specific activity 235 mCi/mmol, was purchased from Moravek Biochemicals, Inc. (Brea, CA). Methanol and potassium phosphate monobasic were of HPLC grade, and were purchased from Aldrich Chemical Co. (Milwaukee, WI) and Fisher Scientific Co. (Pittsburgh, PA), respectively. Ammonium acetate, potassium hydroxide, and 37% hydrochloric acid were of American Chemical Society reagents and were purchased from J. T. Baker, Inc. (Phillipsburg, NJ). Tris(hydroxymethyl)-aminomethane, L-cysteine, and N-acetyl-L-cysteine were purchased from Aldrich Chemical Co. Milli-Q Plus ultra pure water was used throughout the study.

Animals, Dosing, and Sample Collection. Three adult male Beagle dogs (Marshall Laboratories of North Rose, NY), weighing from 8.4 to 9.2 kg, were used in this study. A dosing solution was prepared by dissolving six vials of porfiromycin injection (15 mg) and [methyl-³H]PM (0.7 mg) into a final volume of 45 ml of water to produce a 2 mg/ml [methyl-³H]PM solution (specific activity 5.46 µCi/mg). The dogs were given a single i.v. dose of PM dosing solution at 2 mg/kg (40 mg/m²) and housed individually in metabolism cages with access to water and food. Urine and feces of the test animals were collected at 0 to 4, 4 to 8, and 8 to 24 h, and at 24-h intervals thereafter for 7 days. The samples were stored immediately at 20°C.

Human Urine Sample. A human urine sample obtained from a patient who received a PM/radiation therapy trial at Yale New Haven Hospital is presented in this article. The patient was given a single i.v. dose of PM at 40 mg/m². The urine sample was collected at 0 to 2.5 h after the dose and stored at 80°C until used.

Urine Sample Preparation. After thawing, each 10-ml urine sample was filtered through a 0.45-µm nylon filter. The filtrate was loaded onto two Waters Sep-Pak Plus C18 cartridges connected in tandem (sorbent weight of 360 mg each). The cartridge set, preconditioned with 5 ml of methanol and 10 ml of water sequentially, was washed with 10 ml of water, and then eluted with 5 ml of methanol/water (2:1, v/v). The methanolic effluent was collected and concentrated under a stream of nitrogen. Aliquots of 100 µl of the resulting solution were repeatedly injected onto HPLC for isolation.

HPLC. A Hewlett-Packard 1100 liquid chromatography system (Wilmington, DE) was used for isolation of urinary metabolites and the determination of PM in dog urine. The LC system consisted of a quaternary pump G1311A, an autosampler G1313A, a diode array detector G1315A, and a column thermostat G1316A. Hewlett-Packard LC3D ChemStation software was used for system control and data processing. HPLC isolation and analysis were carried out on a Supelcosil LC-18 column, 5 µm, 150 × 4.6 mm (Supelco, Bellefonte, PA) at 40°C. Detection wavelength was set at 550 nm for the isolation of metabolites and at 360 nm for the quantitation of PM. The mobile phases were methanol (A) and 50 mM potassium phosphate buffer, pH 6.0 (B). A linear gradient elution was performed via increase of 0 to 60% A (v/v) over 30 min at a flow rate of 1.0 ml/min. The effluent was collected with a Gilson fraction collector (model FC204; Gilson, Middleton, WI) at 0.25 ml/vial intervals. Fractions containing the same chromatographic component(s) were pooled. Solvent in each pooled fraction was removed under a stream of nitrogen. Residue was reconstituted with the appropriate amount of water, and the resulting solution was subjected to liquid chromatography/mass spectrometry (LC/MS) analysis.

LC/MS Analysis. Mass spectra of metabolites were obtained on a Navigator single quadrupole mass spectrometer equipped with an atmospheric pressure chemical ionization (APCI) probe (Finnigan, San Jose, CA). An HP 1100 LC system was interfaced to the MS detector via a heated nebulizer. Before analysis of metabolites, instrumental parameters were optimized for detection and fragmentation via infusion of 2 µg/ml PM in methanol/water (1:1, v/v) at a speed of 10 µl/min. The APCI source was operated at 150°C with the corona pin voltage of 2.6 kV and cone voltage of 10 to 30 V. The flow rate of drying gas (N₂) was 5 liter/min. The nebulizer coil temperature was set at 550°C. MS spectra were acquired over m/z 100 to 800 at 2.5 s/scan in the positive ion mode. HPLC was performed on a BDS Hypersil C8 column, 3 µm, 100 × 2 mm, (Keystone Scientific, Bellefonte, PA) at 40°C. An isocratic elution was carried out with 15% (v/v) methanol in an aqueous 20 mM ammonia acetate solution (pH 6.1) at a flow rate of 0.3 ml/min.

Liquid Scintillation Counting. Total radioactivity in dog urine was assayed directly on a LC 6000TA or a LS 6000IC Liquid Scintillation Counter (Beckman, Fullerton, CA) in 5 ml of Ready Value cocktail (Beckman, Palo Alto, CA). All urine samples were counted in duplicate to a statistical accuracy of 2% or for a maximum of 10 min. For comparison of the radioactivity profiles of HPLC with those by UV detection for dog urine, the effluent collected from HPLC at 0.25 ml/vial was counted after mixing with 3 ml of ScintiSafe Gel cocktail (Fisher Scientific, Pittsburgh, PA).

¹H NMR Analysis. ¹H NMR spectra of metabolites in deuteromethanol (CD₃OD) were recorded on a Bruker AC300 NMR spectrometer at 300 MHz. Either internal tetramethylsilane or residual methanol signal was used as a chemical shift reference. The following terminology is used: , chemical shift in parts per million; s, singlet; d, doublet; dd, doublet of doublets; q, quartet; m, multiplet; and J, coupling constant in hertz.

Chemical Syntheses of Metabolites.

10-Decarbamoyl-2-methylamino-10-(L-cysteine-S-yl)-7-aminomitosene (M1) L-Cysteine (17.8 mg) and PM (5.1 mg) were dissolved in 2.0 ml of an aqueous 0.1 M Tris/160 HCl buffer, pH 7.0. The solution was deaerated by bubbling He gas for 5 min, then 2.1 mg of platinum oxide was added. The reaction was carried out by bubbling H₂ gas at 60-65°C for 2 h. The reaction mixture was then exposed to air, filtered, and concentrated in vacuo. The product was purified by reversed-phase high performance liquid chromatography (RP-HPLC). Electrospray ionization-mass spectrometry (ESI-MS): m/z (intensity %) 379 (MH, 100), 258 (56); ¹H NMR (methanol-d₄):  1.78 (s, C(6)CH₃), 2.42 (s, NCH₃), 2.72 (dd, J = 16.4, 4.48 Hz, C(1)H), 2.77 (dd, J = 14.4, 9.9 Hz, C(1')H), 3.14 (dd, J = 14.5, 3.7 Hz, C(1')H), 3.18 (dd, J = 16.5, 7.3 Hz, C(1)H), 3.75 (dd, J = 9.7, 3.7 Hz, C(2')H), 3.83 (1/2ABq, J = 13.3, C(10)H), 3.94 (m, C(2)H), 3.96 (1/2ABq, J = 13.4, C(10)H), 4.02 (dd, J = 13.1, 4.6 Hz, C(3)H), and 4.37 (dd, J = 12.1, 6.7 Hz, C(3)H).

1,2-trans and 1,2-cis-1-Hydroxy-2-methylamino-7-aminomitosenes (M2 and M4). PM (14.0 mg) was dissolved into 10 ml of an aqueous 0.05 N HCl solution and stirred at room temperature for 3 h. The reaction solution was neutralized with a saturated aqueous NaHCO₃ solution and concentrated in vacuo. The crude diastereomers (M2 and M4) were separated and purified by semipreparative HPLC. The purification was performed on a Ultracarb ODS (30) column, 5 µm, 250 × 10 mm, at 50°C. The column was eluted with methanol/50 mM KH₂PO₄, pH 6.0, (3:7, v/v) at a flow rate of 2.0 ml/min. The effluent was collected with a Gilson fraction collector. Fractions containing the same product were pooled and concentrated in vacuo. The final solution of each product was purified using Sep-Pak Plus C18 cartridges to remove salt. The yield was 18% for M2 and 50% for M4. M2ESI-MS: m/z (intensity %) 335 (MH, 80), 317 (7) and 274 (100); ¹H NMR (methanol-d₄):  1.80 (s, C(6)CH₃), 2.45 (s, NCH₃), 3.62 (m, C(2)H), 4.01 (dd, J = 13.4, 3.2 Hz, C(3)H), 4.47 (dd, J = 13.3, 6.3 Hz, C(3)H), 4.96 (d, J = 2.3 Hz, C(1)H), 5.16 (1/2ABq, J = 12.6 Hz, C(10)H), 5.24 (1/2ABq, J = 12.6 Hz, C(10)H). M4ESI-MS: m/z (intensity %) 335 (MH, 100), 317 (2) and 274 (21); ¹H NMR (methanol-d₄):  1.79 (s, C(6)CH₃), 2.50 (s, NCH₃), 3.60 to 3.70 (m, C(2)H), 3.76 (dd, J = 12.0, 8.8 Hz, C(3)H), 4.54 (dd, J = 12.0, 7.1 Hz, C(3)H), 5.10 (d, J = 5.1 Hz, C(1)H), 5.17 (1/2ABq, J = 12.9 Hz, C(10)H), 5.25 (1/2ABq, J = 12.9 Hz, C(10)H).

10-Decarbamoyl-2-methylamino-10-(N-acetyl-L-cysteine-S-yl)-7-aminomitosene (M5). N-Acetyl-L-cysteine (39.6 mg) and Tris base (30.2 mg) were dissolved in 2.0 ml of water to form a solution (pH 6.5), then 11.7 mg of PM was added. The solution was deaerated by bubbling He gas for 5 min, and then 2.1 mg of platinum oxide was added. Hydrogenation proceeded by bubbling H₂ gas at 50-55°C, and was terminated at 2 h. The reaction mixture was exposed to air and filtered through a 0.45-µm nylon filter. The filtrate was concentrated in vacuo, and the residue was further purified by RP-HPLC. ESI-MS: m/z (intensity %) 421 (MH, 100); ¹H NMR (methanol-d₄):  1.79 (s, C(6)CH₃), 1.94 (s, C(O)CH₃), 2.72 (s, NCH₃), 2.67 (dd, J = 13.2, 7.0 Hz, C(1')H), 3.09 (dd, J = 13.2, 4.4 Hz, C(1')H), 3.15 (dd, J = 18.1, 2.9 Hz, C(1)H), 3.21 (dd, J = 17.7, 7.0 Hz, C(1)H), 3.62 (1/2ABq, J = 14.2, C(10)H), 4.07 (d, J = 14.2, C(10)H), 4.23 (dd, J = 6.0, 4.4 Hz, C(2')H), 4.25 (m, 1H, C(2)H), 4.33 (dd, J = 14.2, 2.9 Hz, C(3)H), 4.46 (dd, J = 14.2, 6.1 Hz, C(3)H).

2-Methylamino-7-aminomitosene (M6). PM (24.7 mg) was dissolved in 10 ml of an aqueous 0.1 M KH₂PO₄ buffer, pH 4.55, and then 5.7 mg of platinum oxide was added. The mixture was deaerated by bubbling He gas for 5 min and followed by bubbling H₂ gas at room temperature for 1 h. The reaction mixture was exposed to air, and then filtered. The filtrate was concentrated in vacuo and the residue was purified by RP-HPLC. The purification was performed on a Ultracarb ODS (30) column, 5 µm, 250 × 10 mm, at 50°C. The column was eluted with MeOH/50 mM KH₂PO₄, pH 7.30 (45:55, v/v) at a flow rate of 2.0 ml/min. ESI-MS: m/z (intensity %) 319 (MH, 100), 258 (9); ¹H NMR (methanol-d₄):  1.78 (s, C(6)CH₃), 2.40 (s, NCH₃), 2.71 (dd, J = 16.5, 4.8 Hz, C(1)H), 3.17 (dd, J = 16.5, 7.3 Hz, C(1)H), 3.90 (m, C(2)H), 3.98 (dd, J = 12.9, 4.7 Hz, C(3)H), 4.37 (dd, J = 12.8, 6.7 Hz, C(3)H), 5.14 (s, C(10)H₂).

	Results

Top Abstract Introduction Results Discussion References

Chemical Syntheses. Urinary metabolites M1, M2, M4, M5, and M6 were synthesized in analytical or semipreparative scales using PM as a starting material. Of these compounds, M1, M5, and M6 were prepared under a reductive condition using catalytic hydrogenation by platinum oxide. The procedures for synthesis of M1 and M5, in which an S-nucleophile displaced the C-10 carbamate moiety of PM at 50-65°C, were novel and efficient. The higher temperatures (50-65°C) were applied to facilitate the removal and activation of the C-10 functionality of PM due to its thermal lability. M6 was prepared by catalytic hydrogenation under a mildly acidic condition (pH 4.55). Tomasz and Lipman (1981) have previously described the mechanism of the formation of M6 via an internal redox reaction. A pair of diastereomeric 1-hydroxymitosenes (M2 and M4) were synthesized in a single step by acidic activation of PM in an aqueous 0.05 N HCl solution as described by Hong and Kohn (1991). The attempt to synthesize metabolites M3 and M7 was unsuccessful. All final products were purified by RP-HPLC and characterized by UV-Vis, MS, and ¹H NMR. The synthetic compounds were used for structural confirmation of the metabolites identified in urine by comparison of their HPLC retention times (R_Ts) and MS fragmentation patterns.

Urinary Excretion in Dogs. Cumulative urinary excretions of radioactivity derived from [methyl-³H]PM and the recoveries of PM in three Beagle dogs after a single i.v. dose over 7 days are given in Table 1. An average of 48.0% of total radioactivity given to the dogs was cumulatively excreted in urine over a period of 7 days after the dose. In the first 24 h, about 43.6% of the administered dose was excreted in urine; this accounted for approximately 92% of total radioactivity excreted in urine over the entire period studied. Unchanged parent drug excreted in urine accounted for 10.8% of the administered dose over 7 days. The findings indicated that the majority of excreted dose in dog urine was in the form of metabolites, and urinary excretion of the metabolites was rapid and intensive in dogs.

Identification of Urinary Metabolites. Representative HPLC profiles showing human and dog urinary metabolites are given in Fig. 2. PM and its metabolites were selectively detected at 550 nm due to their characteristic indoquinone chromaphores. Therefore, PM metabolites could be readily differentiated from other complex components present in urine matrices. A total of seven major metabolites M1 to M7 were observed in human urine (Fig. 2A), and a similar HPLC pattern was obtained for dog urine (Fig. 2B). A major peak M1 at R_T 12.70 min observed in both human and dog urine had a typical mitosene UV absorption pattern (_max at 250, 310, and 550 nm). The APCI mass spectrum of M1 showed a cluster of molecular ions at m/z 379 (M+H)⁺, 401 (M+Na)⁺, and 417 (M+K)⁺ in Fig. 3A. The observation of two fragment ions at m/z 335 (MH- CO₂)⁺ and 317 (MH- COOH- NH₃)⁺ suggested the existence of carboxylic acid and amino groups in the molecule. A base peak at m/z 292 (MH-CH₂=C(NH₂)COOH)⁺, and two related fragment ions at m/z 260 (MH- HSCH=C(NH₂)COOH)⁺ and 258 (MH- cysteine)⁺ indicated that M1 was a cysteine conjugate. Further confirmation of the structure of M1 was obtained by comparison with our synthetic reference standards. It was found that the HPLC R_T and MS fragmentation pattern of M1 matched those of 10-decarbamoyl-2-methylamino-10-(L-cysteine-S-yl)-7-aminomitosene. The findings supported our structural assignment of M1.

View larger version (23K): [in this window] [in a new window]   Fig. 2.   Representative chromatograms (550 nm) of human urine (0-2.5 h) after administration of a single i.v. dose of 40 mg/m2 of PM (A) and dog urine (0-8 h) after administration of a single i.v. dose of 2 mg/kg of [3H]PM (B). HPLC conditions are described in Experimental Procedures.

View larger version (21K): [in this window] [in a new window]   Fig. 3.   APCI mass spectra of metabolites M1, 10-decarbamoyl-2-methylamino-10-(L-cysteine-S-yl)-7-aminomitosene) (A) and M3, 10-decarbamoyl-1-(L-cysteine-S-yl)-2-methylamino-7-aminomitosene) (B).

View larger version (13K): [in this window] [in a new window]   Fig. 4.   A representative APCI mass spectrum for metabolites M2 (1,2-trans-1-hydroxy-2-methylamino-7-aminomitosene) and M4 (1,2-cis-1-hydroxy-2-methylamino-7-aminomitosenes).

View larger version (27K): [in this window] [in a new window]   Fig. 5.   Proposed metabolic pathways of PM and the structures of urinary metabolites.

View larger version (11K): [in this window] [in a new window]   Fig. 6.   APCI mass spectrum of metabolite M7 (10-decarbamoyl-2-methylamino-1-(N-acetyl-L-cysteine-S-yl)-7-aminomitosene).

	Discussion

Top Abstract Introduction Results Discussion References

Recently, we reported the identification of metabolites of PM formed in the presence of a rat liver preparation (Lang et al., 2000). Several C-1 and C-10 metabolites of PM were characterized as mitosene analogs in the in vitro system. The findings suggest that the formation of the metabolites is initiated by enzymatic activation of PM to a reactive intermediate with two reactive centers, which subsequently link with environmental nucleophiles in the aqueous medium. The current in vivo studies were our continuing effort to explore the metabolic fate of PM.

Tritium labeled to the N-methyl group of PM was used in the dog study. The isotope is not exchangeable with mobile protons in an aqueous medium. The potential loss of the label by demethylation catalyzed by demethylases has been previously investigated in P388 cell homogenates and liver microsomes (Pan, 1990). The results have demonstrated that PM is a poor substrate for the enzymes. So far, no demethyl metabolites of PM have been found. During the dog study, the effluents from HPLC column and Sep-Pak Plus C18 cartridges were collected, and were subjected to liquid scintillation counting for evaluation of the recoveries of radioactivity in each step. The radiochromatograms of the dog urine samples (data not shown) matched the metabolite profiles of HPLC monitored at 550 nm.

Three phase I metabolites (M2, M4, and M6) were identified in both human and dog urine. The findings are consistent with those previously obtained in vitro (Pan and Iracki, 1988; Lang et al., 2000). Most importantly, we found that M6 is a reactive metabolite that can be directly converted to cysteinyl S-conjugates (M1 and M5) via further biotransformation at the C-10 position (Fig. 5). The further metabolism of M6 is believed to be through the loss of the C-10 carbamate group to generate a positive charge, which is stabilized by the indole nitrogen (Iyer and Szybalski, 1964). The confirmation of the existence of such a metabolic pathway of M6 was conducted in vitro by direct incubation of M6 with L-cysteine to produce M1 under a reductive condition via catalytic hydrogenation (data not shown). No corresponding C-10 cysteine S-conjugates for 1-hydroxy-mitosenes (M2 and M4) were identified under the same conditions.

A pair of isomeric phase II metabolites (M1 and M3) as cysteine S-conjugates of mitosene analogs at C-1 and C-10 were identified in both human and dog urine, whereas the pair of metabolites (M5 and M7) as N-acetylcysteine S-conjugates (mercapturic acids) were only observed in human urine (Fig. 2). The formation of mercapturic acids in general is believed to be through N-acetylation of cysteine conjugates in the kidney. The reaction may be catalyzed by the cysteine conjugate N-acetyl-S-transferase (Jakoby, 1978; Elfarra and Anders, 1984). The observation of the differences in the formation of mercapturic acids between humans and dogs indicates that the enzymatic activity present in humans may be much higher than that in dogs. In addition, the identification of the phase II metabolites in urine suggests that GSH conjugation may be involved in the biotransformation of PM. As expected, the significant amount of total radioactivity (31.8%) was found in feces after i.v. administration of [methyl-H³]PM to three dogs (Table 1) because GSH conjugates are commonly excreted bile by a carrier-mediated transport system (Ketterer et al., 1983). Two cysteine S-conjugates of PM identified in both dog and human urine may be the products of subsequent metabolism of the GSH conjugates (Igwe, 1986). In summary, the identification of the phase II metabolites of PM linked to the thiol species in dog and human urine at the C-1 and C-10 positions is a novel finding. On the basis of the metabolite profiles, we propose the in vivo metabolic pathways of PM in Fig. 5. The identification of the urinary metabolites of PM is important to understand its in vivo metabolic fate and disposition of PM in dogs and humans.