Drug residue formation from ronidazole, a 5-nitroimidazole. VII. Comparison of protein-bound products formed in vitro and in vivo

Chem.-Biol. Interactions, 50 (1984) 189--202

189

Elsevier Scientific Publishers Ireland Ltd.

DRUG RESIDUE FORMATION FROM RONIDAZOLE, A 5-NITROIMIDAZOLE. VII. COMPARISON OF PROTEIN-BOUND PRODUCTS FORMED IN VITRO AND IN VIVO GERALD T. MIWA, RAUL F. ALVARO, JOHN S. WALSH, REGINA WANG and ANTHONY Y. H. LU

Department of Animal Drug Metabolism, Merck Sharp & Dohme Research Laboratories, Rahway, NJ, 07065 (U.S.A.) (Received December 1st, 1983) (Revision received March 8th, 1984) (Accepted March 14th, 1984)

SUMMARY

In vivo experiments were conducted with ronidazole radiolabelled in the 2-14CH~-, 4,5-14C-, N-~4CH3- and 4-3H-positions. The hepatic protein-bound residues, assessed by the radioactivity of exhaustively washed protein samples, were independent of the radiolabel position and occurred with 4 ) H loss (>80%) in excellent agreement to previous results obtained in vitro with anaerobic incubations of liver microsomes (Miwa et al., Chem. Biol. Interact., 41 {1982) 297). HPLC analysis of acid hydrolyzed in vivo protein-bound residues, obtained from [2-~4CH2]ronidazole, produced a radiochromatographic profile which was virtually identical to that obtained from a similarly treated in vitro sample. Moreover, almost quantitative (76--96%) liberation of radiolabeUed methylamine was obtained from hydrolysates of in vivo and in vitro residue samples formed from [N-14CH3] ronidazole. With 4,5-ring labeled ronidazole the distribution of total radioactivity of the protein hydrolysate on cation exchange resin and the fraction of the residue recovered as oxalic acid were nearly identical for the in vivo and in vitro products. We interpret these data to indicate that ronidazole alkylates proteins with retention of most of the carbon framework of the molecule, in vivo. It is also concluded that the in vitro model, previously used to examine the mechanism of protein alkylation, accurately reflects the salient process initially occuring in the intact animal during the formation of protein-bound residues of this drug.

Key words: Nitroimidazole -- Protein-binding -- Drug residue - - P r o t e i n alkylafion -- Metabolic activation

0009-2797/84/$03.00 © 1984 Elsevier Scientific Publishers Ireland Ltd. Printed and Published in Ireland

190 INTRODUCTION The nature and potential toxicity of protein-bound residues in food producing animals is of major concern to the pharmaceutical industry, governmental regulatory agencies and the public since these residues pose u n k n o w n hazards to the consumers of the foodstuffs produced by these animals [1]. This concern becomes even more grave when the parent drug possesses known toxicity characteristics such as mutagenicity or carcinogenicity. Unfortunately, knowledge a b o u t the toxicity and chemical nature of these drug residues, which are often n o t due to the parent drug, is lacking [2]. For this reason, we have initiated a research program to evaluate the mechanism of formation and toxicity of drug residues from ronidazole, a mutagenic 5-nitroimidazole. Previous studies to characterize the protein-bound residues of this comp o u n d have employed anaerobic in vitro liver microsomal incubations to model the residue forming processes occurring in vivo [3--5]. In vitro protein-bound residues are formed some two to three orders of magnitude greater than those typically formed in vivo. Consequently, the in vitro studies have permitted investigations n o t feasible in the intact animal. These studies have established that hepatic microsomal enzymes are capable of activating ronidazole to reactive intermediates which alkylate principally cysteine thiols moieties on proteins. Radiolabel studies have also suggested that the imidazole nucleus is b o u n d to proteins with loss of the carbamoyl group [ 5]. In addition, the structure of two cysteine adducts [6] and the absence of mutagenicity of one of these adducts have been established [ 7]. In spite of the knowledge gained with the model in vitro system no evidence has been obtained that the protein-bound residues formed in vitro are relevant to those formed in the intact animal. Since this is an essential criterion for extrapolating in vitro findings to the whole animal, we have examined the nature of the protein-bound residues formed by b o t h systems. Positional radiolabel studies, partition and chromatographic characteristics of hydrolyrically formed imidazole ring fragments are virtually identical for the residues formed in both systems. From these data we conclude that the hepatic protein-bound drug residues initially formed (6 h after dosing) in the intact animal are identical to those formed under anaerobic incubations with liver microsomes. MATERIALS AND METHODS Su bstrates and reagents The radiolabeUed ronidazole substrates, 1- [ 14C] methyl-5-nitroimidazole-2methanol carbamate (spec. act., 7.0 uCi/mg), 1-methyl-5-nitroimidazole-2[14C]methanol carbamate (spec. act., 75.6 pCi/mg) and 1-methyl-5-nitro[4,5-'4C]imidazole-2-methanol carbamate (spec. act., 14.0 uCi/mg)were synthesized by Dr. R o b e r t Ellsworth and Mr. Gregory Gatto of the Merck Sharp & Dohme Research Laboratories (Rahway, NJ). All substrates were

191 >97% pure radiochemically and chemically. The 1-methyl-5-nitro-[4-3H]imidazole-2-methanol carbamate (spec. act., 67.5 pCi/mg) was prepared with the assistance of Dr. Avery Rosegay. Methylamine p-toluenesulfonate was prepared by dissolving 10 g of methylamine hydrochloride (Aldrich Chemical Co., Milwaukee, WI) in 25 ml of H20. Ethyl acetate (50 ml) was added and the pH adjusted to 12 with NaOH. p-Toluenesulfonic acid (Aldrich) was added until the methylamine salt precipitated and the solution became acidic. The methylamine p-toluenesulfonate salt was then recrystallized from hot acetonitrile.

In vivo samples Male, Sprague--Dawley rats (Charles River), weighing 250--300 g, were dosed by gavage with 2-14CH2-, N-14CH3- or 4,5-14C-radiolabelled ronidazole (10 mg/kg, spec. act., 75.6, 7.0 and 14.0 ~Ci/mg, respectively). The rats were killed 6 h after dosing and the livers and most of muscle tissue removed for analysis. The tissues were each homogenized with water (1 : 3, w/w) and the resulting homogenate used as in vivo samples. In vitro samples Liver microsomes from male, Sprague-Dawley rats (Charles River), weighing 200--250 g, were prepared and stored at -70°C [3]. Liver microsomes (10 mg) were incubated anaerobically at 37°C with an NADPH-generating system, 0.1 M phosphate buffer (pH 7.4), 0.06 M MgCl2 and 1 mM of the appropriately labelled ronidazole in a total volume of 5 ml as previously described [3]. Sample analysis Steps 1--5 refer to procedures or samples depicted in Fig. 1. Step 1. Following the in vitro incubations sufficient untreated liver homogenate was added to the samples to dilute them to approximately the same specific activity as the in vivo samples*. Solid NaCl was ~/dded to the diluted in vitro samples and the homogenates of the in vivo liver or muscle samples to form a 10% solution (w/w). The samples were boiled at 100°C for 30 min and the resulting denatured proteins were sedimented by centrifugation at 10 000 Xg for 30 min. The protein pellet was suspended in 10% NaCl (30 ml in vitro, 100 ml in vivo) and heated again at 100°C for 30 min. Following centrifugation (10 000 × g for 30 min) the proteins were suspended in acetone (100 ml in vivo and in vitro) and filtered over a sintered glass funnel. The protein residue was washed with additional acetone until the radioactivity in the acetone filtrate was less than 200 dpm/ml. The filtrates were pooled and counted to determine the total radioactivity recovered. The samples were air dried and a small aliquot taken for radioactivity measurements by combustion (Residue 1 ) on a Packard Tri-carb sample oxidizer (Model B306). *We have o b s e r v e d t h a t u n d i l u t e d m i c r o s o m a l residues are h y d r o l y z e d t o m o r e polar prod u c t s w i t h l o w e r r e c o v e r y f r o m t h e HPLC c o l u m n .

192

Step 2. The remaining protein was hydrolyzed, under vacuum, with 6 N HC1 for 16 h at 120°C. Following hydrolysis, the samples were cooled in an ice bath and the contents of the vacuum tube swept with a stream of N2 which was passed through a gas trap containing Carbosorb (Packard) to trap radioactive CO2(14CO2). An aliquot of the remaining residue was counted (Residue 2). Step 3. The hydrolyzed protein liquid was evaporated on a rotovap to remove the bulk of the water and HC1. This procedure was repeated four additional times with 5 ml volumes of water and one time with 5 ml of ethanol. All distillates were pooled and counted to determine the total radioactivity in the distillate. The residue was then suspended in 5 ml of ethanol and an aliquot counted to determine the total radioactivity in Residue 3. Step 4. The ethanol was removed on a rotovap and the residue suspended in water and loaded onto a Sep Pak C18 cartridge (Waters Assoc., Milford, MA.) and eluted stepwise with H20, 30% MeOH/H20 and 50% MeOH/H20. Step 5. The pooled M e O H / H 2 0 fractions from [2-14CH2] ronidazole sampies were dissolved in MeOH. An aliquot was injected onto a Partisil PAC column (Whatman Chemical Separations, Inc., Clifton, NJ) equilibrated with acetonitrile. The sample was eluted with acetonitrile for 5 min followed by a 15-min linear gradient to 100% MeOH. Fractions were taken at 1-min intervals (flow rate 1 ml/min) and counted. The protein-bound residues obtained for [N-14CH3] ronidazole were hydrolyzed and analyzed for ['4C] methylamine by the following isotope dilution method. Non radioactive methylamine p-toluenesulfonate (250 mg) was added to samples of Residue 1 and subsequently hydrolyzed as described for the [2J4CH2] residue samples. Following hydrolysis, NaOH was added to adjust the pH to 10--11. The basic mixture was transferred to a 3-neck flask and flushed with a stream of N2. The outlet stream of N2 was passed through a gas trap containing 150 mg p-toluenesulfonic acid dissolved in 100 ml of ethanol. Following a 2-h sample trapping period, an aliquot of ethanol solution was counted. The sample was evaporated to dryness on a rotovap and redissolved in h o t CH3CN. The [14C] methylamine, crystallized as the sulfonic acid salt, was filtered dry and counted. An additional recrystallization from CH3CN confirmed the constant specific activity of the methylamine salt. Residues from [4,5-14C] ronidazole were analyzed for production of oxalic acid by isotope dilution by adding oxalic acid to the protein sample before hydrolysis with 6 N hydrochloric acid. After hydrolysis, the solution was passed through a Dowex 50W-X8 column (Biorad Labs., Richmond, CA) and the column washed with water. The effluent and water washes were combined and adjusted to pH 5 with sodium hydroxide. Calcium oxaiate was precipitated by adding a 10% solution of calcium chloride until no further precipitate was formed. The calcium oxalate precipitate was filtered, washed several times with water followed by acetone and air dried. The specific activity was determined by radioactivity measurement.

193

Tritium release during the in vivo metabolism of [4-3H] ronidazole The quantity of 3H20 released during in vivo drug residue formation from [4-3H] ronidazole was determined in the rat. Rats were dosed (10 mg/kg) by gavage with a mixture of [4-3H] and [ 2 - 1 4 C H 2 ] ronidazole. The livers were taken 6 h after dosing and homogenized. An aliquot of the homogenate was lyophilized to dryness and the dry residue resuspended in H:O and relyophilized. The total 3H20 in the lyophilisate was quantitated (Table X). A second aliquot of the homogenate was dried under a warm lamp and cornbusted to determine the total non-volatile 14C and 3H in the homogenate while a third aliquot of the homogenate was exhaustively washed with TCA and acetone (pellet). Combustion analysis of the pellet gave the quantities of covalently b o u n d 14C and 3H. A known quantity of non-radioactive ronidazole was added to a fourth aliquot of the homogenate and assayed for residual ronidazole by reverse isotope dilution analysis. RESULTS AND DISCUSSION

The analysis procedure used to characterize the tissue samples for 2-14CH2 labeled ronidazole is illustrated in the flow diagram (Fig. 1). Since the protein residue data in swine were originally evaluated after NaCl acetone washes [8] while all the in vitro protein residue data were obtained after TCA washings [3--5], an experiment was conducted to determine if the different washing methods would result in identical protein-bound residue levels. Table I shows that the radioactivity extracted by the washes and that recovered in Residue 1 from in vivo samples are unaffected by the m e t h o d of washing or the position of the radiolabel. Moreover, the total recoveries were quantita-

STEP 1 (Washes)

TISSUE SAMPLE I [

/

[ 1. NaCI/Acetone or TCA Washes 2. Air Dry 1.6 N HCI, 120°C x 16 Hours 2. Nz Flush of CO2

~ STEP 2 (Hydrolysis) STEP 3

(Distill Volatiles) STEP 4

~

i

~

-~,

i

~

~

1. Rotovap, 4 x 5 m~H~O 1 x 5 rnl Ethanol 2. Suspend in 5 ml Ethanol

1.RotovaD pry

2. Resuspend in HzO 3. C18 Sep Pak Elute With 0-50% MeOH

,SepPa,I

t.,otovapDry

STEP 5 (HPLC)

2. Dissolve in MeOH 3. HPLC on Partisil PAC

Fig. 1. F l o w diagram for [ 2-~"CH= ]ronidazole residue analysis. In vivo liver samples from rats were o b t a i n e d 6 h after a single gavage dose o f [2-24CH=]ronidazole while in vitro microsomal samples were prepared as described in Materials and Methods. The samples were prepared for HPLC analysis as shown in this figure and as described in Materials and Methods.

194 TABLE I C O M P A R I S O N O F T C A A N D N a C 1 / A C E T O N E W A S H I N G M E T H O D S ( S T E P 1) ON IN VIVO RESIDUE DISTRIBUTION R e c o v e r y (%) Label position

Wash a

Wash

Residue 1

Total recovery

2-~4CH~ 2-'4CH~

A B

79 76

16.4 15.4

95.4 91.4

4,5-'4C 4,5-'4C

A B

81 80

20.7 22.7

102.0 103.0

aA, 3 M T C A ; B, 10% NaCl a n d 100°C f o l l o w e d b y a c e t o n e .

tire for both methods. Residues obtained by the two wash methods are directly comparable permitting the use of either method to give residues that could be compared to those in previous in vitro experiments [3--5]. All subsequent experiments were conducted with the NaCl/acetone wash method. Table II compares the radioactivity distribution of [2-'4CH2] ronidazole in in vivo liver and muscle samples with in vitro liver microsomes during the T A B L E II R A D I O A C T I V I T Y D I S T R I B U T I O N O F [ 2-'4CH2 ] R O N I D A Z O L E D U R I N G W A S H I N G S (STEP 1 ) In vivo

I n vitro b

Sample

Total radioactivity

%

Total radioactivity

%

Liver h o m o g e n a t e a

9.16 x 106

100

1.68 X 10 s

100

Washings Washed p r o t e i n ( R e s i d u e 1)

7.56 X 106 1.41 x 106

82.5 15.4

1.12 X l 0 s 2.49 X l 0 t

66.5 14.8

Total recovery

8.97 x 106

97.9

1.37 X 10 s

81.3

Muscle h o m o g e n a t e a

1.65 x 10 ~

Washings Washed p r o t e i n ( R e s i d u e 1)

1.47 x 107 1.09 x l 0 s

89.0 0.7

Total recovery

1.48 x 107

89.7

100

aLiver a n d m u s c l e h o m o g e n a t e s were p r e p a r e d f r o m 1 g tissue a n d 3 g H~O. T h e t o t a l s a m p l e sizes were: liver, 12.3 g; muscle, 23.4 g. b T h e in vitro s a m p l e was i n c u b a t e d as d e s c r i b e d in M e t h o d s b u t 1.0 g p r o t e i n f r o m r a t liver h o m o g e n a t e was a d d e d t o d i l u t e t h e residue t o t h e a p p r o x i m a t e level o b t a i n e d in in vivo samples.

195 TABLE III IN VIVO NON-EXTRACTABLE PROTEIN-BOUND RESIDUES OF [2-~4CH~]RONIDAZOLE (RESIDUE 1) Residue Tissue

(nmol/mg)a

Liver/muscle

Liver Muscle

0.0168 0.00065

25.7 --

aResidue level expressed as nmol ronidazole equivalence bound per mg dry R e s i d u e 1.

washing procedure (Fig. 1, Step 1). The results of these experiments accurately reproduce those shown in Table I in that approx. 80% of the radioactivity was recovered in the washes while a b o u t 15--20% was recovered as Residue 1 for the in vivo liver samples. The in vitro microsomal sample also demonstrated a very similar distribution of radioactivity. In contrast, the in vivo muscle demonstrated considerably lower radioactivity in the protein Residue 1. Essentially all of the radioactivity could be recovered in the washes. Table III is derived from the in vivo liver and muscle data of Table II and expresses the protein-bound Residue 1 in terms of ronidazole equivalents bound per mg of dry protein residue. The bound residue is considerably higher in the liver than in the muscle (liver/muscle ratio 25.7). Accordingly, all subsequent comparisons between in vitro and in vivo residues were done with liver samples. The protein-bound ronidazole expressed as Residue 1 was examined as a function of radiolabel position with in vivo experiments (Table IV). The levels obtained at this early time interval following dosing (6 h) were nearly independent of the position of the radiolabel. This is in excellent agreement with results obtained with in vitro microsomai experiments [ 5] and suggests that the initial protein-binding occurs without fragmentation of the imidazole nucleus in both systems. These in vivo data obtained 6 h after dosing, should be contrasted to similar studies from samples obtained 2--14 days TABLE IV IN VIVO NON-EXTRACTABLE LIVER RESIDUES OF VARIOUS ~4C-LABELLED RONIDAZOLES Rats were dosed by gavage with ronidazole (10 mg/kg) radiolabelled and at specific activity indicated. Six hours after dosing the rats were killed and the livers removed and processed as described in Materials and Methods. Label position

Spec. act. (~Ci/mg)

Residue 1 (nmol/mg) a

N'14CH3 2"4CH2 4,5"4C

7.0 75.6 14.0

0.012 0.016 0.012

aResidue level expressed as nmol ronidazole equivalence bound per mg of dry R e s i d u e 1.

196 after dosing which demonstrated a time-dependent divergence in liver residues produced by the different radiolabels [ 9]. Thus, a relatively large fraction of the residues formed in vivo must initially retain an intact imidazole nucleus, This is confirmed by the high fraction of the in vivo residue which liberates methylamine and by the similarity of the chromatographic data discussed below. In order to further compare the nature of the macromolecular residues, the insoluble fraction from both the in vitro and in vivo experiments was hydrolysed with 6 N HCI (Fig. 1, Step 2) as described in Materials and Methods. Since the radioactive products resulting from the different label sites are different, separate methods were derived to evaluate the hydrolysis products from each label site.

Samples ob rained from [2-14CH2]ronidazole For the 2-14CH2-1abel the radioactivity distribution and recoveries of these hydrolysates were very similar for the in vitro and in vivo samples and are summarized in Table V. Very little (<2%) of the protein-bound radioactivity TABLE V R A D I O A C T I V I T Y DISTRIBUTION O F [2-'4CH~]RONIDAZOLE D U R I N G H Y D R O L Y S I S O F L I V E R S A M P L E S (STEPS 2 A N D 3) The washed and dried liver samples ( R e s i d u e 1) were hydrolyzed, under vacuum, with 5 ml of 6 N HCI for 16 h at 120°C. Total radioactivity (dpm) I n vivo R e s i d u e 1 before hydrolysis After hydrolysis (a) CO 2 (b) Residue 2

(c) Hydrolysis recovery After Rotovap (a) Residue 3 (b) Distillate

In vitro a R e s i d u e 1 before hydrolysis After hydrolysis (a) Residue 2 + CO 2 After Rotovap (a) Residue 3 (b) Distillate

564 000

%

I00

6820 427 100

1.21 76,0

433 920

77.2

313 200 77 760

73.3 18.2

390 960

91.5

24 876 000

100

23 856 000

96

19 300 000 4 176 900

81 17.5

23 477 500

98.5

aIn vitro sample was prepared as described in Materials and Methods except that following the incubation, liver homogenate was added to dilute the specific activity (dpm/mg protein) to approximately that found in the in vivo samples.

197 was liberated as 14C02 during hydrolysis. Most (>75%) of the hydrolyzed radioactivity remained in the acidic liquid Residue 2. Less than 20% of this residue could be removed as the acidic distillate (Step 3) while the greater part (>70%) was retained in a black Residue 3. The recovery in each step was almost quantitative (>90%) except in Step 2 of the in vivo sample (77%) suggesting the liberation of a small quantity of volatile radioactivity, not identified as CO2, from this sample. Similar, qualitative characterization was carried o u t with an in vivo muscle sample (Table VI) which also demonstrated that most of the radioactivity {>79%) was insufficiently volatile to be recovered as ~4CO: (1.2%) or as distillate (1.2%). Before chromatographic characterization of the black Residue 3 could be carried out, a method for clarifying the sample had to be devised. Of several methods examined, Sep Pak C18 cartridges were found to offer the best combination of clarification and recovery for the residues obtained from these samples. Table VII demonstrates that approximately 90% of the radioactivity of Residue 3 could be recovered in H20 and MeOH/H20 fractions from these cartridges while almost all of the colored material was retained on the Sep Pak. The clarified samples were dried, taken up in MeOH and injected onto a Partisil PAC column as described in Materials and Methods and Fig. 1 (Step 5). Figure 2 (right panel) illustrates a radiochromatogram obtained with the in vivo liver sample. There is one major peak eluting with a retention time of 19 minutes. Also evident are three minor peaks with retention times of 4, 16 and 23--24 min. Total recovery from the HPLC run was 91%. An in vitro liver microsomal sample was prepared as described in Materials and Methods by adding sufficient liver homogenate to dilute the residues to a level comparable to the in vivo sample. HPLC analysis of the hydrolyzed sample (Fig. 2, left panel) gave a radiochromatogram virtually identical to T A B L E VI R A D I O A C T I V I T Y D I S T R I B U T I O N OF [ 2-14CH 2] R O N I D A Z O L E D U R I N G H Y D R O L Y S I S O F M U S C L E SAMPLES (STEPS 2 A N D 3) Hydrolysis c o n d i t i o n s as described in Table V. In vivo

Total radioactivity (drop)

%

Residue 1 before hydrolysis

44 000

100

A f t e r hydrolysis (a) CO 2 (b) Residue 2

533 40 300

1.2 91.6

40 833

92.8

31 800 500

79.0 1.2

32 200

80.2

(c) Hydrolysis recovery After Rotovap (a) Residue 3 (b) Distillate

198 TABLE VII SEP PAK C18 RECOVERIES (STEP 4) Recovery % Solvent

In vivo

In vitro a

H20 30% MeOH/H20 50% MeOH/H20

46 32 13

38 46 5

91

89

Total recovery

~I'he in vitro sample is described in Table V.

that observed with the in vivo sample (Fig. 2, right panel). A single, major peak (19 min) along with three minor peaks (4, 16 and 23--24 min) were observed. The recovery from the column was 83%. As the overall recovery through the entire analyses are comparable and good (45--57%) and the chromatographic profiles are identical, it is concluded that the nature of the imidazole fragments produced by hydrolysis of tissue residues from in vitro and in vivo experiments with 2-14CH=-labeled ronidazole is identical. Moreover, subsequent studies have shown that a c o m in Vitro I

F

I

5

10

15

in Viva I

I

I

I

I

I

5

10

15

I

25

.~

zo

80 cr

15

#-

20

25

50

20

25

30

Fraction Number

Fig. 2. Radiochromatograms of in vitro and in vivo [2-1'CH2 ]ronidazole residue samples. In vivo and in vitro samples were prepared for HPLC analysis as described in Materials and Methods and in Fig. 1.

199 pound with chromatographic properties identical to carboxymethylcysteine is contained in the major peak (19 min) providing the first direct evidence that protein cysteine thiols add to the 2-methylene position of ronidazole during protein alkylation (R. Alvaro, P. Wislocki and G. Miwa, unpublished observations).

Samples obtained from [N-14CH3] ronidazole Table VIII summarizes the experiments of methylamine liberation following hydrolysis from in vitro and 6-h in vivo samples. This experiment provides a direct comparison of the two systems. In both systems, the fractions of methylamine released upon hydrolysis, relative to the total Residue 1, were high (76--97%) although there was slightly less methylamine recovered from the 6-h in vitro sample. In contrast, hydrolysis of in vivo tissue residue samples obtained from animals with longer withdrawal intervals resulted in a time-dependent decrease in methylamine liberation [8,9]. Consequently, the slightly lower methylamine levels observed from the 6-h in vivo sample may suggest that, even at this short period after dosing, some breakdown of the initial protein-bound products has occurred.

Samples obtained from [4,5-14C] ronidazole When the 4,5-14C-label is used in vivo, the presence of oxalic acid in a rat muscle tissue sample was observed (R. Alvaro, unpublished observations). Consequently the acid hydrolysates from the in vitro and in vivo residues were analyzed for oxalic acid by the reverse isotope dilution analysis method described in Materials and Methods. The results, summarized in Table IX, show that the distribution of the total radioactivity on cation exchange resin is essentially identical. Moreover, the fraction of the total residue which produces oxalic acid is nearly identical (8.7--10%). The low yield of oxalic acid from the residues may indicate that only a small fraction of a single drug res-

TABLE VIII [ 1'C]METHYLAMINE ANALYSIS OF LIVER RESIDUE SAMPLES FROM [N-I'CH~ ]RONIDAZOLE TREATED RATS [ 14C] Methylamine liberated a In vivo

First crystallization Second crystallization

Average

In vitro

Sample A

Sample B

80.4 79.7

74 72

76.3%

96.0 98.3

97.1%

aMethylamine liberation is expressed as the percentage of total radioactivity of Residue 2.

200 TABLE

IX

RADIOACTIVITY DISTRIBUTION FROM ACID HYDROLYSATES OF A RONIDAZOLE-CYSTEINE STANDARD A N D IN V I V O A N D IN V I T R O T I S S U E RESIDUES [4,5-'4C] Ronidazole was used to form adducts.

R e t a i n e d by D o w e x 50W-X8 Oxalic Acid

In vitro

In vivo

Ronidazole-cysteine adduct standard a

50% 10%

46% 8.7%

-17%

aThe r o n i d a z o l e - c y s t e i n e a d d u c t s t a n d a r d u s e d was 4-S-cysteinyl-1, 2-dimethyl-5-arninoimidazole.

idue is hydrolyzed under these conditions since only a small quantity (17%) of oxalic acid is formed when a pure ronidazole-cysteine adduct, 4-Scysteinyl-l,2-dimethyl-5-aminoimidazole is hydrolyzed (Table IX).

Tritium release during the in vivo metabolism of [4-3H] ronidazole The disposition of [4-3H] ronidazole in the liver of rats was examined six hours after a garage dose of [4-3H] ronidazole. Table X demonstrates that the non-volatile 14C in the liver homogenate (2.44 nmol/ml) was approximately an order of magnitude greater than the covalently b o u n d 14C (0.268 nmol/ml) indicating that approx. 90% of the 14C in the homogenate was due to ronidazole and its metabolites. The composition of the radioactivity in the homogenate could be further assessed from the 3H and 14C in the dried homogenate and lyophilisate. The total non-volatile 3H residue in the dried homogenate was only 24% of the ~4C indicating that no more than 24% TABLE

X

IN V I V O SH R E L E A S E

FROM

[4-SH ]R O N I D A Z O L E Radioactivity Content (nmol/ml) a 4C

Dried h o m o g e n a t e b Lyophilisatec Pellet d

2.44 0 0.268

3H

3H/~4C

0.58 4.22 0.042

0.24 -0.157

aBoth ~4C and 3H radioactivity are expressed as nmol of residue equivalent to ronidazole per ml of liverhornogenate. bAn aliquot of liver hornogenate was dried as described in Materials and Methods. The quantity of radioactivity in the dried homogenate was used to quantitate the nonvolatile drug residues in the sample. CAn aliquot of liver homogenate was lyophilized and the quantity of radioactivity in the lyophilisate was quantitated as described in Materials and Methods. dan aliquot of liver hornogenate was exhaustively washed to give a pellet containing the covalently bound drug residue.

201 (0.58 nmol/ml) of the ~4C observed in the sample was due to unchanged drug. This was also confirmed by reverse isotope dilution analysis which showed less than 0.3 nmol/ml of ronidazole in the homogenate. The 3H distribution between the dried homogenate (0.58 nmol/ml) and the lyophilisate (4.22 nmol/ml) indicates that approx. 88% of the total [4-3H] ronidazole was metabolized with release of 3H. This was also observed in the protein-bound residue since the 3H/~4C ratio (0.157) demonstrated that extensive (84%) loss of the 4-3H had occurred in vivo during the 6-h interval between dosing and sacrifice which resulted in protein-bound products containing very little 4-3H. These data are in excellent agreement with similar studies conducted in vitro (G. Miwa, J. Walsh, R. Wang, R. TuUman and A.Y.H. Lu, in prep.) and provide additional evidence that anaerobic incubations of liver microsomes are an accurate model of the residue forming activity in the intact animal In summary, compelling evidence for the identity of the drug residues initially formed (6 h after dosing) in the intact animal and those produced during anaerobic incubations of ronidazole with liver microsomes and NADPH is provided by the following: (a) the distribution characteristics of the residues during work-up, (b) the identical results from positional label studies, (c) the similar chromatographic profiles of hydrolysis products of residues formed by 2-~4CH2-1abeUedronidazole, (d) the comparable amount of methylamine liberated from N-~4CH3-1abelled residues, (e) the retention characteristics on a cation exchange column and quantity of [14C] oxalic acid formed from residues of [4,5-~4C] ronidazole and (f) comparable quantities of 4-3Hrelease during metabolism. These results are also in agreement with a study that suggested that the mammalian enzymes were principally responsible for protein-bound residue formation [ 10]. The potential contribution of gut microflora to drug residue formation can be discounted, or at least can not be qualitatively different from the host enzymes, since gut bacteria are absent in microsomal incubations. Chemical and enzymatic studies with ronidazole and cysteine have yielded two cysteine adducts whose structures have been established [6]. One of these metabolites, 4~-cysteinyl-2-methylene~-cysteinyl-l-methyl-5-aminoimidazole, demonstrates that cysteine addition can occur at both the C2 methylene and C4 positions. Preliminary evidence also suggests that a protein cysteine thiol has added to the C2 methylene position of the protein-bound residues (R. Alvaro, P. Wislocki and G. Miwa, unpublished observations), providing a direct link between model studies and the protein-bound residues formed by biological systems. Studies are in progress to evaluate the relative importance of the C2 methylene and C4 positions in the protein alkylation, mutagenicity and the therapeutic activity of this nitroimidazole. ACKNOWLEDGEMENTS

The authors would like to thank Dr. Peter Wislocki for a sample of 5amino-4-S-cysteinyl-1,2-dimethylimidazole.

202 REFERENCES 1 The New York Times 132 No. 45,618, p.1 March 15, 1983. 2 M.K. Cordle, J. Environ. Pathol. Toxicol., 3 (1980) 135. 3 S.B. West, P.G. Wislocki, K.M. Fiorentini, R. Alvaro, F.J. Wolf and A.Y.H. Lu, Drug residue formation from ronidazole, a 5-nitroimidazole. I. Characterization of in vitro protein alkylation, Chem.-Biol. Interact., 41 (1982) 265. 4 S.B. West, P.G. Wislocki, F.J. Wolf and A.Y.H. Lu, Drug residue formation from ronidazole, a 5-nitroimidazole. II. Involvement of microsomal NADPH-cytochrome P-450 reductase in protein alkylation in vitro, Chem.-Biol. Interact., 41 (1982} 281. 5 G.T. Miwa, S.B. West, J.S. Walsh, F.J. Wolf and A.Y.H. Lu, Drug residue formation from ronidazole, a 5-nitroimidazole. III. Studies on the mechanism of protein alkylation in vitro, Chem.-Biol. Interact., 41 (1982) 297. 6 P.G. Wislocki, E.S. Bagan, W.J.A. VandenHeuval, R.W. Walker, R.F. Alvaro, B.H. Arison, A.Y.H. Lu and F.J. Wolf, Drug residue formation from ronidazole, a 5-nitroimidazole. V. Cysteine adducts formed upon reduction of ronidazole by dithionite or rat liver enzymes in the presence of cysteine, Chem.-Biol. Interact., 49 (1984) 13. 7 P.G. Wislocki, E.S. Bagan, M.M. Cook, M.O. Bradley, F.J. Wolf and A.Y.H. Lu, Drug residue formation from ronidazole, a 5-nitroimidazole. VI. Lack of mutagenic activity of reduced metabolites and derivatives of ronidazole, 49 (1984) 27. 8 F.J. Wolf, F.P. Baylis, G.E. Smith, C. Rosenblum, H.T. Meriwether, R.F. Aivaro, D.E. Wolf, F.R. Koniuszy and T.A. Jacob, Disposition of ronidazole by swine. I. Radiocarbon content of tissues, J. Agric. Fd. Chem., 31 (1983) 559. 9 F.J. Wolf, R.F. Alvaro, L.R. Chapin, M.L. Green, D.E. Wolf, F.R. Koniuszy, J.J. Steffens and T.A. Jacob, Tissue residue studies with ronidazole, effect of label site on total radioactivity content of rat tissue, J. Agric. Fd. Chem., (1984} in press. 10 F.J. Wolf, R.F. Alvaro, K. Fiorentini, M. Green, A.Y.H. Lu, G.T. Miwa, P.G. Wislocki, R.T. Koch, M.A. McLafferty and P. Goldman, Drug residue formation from ronidazole, a 5-nitroimidazole. IV. The role of the microflora, Chem.-Biol. Interact., 45 (1983) 7.