EUROPEAN JOURNAL OF DRUG METABOLISM AND PHARMACOKINETICS, 1992, Vol. 17, No.1, pp. 51-59

Metabolism of diazepam and related benzodiazepines by human liver microsomes W.D. HOOPER, I.A. WAIT, G.E. McKINNON and P.E.B. REILLY Departments ofMedicine andBiochemistry, The University ofQueensland, Brisbane, Queensland, Australia

Received for publication: April 29, 1991 Keywords: Diazepam, benzodiazepines, microsomes,cytocbrome P450, metabolism

SUMMARY The metabolism of diazepam has been studied in vitro using microsomal preparations from five human livers. An HPLC method was developed for the assay of diazepam, its congeners and its metabolites. Various methods for the incorporation of diazepam into the incubation medium were explored. It was shown that the use of organic solvents or small quantities of hydrochloric acid enhanced the solubility of this substrate. However all of the organic solvents tested were associated with substantial (around 50%) inhibition of metabolism of diazepam by both major pathways ( N-demethylation and C3-hydroxylation). The use of hydrochloric acid gave satisfactory solubilization of diazepam, but not of pinazepam, prazepam or halazepam. Detailed metabolic studies were conducted only for diazepam, using neither hydrochloric acid nor organic solvents in the incubation medium. Formation of N-desmethyldiazepam increased approximately linearly with diazepam concentration to 200 f1M, and did not show saturation. Formation of temazepam gave a curved profile over the same range of diazepam concentrations, suggestive of a sigmoidal relationship. Michaelis-Menten parameters could not be determined for either reaction, but intrinsic clearances for N-demethylation varied over a 6-fold range. Diazepam N-demethylation was apparently promoted by the inclusion of temazepam in the incubation medium, while C3-hydroxylation of diazepam was enhanced in the presence of N-desmethyldiazepam. Mephenytoin in the incubation mixture had no effect on diazepam metabolism by either pathway. The present studies have defined some of the methodological problems inherent in in vitro metabolic studies with benzodiazepines, and have shed further light on the metabolism of diazepam in vitro by human liver.

INTRODUCTION The first marketed benzodiazepine (chlordiazepoxide) was introduced into therapy in the USA in 1960, and was soon followed by a series of related compounds including diazepam in 1963 and oxazepam in 1965 (1). Many other compounds of the benzodiazepine class have subsequently achieved therapeutic application on account of their hypnotic, anxiolytic, anticonPlease send reprint requests to : Dr W.O. Hooper, Department of Medicine, Clinical Sciences Building, Royal Brisbane Hospital, Herston, Queensland 4029, Australia

vulsant or muscle-relaxant properties. The metabolic disposition of the benzodiazepines has been extensively studied, but is still only incompletely documented. The early studies have been well summarized (2, 3). For diazepam and several congeners (e.g. prazepam, pinazepam and halazepam) (Fig. 1) the major metabolic events are believed to be N-dealkylation and hydroxylation at carbon-3, though quantitative data on the disposition of these compounds in man are scant. There are wide inter-species differences of both a qualitative and a quantitative nature in the metabolism of the benzodiazepines. Most of the published work on in vitro metabolism has been performed with

52

Eur. J. Drug Metab. Pharmacokinet., 1992, No.1

Cl

R1

Compound

Rs

Diazepam

CH s

H

Metabolites: Temazepam N-Desmethyldiazepam Oxazepam

CHs H H

OH H OH

Congeners: Camazepam Halazepam Pinazepam Prazepam

CH s OCON(CHs)z H CFsCH z Propargyl H CyclopropylH methyl

Fig. 1 : Chemical structures of diazepam, its major metabolites, several congeners, and camazepam

animal tissues, principally rat liver. The relevance of these data to understanding of benzodiazepine metabolism in man is limited. The present studies were init-

iated to examine the metabolism of benzodiazepines using human liver microsomes, both to assess the inter-subject variability in the metabolism of diazepam in man, and to investigate the comparative metabolism of diazepam, pinazepam, prazepam and haIazepam (which differ only in the substituent at N-l; Fig. 1). However a review of the literature showed that many in vitro metabolism studies involved methodology which was open to criticism. Some examples are the use of organic solvents to solubilize benzodiazepines in the incubation medium (4-6), inappropriately long incubation times (e.g. 60 min (7», and addition of nicotinamide, a monooxygenase inhibitor, to the incubating medium (8, 9). Diazepam metabolism was studied in human liver microsomes (adult and fetal) by Ackermann and Richter (10), who showed that production of both N-desmethyldiazepam and temazepam (3-hydroxydiazepam) increased linearly with time. The only other in vitro study using human liver is that of Inaba et al. (11) who showed that those two pathways were catalysed by different cytochrome P450 isoenzymes, a conclusion which was recently supported by Bertilsson et aI. (12). Shimada and co-workers (13) demonstrated that the purified human liver cytochrome P450 which catalysed the 4-hydroxylation of S-mephenytoin also catalysed the N-demethylation of diazepam. We present here studies on the metabolism of diazepam by human liver microsomes, with particular reference to methodological issues. Inhibition studies directed at further characterizing the isoenzymes of cytochrome P450 involved in diazepam metabolism,

Table I : Particulars of subjects from whom liver specimens were obtained

Subject

Gender

Age

A

F

30

B

F

61

Medical history

Drugsreceived prior to specimen collection

Asthma; died of cardio-

albuterol, aminophylline, epinephrine, hydrocortisone.

respiratory arrest

isoproterenol

Asthma; died of cardio-

aminophylline, dopamine, folic acid, heparin.

respiratory arrest

hydrocortisone, methylprednisolone, phenytoin, ranitidine

C

D

E

M

F

M

68

63

20

Hepatoma; right lobe

amitriptylline, cloxacillin, methyldopa, metoprolol,

liver resection

temazepam

Hepatoma;

allopurinol, atenolol, captopril, digoxin, furosemide,

hemihepatectomy

oxazepam, thyroxine

Died in motor

ethanol

vehicle accident

W. D. Hooper et al., Diazepam & related benzodiazepines and feasibility studies with prazepam, pinazepam and halazepam are also reported.

MATERIALS AND METHODS Microsomal preparations Human liver specimens were obtained from organ donors (A, B, E) or from patients undergoing liver resection (C, D). The personal particulars, medical notes and drugs received by each tissue donor are summarized in Table 1. Tissue samples were obtained as soon as possible after removal from the donor; periods of possible hypoxia ranged from approximately 1-8 h. Tissue samples were sliced, frozen in liquid nitrogen, and stored at -75°C. Microsomal suspensions of these samples were prepared following the method of Guengerich (14), modified by the addition of two wash steps using the final suspension buffer to ensure the removal of drugs and metabolites. The washed microsomes contained no compounds which interfered in the HPLC assay for benzodiazepines. Washed microsomes were resuspended in Tris-HCI buffer (0.01 M; pH 7.4) containing EDTA (1.0 roM) and glycerol (20% w/v), and stored at -75°C. A number of general characteristics of the hepatic cytochrome P450 system were assayed to determine the state of preservation of microsomal preparations employed in the study. Cytochrome P450 concentrations were measured using the procedure of Estabrook et al. (15), protein determinations by the method of Lowry et al. (16), and NADPH cytochrome c (P450) reductase activities by the method of Phillips and Langdon (17).

Ethylmorphine demethylase activity ~-NADP and horse heart cytochrome c (type ill) were products of the Sigma Chemical Company (St Louis, MO, USA); glucose-6-phosphate dehydrogenase (yeast enzyme, Grade 1, 5 mg protein/ml, 350 ill/mg protein) and glucose-6-phosphate disodium salt were obtained from Boehringer Mannheim Australia Pty. Ltd, (Sydney, NSW, Australia). Ethylmorphine demethylase activity of each preparation was measured according to the method described by Reilly and Winzor (18) using 5 roM substrate, 10 min incubations at 37°C and 0.5 JiM cytochrome P450 using the Nash procedure (19) for quantification of fonnaldehyde.

53

Substrate solubility studies The solubilities of the four substrates (diazepam, pinazeparn, prazepam, halazepam) in the incubating medium were determined using three procedures: (i) dissolution of dried substrate residue in incubating medium using an ultrasonic bath; (ii) addition of a solution of substrate in dilute hydrochloric acid to incubating medium; and (iii) addition of a solution of substrate in acetone-propylene glycol (2:5 v/v) to incubating medium. (In separate studies acetone, methanol, and dimetbylsulphoxide were each used in place of acetone-propylene glycol.) In case (i) an aliquot of methanolic solution of substrate was dispensed into a 1.5 ml Eppendorf tube, the solvent removed in vacuo, and the substrate reconstituted in 150 ul incubating medium (see below) with intermittent sonication. The tube was centrifuged at 18,000 g for 10 min to sediment microcrystalline substrate, and two 50 ul aliquots of the supernatant were removed for assay of the substrate by HPLC. The residue in the tube was assayed in situ to confirm complete analytical recovery of added drug. The limiting solubility of substrate was determined as the highest concentration at which homogeneity of the three aliquots was demonstrated. Similar studies were carried out in cases (ii) and (iii), except that incubation medium was added to an appropriate solution rather than to a dried residue.

Diazepam metabolism studies Aliquots of a solution of diazepam in methanol, to give a final concentration of diazepam in the incubation medium in the range of 5-200 JiM, were dispensed into 1.5 ml plastic Eppendorf tubes, and the solvent removed in vacuo. The substrate was reconstituted with sonication in Tris-HCI buffer (0.1 M, pH 7.4) containing glucose-6-phosphate to be 2.5 roM in the final volume. The remaining constituents of the incubating mixture were added and allowed to equilibrate for 2 min at 37°C in a shaking water bath (160 oscillations/min). These were glucose-6-phosphate dehydrogenase (1 ul containing 1.75 ill) and NADPH to be 1.0 roM in the fmal volume. The reaction was initiated by addition of microsomal suspension which was 0.4 JiM with respect to cytochrome P450 in the fmal mixture. The total volume of the incubation mixture was 200 ul. Blank incubations contained all ingredients except NADPH. The reactions were terminated at 10 min by addition of tetrabydrofuran (50 ul) which contained the internal standard for the HPLC

54

Eur. J. Drug Metab. Pharmacokinet., 1992, No.1 E as

Q. Q)

N

I'll

EO

1'll"Ot 0.0 Q) •

N .....

I'll-

E as Q. Q)

N

I'll

E

I'll

0

en ~

(')

C\l

E '6 >.

-

I'll ,c. 0. Q)

N

I'll

E Q)

....

Q)

E II) Q)

"0 I

- -en

"Ot

0

~

(')

..;

Fig. 2 : Chromatogram from HPLC analysis of microsomal incubation mixture containing diazepam (initial concentration 100 J.IM), and showing metabolites (temazepam, N-desmethyldiazepam) and the internal standard (camazepam). Retention times are given in the figure

assay (camazepam, 500 ng). In preliminary experiments, it was established that both N-demethylation and Cs-hydroxylation rates were linear up to 10 min incubation at several substrate concentrations, and that velocities of both reactions increased linearly with cytochrome P450 concentration over the range of 0.1-Q.41JM.

HPLC assay for benzodiazepines Incubation mixtures were extracted by the addition of 0.2 ml carbonate buffer (1.0 M; pH 10.0) and 0.8 ml dichloromethane. After vigorous vortex mixing for 30 s and centrifuging at 2000 g for 4 min, the aqueous layer was aspirated to waste. The dichloromethane

was transferred to a tapered plastic test tube and the solvent removed under a gentle air stream at 30·C. The residue was reconstituted in HPLC mobile phase (100 J.1l) and a 50 J.Ll aliquot was injected onto the HPLC column. It was shown that ethyl acetate gave slightly higher extraction recovery of the benzodiazepines than did dichloromethane, but equivalent results were obtained with either solvent using appropriate calibration standards. The HPLC system comprised a Waters 6000A solvent delivery system, a Kortee K65B autoinjector, a Waters model 481 ultraviolet-visible spectrophotometer set to monitor at 236 om, and a Shimadzu C-R3A integrator with FDD-IA floppy disc drive and CRT monitor. The column was an RCM-I00 radial compression module (Waters) containing a Novapak CIS cartridge (5 micron). The mobile phase was 55% methanoVwater containing triethylamine (1 mYI) and orthophosphoric acid (1.0 M, 7 mVI), with a flow rate of 1.8 ml/min. A chromatogram of an extract of incubation mixture is shown in Figure 2, which shows the retention characteristics of temazepam, N-desmethyldiazepam, diazepam and camazepam. The method was readily applicable to the assay of oxazepam, another metabolite of diazepam, and of the congeners pinazepam, prazepam and halazepam (data not shown).

Inhibition studies Incubation mixtures were set up as described above, with the addition of compounds to be tested as possible inhibitors of diazepam metabolism. In one such experiment using microsomes from liver C, 50 IJM or 150 IJM temazepam was added to diazepam incubations, and only the N-dealkylation reaction was studied Conversely 50 IJM or 150 IJM N-desmethyldiazepam was added to incubation mixtures, and in these instances only the Cs-hydrcxylation reaction was measured. The effect of mephenytoin on diazepam metabolism in three human livers (A, B, C) was also investigated by the same procedure, allowing for both N-dealkylation and C3-hydroxylation to be measured.

RESULTS Table II shows the protein concentrations, cytochrome P450 specific contents, NADPH cytochrome c (P450) reductase specific activities and ethylmorphine demethylase activities of the human liver microsomal

w. D. Hooper et at., Diazepam & related benzodiazepines preparations used here. All of the values are within published ranges for microsomes from similar sources (20-23), which confmns that the preparation and use of these reconstituted enzyme systems was appropriate in the present studies. The limiting solubilities of the four benzodiazepine substrates, using either dissolution of dried residue or solubilization with hydrochloric acid or acetone! propylene glycol are shown in Table III. Neither method gave concentrations of pinazepam, prazepam or halazepam sufficient for the conduct of useful metabolic studies. Diazepam was the most soluble substrate, and the highest concentrations were attained by addition of hydrochloric acid. Slightly enhanced solubilization could be achieved with a variety of organic solvents (e.g, Table III), but all caused an inhibition of cytochrome P450 catalysed reactions. The extent of inhibition by organic solvents is illustrated in Table IV, which presents the data for 1% acetone plus 2.5% propylene glycol, both of which are commonly used constituents for cytochrome P450 reactions (24, 25). For both metabolic pathways, formation of metabolites of diazepam was reduced by almost 50%. This degree of inhibition was typical of the range of solvents tested (acetone, methanol, dimethylsulphoxide).

55

The formation of temazepam as a function of increasing concentrations of diazepam, solubilized by sonication alone (i.e. no added organic solvent or hydrochloric acid) is shown in Figure 3. Microsomes from each liver gave a curved profile which suggested that a sigmoidal relationship may have been observed had higher substrate concentrations been attained. The formation of N-desmethyldiazepam, on the other hand, increased approximately linearly with substrate concentration to 200 J.IM (Fig. 4). This reaction also showed no clear evidence of saturation, though a trend towards it was seen in two of the five cases (Fig. 4). It is clear from Figures 3 and 4 that precise MichaelisMenten parameters for diazepam metabolism could not be determined. The N-dealkylation reaction could be interpreted in terms of VmaxlKm, the tangent to the first-order region of the putative hyperbolic substrate saturation curve (Table V). The substantial intersubject variability is emphasized by these ratios, which covered a 6-fold range in metabolic capacity. Attempted inhibition studies with temazepam (50 and 150 J.IM) produced the unexpected result of a moderate increase in N-desmethyldiazepam formation (approximately 2-fold with 50 J.IM temazepam, and 2.5-fold with 150 J.IM temazepam). Similarly N-desmethyldiazepam slightly enhanced the formation of

Table 11: Characteristics of human liver microsomal preparations . Ethylmorphine demethylase (J.l17WI HCHO/min per JUnol P450)

Liver

Protein concentration (mg/ml)

Cytochrome P450 specificcontent (nmollmg protein)

NADPH cytochrome c reductase (IU/mg protein)

A

B

8.7 ll.8

0.223 0.463

0.0142 0.0437

5.14 9.89

C D

9.2 11.6

0.351 0.241

0.0264

4.24 6.28

E

21.9

0.158

0.0164 0.0286

7.18

TableIII : Limiting solubilities of four benzodiazepines in microsomal incubating medium, when introduced by three different methods

Limiting solubility(pM) Compound

No solubilizing

Hydrochloric

Acetone-

agent

acid

propylene glycol

-200

-600

Prazepam