Copyright © 2003 by Institute of Pharmacology Polish Journal of Pharmacology Polish Academy of Sciences Pol. J. Pharmacol., 2003, 55, 1045–1053 ISSN 1230-6002

EFFECTS OF CLASSIC AND NEWER ANTIDEPRESSANTS ON THE OXIDATION PATHWAYS OF CAFFEINE IN RAT LIVER. IN VITRO STUDY

W³adys³awa A. Daniel#, Marta Kot, Jacek Wójcikowski

Department of Pharmacokinetics and Drug Metabolism, Institute of Pharmacology, Polish Academy of Sciences, Smêtna 12, PL 31-343 Kraków, Poland

Effects of classic and newer antidepressants on the oxidation pathways of caffeine in rat liver. In vitro study. W.A. DANIEL, M. KOT, J. WÓJ- CIKOWSKI. Pol. J. Pharmacol., 2003, 55, 1045–1053.

Caffeine undergoes 3-N-demethylation via CYP1A2, as well as 1-N- demethylation, 7-N-demethylation and 8-hydroxylation, which may involve other CYP isoenzymes. The aim of the present study was to investigate the influence of clomipramine, desipramine, sertraline, nefazodone and mirtaza- pine on cytochrome P-450 activity measured by caffeine oxidation in rat liver microsomes. The obtained results showed that all the investigated anti- depressants, with an exception of mirtazapine, added in vitro to liver micro- somes had an inhibitory effect on caffeine metabolism (via competitive or mixed mechanism), though their potency towards particular metabolic path- ways was different. Dixon analysis of caffeine metabolism carried out in the control liver microsomes, in the absence and presence of the antidepressant drugs showed that desipramine and clomipramine exerted the most potent inhibitory effect on caffeine metabolism. Desipramine decreased the rates of 1-N-, 3-N- and 7-N-demethylations, and 8-hydroxylation of caffeine (Ki = 23.3, 36.6, 23.3 and 63.3 mM, respectively), the effect on 1-N- and 7-N-de- methylation being the most pronounced. Clomipramine showed distinct ini- bition of 1-N- and 3-N-demethylation and 8-hydroxylation of caffeine, the effects on N-demethylations being the most pronounced (Ki = 38.6, 34.8, m 45.6 M, respectively). Its effect on 7-N-demethylation was rather weak (Ki = 97.8 mM). Sertraline decreased significantly the rate of 1-N- and 3-N-de- m methylation and 8-hydroxylation (Ki = 37.3, 69.3 and 64 M, respectively), while its effect on 7-N-demethylation of caffeine was less pronounced (Ki = 92.1 mM). Nefazodone displayed clear effect on 3-N- and 7-N-demethyla- m tion (Ki = 68.8 and 66.4 M, respectively), but was weak in inhibiting 1-N- m demethylation and 8-hydroxylation of caffeine (Ki = 110 and 186 M, re- spectively). In contrast to the above-tested antidepressants, mirtazapine did not decrease significantly the oxidation rates of 3-N-demethylation or 8-hy- m droxylation (Ki = 264 and 455 M, respectively) and had no effect on other oxidation pathways of caffeine. In summary, we have observed intra- and inter-drug differences in the inhibitory effects of the antidepressants on the four oxidation pathways of caffeine in rat liver microsomes. The tested anti- depressants (with an exception of mirtazapine) may lead to drug-drug meta- bolic interactions at a level of a few CYP isoforms. The obtained results pro- vide further indirect evidence that apart from CYP1A2, other CYP isoforms are also important for the metabolism of caffeine.

Key words: caffeine metabolism, rat, cytochrome P-450 activity, clomi- pramine, desipramine, sertraline, nefazodone, mirtazapine

# correspondence W.A. Daniel, M. Kot, J. Wójcikowski

INTRODUCTION 1-N- and 3-N-demethylation, while fluoxetine was effective in this respect towards 8-hydroxylation Caffeine (1,3,7-trimethylxanthine), a purine al- and 7-N-demethylation of caffeine. kaloid and a component of coffee, tea and many The aim of our present study was to estimate an drugs, has some essential pharmacological proper- influence of two other tricyclic antidepressant ties, e.g. stimulation of the central nervous system, drugs, clomipramine and desipramine, one selec- diuretic and biochemical effects based on inhibition tive serotonin reuptake inhibitor (SSRI), sertraline, of phosphodiestrerase leading to the accumulation and two new antidepressants, nefazodone and mir- of cAMP and cGMP in the cells. A number of stud- tazapine, on 1-N-, 3-N- and 7-N-demethylations ies have indicated that phase I of caffeine biotrans- and 8-hydroxylation of caffeine in vitro in rat liver formation in the liver proceeds mainly via mono- microsomes. The direct inhibitory effect of those oxygenases, but its particular metabolic pathways antidepressant drugs on CYP (via binding to en- are not well characterized in respect of contribution zyme protein) has not been thoroughly studied in of cytochrome P-450 (CYP) isoenzymes. Because humans and laboratory animals as yet. While their of its natural character, caffeine is used as a test direct effects on CYP2D are relatively well known substance for estimation of phenotype with regard [5, 7], their interactions with other CYP isoforms to the activity of N-acetyltransferase NAT2 and have not been fully tested. CYP1A2 (3-N-demethylation to paraxanthine is a specific, marker reaction for CYP1A2) in humans MATERIALS and METHODS and rats [4, 16, 18]. However, caffeine is oxidized in a few positions Chemicals of its structure. Apart from 3-N-demethylation cata- lyzed by CYP1A2, the compound undergoes 1-N- Clomipramine hydrochloride was provided by demethylation, 7-N-demethylation and 8-hydroxy- RBI (Natick, MA, USA) and desipramine hydro- lation. The literature data suggest that the last three chloride by Ciba-Geigy (Wehr, Germany). Sertra- oxidation pathways of caffeine may be mediated by line hydrochloride was obtained from Pfizer Corp. CYP isoenzymes other than CYP1A2, both in hu- (Brussels, Belgium). Mirtazapine hydrochloride was mans [1, 12, 13, 21] and rats [2, 3, 4, 17]. It seems donated by Organon (The Netherlands) and nefazo- that the CYP3A subfamily is the main isoenzyme done by Bristol-Meyers Squibb International, Ltd. catalyzing 8-hydroxylation to 1,3,7-trimethyluric (Uxbridge, UK). Caffeine and its metabolites, NADP, acid [17, 21]. Moreover, there are some suggestions DL-isocitric acid (trisodium salt) and isocitric de- that 1-N-demethylation and in particular 7-N-de- hydrogenase were purchased from Sigma (St. Louis, methylation leading to the formation of theobro- USA). All organic solvents with HPLC purity were mine and theophylline (respectively) engage proba- supplied by Merck (Darmstadt, Germany). bly isoenzymes CYP2B and/or CYP2E1 [2, 21]. Animals However, the relative contribution of CYP iso- forms to the catalysis of the four oxidation path- The experiment was carried out on male Wistar ways of caffeine still requires investigation, since rats (230–260 g) kept under standard laboratory no detailed, comprehensive comparative studies conditions. Liver microsomes were prepared by concerning individual CYP isoforms have been differential centrifugation in 20 mM Tris/KCl buffer carried out so far. (pH = 7.4) including washing with 0.15 M KCl, ac- Our earlier results on the influence of psycho- cording to a conventional method. tropic drugs on the rate of caffeine oxidation in the In vitro studies into caffeine metabolism in rat liver showed distinct, but differential changes in liver microsomes the rates of individual oxidation reactions, which additionally confirmed the possibility of engage- Pooled liver microsomes from six control rats ment of different CYP isoenzymes in caffeine me- were used. Each sample was prepared in duplicate. tabolism [8, 9]. Out of the antidepressants studied The caffeine metabolism in liver microsomes was (imipramine, amitriptyline and fluoxetine), imi- studied at linear dependence of the product forma- pramine distinctly inhibited the four oxidation tion on time, and protein and substrate concentra- pathways, amitriptyline exerted such an effect on tions. The rates of 1-N-, 3-N- and 7-N-demethyla-

1046 Pol. J. Pharmacol., 2003, 55, 1045–1053 INHIBITION OF CAFFEINE METABOLISM BY ANTIDEPRESSANTS tions and 8-hydroxylation of caffeine (caffeine con- The intra- and inter-assay coefficients of variance centrations: 100–800 nmol/ml) were assessed in were below 4% and 6%, respectively. The results of the absence and presence of one of the antidepres- the above measurements were evaluated using sants added in vitro (antidepressant concentrations: Dixon and Lineweaver-Burk analysis. The values 50–800 nmol/ml). Incubations were carried out in of Ki,Km and Vmax were obtained graphically. a system containing liver microsomes (ca. 1 mg of protein/ml), phosphate buffer (0.1 M, pH = 7.4), RESULTS and DISCUSSION MgCl2 ×6H2O (6 mM), NADP (1.2 mM), DL-iso- citric acid (6 mM) and isocitric dehydrogenase (1.2 The obtained results showed intra- and inter- U/ml). The final incubation volume was 0.5 ml. drug differences in the inhibitory effects of the in- After a 2-min preincubation, the reaction was initi- vestigated antidepressants on the four oxidation ated by adding NADPH generating system and the pathways of caffeine in rat liver microsomes. The incubation lasted for 50 min. Afterwards, the reac- tricyclic antidepressants clomipramine and desi- m tion was stopped by adding 350 lof2%ZnSO4 pramine displayed distinct inhibitory effects on caf- and 25 mlof2MHCl. feine metabolism via competitive or mixed mecha- Determination of caffeine and its metabolites nisms, but their potency towards particular meta- bolic pathways was different. Dixon analysis of Caffeine and its four primary metabolites were caffeine metabolism carried out on control liver mi- assessed using the HPLC method with UV detec- crosomes, in the absence and presence of the anti- tion as described previously [9]. After incubation, depressants showed that desipramine exerted the samples were centrifuged for 10 min at 2000 × g. most potent inhibitory effect on caffeine metabo- A water phase containing caffeine and its metabo- lism. Desipramine decreased the rate of both the lites was extracted with 6 ml of organic mixture three N-demethylations and 8-hydroxylation of caf- consisting of ethyl acetate and 2-propanol (8:1, feine (Fig. 1A, B, C, D), the effect on 1-N- and v/v). The residue obtained after evaporation of mi- 7-N-demethylations being the most pronounced crosomal extract was dissolved in 100 ml of the mo- (Tab. 1). This indicates inhibition of CYP1A2 (inhi- bile phase described below. An aliquot of 20 mlwas bition of 3-N-demethylation) and possibly other injected into the HPLC system. A Merck-Hitachi CYP isoenzymes (inhibition of 7-N-demethylations chromatograph, “LaChrom” (Darmstadt, Germa- and 8-hydroxylation) by the antidepressant. As men- ny), equipped with a L-7100 pump, an UV detector tioned elsewhere, 7-N-demethylation is suggested to and a D-7000 System Manager was used. The ana- be catalyzed mainly by other CYP isoenzymes than lytical column (Supelcosil LC-18, 15 cm × 4.6 mm, CYP1A2, i.e. CYP2B and/or CYP2E1 [2, 21], or by 5 mm) was from Supelco (Bellefonte, USA). The subfamily CYP2C isoforms as indicate our not yet mobile phase consisted of 0.01 M acetate buffer published results. In the case of 8-hydroxylation, (pH = 3.5) and methanol (91:9, v/v). The flow rate CYP3A seems to play an important role [17, 21]. was 1 ml/min (0–26.5 min) followed by 3 ml/min Clomipramine decreased the rates of 1-N- and (26.6–35 min). The column temperature was main- 3-N-demethylations and 8-hydroxylation of caf- tained at 30°C. The absorbance of caffeine and its feine which indicates an inhibition of CYP1A2 and metabolites was measured at a wavelength of 254 perhaps also other CYP isoforms (e.g. CYP3A). nm. The compounds were eluted in the following The above effects were of similar potency compared order: theobromine (9.7 min), paraxanthine (15.8 to those produced by desipramine (Fig. 2). However, min), theophylline (16.9 min), 1,3,7-trimethyluric the influence of clomipramine on 7-N-demethyl- acid (23.4 min), caffeine (30.5 min). The sensitivity ation of caffeine was weaker than on the other of the method allowed for quantification of theo- metabolic pathways and was less pronounced than bromine as low as 0.001 nmol, paraxanthine as low the effect of desipramine on 7-N-demethylation as 0.004 nmol, theophylline as low as 0.005 nmol, (Fig. 1), suggesting a narrower spectrum of CYP 1,3,7-trimethyluric acid as low as 0.01 nmol and inhibition by clomipramine (involving mainly caffeine as low as 0.005 nmol in one sample. The CYP1A2) than by desipramine (Tab. 1). Thus, be- accurracy of the method amounted to 1.2% (para- sides their known ability to inhibit CYP2D [5, 7], xanthine), 1.3% (theophylline), 2.1% (theobromine), tricyclic antidepressants are also able to inhibit 2.3% (1,3,7-trimethyluric acid) and 2.9% (caffeine). CYP1A2 and other CYP isoforms, and that is why

ISSN 1230-6002 1047 W.A. Daniel, M. Kot, J. Wójcikowski

A. 1-N-demethylation 1/V 1/V B. 3-N-demethylation 2000 caffeine [100 m M] 2000

-1 m -1 caffeine [100 M] 1500 1500

1000 1000 Paraxanthine Theobromine

500 m caffeine [400 m M]

caffeine [400 M] [nmol/mg of protein/min] 500 [nmol/ mg of protein/min] caffeine [800 m M] caffeine [800 m M]

-200 0 200 400 600 800 -200 0 200 400 600 800 I I Desipramine [m M] Desipramine [m M]

1/V C. 7-N-demethylation 1/V D. 8-hydroxylation

m -1 -1 1500 caffeine [100 M] caffeine [100 m M] 300

1000 200 Theophylline caffeine [400 m M] 500

1,3,7-trimethyluric acid 100 [nmol/mg of protein/min] [nmol/mg of protein/min] caffeine [400 m M] caffeine [800 m M] caffeine [800 m M]

-200 0 200 400 600 800 I -200 0 200 400 600 800 I Desipramine [m M] Desipramine [m M] Fig. 1. The influence of desipramine on the metabolism of caffeine in rat liver microsomes (Dixon plots). Kinetics of the inhibition of caffeine 1-N-demethylation: Km = 470 mM, Vmax = 0.045 nmol/mg protein/min, Ki = 23.3 mM (A); 3-N-demethylation: Km = 380 mM, Vmax = 0.039 nmol/mg protein/min, Ki = 36.6 mM (B); 7-N-demethylation: Km = 690 mM, Vmax = 0.05 nmol/mg protein/min, Ki = 23.3 mM (C); 8-hydroxylation: Km = 640 mM, Vmax = 0.19 nmol/mg protein/min, Ki =45mM (D). V – velocity of the reaction, I – concentration of inhibitor

Table 1. The influence of antidepressant drugs on the metabolism of caffeine. The presented inhibition constants (Ki) for inhibition of particular metabolic pathways were calculated using Dixon analysis. The character of inhibition was evaluated using Lineweaver- Burk analysis

Inhibition of caffeine metabolism

Antidepressants Theobromine (caffeine Paraxanthine (caffeine Theophylline (caffeine 1,3,7-trimethyluric acid (inhibitors) 1-N-demethylation) 3-N-demethylation) 7-N-demethylation) (caffeine C-8-hydroxylation)

Ki [mM] Ki [mM] Ki [mM] Ki [mM] Clomipramine 38.6 34.8 97.8 45.6 mixed competitive competitive competitive Desipramine 23.3 36.6 23.3 63.3 mixed competitive competitive competitive Sertraline 37.3 69.3 92.1 64.0 competitive mixed mixed mixed Nefazodone 110.0 68.8 66.7 186.7 mixed mixed competitive mixed Mirtazapine no effect 264.0 no effect 455.8 mixed mixed

1048 Pol. J. Pharmacol., 2003, 55, 1045–1053 INHIBITION OF CAFFEINE METABOLISM BY ANTIDEPRESSANTS

1/V A. 1-N-demethylation 1/V B. 3-N-demethylation caffeine [100 m M] caffeine [100 m M] 1000 2000 -1 -1

800 1500 caffeine [200 m M] 600 caffeine [200 m M] 1000 Paraxanthine Theobromine 400 caffeine [400 m M] caffeine [400 m M] [nmol/mg of protein/min]

[nmol/mg of protein/min] 500 200

-100 -50 0 50 100 150 200 I -100 -50 0 50 100 150 200 I Clomipramine [m M] Clomipramine [m M]

C. 7-N-demethylation D. 8-hydroxylation 1/V 1/V caffeine [100 m M] m caffeine [100 M] 120 800 -1 -1 100 m 600 caffeine [200 M] caffeine [200 m M] 80

400 60

Theophylline m caffeine [400 m M] caffeine [400 M] 40 1,3,7-trimethyluric acid [nmol/mg of protein/min] 200 [nmol/mg of protein/min] 20

-100 -50 0 50 100 150 200 I -100 -50 0 50 100 150 200 I Clomipramine [m M] Clomipramine [m M] Fig. 2. The influence of clomipramine on the metabolism of caffeine in rat liver microsomes (Dixon plots). Kinetics of the inhibition of caffeine 1-N-demethylation: Km = 461 mM, Vmax = 0.014 nmol/mg protein/min, Ki = 38.6 mM (A); 3-N-demethylation: Km = 310 mM, Vmax = 0.017 nmol/mg protein/min, Ki = 34.8 mM (B); 7-N-demethylation: Km = 353 mM, Vmax = 0.013 nmol/mg protein/min, Ki = 97.8 mM (C); 8-hydroxylation: Km = 513 mM, Vmax = 0.18 nmol/mg protein/min, Ki = 45.6 mM (D). V – velocity of the reaction, I – concentration of inhibitor they frequently lead to pharmacokinetic interac- strate and an inhibitor of CYP3A4 [11, 19]. There- tions. fore, nefazodone may cause pharmacokinetic inter- Sertraline showed a similar profile of inhibition actions in humans in vivo [11]. of caffeine metabolism to that of clomipramine, but Hence, many of the above-described inhibitory it was a weaker inhibitor of 3-N-demethylation effects of antidepressants on CYP activity in vitro (Fig. 3, Tab. 1). The observed inhibition of CYP1A2 may be important in situations in vivo, since the by sertraline in vitro may not be of clinical signifi- calculated Ki values were lower than the respective cance in vivo, since the antidepressant does not in- Km values (Figs. 1–4), and were within the pre- hibit the metabolism of atypical neuroleptic olan- sumed concentration range of drugs (in particular zapine in psychiatric patients, compared to fluvox- of tricyclic antidepressants) in the liver in vivo after amine, a CYP1A2 inhibitor [22]. chronic treatment (i.e. below 100 mM), both in Nefazodone displayed a moderate inhibitory ef- pharmacological experiments and in psychiatric fect on 3-N- and 7-N-demethylations and appeared patients [6, 10, 14, 23]. Considering the possibility to be a weak inhibitor of 1-N-demethylation and that those antidepressants inhibit also CYP iso- 8-hydroxylation (Fig. 4, Tab. 1). In our previous forms other than CYP1A2, e.g. CYP2D and/or studies, nefazodone showed also a moderate inhibi- CYP3A, those drugs are prone to drug-drug meta- tory effect on CYP2D in rats [7]. On the other bolic interactions. hand, recent studies carried out on human liver mi- In contrast to the above-tested antidepressants, crosomes showed that nefazodone was both a sub- mirtazapine did not exert any considerable effects

ISSN 1230-6002 1049 W.A. Daniel, M. Kot, J. Wójcikowski

1/V A. 1-N-demethylation 1/V B. 3-N-demethylation 1000 caffeine [100m M] -1 -1 1000 800 caffeine [100 m M] 600

400 500 Theobromine Paraxanthine caffeine[400 m M] caffeine [400 m M] [nmol/mg of protein/min] 200 [nmol/mg of protein/min] caffeine [800 m M] caffeine [800 m M]

-200 0 200 400 600 I -200 0 200 400 600 I Sertraline [ m M] Sertraline [m M]

1/V C. 7-N-demethylation 1/V D. 8-hydroxylation caffeine [100 m M] 1000 caffeine [100 m M] 150 -1 -1 800

600 100

400 Theophylline m caffeine [400 M] caffeine [400 m M] 50 1,3,7-trimethyluric acid [nmol/mg of protein/min] [nmol/mg of protein/min] 200 caffeine [800 m M] caffeine [800 m M]

-200 0 200 400 600 I -200 0 200 400 600 I Sertraline [m M] Sertraline [ m M] Fig. 3. The influence of sertraline on the metabolism of caffeine in rat liver microsomes (Dixon plots). Kinetics of the inhibition of caffeine 1-N-demethylation: Km = 220 mM, Vmax = 0.029 nmol/mg protein/min, Ki = 37.3 mM (A); 3-N-demethylation: Km = 330 mM, Vmax = 0.037 nmol/mg protein/min, Ki = 69.3 mM (B); 7-N-demethylation: Km = 960 mM, Vmax = 0.048 nmol/mg protein/min, Ki = 92.1 mM (C); 8-hydroxylation: Km = 500 mM, Vmax = 0.23 nmol/mg protein/min, Ki =64mM (D). V – velocity of the reaction, I – concentration of inhibitor

Table 2. The influence of antidepressant drugs of different chemical structures on 1-N-, 3-N- and 7-N-demethylations and 8-hydroxylation of caffeine. The antidepressants are listed in each column according to the ascending Ki values. CLO – clomi- pramine, IMI – imipramine, AMI– amitriptyline, DMI – desipramine, FLX – fluoxetine, SRT – sertraline, NEF – nefazodone, MRT – mirtazapine

Inhibition of caffeine metabolism

Theobromine (caffeine Paraxanthine (caffeine Theophylline (caffeine 1,3,7-trimethyluric acid 1-N-demethylation) 3-N-demethylation) 7-N-demethylation) (caffeine C-8-hydroxylation)

Ki [mM] Ki [mM] Ki [mM] Ki [mM] DMI: 23.3 IMI: 33.0(1) DMI: 23.3 FLX: 40.0(1) SRT: 37.3 CLO: 34.8 NEF: 66.7 IMI: 45.0(1) CLO: 38.6 DMI: 36.6 FLX: 72.0(1) CLO: 45.6 IMI: 47.0(1) AMI: 57.0(1) IMI: 73.0(1) DMI: 63.3 AMI: 61.0(1) NEF: 68.8 SRT: 92.1 SRT: 64.0 NEF: 110.0 SRT: 69.3 CLO: 97.8 AMI: 108.0(1) FLX: 196.0(1) FLX: 152.0(1) AMI: 190.0(1) NEF: 186.7 MRT: no effect MRT: 264.0 MRT: no effect MRT: 455.8

(1) According to Daniel et al. [9]

1050 Pol. J. Pharmacol., 2003, 55, 1045–1053 INHIBITION OF CAFFEINE METABOLISM BY ANTIDEPRESSANTS

1/V m 1/V caffeine [100 m M] A. 1-N-demethylation caffeine [100 M] B. 3-N-demethylation 2500 3000 -1 -1 2500 2000

2000 1500 1500 caffeine [400 m M]

1000 m Paraxanthine Theobromine caffeine [400 M] 1000 caffeine [800 m M] m [nmol/mg of protein/min] caffeine [800 M] [nmol/mg of protein/min] 500 500

-200 0 200 400 600 800 I -200 0 200 400 600 800 I Nefazodone [m M] Nefazodone [m M]

1/V 1/V D. 8-hydroxylation m C. 7-N-demethylation caffeine [100m M] 350 caffeine [100 M] 1000 -1 -1 300

800 250

600 200

150 caffeine [400 m M] Theophylline 400 caffeine [400 m M] 100 1,3,7-trimethyluric acid [nmol/mg of protein/min] 200 [nmol/mg of protein/min] caffeine [800 m M] caffeine [800 m M] 50

-200 0 200 400 600 800 I -200 0 200 400 600 800 I Nefazodone [m M] Nefazodone [m M] Fig. 4. The influence of nefazodone on the metabolism of caffeine in rat liver microsomes (Dixon plots). Kinetics of the inhibition of caffeine 1-N-demethylation: Km = 360 mM, Vmax = 0.014 nmol/mg protein/min, Ki =110mM (A); 3-N-demethylation: Km = 403 mM, Vmax = 0.016 nmol/mg protein/min, Ki = 68.8 mM (B); 7-N-demethylation: Km = 694 mM, Vmax = 0.087 nmol/mg protein/min, Ki = 66.7 mM (C); 8-hydroxylation: Km = 710 mM, Vmax = 0.12 nmol/mg protein/min, Ki = 186.7 mM (D). V – velocity of the reaction, I – concentration of inhibitor

on the oxidative metabolism of caffeine as reflected the ascending Ki values for the four metabolic reac- by Ki values (Fig. 5, Tab. 1), which indicates its tions of caffeine. It shows that tricyclic antidepres- low affinity for CYP1A2 and probably for other sants are more potent inhibitors of CYP1A2 (3-N- CYP isoforms. Our earlier studies showed that mir- demethylation) than the other investigated drugs. tazapine was also a weak inhibitor of CYP2D in On the other hand, desipramine is the strongest in- m rats with a Ki value of 107 M [7]. Thus, assuming hibitor of CYP isoforms engaged in the metabolism that the four oxidation pathways of caffeine are of caffeine via 7-N-demethylation, while tricyclic catalyzed by more than one CYP isoform (CYP1A2), antidepressants and SSRIs inhibit 8-hydroxylation, mirtazapine may be considered as a safe drug in re- the reaction supposed to involve CYP3A. spect of drug-drug interactions. The obtained re- In summary, we have observed intra- and inter- sults are consistent with recent data concerning drug differences in the inhibitory effects of the anti- human liver microsomes, which showed a weak depressants on the four oxidation pathways of caf- ability of mirtazapine to inhibit CYP1A2 activity feine in rat liver microsomes. CYP1A2 activity measured as 7-ethoxyresorufin O-deethylation (Ki represented by 3-N-demethylation of caffeine was = 159 mM), as well as the activities of CYP2C9, clearly inhibited by clomipramine, desipramine, CYP2C19, CYP2D6, CYP2E1 and CYP3A4 [15, sertraline and nefazodone, while 7-N-demethyla- 20]. tion catalyzed by other CYP isoenzymes (CYP2B, Table 2 presents antidepressants tested in the CYP2E1 or CYP2C) was diminished by desipra- present and previous studies [9] listed according to mine and nefazodone. Moreover, clomipramine,

ISSN 1230-6002 1051 W.A. Daniel, M. Kot, J. Wójcikowski

1/V 1/V A. 1-N-demethylation B. 3-N-demethylation 500 caffeine [200 m M] 700 caffeine [200 m M] -1 -1 600 400 caffeine [400 m M] 500 300 caffeine [800 m M] 400 caffeine [400 m M]

200 300 caffeine [800 m M] Theobromine Paraxanthine 200 [nmol/mg of protein/min]

100 [nmol/mg of protein/min] 100

-200 0 200 400 600 800 I -400 -200 0 200 400 600 800 I Mirtazapine [m M] Mirtazapine [m M]

1/V C. 7-N-demethylation 1/V 800 D. 8-hydroxylation caffeine [200 m M] 80

-1 70 caffeine [200 m M] -1 600 60 caffeine [400 m M] 50 m 400 caffeine [400 M] 40 caffeine [800 m M] Theophylline 30 caffeine [800 m M] 200 1,3,7-trimethyluric acid

[nmol/mg of protein/min] 20 [nmol/mg of protein/min] 10

-200 0 200 400 600 800 I -600 -400 -200 0 200 400 600 800 I Mirtazapine [m M] Mirtazapine [m M] Fig. 5. The influence of sertraline on the metabolism of caffeine in rat liver microsomes (Dixon plots). Kinetics of the inhibition of caffeine 1-N-demethylation: Km = 400 mM, Vmax = 0.007 nmol/mg protein/min (A); 3-N-demethylation: Km = 350 mM, Vmax = 0.0126 nmol/mg protein/min, Ki = 264 mM (B); 7-N-demethylation: Km = 1595 mM, Vmax = 0.017 nmol/mg protein/min (C); 8-hydroxylation: Km = 348 mM, Vmax = 0.083 nmol/mg protein/min, Ki = 455.8 mM (D). V – velocity of the reaction, I – concentra- tion of inhibitor desipramine and sertraline moderately decreased of several cytochromes P-450 in the first steps of caf- the rate of 8-hydroxylation, possibly involving feine metabolism by human liver microsomes. Drug CYP3A. Thus, the above-described antidepressants Metab. Disposition, 1991, 19, 561–567. 2. Berthou F., Goasduff T., Dreano Y., Menez J.-F.: Caf- (in particular tricyclic ones) may lead to drug-drug feine increases its own metabolism through cyto- metabolic interactions at a level of a few CYP iso- chrome P4501A induction in rats. Life Sci., 1995, 57, forms. In contrast, mirtazapine did not produce any 541–549. significant inhibitory effects on caffeine metabo- 3. Berthou F., Guillois B., Riche C., Dreano Y., Jacqz- lism in vitro and, therefore, it is unlikely to cause Aigrain E., Beaune P.: Interspecies variations in caf- any serious drug interactions in vivo. Finally, the feine metabolism related to cytochrome P4501A en- obtained results provide further indirect evidence zymes. Xenobiotica, 1992, 22, 671–680. 4. Chung W.G., Roch H.K., Kim H.M., Cha Y.N.: In- that apart from CYP1A2, other CYP isoforms are volvement of CYP3A1, 2B, and 2E1 in C-8 hydroxy- also important for the metabolism of caffeine. lation and CYP1A2 and flavin-containing monooxy- genase in N-demethylation of caffeine; identified by Acknowledgment. This study was supported by grant using inducer treated rat liver microsomes that are no. 4 PO5F 010 15 from the State Committee for Scientific characterized with testosterone metabolic patterns. Research, Warszawa, Poland. Chem.-Biol. Inter., 1998, 113, 1–14. 5. Crewe H.K., Lennard M.S., Tucker G.T., Woods F.R., REFERENCES Haddock R.E.: The effect of selective re-uptake in- hibitors on cytochrome P4502D6 (CYP2D6) activity 1. Berthou F., Flinois J.P., Ratanasavanh D., Beaune P., in human liver microsomes. Brit. J. Clin. Pharmacol., Riche C., Guillouzo A.: Evidence for the involvement 1992, 34, 262–265.

1052 Pol. J. Pharmacol., 2003, 55, 1045–1053 INHIBITION OF CAFFEINE METABOLISM BY ANTIDEPRESSANTS

6. Daniel W., Adamus A., Melzacka M., Szymura J., 16. Kilban A.J., Silbart L.K., Manis M., Beitins O.Z., We- Vetulani J.: Cerebral pharmacokinetics of imipramine ber W.W.: Human N-acetylation genotype determina- after single and multiple dosages. Naunyn-Schmied. tion with urinary caffeine metabolites. Clin. Pharma- Arch. Pharmacol., 1981, 317, 209–213. col. Ther., 1990, 47, 470–477. 7. Daniel W.A., Haduch A., Wójcikowski J.: Inhibition 17. Morita K., Maeda Y., Masuda M., Kazusaka A., and possible induction of rat CYP2D after short- and Imaoka S., Funae Y., Fujita S.: Strain differences in long-term treatment with antidepressants. J. Pharm. CYP3A-mediated C-8 hydroxylation (1,3,7-trimethyl- Pharmacol., 2002, 54, 1545–1552. uric acid formation) of caffeine in Wistar and Dark 8. Daniel W.A., Syrek M., Ry³ko Z., Kot M.: Effects of Agouti rats. Biochem. Pharmacol., 1998, 55, 1405– phenothiazine neuroleptics on the rate of caffeine de- 1411. methylation and hydroxylation in the rat liver. Pol. J. 18. Rostami-Hodjegan A., Nurminen S., Jackson P.R., Pharmacol., 2001, 53, 615–621. Tucker G.T.: Caffeine urinary metabolite ratios as 9. Daniel W.A., Syrek M., Ry³ko Z., Wójcikowski J.: markers of enzyme activity: a theoretical assessment. Effects of antidepressant drugs on the activity of cyto- Pharmacogenetics, 1996, 6, 121–149. chrome P-450 measured by caffeine oxidation in 19. Rotzinger S., Baker G.B.: Human CYP3A4 and the rat liver microsomes. Pol. J. Pharmacol., 2001, 53, metabolism of nefazodone and hydroxynefazodone by 351–357. human liver microsomes and heterologously ex- 10. Daniel W.A., Wójcikowski J.: The role of lysosomes pressed enzymes. Eur. Neuropsychopharmacol., 2002, in the cellular distribution of thioridazine and poten- 12, 91–100. tial drug interactions. Toxicol. Appl. Pharmacol., 20. Störmer E., von Moltke L.L., Shader R.I., Greenblatt 1999, 158, 115–124. D.J.: Metabolism of the antidepressant mirtazapine in 11. Davis R., Whittington R., Bryson H.M.: Nefazodone. vitro: contribution of cytochromes P-450 1A2, 2D6 A review of its pharmacology, and clinical efficacy in and 3A4. Drug Metab. Disposition, 2000, 28, 1168– the management of major depression. Drugs, 1997, 1175. 21. 53, 608–636. Tassaneeyakul W., Birkett D.J., McManus M.E., Tas- 12. Eugster H.P., Probst M., Wurgler F.E., Sengstag C.: saneeyakul W., Veronese M.E., Andersson T., Tukey Caffeine, estradiol, and progesterone interact with hu- R.H., Miners J.O.: Caffeine metabolism by human he- man CYP1A1 and CYP1A2. Evidence from cDNA- patic cytochromes P450: contributions of 1A2, 2E1 directed expression in Saccharomyces cerevisiae. and 3A isoforms. Biochem. Pharmacol., 1994, 47, Drug Metab. Disposition, 1993, 21, 43–49. 1767–1776. 22. Weigmann H., Gerek S., Zeisig A., Muller M., Hartter 13. Fuhr U., Doehmer J., Battula N., Wolfel C., Kudla C., S., Hiemke C.: Fluvoxamine but not sertraline inhibits Keita Y., Staib A.H.: Biotransformation of caffeine the metabolism of olanzapine: evidence from thera- and theophylline in mammalian cell lines genetically peutic drug monitoring service. Ther. Drug Monit., engineered for expression of single cytochrome P450 2001, 23, 410–413. isoforms. Biochem. Pharmacol., 1992, 43, 225–235. 23. Wójcikowski J., Daniel W.A.: Distribution interac- 14. Fuller R.W., Hemrick-Luecke S.K., Littlefield E.S., tions between perazine and antidepressant drugs. In Audia J.E.: Comparison of desmethylsertraline with vivo studies. Pol. J. Pharmacol., 2000, 52, 449–457. sertraline as a monoamine uptake inhibitor in vivo. Prog. Neuro-Psych. Biol. Psych., 1995, 19, 135–149. 15. Holm K.J., Markham A.: Mirtazapine. A review of its Received: September 1, 2003; in revised form: October 23, use in major depression. Drugs, 1999, 57, 607–631. 2003.

ISSN 1230-6002 1053