CYP2D6 Is Primarily Responsible for the Metabolism of Clomiphene

Home , Zuclomifene

Drug Metab. Pharmacokinet. 23 (2): 101–105 (2008). Regular Article CYP2D6 is Primarily Responsible for the Metabolism of Clomiphene

Cyrus GHOBADI1, Anne GREGORY1,H.KimCREWE1,2, Amin ROSTAMI-HODJEGAN1,2 and Martin S. LENNARD1,* 1Academic Unit of Clinical Pharmacology, School of Medicine, University of Sheffield, Royal Hallamshire Hospital, Sheffield, UK 2Simcyp Limited, Sheffield, UK

Full text of this paper is available at http://www.jstage.jst.go.jp/browse/dmpk

Summary: Clomiphene is a first line treatment for anovulation, a common cause of infertility. Response to clomiphene is variable and unpredictable. Tamoxifen is structurally related to clomiphene, and also shows considerable variation in response. CYP2D6 and CYP3A4 are major contributors to the metabolism of tamoxifen. The aim of the present work was to define the role of CYP2D6 and CYP3A4 in the in vitro metabolism of enclomiphene, regarded by some as the more active isomer of clomiphene. Enclomiphene (25 mM) was incubated with human liver microsomes (from 4 extensive (EM) and 1 poor metaboliser with respect to CYP2D6) and with microsomes from lymphoblastoid cells expressing CYP2D6. Microsomes from all the EM livers and recombinant CYP2D6 metabolised enclomiphene (the disappearance of drug ranged from 40–60%). No metabolism was detected in microsomes from the PM liver. Quinidine (1 mM) completely inhibited the metabolism of enclomiphene by all the EM livers and by recombinant CYP2D6 (pº0.001, one way ANOVA). Ketoconazole (2 mM) had no significant effect on enclomiphene metabolism in 3 out of the 4 EM livers. The extent of enclomiphene metabolism was correlated with the amount of CYP2D6 present (pº0.001, Pearson correlation test). The findings indicate that CYP2D6 is primarily responsible for the metabolism of enclomiphene.

Keywords: clomiphene; enclomiphene; zuclomiphene; CYP2D6, metabolism

effects of the drug.7–11) The aim of this work was to define Introduction theroleofCYP2D6andCYP3A4inthemetabolismofen- Clomiphene is the best initial treatment for anovulation, clomiphene and zuclomifene. a common cause of infertility. Response to clomiphene is Materials and Methods variable and unpredictable, with 60–85% of patients ovulat- ing over a dose range of 50–150 mg.1). The anti-breast can- Drugs and Chemicals: Enclomiphene and zuclomi- cer drug tamoxifen, which is structurally related to fene were gifts from Aventis Pharma (Frankfurt, Germany). clomiphene, also shows considerable variation in response.2) N-didesmethyl tamoxifen was obtained from Klinge Pharma Although CYP2D6 and CYP3A4 contribute to the me- (Munich, Germany). Ketoconazole and quinidine were pur- tabolism of tamoxifen (Crewe et al., 1997), only CYP2D6 chased from Sigma-Aldrich Company Ltd (Gillingham, UK). gentotype has been shown to be an independent predictor All other chemicals were of analytical grade and used of treatment outcome.3) Clomiphene is metabolised by rat without further purification. and rabbit liver microsomes,4,5) but the role of individual Human liver samples and recombinant CYP2D6: forms of cytochrome P450 has not been defined in humans. Samples of human liver from one poor metaboliser (PM) and Clomiphene is administered as a mixture of two geometric 4 extensive metabolisers (EM) with respect to CYP2D6, isomers enclomiphene and zuclomiphene in the ratio were used. The activities of all the main CYPs had been 62:38.6) However, there is no consensus as to which is the characterised in these livers, and the liver from the PM sub- more potent isomer, primarily responsible for the ovulatory ject showed activity with respect to all the other CYPs. The

Received; September 24, 2007; Accepted; October 29, 2007 *To whom correspondence should be addressed: Dr.M.S.LENNARD, Academic Unit of Clinical Pharmacology, Floor M, The Royal Hallamshire Hospital, Sheffield S10 2JF, UK. Tel: ＋441142712578, E-mail: m.s.lennard＠sheffield.ac.uk Non standard abbreviations: CYP, cytochrome P450; rCYP2D6, recombinantly expressed CYP2D6; HLM, human liver microsomes; PM, poor metaboliser; EM, extensive metaboliser; LC-MS, Liquid Chromatography-Mass Spectrometry.

101 102 Cyrus GHOBADI, et al.

work was approved by the local Hospital Ethics Committee. inhibtor were made using a one way ANOVA (SPSS for Win- Details of the liver donors and the method of preparation of dowsXP,version12).Pvaluesoflessthan0.05werere- the microsomes have been described previously.12) Micro- garded as statistically significant. somes were prepared prior to each experiment. Micro- The in vitro metabolic data were used to estimate the in somes from lymphoblastoid cells expressing CYP2D6 vivo hepatic clearance of clomiphene. Assuming that the (rCYP)s were purchased from Gentest Corporation (Wol- substrate concentration of 25 mM used in the incubation is burn, USA). Values of the CYP content of these were sup- below the Km value for enclomiphene metabolism, a first plied by the manufacturer. order disappearance of drug was assumed and intrinsic Incubation Condition: All the assays were conducted clearance was calculated using the following equation:15) in triplicate. The incubation mixture (pre-warmed for 5 min Ln (EC20/EC0) V (ml) at 379C in a shaking water bath) consisted of 25 mMofsub- CLint (ml/min/mg)＝－ × (1) strate, an NADPH-generating system (4 mmol G6P, 0.4 t (min) P (mg) mmol NADP, 0.4 U G6PD and 2 mmol MgCl2 dissolved in where enclomiphene20/enclomiphene0 is the fractional loss 0.2 M potassium phosphate buffer, pH 7.4) and the of drug during the 20 min incubation period ``t'', ``V'' is the microsomal suspension (0.2 ml). The human liver incuba- incubation volume, and ``P'' is the amount of microsomal tions contained 0.1 mg/ml of microsomal protein and the protein present. The unbound in vivo hepatic intrinsic rCYP incubations 5 pmol of CYP2D6. The final reaction CLint, H was estimated using equation 2: volume was made up to 1 ml with 1.15% (w/v) KCL in human liver microsomes or 0.2 M potassium phosphate buffer CLuint, H (ml/min)＝[CLint (ml/min/mg)/fumic] in rCYP2D6. ×MPPGL (mg/g)×LW (g) (2) Inhibition studies: Incubations of enclomiphene with microsomes were carried out in the absence and presence where fu(mic) is the fraction of unbound enclomiphene in hu- of quinidine (1 mM) or ketoconazole (2 mM), selective inhi- man liver microsomes, MPPGL is the scaling factor allowing bitors of CYP2D6 and CYP3A4, respectively.12–14) The reac- clearance to be expressed per gram of liver, and LW is the tion was initiated by the addition of microsomes and liver weight. stopped by adding 5ml of methyl t-butyl ether. Attempts to measure fu(mic) by ultrafiltration using Amicon Sample preparation and LC-MS analysis of en- Centrifree Micropartition devices (Amicon Ltd., Stone- clomiphene: Samples were spiked with 600 pg of N- house, U.K.) failed due to a high extent of binding to the didesmethyl tamoxifen, mixed using a vortex device for 1 filter, leading to undetectable concentrations in the collect- min and centrifuged at 600 g for 10 min. The supernatant ing tube. Thus, the fraction of unbound enclomiphene in was removed and evaporated to dryness under nitrogen at human liver microsomes was estimated using a QSAR room temperature. The samples were stored in the dark at model implemented in Simcyp (version 7.1) (www. 49CbeforeanalysisbyLC-MS.12) A mobile phase of simcyp.com). A value of 0.01 was obtained using entry methanol-water containing 0.05% trifluoracetic acid (72:28 values of the lipophilicity (clogP＝6.58) and ionisation char- v/v) was delivered isocratically at 1 ml/min. Chromatograph- acteristics (pKa＝9). ic separation was performed on a Luna C18 analytical MPPGLwasestimatedusingthefollowingequation:16) column (3 mmparticlesize,100mm×4.6 mm I.D.) cou- MPPGL (mg/g)＝10(－0.3×log Age(yr)＋2.04) (3) pled to a guard C18 precolumn (Phenomenex, Macclesfield, UK). Analysis was performed on a WatersTM 2690 separa- Liver weight was estimated using the following equation module (Waters, Watford, UK) coupled to a platform tion:17) LC single quadruple mass spectrometer (Micromass, Altrin- 2 1.176 2 cham, U.K), equipped with an atmospheric pressure ionisa- LW (g)＝0.722 (ml/m )×BSA (m ) tion (API) source and a cross-flow counter electrode. An ×Liver Density (g/ml) (4) API pin voltage of 4.2 kV and a cone voltage of 17 V were used. The source heater was set at 1509C and the API heat- where the liver density was set at 1.080 g/ml18) and the body er at 5009C. The drying nitrogen flow rate was 402 l/h.The surface area (BSA) was calculated from the formula of:19) inter and intra-assay coefficients of variation of the method BSA (m2)＝Wt 0.425 (kg)×Ht 0.725 (cm)×0.0072 (5) were 14–20% and 5.5–6.8%, respectively, and the limit of determination was 0.25 ng/ml. All demographic data including age, weight (Wt) and height Data analysis: The percentage of remaining zuclomi- (Ht) were extracted from,20) who have described the phar- fene and enclomiphene after 20 min of incubation was used macokinetics of clomiphene isomers. as a measure of the rate of metabolism. The relationship be- In the final stage of the in vitro–in vivo extrapolation, in- tween the extent of metabolism and CYP content was as- trinsic clearance was converted to hepatic clearance utiliz- sessed by the Pearson correlation test. Comparisons be- ing the well-stirred model:21) tween rates of metabolism in the presence and absence of CYP2D6 is primarily responsible for the metabolism of clomiphene 103

Q×fu, B×CLuint CLH (ml/min)＝ (6) Q＋( fu, B×CLuint)

Where fu, B is the free fraction of enclomiphene in blood. The value of fu, B was the quotient of fu (fraction unbound in plasma) and B/P (blood to plasma concentration ratio of the compound). Experimental values were not available from the literature on B/P or fu. Thus, we assumed B/P to be 1 and an fu value of 0.01 was estimated using QSAR models implemented within Simcyp (version 7) according to the method described by Lobell and Sivarajah.22) Q is the hepatic Fig. 1. The metabolism of enclomiphene by human liver micro- blood flow, established to be 1500 ml/min in humans. somes and recombinant CYP2D6 Values are the mean (±s.d) data from three separate incubations. Finally, the predicted in vivo oral clearance (CLpo-predicted)of HL＝Human liver enclomiphene was estimated from the following equation, rCYP＝Recombinant CYP2D6 assuming negligible renal clearance based on the very high PM＝CYP2D6 Poor metaboliser lipophilicity of the drug: EM＝CYP2D6 Extensive metaboliser

CLpo-predicted (ml/min)＝CLH/(FH×FH×Fa×Fg)(7)

where Fg and FH are the fractions escaping first-pass gut (Fg) and liver (FH) metabolism, and Fa is fraction of dose ab- sorbed into enterocytes. It was assumed that Fa×Fg＝1, based on the bioavailability (BA) calculated from the ratio of AUC values after oral20) and iv administration23) and com- 23) paring this value to that of FH estimated from iv data (equ- ations 8 and 9, respectively):

FH＝1－(CLiv/QH)(8) Fig. 2. Relationship between the extent of enclomifene Fa×Fg＝BA/FH (9) metabolism and microsomal CYP2D6 content Values are the mean (±s.d) data from three separate incubations. The predicted oral clearance, CL(po-predicted), was compared HL＝Human liver with the observed clearance, CL(po-observed) of the drug, calcu- rCYP＝Recombinant CYP2D6 lated from previously published plasma concentration-time profiles. Results Discussion When zuclomifene was incubated with human liver Incubation of enclomiphene with human liver micro- micrsomes for up to one hour, no disappearance of drug somes led to significant disappearance of the drug, indicat- was detected (data not shown). ing that its metabolism is catalysed by cytochrome P450. Microsomes from all the EM livers and from rCYP2D6 The findings that (a) enclomiphene was metabolised by metabolised enclomiphene, the disappearance of the drug CYP2D6 EM livers and recombinant CYP2D6, but not by a ranging from 35–65% (pº0.001) (Fig. 1). No metabolism PM liver, (b) quinidine abolished the metabolism of the was detected in microsomes from the PM liver. drug, and (c) the extent of metabolism correlated with the Quinidine completely inhibited the metabolism of en- amount of CYP2D6, provide strong evidence that CYP2D6 clomiphene in all the EM livers and in rCYP2D6 (pº0.001) is primarily responsible for the metabolism of enclomi- (Fig. 1). Ketoconazole had no significant effect on en- phene. clomiphene in 3 out of the 4 EM livers (pÀ0.05). We cannot explain the unexpected observation in micro- The extent of enclomiphene metabolism was highly cor- somes HL21 that in the presence of quinidine the concen- related with the amount of CYP2D6 (pº0.001, r2＝0.43) tration of enclomifene was 50% greater than that of con- (Fig. 2). trol. Experiments were performed in triplicate and the Based on these in vitro data and on the assumption that results from all the livers were highly consistent. Thus, we the metabolism of enclomiphene was solely catalysed by see no reason to question their interpretation. It is feasible CYP2D6, the predicted value of the in vivo oral clearance of that the unexpected finding in HL21 may have been caused clomiphene was 723 l/hr. by an interfering peak, possibly a metabolite of quinidine, generated to a much greater extent in this compared to the other livers. 104 Cyrus GHOBADI, et al.

Data from this in vitro work was then used to calculate an (2007). in vivo oral clearance for enclomiphene, based on the as- 4) Ruenitz, P. C.: Rabbit liver microsomal metabolism of en- sumption that its metabolism is catalysed solely by CYP2D6. clomiphene. Drug Metab Dispos. 9: 456–460 (1981). The value of 723 L/h obtained is very similar to the ex- 5) Ruenitz, P. C., Bagley, J. R. and Mokler, C. M.: Metabolism of perimental value of 964 L/h,20) providing further support for clomiphene in the rat. Estrogen receptor affinity and antiestro- a predominant role for CYP2D6 in the metabolism of en- genic activity of clomiphene metabolites. Biochem Pharmacol. 32: 2941–2947 (1983). clomiphene. Because CYP2D6 appeared to metabolise 6) Ernst, S., Hite, G., Cantrell, J. S., Richardson, A., Jr. and Benson, clomiphene exclusively, no further work was undertaken to H. D.: Stereochemistry of geometric isomers of clomiphene: a identify the other forms of CYP that might be involved. correction of the literature and a reexamination of structure-ac- A potential limitation of our work is the very high sub- tivity relationships. J Pharm Sci. 65: 148–150 (1976). strate concentration (25 mM)wehadtouse,whichisabout 7) VanCampenhout,J.,Borreman,E.,Wyman,H.andAntaki,A.: 50 times the maximum plasma concentration achieved dur- Induction of ovulation with cisclomiphene. Am J Obstet Gynecol. ing treatment with clomiphene (Ghobadi et al., in prepara- 115: 321–327 (1973). tion). In the present work, when therapeutic concentrations 8) Clark, J. H. and Guthrie, S. C.: Agonistic and antagonistic ef- of the drug were incubated with microsomes, no disappear- fects of clomiphene citrate and its isomers. Biol Reprod. 25: ance was observed. This was attributed to extensive non- 667–672 (1981). specific binding of clomiphene to microsomes, leaving very 9) Glasier, A. F., Irvine, D. S., Wickings, E. J., Hillier, S. G. and Baird, D. T.: A comparison of the effects on follicular develop- little free drug available for metabolism. Disappearance of ment between clomiphene citrate, its two separate isomers and drug occurred only at much higher substrate concentra- spontaneous cycles. Hum Reprod. 4: 252–256 (1989). tions, presumably because any binding to microsomes 10) Schmidt,G.E.,Kim,M.H.,Mansour,R.,Torello,L.andFried- became saturated. It could be argued that different forms of man, C. I.: The effects of enclomiphene and zuclomiphene ci- CYP may metabolise clomiphene under clinical conditions, trates on mouse embryos fertilized in vitro and in vivo. Am J Ob- where concentrations are much lower. However, the obser- stet Gynecol. 154: 727–736 (1986). vation that the predicted value of in vivo clearance based on 11) Young, S. L., Opsahl, M. S. and Fritz, M. A.: Serum concentra- these in vitro data (and assuming that clomiphene is tions of enclomiphene and zuclomiphene across consecutive cy- metabolised entirely by CYP2D6) is similar to that obtained cles of clomiphene citrate therapy in anovulatory infertile wo- experimentally20) suggests that our findings can be applied men. Fertil Steril. 71: 639–644 (1999). to the clinical setting. 12) Crewe, H. K., Notley, L. M., Wunsch, R. M., Lennard, M. S. and 3) Gillam, E. M.: Metabolism of tamoxifen by recombinant human Goetz et al. have shown that breast cancer patients with cytochrome P450 enzymes: formation of the 4-hydroxy, 4?- impaired CYP2D6 metabolism had a significantly shorter hydroxy and N-desmethyl metabolites and isomerization of time to recurrence and worse relapse free survival following trans-4-hydroxytamoxifen. Drug Metab Dispos. 30: 869–874 treatment with tamoxifen, a drug closely related to (2002). clomiphene. The therapeutic effect of tamoxifen has been 13) Baldwin, S. J., Bloomer, J. C., Smith, G. J., Ayrton, A. D., Clarke, attributed to one of its metabolites, endoxifen,24) which is S. E. and Chenery, R. J.: Ketoconazole and sulphaphenazole as the product of 4-hydroxlyation, catalysed by CYP2D6 and the respective selective inhibitors of P4503A and 2C9. Xenobiot- N-demethylation, catalysed by CYP3A4.12) CYP2D6 PMs ica. 25: 261–270 (1995). produce less endoxifen,24) thus providing a possible explana- 14) Otton, S. V., Crewe, H. K., Lennard, M. S., Tucker, G. T. and tion for the genotype difference in effect. Clomiphene is Woods, H. F.: Use of quinidine inhibition to define the role of also 4-hydroxylated and N-dealkylated, at least in rats5) and the sparteine/debrisoquine cytochrome P450 in metoprolol oxi- 4) dation by human liver microsomes. J Pharmacol Exp Ther. 247: rabbits. We were unable to identify the products of 242–247 (1988). clomiphene metabolism in the present work in human liver. 15) Baranczewski, P., Stanczak, A., Sundberg, K., et al.: Introduc- WhetherornotCYP2D6 genotype is a determinant of tion to in vitro estimation of metabolic stability and drug interac- response to clomiphene remains to be investigated. tions of new chemical entities in drug discovery and develop- References ment. Pharmacol Rep. 58: 453–472 (2006). 16) Barter,Z.E.,Bayliss,M.K.,Beaune,P.H.,et al.: Scaling factors 1) Hughes, E., Collins, J. and Vandekerckhove, P.: Clomiphene ci- for the extrapolation of in vivo metabolic drug clearance from in trate for ovulation induction in women with oligo-amenorrhoea. vitro data: reaching a consensus on values of human microsomal Cochrane Database Syst Rev. CD000056 (2000). protein and hepatocellularity per gram of liver. Curr Drug 2) Rostami-Hodjegan,A.,Lennard,M.S.,Tucker,G.T.andLedg- Metab. 8: 33–45 (2007). er, W. L.: Monitoring plasma concentrations to individualize 17) Johnson, T. N., Tucker, G. T., Tanner, M. S. and Rostami-Hod- treatment with clomiphene citrate. Fertil Steril. 81: 1187–1193 jegan, A.: Changes in liver volume from birth to adulthood: a (2004). meta-analysis. Liver Transpl. 11: 1481–1493 (2005). 3) Goetz, M. P., Knox, S. K., Suman, V. J., et al.: The impact of 18) Price, P. S., Conolly, R. B., Chaisson, C. F., et al.: Modeling in- cytochrome P450 2D6 metabolism in women receiving ad- terindividual variation in physiological factors used in PBPK juvant tamoxifen. Breast Cancer Res Treat. 101: 113–121 models of humans. Crit Rev Toxicol. 33: 469–503 (2003). CYP2D6 is primarily responsible for the metabolism of clomiphene 105

19) Du Bois, D. and Du Bois, E. F.: A formula to estimate the ap- 22) Lobell, M. and Sivarajah, V.: In silico prediction of aqueous solu- proximate surface area if height and weight be known. 1916. bility, human plasma protein binding and volume of distribution Nutrition. 5: 303–311; discussion 312–303 (1989). of compounds from calculated pKa and AlogP98 values. Mol 20) Mikkelson,T.J.,Kroboth,P.D.,Cameron,W.J.,Dittert,L.W., Divers. 7: 69–87 (2003). Chungi, V. and Manberg, P. J.: Single-dose pharmacokinetics of 23) Szutu,M.,Morgan,D.J.,McLeish,M.,et al.: Pharmacokinetics clomiphene citrate in normal volunteers. Fertil Steril. 46: of intravenous clomiphene isomers. Br J Clin Pharmacol. 27: 392–396 (1986). 639–640 (1989). 21) Wilkinson, G. R. and Shand, D. G.: Commentary: a physiological 24) Beverage,J.N.,Sissung,T.M.,Sion,A.M.,Danesi,R.andFigg, approach to hepatic drug clearance. Clin Pharmacol Ther. 18: W. D.: CYP2D6 polymorphisms and the impact on tamoxifen 377–390 (1975). therapy. J Pharm Sci. 96: 2224–2231 (2007).