Journl of Chromtogrphy B, 741 (2000) 195 203 www.elsevier.com/ locte/ chrom Amounts nd vrition in grpefruit juice of the min components cusing grpefruit drug interction,, Ktsuyuki Fukud *, Linqing Guo, Noriko Ohshi, Msyoshi Yoshikw, Ysushi Ymzoe Division of Drug Metolism nd Moleculr Toxicology, Grdute School of Phrmceuticl Sciences, Tohoku University, Armki-Ao, Ao-ku, Sendi 980-8578, Jpn Discovery Reserch Lortory, Tne Seiyku Co., Ltd., 2-2-50, Kwgishi, Tod-shi, Sitm 335-8505, Jpn Received 18 Octoer 1999; ccepted 27 Jnury 2000 Astrct A method for the determintion of three furocoumrins contining two new chemicls (GF-I-1 nd GF-I-4) in commercilly ville grpefruit juice nd grpefruit itself ws developed using high-performnce liquid chromtogrphy (HPLC). These components isolted from grpefruit juice hve 5-gernyloxyfurocoumrin dimer structures showing extremely high ffinities for form of cytochrome P450 (CYP3A4). Considerle differences were oserved on the contents mong commercil rnds nd lso tches. The contents were determined to e 321.4695.2 ng/ ml GF-I-1, 5641.261538.1 ng/ ml GF-I-2 nd 296.3684.9 ng/ ml GF-I-4 in twenty-eight white grpefruit juices. These chemicls were not detected in everges from ornge, pple, grpe nd tngerine, except tht trce mount of GF-I-2 nd GF-I-4 were found in lemon juice. The verge levels of these furocoumrins were lower in the juice from red grpefruit thn white one. The highest level of these components were found in the fruit met. 2000 Elsevier Science B.V. All rights reserved. Keywords: Grpefruit juice; Cytochrome P450 inhiitors 1. Introduction drug interction suggest severl essentil properties of drug influenced: All the drugs re lipophilic nd Grpefruit juice hs een shown to elevte the require the extensive oxidtive iotrnsformtion phrmcologicl efficcy of orlly dministered prior to the excretion. In ddition, humn studies in drugs, such s felodipine [1], cyclosporin [2], ter- vitro [6,7] nd in vivo [8,9] indicte the mjor role of fendine [3], midzolm [4] nd lovsttin [5], nd CYP3A4 on the oxidtive metolisms. These lines sometimes evoke side rections through yielding of evidence suggest the involvement of specific conditions with the extrordinry high plsm con- grpefruit juice component on the first pss metocentrtion. Biovilility of these drugs fter orl lism of drugs which is medited y intestinl dministrtion re incresed 1.5 to 15-fold y the CYP3A4. ingestion of grpefruit juice. Dt compiled on this Flvonoid components, prticulrly nringin, re contined undntly in this juice, s the mjor itter *Corresponding uthor. Fx: 181-484-338-170. principle [10], nd thus their inhiitory properties E-mil ddress: k-fukud@tne.co.jp (K. Fukud) were exmined. Nringenin, n glycone generted 0378-4347/ 00/ $ see front mtter 2000 Elsevier Science B.V. All rights reserved. PII: S0378-4347(00)00104-3
196 K. Fukud et l. / J. Chromtogr. B 741 (2000) 195 203 y hydrolysis of nringin in intestine [11], inhiit respectively. ESI/ MS ws crried out using nitrogen felodipine metolism in vitro [6,7]. However, nrin- to ssist neuliztion. Collisionlly-induced dissocigin is shown to produce negligile interction with tion (CID) ws done with xenon t collision energy felodipine in vivo despite eing given t similr of 1 kv. For hydrogen deuterium (H D) exchnge mount s tht found in the grpefruit juice [12]. experiment [15], the smple ws dissolved in the Consistent with the result, the inhiitory effect of moile phse contining deuterium oxide insted of ethyl cette-extrcts of grpefruit juice on micro- wter. soml CYP3A4 ctivity ws not ccounted for y the HPLC nlysis ws performed with Chemcosor mounts of nringenin in the juice in our previous 5-ODS-H (5 mm, 15036.0 mm I.D., Chemco Scistudy [13]. To identify the component responsile entific Co., Ltd., Osk, Jpn) equipped with for the inhiition of CYP3A4 ctivity, we exmined precolumn pcked with Nucleosil 120-5C 18, (5 mm, the inhiitory effect of the extrct y frctiontion to 3034.6 mm I.D., Chemco Scientific Co., Ltd., isolte its inhiitory components. As the results, two Osk, Jpn) using Jsco Model PU-980 HPLC new chemicls showing extremely high ffinities for pump. Smples were introduced on the column vi form of cytochrome P450 (CYP3A4) were iden- Wters Model 712 Wisp utosmpler. Metolites tified s 4-[[6-hydroxy-7-[[1-[(1-hydroxy-1-methyl)- were detected y the sornce t 240 nm using ethyl]- 4- methyl- 6-(7- oxo- 7H-furo[3,2-g][1]enzo- Jsco Model UV-970 vrile wvelength UV vispyrn-4-yl)-4-hexenyl]oxy]-3,7-dimethyl-2-octenyl]- ile detector t room temperture. The moile phse oxy]-7h-furo[3,2-g][1]enzopyrn-7-one (GF-I-1) consisted of multiple grdient of solvent A (wter) nd 4-[[6-hydroxy-7-[[4-methyl-1-(1-methylethenyl)- nd solvent B (methnol): 0 min, 40% (A); 30 min, 6-(7-oxo-7H-furo[3,2-g][1]enzopyrn-4-yl)-4-10%; 50 min, 5%; 60 min, 5%; 70 min, 40%, nd set hexenyl] oxy] - 3, 7 - dimethyl - 2 - octenyl]oxy] - 7H - t the flow-rte of 1 ml/min. furo[3,2-g][1]enzopyrn-7-one (GF-I-4) to hve 5- gernyloxyfurocoumrin dimer structures [14]. Grpefruits re nturl products nd, thus, the 2.2. Mterils furocoumrin contents re expected to show considerle chnges mong the juices. In ddition, wide Two new chemicls, GF-I-1 nd GF-I-4 were vrition re known on the consequence of grpefruit seprted from grpefruit juice s descried in our juice-induced drug interction [8,9]. Grpefruit juices previous study [14]. The mss spectr of GF-I-1 nd re dily consumed nd often tken together with GF-I-4 re shown in Fig. 1 nd. The protonted drugs. Therefore, differences in furocoumrin con- moleculr ions produced t m/z 727 (GF-I-1) nd tents in commercil grpefruit juice nd grpefruit m/z 709 (GF-I-4). These ions shifted t m/z 730 itself hve een investigted s sis of prediction (GF-I-1) nd m/z 711 (GF-I-4) fter hydrogen for grpefruit drug interction in the present study. deuterium exchnges, confirming tht GF-I-1 nd GF-I-4 hve two nd one lile hydrogen in ech molecule, respectively (Fig. 1c nd d). Authentic 2. Experimentl section GF-I-2 (ergmottin) nd n internl stndrd, 8- decnyloxypsorlen, were synthesized y method 2.1. Apprtus of A. McLillop et l. [16] nd identified y NMR nd mss spectrometry. All mss spectr were otined using MSttion Grpefruit juice (totl: Thirty-two smples) includ- 700 tndem type mss spectrometer (JOEL, Jpn) ing two different types (white or red grpefruit) or equipped with n electrospry ioniztion source. An ten different rnds, nd other juices were purchsed HPLC model HP1100 system (Hewlett-Pckrd, from locl commercil sources. The solvent, meth- USA) ws used. For ESI/ MS nlysis, the temper- nol, ws HPLC grde from Nkli (Kyoto, Jpn). tures of the desolvting plte nd orificel were set t Wter used for HPLC ws filtered nd pssed 2008C nd 808C, respectively. The voltges of the through Millipore Norgnic crtridge efore the ring lens nd orificel were set t 90 V nd 50 V, mixing with HPLC-grde methnol.
K. Fukud et l. / J. Chromtogr. B 741 (2000) 195 203 197 Fig. 1. Product ion spectr nd mss spectr of furocoumrins. For smple preprtion nd nlyticl condition, see the Experimentl. (A) 1 Product ion spectrum of GF-I-1 produced y CID of the [M1H] ion (m/z 727); (B) product ion spectrum of GF-I-4 produced y CID of 1 1 the [M1H] ion (m/z 709); (C) mss spectrum of GF-I-1 otined with H D exchnge ([MD1D] : m/z 730), nd (D) mss spectrum of 1 GF-I-4 otined with H D exchnge ([M 1D] : m/z 711). D
198 K. Fukud et l. / J. Chromtogr. B 741 (2000) 195 203 2.3. Smple preprtion 3. Results GF-I-1, GF-I-2 nd GF-I-4 were diluted to mke 3.1. Clirtion nd recovery methnolic stndrd solutions (28.1 900.0 ng/ 50 ml for GF-I-1 nd GF-I-4, nd 281.3 9000.0 ng/ 50 ml An dequte seprtion of ll nlytes were for GF-I-2). These stndrd solutions (50 ml) fter chieved s shown in Fig. 2. Linerity of GF-I-1, the ddition of n internl stndrd, 8-decnylox- GF-I-2 nd GF-I-4 ws evluted over the rnge ypsorlen (800 ng), were dissolved gin in moile 10.1 324.0 ng per injection for GF-I-1 nd GF-I-4, phse (finl volume of 250 ml), nd then n liquot nd 101.3 3240.0 ng per injection for GF-I-2. The (90 ml) ws injected to HPLC nlyses fter filtr- clirtion curves of GF-I-1, GF-I-2 nd GF-I-4 tion. Grpefruit juice or other juices (1 ml) were were y 5 0.00108x 1 0.00368 (r 5 0.999), y 5 extrcted with 4 ml of ethyl cette fter the ddition 0.00115x 1 0.0444 (r 5 0.999) nd y 5 0.00100x 1 of the internl stndrd. For the determintion of 0.00224 (r 5 0.999), respectively. The intr-dy vrifurocoumrins in grpefruit tissue, the fruit ws ility of GF-I-1, GF-I-2 nd GF-I-4 re summrized divided into four prts, the peel, sc, fruit met in Tle 1. The intr-dy C.V. vlues were less thn (flesh) nd seed. Ech prt ws homogenized y 5.4% nd the intr-dy ccurcies were etween polytron fter the ddition of 3 to 10-fold mounts of 25.2 nd 2.7%, within the mount rnge of the wter nd portion (1 ml) of the resultnt homoge- clirtion curves for these nlytes. The limits of nte ws extrcted with 4 ml of ethyl cette fter quntifiction for GF-I-1, GF-I-2 nd GF-I-4 were eing mixed with the internl stndrd. The residue, set to 10.1 ng, 101.3 ng nd 10.1 ng per injection, fter evportion, ws reconstituted in 250 ml of respectively, which re the lowest mounts tht moile phse nd n liquot of 90 ml ws injected could e mesured with cceptle C.V. (,20%) nd into the HPLC. ccurcy (,20%). The recoveries from grpefruit juice spiked were etween 94.2 nd 104.7% for GF-I-1, 92.0 nd 108.4% for GF-I-2 nd 83.1 nd 2.4. Clirtion nd recovery 97.0% for GF-I-4, respectively. Clirtion curves were generted using six differ- 3.2. Concentrtion of GF-I-1, GF-I-2 nd GF-I-4 ent concentrtions of stndrd solutions of GF-I-1, in juices GF-I-2 nd GF-I-4 from lest-squres regressions of pek re of these nlytes versus internl stndrd. To evlute the concentrtion of GF-I-1, GF-I-2 The intr-dy precision nd ccurcy of the quntifi- nd GF-I-4 mong juices of different types (white or ction were evluted y replicte nlyses (n54) of red grpefruit), rnds nd tches, commercilly the stndrds. Precision ws sed on the clcultion ville, thirty-two different grpefruit juices were of the coefficient of vrition (C.V.). Accurcy ws used in the nlysis. Considerle differences were sed on the clcultion of the reltive error (R.E.) oserved on the mounts mong rnds nd lso of the found mount compred to theoreticl one. tches. White grpefruit contined 321.4695.2 ng/ For recovery study, grpefruit juice were divided ml of GF-I-1, 5641.261538.1 ng/ ml of GF-I-2 nd into two portions. Hlf of the juice ws spiked with 296.3684.9 ng/ml of GF-I-4 (Tle 2). The verge 40, 180 or 360 ng of GF-I-1 nd GF-I-4, nd 1000, contents of furocoumrins in pink grpefruit juices, 2000 or 4000 ng of GF-I-2 nd then processed s which were mde from ruyred grpefruit, were descried ove. The pek res corresponding to determined s 184.8696.7 ng/ ml GF-I-1, spiked GF-I-1, GF-I-2 nd GF-I-4 were clculted 3740.061443.3 ng/ ml GF-I-2 nd 209.16103.5 ng/ from the differences of pek res etween spiked ml GF-I-4. To know whether these furocoumrins nd non-spiked juices. The recovery ws determined re contined selectively in grpefruit juices, other y compring the pek re of spiked GF-I-1, GF-I-2 fruit juices were lso nlyzed. As shown in Fig. 3, nd GF-I-4 with the stndrd solution. GF-I-1, GF-I-2 nd GF-I-4 were not detected in
K. Fukud et l. / J. Chromtogr. B 741 (2000) 195 203 199 Tle 1 Intr-dy precision nd ccurcy for nlyses of GF-I-1, GF-I-2 nd GF-I-4 (n54) Compounds Amounts C.V. Accurcy (ng/injection) (%) (%) GF-I-1 10.1 5.4 25.1 20.3 1.1 21.8 40.5 3.3 20.4 81.0 2.5 1.0 162.0 1.0 0.3 324.0 0.3 20.1 GF-I-2 101.3 1.0 24.9 202.5 0.4 0.8 405.0 2.4 2.3 810.0 1.7 1.0 1620.0 0.4 21.3 3240.0 1.7 0.2 GF-I-4 10.1 2.1 25.2 20.3 2.2 24.5 40.5 5.3 20.4 81.0 1.3 1.0 162.0 2.0 0.8 324.0 1.4 0.1 C.V.5coefficient of vrition. Accurcy5reltive error. juices from ornge, pple, grpe nd tngerine, except tht trivil mount of GF-I-2 nd GF-I-4 were detected in lemon juice (Fig. 3). 3.3. Locliztion of GF-I-1, GF-I-2 nd GF-I-4 in grpefruit Fig. 2. HPLC seprtion of furocoumrins. For smple preprtion nd nlyticl condition, see the Experimentl. Peks: 15GF- I-1, 25GF-I-2, 35GF-I-4, 45internl stndrd. (A) Stndrd furocoumrins, (B) grpefruit juice, nd (C) grpefruit juice spiked with internl stndrd. To ssess tissue locliztion of furocoumrins in grpefruit, the fruit ws divided into four prts, the peel, sc, fruit met (flesh) nd seed. Results on the qulifiction of GF-I-1, GF-I-2 nd GF-I-4 in white grpefruit re shown in Fig. 4. These three furocoumrins were found to e contined in the flesh portion, followed y the sc, peel nd seed in decresing order, except tht GF-I-2 in the peel ws higher thn in the sc. Amounts of GF-I-1, GF-I-2 nd GF-I-4 in white grpefruit were clculted s 207.7655.0 mg, 4678.461276.0 mg nd 167.7644.8 mg per grpefruit, respectively. Contents of these chemicls in ruyred grpefruit were less, ut consistent with the results of commercil grpefruit juices.
200 K. Fukud et l. / J. Chromtogr. B 741 (2000) 195 203 Tle 2 Concentrtion of furocoumrins in commercilly ville grpefruit juices Commercil n White/ Pink GF-I-1 GF-I-2 GF-I-4 product (ng/ ml) (ng/ ml) (ng/ ml) Tropicn 8 White 363.3634.1 5455.16492.2 292.4641.5 Dole 5 White 348.9615.8 5105.8698.5 288.8619.5 Sunkist 5 White 346.4624.7 5868.16278.3 299.6629.5 Welch 3 White 352.0690.5 6991.761961.8 351.46131.4 Kgome 3 White 202.86173.6 4416.362812.6 250.46167.7 Sunpokk 1 White 392.0 7400.3 374.6 Berri 1 White 312.5 9985.2 436.4 Tkno 1 White 210.0 4136.6 162.7 Liy 1 White 54.3 4922.1 164.4 Totl 28 White 321.4695.2 5641.261538.1 296.3684.9 Welch 1 Pink 138.4 3888.5 272.2 Sunpokk 1 Pink 296.0 5452.1 308.5 Texun 1 Pink 120.1 3694.2 178.2 Liy 1 Pink N.D. 1925.2 77.6 Totl 4 Pink 184.8696.7 3740.061443.3 209.16103.5 N.D.5Not detected. This result is reported s the men6sd of three smples. 4. Discussion inhiited CYP3A4-dependent oxidtion in humn liver microsomes, ut the effect is t lest one order Codministrtion of grpefruit juice with drugs, of mgnitude weker thn GF-I-1 or GF-I-4 (unmetolized minly y CYP3A4, hs een reported pulished dt). Cler inhiition of CYP3A4 ctivity to result in sustntil increses in their orl io- ws oserved in the frction contining GF-I-1 nd vililities [1 5]. This phenomenon occurred GF-I-4, wheres only mrginl inhiitory effect through the inhiition of CYP3A4. In our previous ws oserved in the frction contining GF-I-2 in our study [14], GF-I-1 nd GF-I-4 were shown to inhiit previous pper [14]. These results suggest the mjor microsoml testosterone 6-hydroxyltion in humn contriutors re GF-I-1 nd GF-I-4, lthough ll livers. The present study indictes tht GF-I-1 nd furocoumrin components would e ctive ingredi- GF-I-4 re specificlly contined in grpefruit juice, ents responsile for grpefruit juice interction. ut not in other fruit juices of dily consumption. Consistent with selective occurrence of drug inter- As shown in Tle 2, GF-I-1, GF-I-2 nd GF-I-4 ction with grpefruit juice [18,19], these were contined in ll thirty-two different smples of furocoumrins were not detected in other everges grpefruit juices. The contents of these three chemi- from fruits such s ornge, pple, grpe nd cls differed considerly mong smples, prticu- tngerine, except tht trce mounts of GF-I-2 nd lrly on GF-I-1. The verge levels of GF-I-1, GF-I- GF-I-4 were found in lemon juice. These results lso 2 nd GF-I-4 were higher in juices from white support the ide tht GF-I-1 nd GF-I-4 would e grpefruit thn red grpefruit, lthough some red custive components, lthough definitive conclugrpefruit juice contined these furocoumrins t sion should not e mde t this moment. levels similr to the verge level of white grpe- To estimte the locliztion of these fruit. furocoumrins, grpefruit ws divided into four GF-I-2, presumed precursor of GF-I-1 nd GF-I- prts, the peel, sc, fruit met (flesh) nd seed. 4, is known to e n ingredient of grpefruit essentil Results of the nlysis y HPLC showed tht the oil nd ergmot oil [17]. In the present study, three components were found in the fruit t the grpefruit juice is shown to contin 17-fold higher highest level. Concentrtion of these furocoumrins mount of GF-I-2 thn tht of GF-I-1. GF-I-2 in the fruit of white grpefruit were determined to e
K. Fukud et l. / J. Chromtogr. B 741 (2000) 195 203 201 Fig. 3. HPLC seprtion of furocoumrins in citrus juices. For smple preprtion nd nlyticl condition, see the Experimentl. Peks: 15GF-I-1, 25GF-I-2, 35GF-I-4, 45internl stndrd. (A) Grpefruit, (B) ornge, (C) lemon, (D) tngerine, (E) pple nd (F) grpe.
202 K. Fukud et l. / J. Chromtogr. B 741 (2000) 195 203 dihydroergmottin isolted from grpefruit juice, lthough the solute configurtion remins undefined. The enntiomer excess ws 95% s compred with 50% in n synthetic stndrd of this chemicl (dt not shown). The determintion of the solute configurtion nd the chemicl synthesis for GF-I-1 nd GF-I-4 re in progress. 5. Definition list Fig. 4. Locliztion of furocoumrins in grpefruit tissue. For smple preprtion nd nlyticl condition, see the Experimentl. (A) GF-I-1, (B) GF-I-2, (C) GF-I-4.The results re represented s the men6sd of three grpefruits. HPLC high-performnce liquid chromtogrphy CYP nd P450 cytochrome P450 LC/ MS/ MS liquid chromtogrphy/ tndem mss spectrometry NMR nucler mgnetic resonnce CID collisionlly induced dissocition H D hydrogen deuterium GF-I-1 4-[[6-hydroxy-7-[[1-[(1-hydroxy-1- methyl)ethyl]-4-methyl-6-(7-oxo- 7H-furo[3,2-g][1]enzopyrn-4- yl)-4-hexenyl]oxy-3,7-dimethyl-2- octenyl]oxy]-7h-furo[3,2- g][1]enzopyrn-7-one GF-I-4 4-[[6-hydroxy-7-[[4-methyl-1-(1- methylethenyl)-6-(7-oxo-7h- furo[3,2-g][1]enzopyrn-4-yl)-4- hexenyl]oxy]-3,7-dimethyl-2-oc- tenyl]oxy]-7h-furo[3,2-g][1]enzopyrn-7-one 599 ng/g, 11 418 ng/g nd 489 ng/g, for GF-I-1, GF-I-2 nd GF-I-4, respectively. Consistent with the References results of commercilly ville grpefruit juice, [1] D.G. Biley, B. Edger, J.D. Spence, C. Munoz, J.M. Arnold, contents of these furocoumrins were less in ruyred Clin. Phrmcol. Ther. 47 (1990) 180. grpefruit thn in white ones. It might e clculted [2] M.P. Duchrme, L.H. Wrsse, D.J. Edwrds, Clin. Phrmcol. tht ingestion of 100 g of the fruit flesh, corresponding Ther. 57 (1995) 485. to out third of the whole fruit, would [3] R. Benton, P. Honig, K. Zmni, J. Hewett, L.R. Cntilen, R.L. Woosley, Clin. Phrmcol. Ther. 55 (1994) 146. e equivlent to 200 ml of regulr strength grpefruit [4] H.H. Kupferschmidt, H.R. H, W.H. Ziegler, P.J. Meier, S. juice. The mount is known to increse plsm Krhenuhl, Clin. Phrmcol. Ther. 58 (1995) 20. concentrtion of felodipine [12]. [5] T. Kntol, K.T. Kivisto, P.J. Neuvonen, Clin. Phrmcol. The C-1 nd C-6 positions in two gernyl moieties Ther. 63 (1998) 397. of GF-I-1 nd GF-I-4 re chirl centres. In our [6] F.P. Guengerich, D.H. Kim, Crcinogenesis 11 (1990) 2275. [7] A. Minisclco, J. Lundhl, C.G. Regrdh, B. Edgr, U.G. preliminry experiment using reversed-phse Eriksson, J. Phrmcol. Exp. Ther. 261 (1992) 1195. HPLC equipped with chirl column, predominnce [8] D.G. Biley, J.M. Arnold, J.D. Spence, Clin. Phrmcokinet. of stereoisomer ws detected y nlysis of 69,79-26 (1994) 91.
K. Fukud et l. / J. Chromtogr. B 741 (2000) 195 203 203 [9] B. Ammer, R.A. Weintru, Clin. Phrmcokinet. 33 (1997) [15] K.E. Krlsson, J. Chromtogr. 647 (1993) 31. 103. [16] A. McKillop, J.-C. Fiud, R.P. Hug, Tetrhedron 30 (1974) [10] R.H. Higy, J. Am. Chem. Soc. 60 (1938) 3013. 1379. [11] U. Fuhr, A.L. Kummert, Clin. Phrmcol. Ther. 58 (1995) [17] M.A. Pthk, F. Dniels, T.B. Fitzptrick, J. Invest. Derm. 365. 39 (1962) 225. [12] D.G. Biley, J.M. Arnold, C. Munoz, J.D. Spence, Clin. [18] D.G. Biley, J.D. Spence, C. Munoz, J.M. Arnold, Lncet Phrmcol. Ther. 53 (1993) 637. 337 (1991) 268. [13] K. Fukud, T. Oht, Y. Ymzoe, Biol. Phrm. Bull. 20 [19] G.C. Yee, D.L. Stnley, L.J. Pess, C.T. Dll, S.E. Beltz, J. (1997) 560. Ruiz, D.T. Lowenthl, Lncet 345 (1995) 955. [14] K. Fukud, T. Oht, Y. Oshim, N. Ohshi, M. Yoshikw, Y. Ymzoe, Phrmcogenetics 7 (1997) 391.