The effect of grapefruit juice on drug disposition

Abstract

Introduction

Since their initial discovery in 1989, grapefruit juice-drug interactions have received extensive interest from the scientific, medical, regulatory, and lay communities. Although knowledge regarding the effects of grapefruit juice on drug disposition continues to expand, the list of drugs studied in the clinical setting remains relatively limited.

Areas covered

This article reviews the in vitro effects of grapefruit juice and its constituents on the activity of cytochrome P450 enzymes, organic anion-transporting polypeptides, P-glycoprotein, esterases and sulfotransferases. The translational applicability of the in vitro findings to the clinical setting is discussed for each drug metabolizing enzyme and transporter. Reported area under the plasma concentration-time curve ratios for available grapefruit juice-drug interaction studies are also provided. Relevant investigations were identified by searching the Pubmed electronic database from 1989 to 2010.

Expert opinion

Grapefruit juice increases the bioavailability of some orally-administered drugs that are metabolized by CYP3A and normally undergo extensive presystemic extraction. In addition, grapefruit juice can decrease the oral absorption of a few drugs that rely on organic anion-transporting polypeptides in the gastrointestinal tract for their uptake. The number of drugs shown to interact with grapefruit juice in vitro is far greater than the number of clinically relevant grapefruit juice-drug interactions. For the majority of patients, complete avoidance of grapefruit juice is unwarranted.

1. Introduction

The first grapefruit juice (GFJ)-drug interaction was serendipitously identified by Bailey and colleagues in 1989 [1]. As a result of extensive work conducted since this initial discovery, the scientific literature is now replete with information regarding the effects of GFJ and its constituents on drug disposition, both in vitro and in vivo [2-6]. Pharmacokinetic studies have demonstrated that GFJ can increase the bioavailability of drugs from an array of therapeutic classes. Examples include some calcium channel blockers, benzodiazepines, and statins – Table 1. The elevation in a drug's area under the plasma concentration-time curve (AUC) with GFJ ingestion is the result of irreversible inhibition of cytochrome P450 (CYP) 3A by furanocoumarins present in the juice [7, 8]. Unlike other known CYP3A inhibitors, normal consumption of GFJ only inhibits CYP3A in the enterocyte cells lining the small intestine – hepatic CYP3A activity remains unaffected, except with unrealistically large ingestion of GFJ.

Table 1.

In vivo effects of grapefruit juice on drug exposure.

Central Nervous System Drugs
Benzodiazepines
Drug	AUC Ratio	Reference
Alprazolam	1.12*	[70]
Diazepam	3.24	[71]
Midazolam	1.53	[68]
	1.65	[29]
	1.52	[72]
	2.31	[73]
	2.00 2.15* 5.95*	[74]
	2.39	[75]
	1.49*	[76]
	1.73	[77]
	1.36 – mean concentration at 2 hour ratio	[78]
Quazepam	1.31*	[79]
Triazolam	1.48	[80]
	1.53 1.49* 2.43*	[81]
	1.73*	[79]
	1.51 – acute GFJ exposure 1.60 – extended GFJ exposure	[40]
Antiepileptics
Drug	AUC Ratio	Reference
Carbamazepine	1.41	[82]
Phenytoin	0.89 – male volunteers 0.95 – epileptic patients	[83]
Other Central Nervous System Drugs
Drug	AUC Ratio	Reference
Buspirone	9.21*	[84]
Clozapine	1.01 – mean plasma concentration ratio	[85]
	1.15* – mean trough level ratio	[86]
	0.92	[87]
Fluvoxamine	1.60*	[88]
Haloperidol	0.99* – mean trough level ratio	[89]
Sertraline	1.47 – mean trough level ratio	[90]
	2.04*	[91]
Opioids
Drug	AUC Ratio	Reference
Alfentanil	1.62	[77]
Fentanyl	0.99	[92]
Methadone	1.19	[93]
	1.17	[94]
Oxycodone	1.62*	[95]
Antiinfectives
Antibacterials
Drug	AUC Ratio	Reference
Clarithromycin	1.15*	[96]
Erythromycin	1.49	[97]
Telithromycin	1.05	[98]
Antimalarials
Drug	AUC Ratio	Reference
Artemether	1.90*	[99]
	2.44* – acute GFJ exposure 3.51*–extended GFJ exposure	[100]
Halofantrine	2.21	[101]
Quinine	0.96* 0.77	[102]
Primaquine	1.19	[103]
Antihelmintics
Drug	AUC Ratio	Reference
Albendazole sulfoxide	3.13*	[104]
Praziquantel	1.84	[105]
Antiretrovirals
Drug	AUC Ratio	Reference
Amprenavir	0.90	[106]
Indinavir	0.98*	[107]
	0.94	[108]
Saquinavir	1.50	[109]
Antifungals
Drug	AUC Ratio	Reference
Itraconazole	1.03	[110]
	0.57*	[111]
	1.20*	[112]
	1.30* – female subjects 1.11* – male subjects	[113]
Cardiovascular Agents
6-blockers
Drug	AUC Ratio	Reference
Acebutolol	0.94*	[114]
Celiprolol	0.15*	[115]
Talinolol	0.70 – acute GFJ exposure 0.73* – extended GFJ exposure	[59]
Calcium channel blockers
Drug	AUC Ratio	Reference
Amlodipine	1.14	[116]
	1.08	[117]
Azelnidipine	3.28	[118]
Diltiazem	1.10*	[119]
	1.18	[120]
Felodipine	2.51*	[121]
	1.86	[14]
	1.43	[42]
	1.72	[122]
	2.16* – acute GFJ exposure 3.11* – extended GFJ expsoure	[37]
	1.73 – acute GFJ exposure 1.57 – extended GFJ exposure	[123]
	2.45	[124]
	2.16	[125]
	2.88 – single dose study 2.86 – steady-state study	[126]
	1.94	[127]
	1.81	[35]
	2.30 – median AUC ratio	[128]
	2.04 – median AUC ratio	[7]
	2.01	[129]
	1.35	[36]
	1.93	[130]
	2.85 3.34*	[131]
Manidipine	2.62	[132]
Nicardipine	1.56	[133]
Nifedipine	1.47*	[15]
	2.02*	[134]
	1.58*	[135]
	1.10	[136]
	1.35*	[121]
Nimodipine	1.51	[137]
Nisoldipine	1.76	[13]
	4.11	[41]
Nitrendipine	2.25*	[138]
Pranidipine	1.68	[139]
Verapamil	1.09	[140]
	1.29*	[141]
	1.43*	[142]
Antiarrhythmics
Drug	AUC Ratio	Reference
Amiodarone	1.50*	[143]
Quinidine	1.08	[144]
	1.05* – median Cmax ratio	[145]
HMG-CoA Reductase Inhibitors
Drug	AUC Ratio	Reference
Atorvastatin	1.83*	[146]
	1.33*	[147]
	2.46*	[148]
Lovastatin	15.3*	[149]
	1.91	[75]
Pitavastatin	1.13*	[146]
Pravastatin	1.00*	[147]
	0.92*	[148]
Simvastatin	16.1*	[150]
	13.5*	[43]
	3.56	[151]
Other Cardiovascular Agents
Drug	AUC Ratio	Reference
Aliskiren	0.39*	[152]
Digoxin	1.03*	[62]
	1.09*	[61]
Losartan	1.17	[153]
Warfarin	No difference in INR values*	[154]
Proton pump inhibitors
Drug	AUC Ratio	Reference
Lansoprazole	1.21	[155]
	1.15	[156]
Omeprazole	1.11	[157]
Immunosuppressants
Drug	AUC Ratio	Reference
Cyclosporine	1.23	[158]
	1.37*	[159]
	1.60* – African-American subjects 1.44* – Caucasian subjects	[160]
	1.60*	[161]
	1.85*	[34]
	1.38 – median AUC ratio	[8]
	1.55*	[39]
	1.42*	[162]
	1.24	[163]
	1.08*	[164]
	1.43	[165]
	1.41* – oral solution 1.38* – microemulsion capsules	[166]
	1.43	[167]
	1.24 – mean trough level ratio	[168]
	Increased AUC reported*	[169]
Tacrolimus	2.10* – mean trough level ratio	[170]
Insulin secretagogues
Drug	AUC Ratio	Reference
Glyburide (glibenclamide)	1.05*	[49]
Repaglinide	1.13	[171]
Antineoplastics
Drug	AUC Ratio	Reference
Etoposide	0.76	[172]
Nilotinib	1.18*	[173]
Sunitinib	1.11*	[76]
Steroids and Hormones
Drug	AUC Ratio	Reference
Budesonide	1.70* – immediate release capsules 2.29* – extended release capsules	[174]
17β-estradiol	1.16*	[175]
Ethinylestradiol	1.28*	[176]
Levothyroxine	0.91*	[177]
Methylprednisolone	1.72*	[178]
Prednisone	1.50*	[164]
Antihistamines
Drug	AUC Ratio	Reference
Desloratadine	1.07*	[179]
Fexofenadine	0.67*	[179]
	0.64 0.39*	[180]
	0.37*	[46]
	0.48	[38]
	0.58 0.59	[10]
Terfenadine	2.88	[181]
	1.28* – active metabolite (fexofenadine) ratio	[182]
	1.54* – active metabolite (fexofenadine) ratio	[183]
	1.53 – active metabolite (fexofenadine) ratio 1.57* – active metabolite (fexofenadine) ratio	[184]
Other Drugs
Drug	AUC Ratio	Reference
Caffeine	1.33*	[185]
	1.04	[186]
Chlorzoxazone	No change in 2 hour 6-OH-chlorzoxazone/chlorzoxazone ratio	[187]
Cisapride	1.37	[188]
	2.44*	[189]
	1.49	[190]
	2.55*	[191]
Dextromethorphan	Plasma levels not measured	[192]
	Plasma levels not measured	[193]
Montelukast	0.93*	[194]
Scopolamine	1.35*	[195]
Sildenafil	1.23	[196]
Theophylline	1.02*	[197]

AUC: area under the plasma concentration-time curve; AUC ratio: AUC with GFJ ingestion divided by AUC with control beverage ingestion; C_max: maximum observed plasma concentration.

AUC ratio computed using mean values unless otherwise specified.

^*Indicates administration of GFJ in a manner deemed to be inconsistent with usual dietary consumption – for example, use of double-strength juice or excessive volumes of juice.

Interested readers are encouraged to review the original research reports for specific study details.

Grapefruit juice can also decrease the bioavailability of some drugs (e.g., fexofenadine) [9]. The putative mechanism underlying such interactions is a reduction in drug uptake transport via inhibition of organic anion transporting polypeptides (OATPs) by GFJ flavonoids [10]. Other in vitro studies have suggested that GFJ is capable of inhibiting P-glycoprotein (P-gp), esterases, and sulfotransferases (SULT). However, a lack of clinical evidence makes it difficult to assess the actual impact, if any, GFJ has on these enzymes and transporter. The need for controlled clinical studies is especially important because an in vitro-in vivo disconnect has been noted for many natural products and beverages [3, 5].

This article provides a thorough review of the published data regarding GFJ-drug interactions, with an emphasis on the clinical significance of in vitro findings. Since the interaction potential of GFJ will vary based upon the specific product tested, a discussion of GFJ chemistry has been included. To facilitate clinical decision making, data from published pharmacokinetic investigations have been collated and are presented in Table 1.

2. Grapefruit Juice Composition

Grapefruit juice is rich in a number of phytochemicals, including flavonoids and furanocoumarins. The most abundant flavonoid in the juice is naringin, with reported concentrations ranging from approximately 200 to 2000 μmol/L [11, 12]. Due to their abundance and in vitro inhibition of CYP3A, the flavonoids were originally presumed to be the GFJ constituents responsible for mediating drug interactions. However, administration of naringin capsules, in the same amounts found in GFJ, failed to alter the pharmacokinetics of the CYP3A substrates, felodipine and nisoldipine [13, 14]. A similar study involving another flavonoid, quercetin, also failed to demonstrate an effect on nifedipine disposition [15]. Based on these findings and more recent work, the flavonoids are no longer thought to play a major role in GFJ-drug interactions involving CYP3A substrates.

The furanocoumarins (FCs, Figure 1) are a structurally distinct class of compounds found in GFJ. All FCs contain a three-ring “head” and, except for bergaptol and bergapten, an aliphatic “tail.” The FCs differ in the presence and composition of their aliphatic tail, which in turn influences their inhibitory potency toward CYP3A. In addition, the furan ring moiety has an important role in the generation of a reactive intermediate – a furanoepoxide or γ-ketoenal – that irreversibly binds to CYP apoprotein and eliminates enzymatic activity [16].

Figure 1.

Structures of GFJ constituents implicated in clinical drug interactions. Left panel: The furanocoumarins are potent, mechanism-based, inhibitors of CYP3A. Mechanism-based inhibition of CYP3A is thought to result after binding of CYP protein to either a furanoepoxide or γ-ketoenal intermediate. Right panel: Naringin is the most abundant flavonoid in GFJ and inhibits OATP transport in vivo. Paradisin A is a representative member of a class of dimeric compounds present in GFJ formed by linkage of 6′,7′-dihydroxybergamottin to itself, or to bergamottin.

Of the various FCs present in GFJ, bergamottin and 6′,7′-dihydroxybergamottin (DHB) have been the most-extensively studied regarding their capacity to mediate GFJ-drug interactions (Figure 1). Reported concentration ranges for bergamottin and DHB in GFJ are 1-37 μM and 0.2-52.5 μM, respectively [11, 17, 18]. In addition, GFJ contains FC dimers – also known as spiroesters or paradisins – that are formed through either head-to-tail or tail-to-tail linkage of DHB to itself, or to bergamottin. Although present at lower concentrations than bergamottin and DHB, these dimers are potent inhibitors of CYP3A in vitro [18-21].

The variability in the concentrations of flavonoids and FCs in GFJ may result from factors such as (1) the type, origin, and quality of the grapefruits used to make the juice; (2) the manufacturing process; and, (3) storage conditions [11]. Exposure of GFJ to ultraviolet (UV) light or heat has been shown to alter the concentrations of FCs in the juice [22, 23]. The concentrations of bergamottin and DHB rapidly declined after being irradiated with UV light, whereas the bergaptol levels decreased more slowly. After 6 hours of UV exposure, the concentrations of bergaptol, DHB, and bergamottin were reduced to 6%, 2%, and 2% of their baseline values, respectively [22]. The same investigators also demonstrated that bergamottin and DHB are unstable at elevated temperatures. Although exposure to 4°C and 37°C for one hour failed to change the bergamottin and DHB concentrations in the juice, their levels declined when the juice was heated to 62°C, 72°C, and 95°C [23]. More specifically, 95°C exposure for one hour caused the bergamottin concentration to decrease from 17.9 μM to 3.14 μM, and the DHB concentration to fall from 7.85 μM to 0.16 μM. Bergaptol concentrations concurrently increased by 14.1 μM under the same treatment conditions, suggesting that bergamottin and DHB are degraded to bergaptol by exposure to heat. The UV- and heat-treated GFJ did not interact with nifedipine after administration to rats [22, 23].

We evaluated the effect of prolonged storage at room temperature on the furanocoumarin content and the CYP3A-inhibiting capacity of a single GFJ sample. Aliquots were removed and analyzed at multiple time points over a one year period. Furanocoumarin concentrations were determined by liquid chromatography-mass spectroscopy, and in vitro inhibition of CYP3A in human liver microsomes was determined using the index substrate triazolam. Concentrations of DHB and paradisin C declined substantially with time (Figure 2). Bergamottin concentrations were relatively stable, while bergaptol concentrations increased. In the same samples, reversible and irreversible CYP3A inhibition declined with time in parallel with the decrement in DHB and paradisin C concentrations (Figure 3). The findings confirm the assumption that DHB and paradisins in GFJ are mainly responsible for CYP3A inhibition. The results also suggest that drug interactions with GFJ are less likely with GFJ that has undergone extensive storage at room temperature.

Figure 2.

Time-dependent changes in concentrations of bergaptol, bergamottin, DHB, and paradisin C in aliquots of a GFJ sample stored at room temperature for one year.

Figure 3.

In the same GFJ sample described in Figure 2, time-dependent changes in the capacity of the aliquots to inhibit CYP3A activity – represented as triazolam hydroxylation activity – by human liver microsomes in vitro. Lower values on the y-axis indicate greater inhibitory capacity. Each point is the mean ± SE from four different human liver samples. At all time points, the irreversible component of CYP3A inhibition (indicated by the values obtained following preincubation of GFJ with microsomes prior to addition of substrate) exceeds the reversible component. Note that the irreversible component of inhibition – thought to be responsible for clinical drug interactions – decreases substantially over the one year period. For a detailed description of methodology see references [29, 68, 69].

The complex chemistry of GFJ makes its use in drug-interaction studies challenging. The majority of investigations to date have not provided a phytochemical analysis of the juice, instead reporting that “the same lot of juice was used throughout the study.” This assures some degree of product quality control, but does not allow comparison of results from different studies. Ideally, future studies should report the chemical composition of the GFJ being administered, especially the bergamottin, DHB, and paradisin concentrations when CYP3A inhibition is anticipated. When the mechanism of interaction is presumed to be OATP inhibition, quantifying the amount of naringin in the juice appears warranted [10].

3. Interpretation of Grapefruit Juice-Drug Interaction Studies

Drug interaction studies with GFJ have usually employed a crossover design, in which the substrate drug of interest is given on one occasion under control conditions (e.g., with water), and on another occasion with GFJ. In both trials, the primary outcome is the drug's AUC. The mean AUC ratio – AUC with GFJ ingestion divided by AUC under control conditions – is then calculated from the individual study participant data. A mean AUC ratio that significantly differs from 1.0 indicates the potential for a GFJ-drug interaction. For AUC ratios greater than 1.0, the following ranges are generally utilized to describe the qualitative degree of inhibition: strong, AUC ratio ≥ 5.0; moderate, 2.0 ≤ AUC ratio < 5.0; weak, 1.25 ≤ AUC ratio < 2.0; and, negligible, AUC ratio < 1.25 [24].

Two critical points regarding GFJ-drug investigations are: (1) statistical significance does not necessarily imply clinical significance, and (2) an AUC ratio alone is insufficient to draw conclusions regarding the clinical relevance of an interaction. This judgment can only be made with additional information regarding the drug's exposure-response relationship. Generally speaking, a GFJ-drug interaction is likely to be of clinical importance if the substrate drug has a narrow therapeutic index, or when the drug's pharmacokinetics are dramatically altered by GFJ administration.

4. Effects of Grapefruit Juice on Drug Metabolizing Enzymes and Transporters

4.1 Cytochrome P450 Enzymes

The cytochrome P450 enzyme family is responsible for the phase I metabolism of the majority of drugs used in clinical practice. Of the various CYP isoforms, CYP3A (including CYP3A4 and CYP3A5) is the most important because it is responsible, either entirely or in part, for the metabolism of over 50% of the most commonly prescribed drugs [25]. CYP3A exhibits a broad substrate specificity and comprises the majority of CYP content in both enterocytes and hepatocytes. Since an orally-administered CYP3A substrate is sequentially exposed to enteric and then hepatic CYP3A, the amount of drug that reaches the systemic circulation can be markedly different than the administered dose – a process known as presystemic extraction or first-pass metabolism. In turn, inhibition of CYP3A at the enteric and/or hepatic level will increase the oral bioavailability of drugs that normally undergo extensive presystemic extraction.

4.1.1 In vitro Interactions with Grapefruit Juice

Many in vitro investigations have reported on the ability of the FCs to inhibit CYP activity, both reversibly and irreversibly, with inhibitory constants in the nanomolar to low micromolar range [16-18, 20, 21, 26-33]. Bergamottin inhibits multiple CYP isoforms, including CYP1A2, CYP1B1, CYP2A6, CYP2B6, CYP2C9, CYP2C19, CYP2D6, CYP2E1, CYP3A4, and CYP3A5. The paradisins and DHB inhibit the in vitro activity of CYP1A2, CYP1B1, CYP2C9, CYP2C19, CYP2D6, and CYP3A4. Some notable differences amongst the FCs exist, especially for inhibition of CYP3A4. The order of potency for in vitro inhibition of CYP3A4 activity has been shown to be: paradisins > DHB > bergamottin > bergaptol. This has been demonstrated in studies from our own laboratory (Figure 4). In another in vitro study, CYP3A4-mediated nifedipine oxidation was inhibited by the paradisins, DHB, and bergamottin with IC₅₀ values of 3, 650, and 1500 nanomolar, respectively [21]. In a study exploring the effects of paradisin A, DHB, bergamottin and bergaptol on CYP3A activity, only bergaptol failed to significantly inhibit testosterone hydroxylation at concentrations up to 20 μM [30].

Figure 4.

Inhibition of triazolam hydroxylation activity – an index of CYP3A activity – in human liver microsomes in relation to varying concentrations of paradisin C, DHB, bergamottin, and bergaptol. All studies represent preincubation of inhibitors with microsomal protein prior to addition of the substrate, thereby indicating irreversible (mechanism-based) inhibition. The numbers next to the inhibitor lines are the 50% inhibitory concentrations (IC₅₀). Smaller numbers indicate greater inhibitory potency. For a detailed description of methodology see references [29, 68, 69].

Mechanism-based inhibition of CYP3A by bergamottin and DHB is substrate-independent for both compounds [27]. For reversible inhibition, DHB inhibited midazolam and testosterone hydroxylation by human intestinal microsomes with nearly identical K_i values – 0.9 and 0.7 μM, respectively. In contrast, the reversible inhibition for bergamottin was substrate-dependent, with the K_i value for midazolam hydroxylation (13.3 μM) significantly greater than the K_i for testosterone hydroxylation (1.6 μM). A subsequent study verified the importance of DHB over bergamottin in the inhibition of CYP3A4 [28]. Using Caco-2 cells modified to express CYP3A4, it was found that DHB and bergamottin differ in their time courses for CYP3A4 inhibition. While maximal CYP3A4 inhibition by DHB occurred within 30 minutes, up to 3 hours was needed for bergamottin to significantly reduce midazolam hydroxylation. Similar data for the paradisins are not available. These in vitro studies suggest that the concentration of DHB in GFJ is more important for in vivo inhibition of CYP3A than the presence of bergamottin. The lack of clinically significant interactions with two beverages known to contain high levels of bergamottin corroborates this contention [34, 35]. Furthermore, when pure bergamottin was administered to 11 volunteers, mean felodipine AUC_0-24 was not significantly increased by a 2 mg, 6 mg, or 12 mg bergamottin dose [36].

4.1.2 Clinical Relevance

Definitive evidence that the FCs are the constituents in GFJ responsible for mediating in vivo drug interactions has recently been obtained. Processing of commercial GFJ with a citrus-debittering system yielded a product in which roughly 99% of the FCs were removed [7]. The median AUC for felodipine after administration with 240 mL of this juice was 48 nmol · h/L. This was not significantly different from the value (54 nmol · h/L) obtained with the control beverage, orange juice. In constrast, the median AUC was significantly greater when felodipine was ingested with unprocessed GFJ – 110 nmol · h/L. The furanocoumarin-free GFJ also did not alter the disposition of cyclosporine in a more recent pharmacokinetic trial [8].

Although the FCs inhibit multiple CYP enzymes in vitro, clinically significant inhibition by GFJ has only been unequivocally shown for CYP3A. For example, GFJ does not interact with theophylline (CYP1A2 substrate) or warfarin (CYP2C9 substrate) – Table 1. The vast majority of CYP3A substrate drugs have AUC ratios with GFJ ingestion that are below the “strong inhibition” cut-off value of 5.0. More importantly, AUC ratios > 5 often occurred in studies where GFJ was administered in quantities exceeding usual levels of consumption – for instance, thrice daily ingestion of double-strength GFJ.

Grapefruit juice is unique among CYP3A inhibitors because, when consumed in usual dietary volumes, only enteric CYP3A is affected. Therefore, GFJ has no impact on the pharmacokinetics of CYP3A substrate drugs administered intravenously, and does not prolong the plasma half-life of orally administered agents. Ingestion of GFJ causes an irreversible loss of enteric CYP3A protein, without a decrease in CYP3A mRNA [37-39]. Clinically, implies that prior exposure to GFJ is sufficient to cause an interaction – i.e., GFJ does not have to be physically present in the gastrointestinal tract for an interaction to occur [29, 40-43]. Recovery of CYP3A activity requires de novo enzyme synthesis, and the half-life for this process is approximately 23 hours [29].

4.2 Organic Anion-Transporting Polypeptides

Organic anion-transporting polypeptides (OATPs) are a family of proteins involved in the transport of bile acids, hormones, and drugs. The OATPs contribute to the pharmacokinetic behavior of many drugs. OATPs are found throughout the body, including sites relevant to drug absorption, distribution, metabolism, and excretion. In the small intestine, OATP1A2 and OATP2B1 are the principal OATPs; however, OATP1B1 and OATP1B3 mRNA has been detected in human intestinal biopsy samples [38].

4.2.1 In vitro Interactions with Grapefruit Juice

Flavonoids are capable of inhibiting the OATP1B1-mediated uptake of dehydroepiandrosterone sulfate in stably transfected HeLa cells [44]. When tested at the same concentration (50 μM), naringin and its aglycone, naringenin, decreased dehydroepiandrosterone sulfate uptake to 44% and 47% of control values, respectively.

Another study examined the effects of GFJ and its constituents on OATP2B1 activity using estrone-3-sulfate and glyburide (glibenclamide) as probe substrates [45]. In transfected HEK293 cells, 5% GFJ decreased estrone-3-sulfate uptake by 82%, and glyburide uptake by approximately 60%. Of the GFJ constituents tested at a concentration of 10 μM (naringin, naringenin, quercetin, bergamottin, DHB), only bergamottin reduced estrone-3-sulfate uptake by at least 50%. However, bergamottin did not inhibit glyburide uptake. For this substrate, DHB (10 μM) and naringenin (10 μM) exhibited the greatest inhibitory capacity, reducing transport by approximately 65% and 50%, respectively.

The uptake of fexofenadine by OATP1A2-expressing HeLa cells was inhibited by GFJ in a concentration-dependent manner, with 0.5% juice causing a 50% decrease in transport [46]. In a subsequent study, naringin inhibited OATP1A2-mediated fexofenadine uptake with an IC₅₀ value of 3.6 μM [10]. In a Xenopus laevis oocyte system, naringin inhibited talinolol transport by OATP1A2 (IC₅₀ = 343 μM), but not by OATP2B1 [47].

4.2.2 Clinical Relevance

The number of reported drugs whose AUC values are reduced in the presence of GFJ is limited – Table 1 [9]. Fexofenadine was the first drug reported to interact with GFJ through OATP inhibition and has been the most extensively studied [46]. The reported AUC ratios for fexofenadine are between 0.37 and 0.67, with lower values associated with higher intake of GFJ (1200 mL). Unlike its effect on CYP3A, GFJ-mediated inhibition of OATP is short-lived. One study demonstrated that the AUC ratio of fexofenadine was 0.48 when coadministered with 300 mL of GFJ [38]. However, when the same volume of juice was given 2 hours and 4 hours before fexofenadine, the ratios were 0.62 and 0.96, respectively. Furthermore, immunohistochemical analysis of intestinal biopsy samples from study volunteers revealed no changes in OATP1A2 expression as a result of GFJ ingestion, in contrast to the observed reduction in intestinal CYP3A protein levels [38].

In vitro and in vivo data indicate that the flavonoids are the constituents in GFJ primarily responsible for OATP inhibition. Administration of a 1,210 μM aqueous solution of naringin with fexofenadine resulted in an AUC ratio of 0.78 [10] . The AUC ratio was 0.58 when fexofenadine was ingested with GFJ (naringin concentration of 1,234 μM). A furanocoumarin-rich fraction of GFJ concomitantly with, or two hours prior to, fexofenadine failed to alter drug exposure – AUC ratios of 0.98 and 0.95, respectively [10]. As further evidence implicating the flavonoids in the inhibition of OATP, orange juice and apple juice, which do not contain FCs, also reduce plasma concentrations of OATP substrates [46, 48].

Other OATP substrates that have their bioavailability notably reduced by GFJ include celiprolol, talinolol, aliskiren, and etoposide – Table 1. However, Table 1 also contains many drugs that are putative OATP substrates, but do not exhibit reduced exposure when ingested with GFJ. An example is glyburide. Although OATP-mediated uptake of glyburide is inhibited by GFJ in vitro, the only clinical study to date reported an AUC ratio of 1.05 [49]. The statins are also known OATP substrates, but coadministration with GFJ either causes no change, or an increase, in their exposure.

There are many reasons for these discrepancies. The rate and extent of drug absorption is influenced by a number of factors related to drug dissolution, drug transport, and enteric/hepatic first-pass metabolism. Furthermore, many drugs are substrates for OATPs, P-glycoprotein, and CYP3A – all of whom are reported to be influenced by GFJ. As a result, OATP-mediated transport may only play a small role in determining the systemic exposure for a given drug [9]. This implies that the only way to identify the potential of GFJ to interact with a particular therapeutic agent is to conduct a clinical pharmacokinetic study.

4.3 P-glycoprotein

P-glycoprotein (P-gp) is a 170 kDa transmembrane protein that belongs to the ATP-binding cassette superfamily of transporters. It is located throughout the body, including multiple sites involved in drug disposition. P-gp is localized to the apical membrane of hepatocytes and enterocytes, where it serves as a barrier to drug absorption by transporting drugs into the bile and intestinal lumen, respectively. It colocalizes with OATP1A2 to the apical brush border membrane of small bowel enterocytes [38]. P-glycoprotein possesses a broad substrate specificity, and there is considerable overlap with known CYP3A substrates and inhibitors [50]. Consequently, it is thought that GFJ-mediated inhibition of P-gp efflux transport might augment the increase in drug exposure seen for dual P-gp/CYP3A substrates.

4.3.1 In vitro Interactions with Grapefruit Juice

The majority of in vitro investigations have reported an inhibition of P-gp activity by GFJ, extracts of the juice, and GFJ constituents [8, 30, 31, 39, 46, 47, 51-57]. Nevertheless, the use of different P-gp substrates, inhibitor concentrations, and experimental systems has made it difficult to clearly define the in vitro effects of GFJ and its constituents on this transporter.

Grapefruit juice is a potent inhibitor of P-gp-mediated colchicine transport [56]. GFJ inhibited the secretory transport (basal-to-apical) of colchicine across Caco-2 cell monolayers with an IC₅₀ value of 0.46%. Secretory colchicine transport was also dose-dependently reduced by naringenin, DHB, and naringin – IC₅₀ values of 12, 90, and 592 μM, respectively. However, it should be noted that DHB decreased P-gp activity to a greater extent than naringin or naringenin. Naringin and naringenin are weak inhibitors of P-gp-mediated talinolol transport, with calculated IC₅₀ values in the millimolar range [47, 54]. Although the mechanism of P-gp inhibition by naringin has not been directly examined, the fact that naringin is a substrate for this transporter suggests it is a competitive inhibitor [58].

A study by Ohnishi and colleagues [30] compared the effect of five FCs on vinblastine uptake by Caco-2 cells. Paradisin A, bergamottin, DHB, bergapten and bergaptol all increased vinblastine uptake – by decreasing P-gp-mediated vinblastine efflux – at concentrations ranging from 0.1 to 20 μM. The estimated concentrations necessary to increase uptake by 100% over baseline values were 0.035 μM for paradisin A and bergamottin, 0.040 μM for DHB, 0.050 μM for bergapten, and 0.30 μM for bergaptol. While paradisin A demonstrated a marked concentration-dependent increase in vinblastine uptake, the concentration-response relationship for the other FCs appeared to plateau in concentration range tested. Additional studies with other cell systems and substrates will be necessary to further determine if the paradisins are the primary P-gp inhibitors in GFJ.

4.3.2 Clinical Relevance

Definitive evidence supporting in vivo inhibition of P-gp by GFJ is limited. Three studies of intestinal biopsy samples demonstrated that GFJ ingestion does not alter the mRNA and protein expression levels of P-gp [37-39, 59]. This is similar to GFJ's effect on OATP expression, and suggests that GFJ-mediated inhibition of P-gp would be short-lived.

Cyclosporine is a well-known P-gp/CYP3A substrate, but P-gp may be a more important determinant of cyclosporine absorption than enteric CYP3A [60]. Numerous studies have examined GFJ's effect on cyclosporine disposition, with reported AUC ratios ranging from 1.08 to 1.85 – Table 1. Therefore, it is possible that GFJ inhibition of P-gp may be occurring – albeit in the context of enteric CYP3A inhibition. The results of one study in seven healthy volunteers support this hypothesis [39]. When cyclosporine was coadministered with Seville orange juice or GFJ, only GFJ significantly increased the mean AUC of cyclosporine, even though the concentration of DHB was similar in the two juices and both juices reduced intestinal CYP3A content. The authors concluded that the elevated levels of cyclosporine were a result of P-gp inhibition by constituents in GFJ that were not found in Seville orange juice. However, Paine and Oberlies [3] noted that the study may have been underpowered (n=7) and/or the Seville orange juice administered may have differed from GFJ in total furanocoumarin content. Thus, the findings might reflect less robust inhibition of CYP3A by Seville orange juice as compared with GFJ.

Considerable evidence indicates that GFJ has minimal effect on P-gp. The mean AUC ratios for the P-gp/CYP3A substrates amprenavir and indinavir did not significantly differ from 1 (Table 1). Digoxin transport has also been extensively used as an in vitro P-gp activity marker, and GFJ reduces digoxin transport in cell culture systems. Nevertheless, GFJ had a negligible effect on the pharmacokinetics of digoxin in the two studies published to date [61, 62]. Fexofenadine, talinolol, and celiprolol are transported by both P-gp and OATP. Since GFJ decreases the bioavailability of all three of these drugs, it is likely that GFJ's effect on OATPs is more pronounced than its effect (if any) on enteric P-gp for these compounds.

It is not fully established whether GFJ inhibits in vivo P-gp activity, largely because an ideal P-gp marker substrate has not been identified. Most P-gp transported drugs are also metabolized by CYP3A, thereby making it difficult to assign causality when the levels of a P-gp/CYP3A substrate drug are elevated with GFJ ingestion. Furthermore, many non-metabolized P-gp substrates (e.g. fexofenadine) are transported by OATPs.

4.4 Esterases

Esterases are a ubiquitous class of enzymes present in many body tissues, including those important for drug metabolism – e.g., liver and small intestine. These enzymes are involved in the activation of a number of prodrugs used clinically.

4.4.1 In vitro Interactions with Grapefruit Juice

The hydrolysis of the ester prodrugs, enalapril and lovastatin, was reduced in the presence of GFJ [63]. At a concentration of 40%, GFJ reduced purified porcine esterase-mediated hydrolysis of enalapril and lovatstatin to 31% and 26% of control values, respectively. In similar studies with liver S9 fractions, 10%, 20%, and 40% GFJ reduced the hydrolysis of lovastatin to 54%, 46%, and 38% of control, respectively. Conversely, the liver S9 fraction-mediated hydrolysis of enalapril was resistant to inhibition – 40% GFJ reduced activity to 78% of the level observed in the absence of juice. A subsequent study demonstrated that naringin (1000 μM), hesperidin (200 μM), bergamottin (100 μM), DHB (100 μM), and bergapten (100 μM) were weak inhibitors of esterase activity [64]. However, the flavonoid aglycones inhibited the hydrolysis of p-nitrophenylacetate by human liver microsomes with IC₅₀ values in the micromolar range: naringenin (30 μM), quercetin (43 μM), kaempferol (62 μM), morin (80 μM), and galangin (81 μM).

4.4.2 Clinical Relevance

At the present time, the effect of GFJ on esterase activity in humans is not established. Of the two drugs tested in vitro – enalapril and lovastatin – only the GFJ-lovastatin interaction has been explored in vivo (Table 1). Although the AUC ratio for lovastatin and lovastatin acid were increased with GFJ exposure, the fact that lovastatin is also a CYP3A substrate makes it difficult to characterize the role, if any, of esterase inhibition.

4.5 Sulfotransferases

The sulfotransferase enzymes are involved in the phase II conjugation of both endo- and xenobiotics. SULT1A1 has extensive tissue distribution and is the most abundant SULT isoform in the adult liver. Levels of SULT1A3, on the other hand, are highest in the intestine and are very low in the liver.

4.5.1 In vitro Interactions with Grapefruit Juice

GFJ inhibited SULT1A1 and SULT1A3 activity in a concentration-dependent manner [65, 66]. Ritodrine sulfation by SULT1A1was reduced by approximately 50%, 90%, and 100% in the presence of 1%, 5%, and 10% GFJ, respectively. The degree of inhibition of ritodrine sulfation by SULT1A3 was markedly lower – 10% GFJ reduced activity by only 50%. Similar results were observed with other substrates – p-nitrophenol (SULT1A1) and dopamine (SULT1A3) – indicating the inhibition by GFJ was not substrate dependent [65, 66]. Among individual GFJ constituents, quercetin was the most potent inhibitor of SULT, virtually abolishing SULT1A1 activity at the three concentrations (0.1, 1, 10 μM) tested. The resistance of SULT1A3 to inhibition was again noted, with 10 μM quercetin decreasing activity by only 50%. The effects of naringin, naringenin, bergamottin, and DHB on SULT1A1 and SULT1A3 activity were less pronounced [66]. These findings for recombinant SULT isoforms are consistent with earlier work that characterized the inhibitory effects of quercetin, naringenin, and naringin on human liver cytosol sulfotransferase activity [67].

4.5.2 Clinical Relevance

The potential inhibition of sulfotransferases by GFJ has not been directly studied in humans. Nevertheless, the in vitro observation that the predominant intestinal isoform (SULT1A3) is relatively resistant to GFJ-mediated inhibition would suggest that a clinical interaction is unlikely.

5. Conclusion

Grapefruit juice and its constituents alter the activity of CYP enzymes, esterases, sulfotransferases, OATPs, and P-gp in vitro. Although in vitro studies can be informative, the in vivo environment is complex and cannot be fully replicated in vitro. Consequently, the question of whether GFJ interacts with a given drug can only be definitively answered with data from controlled pharmacokinetic studies.

Currently, clinical evidence for significant GFJ-drug interactions exists for CYP3A and OATPs – Figure 5. For CYP3A substrates, grapefruit juice will only cause a clinically relevant interaction if all of the following conditions are met: (1) the drug is given orally; (2) the drug is significantly metabolized by CYP3A; and (3), the drug normally undergoes extensive first-pass metabolism by enteric CYP3A. Because the effect of GFJ on CYP3A is irreversible, up to 3 days may be necessary for enzymatic activity to return after GFJ ingestion. The number of drugs where GFJ causes a strong degree of CYP3A inhibition (AUC ratio ≥ 5) is quite small – Table 1. In fact, the majority of affected drugs exhibit AUC ratios between 1.0 and 2.0 in the presence of GFJ. Nevertheless, avoidance of GFJ may be prudent in some clinical contexts.

Figure 5.

Schematic representation of intestinal enterocyte cells. In the control condition (without GFJ, left), substrate drugs in the intestinal lumen (below) may undergo uptake transport by OATPs, efflux transport by P-gp, and/or metabolism by CYP3A, all of which may influence the extent to which the drug reaches the portal circulation (above). With exposure to GFJ (middle), uptake transport by OATPs and metabolism by CYP3A are potentially inhibited.

The flavonoids are the GFJ constituents responsible for clinically significant OATP inhibition. In contrast to CYP3A, OATP inhibition by GFJ is short-lived, suggesting that the GFJ-mediated reduction in drug bioavailability can be avoided by separating drug administration from GFJ ingestion. This type of GFJ-mediated interaction has only recently been identified and, as a result, examined only for a few drugs. Since OATPs are increasingly recognized for their role in mediating drug disposition, an increase in the number of GFJ-OATP investigations can be anticipated.

The variability in reported AUC ratios among studies using the same drug should also be appreciated. This variability can be attributed to intrinsic biological variation, as well as, variable concentrations of bioactive constituents in the juices tested. Since the furanocoumarins are the substances responsible for CYP3A inhibition, future GFJ-drug interaction studies will be most useful if the concentrations of these constituents are measured and reported. Similarly, measurement of flavonoids should be conducted for GFJ-OATP interaction studies. Phytochemical characterization of the juice will enable between-study comparisons and could potentially identify the minimum constituent concentrations necessary to cause clinically significant effects.

6. Expert Opinion

The fortuitous finding of pharmacokinetic interactions of prescription drugs with GFJ over 20 years ago precipitated a great deal of important basic and translational research at the interface of nutrition science, molecular pharmacology, and clinical therapeutics. We also have learned some cautionary lessons about the interpretation and misinterpretation of biomedical science as it relates to a widely-consumed and popular beverage.

In vitro metabolic models are useful in providing valuable data for estimating the probability of drug interactions involving prescription drugs, but are far less useful for predicting interactions of prescription drugs with GFJ and other fruit beverages. CYP3A is inhibited in vitro by many fruit beverages and their components, but only GFJ has the potential to produce clinically important drug interactions in vivo. This may be explained by the unique presence in GFJ of furanocoumarin derivatives that irreversibly inhibit CYP3A.

Among clinical pharmacokinetic studies that demonstrate statistically significant drug interactions with GFJ, only a few show interactions large enough to be clinically important, and many of these involve exposure to GFJ that exceed quantities typically ingested by consumers. Although the use of concentrated or unrealistically large quantities of juice increases the likelihood of finding a GFJ-drug interaction should one exist, it makes it difficult to translate the applicability of the findings to the usual clinical setting. As a result, future clinical investigations should involve administration of GFJ in a manner consistent with “real world” consumption by the general public. Investigators might also consider studying the effects of high-dose and low-dose GFJ administration within the same clinical trial.

Variability among studies in the magnitude of drug interactions with GFJ may be partly explained by variations in furanocoumarin content. The value of future research will be enhanced by quantitation of furanocoumarin levels in the GFJ products used. Moreover, our knowledge could be enhanced by drug interaction studies involving administration of individual furanocoumarins. These types of trials might lead to the identification of the furanocoumarin(s) necessary to cause an interaction, and the minimum concentration required for in vivo CYP3A inhibition. To date, only bergamottin has been studied in this manner [36].

The recent findings that GFJ can be rendered “interaction-free” by removal of furanocoumarins via methods such as filtration, heating, or UV exposure suggests that we are one step closer toward eliminating the risk of GFJ-drug interactions. If methods of furanocoumarin removal can be adapted for large-scale production of taste-acceptable GFJ products consumed by the public, the therapeutic problem of drug interactions with GFJ would diminish or disappear. Ideally, this modified GFJ would also provide the same health benefits to the consumer as the unprocessed juice.

List of abbreviations

AUC

area under the plasma concentration-time curve

CYP

cytochrome P450

DHB

6′,7′-dihydroxybergamottin

FC

furanocoumarin

GFJ

grapefruit juice

OATP

organic anion-transporting polypeptide

P-gp

P-glycoprotein

SULT

sulfotransferase

UV

ultraviolet