Abstract and Introduction
Abstract
Traditionally, drug-induced changes in cytochrome P450 isoenzyme activity, causing changes in drug metabolism and bioavailability, have been the main focus of drug interaction studies. Recent research, however, suggests that the drug transporters P-glycoprotein and organic anion transporting peptide (OATP), which can effect the efflux and influx of many classes of drugs, may contribute to drug interactions by mechanisms independent of oxidative metabolism. Experimental models designed to selectively probe the function of P-glycoprotein or OATP have demonstrated that changes in the activities of these transporters may have a significant effect on the bioavailability of clinically important drugs, leading to the potential for adverse drug interactions.
This review focuses on what is known about the P-glycoprotein and OATP drug transporters and their effects on drug bioavailability. Where possible, it uses as examples the second-generation H1-receptor antagonists, where concomitant administration of other drugs or food constituents has been shown to alter the bioavailability of some agents of this class via mechanisms probably mediated by P-glycoprotein and/or OATP.
Introduction
Over the past 30 years there has been an increase in the incidence of allergic diseases in the United States and worldwide.[1] Antihistamines are the mainstay of treatment for many allergic diseases, such as allergic rhinitis, and are among the most widely prescribed medications in the world. The ubiquity of antihistamine therapy, however, increases the risk of potentially serious drug interactions, such as those seen when terfenadine was coadministered with erythromycin[2] or ketoconazole.[3]
The cytochrome P450 (CYP) pathway has been the traditional focus of investigation relating to drug-drug interactions. A wide range of drugs acts as substrates, inhibitors and inducers of CYP enzymes. For example, the interactions between certain H1-receptor antagonists and the anti-microbial agents erythromycin and ketoconazole have been well described and were initially interpreted exclusively in terms of inhibition of CYP3A4.[4] Recent research, however, has revealed that changes in absorption and excretion of drugs independent of CYP metabolism can alter drug disposition and may account for some drug interactions previously attributed to changes in CYP activity. This experience has emphasised the need to understand the mechanisms of potential drug interactions, especially with drug classes (like antihistamines) that are commonly used by large heterogeneous patient populations.
Altered drug disposition through changes in absorption and excretion is particularly apparent for agents that are minimally metabolised by the CYP pathway. These agents are also subject to pharmacokinetic changes when coadministered with certain other drugs. For example, fexofenadine is an H1-receptor antagonist that undergoes minimal hepatic or enteric metabolism. However, pharmacokinetic studies have demonstrated elevated concentrations of fexofenadine in volunteers coadministered fexofenadine together with erythromycin or ketoconazole.[5] A preliminary clinical report has shown that consumption of grapefruit juice significantly decreased the bioavailability of orally administered fexofenadine.[6] Although CYP enzymes have traditionally been the usual suspects in drug-drug interactions, these results and other recent evidence show that drug inter-actions can occur through non-CYP-mediated mechanisms.
A newly recognised class of active drug transporters, including P-glycoprotein (P-gp) and organic anion transporting polypeptide (OATP), are now known to affect the disposition and bioavailability of many drugs. In general, P-gp inhibits the absorption and increases the excretion of drugs.[7] OATP is a bidirectional transporter that facilitates drug absorption and biliary excretion.[8,9] Both transporters are present in the blood- brain barrier, intestinal mucosa, kidney epithelial cells and other tissues. Interestingly, the activities and protein levels of these transporters are altered by many of the same agents that affect the CYP enzymes, although distinct interactions have also been noted.[10]
Because of the broad potential for drug inter-actions via this mechanism, the impact of these transporters on the disposition of clinically important agents, such as H1-receptor antagonists, is the subject of this review.
1. Characteristics of the P-Glycoprotein Transporters
P-gp is a 170kD transmembrane glycoprotein that in humans is encoded by the MDR1 (multidrug resistance) gene. It is the most extensively studied member of the ATP-binding cassette (ABC) transporter superfamily.[7,11,12] It was originally discovered in drug-resistant tumour cells[13] and later identified in normal human tissues.[14] In mice, two genes have been identified that code for the P-gp transporters, mdr1a and mdr1b. In mdr1a/mdrlb
-/- mice, both of the genes coding for P-gp have been removed by genetic engineering (knockout mice), resulting in an absence of P-gp activity. Research on drug transport by the P-gp pathway has been greatly enhanced by the use of cell lines transfected with the human P-gp gene and mice in which the corresponding genes (mdr1a or mdr1b) are either overexpressed or deleted.[7,11,15-17]
P-gp is an ATP-dependent efflux pump that exports drugs and endogenous metabolites out of the cell, thus affecting distribution within the body (fig. 1).[7,12] P-gp is specifically localised on the apical membrane of secretory cells,[14] where it plays an important defensive role in secreting xenobiotics and metabolites into the intestinal lumen, urine and bile, and in protecting the brain from excessive accumulation of toxic drugs and metabolites. In support of these functions, human P-gp is present at high levels in the intestinal mucosa, lumenal membranes of the renal proximal tubules, the biliary canalicular membrane of hepatocytes, the adrenal gland, endometrium and astrocyte foot processes associated with the blood-brain barrier (BBB).[7,11,14,18] However, P-gp also confers drug resistance to certain cell types, which has hindered HIV and anticancer therapy by inhibiting therapeutic drug accumulation in target cells.[7]
Figure 1. Effect of inhibition of P-glycoprotein (P-gp) on drug absorption. (a) In this example, the P-gp transporter is located on the apical membrane of polarised intestinal mucosal cells where it reduces the absorption of P-gp substrates by pumping substrates out of the cell through the apical membrane and into the intestinal lumen. P-gp substrates that enter from the plasma (basolateral) side of these cells are also eliminated through the apical side. P-gp is also found in the kidney, liver, adrenal gland and blood-brain barrier. (b) Inhibition of P-gp allows increased absorption of P-gp substrates; these substrates are no longer pumped out of these cells, allowing increased absorption from the intestinal lumen and decreased excretion from plasma.
The roles of other structurally similar members of the ABC transporter family, such as those encoded by MDR3, multidrug resistance-associated protein (MRP) and the canalicular multispecific organic anion transporter (cMOAT) genes, are less well defined with respect to drug transport.[7] The MDR3 gene product is a phospholipid transporter (also called flippase) that acts mainly as an intra-cellular translocator of lipids and has been shown to transport drugs across cells in vitro.[19] Unlike the MDR1 gene and related murine mdr1a and mdr1b genes, the MDR3 gene (and the corresponding murine mdr2 gene) does not confer multidrug resistance on drug-sensitive cells.[7]
1.1 Drugs Transported by P-Glycoprotein
P-gp plays a significant role in the transport and efflux of a wide range of drugs in different tissues. Generally, P-gp substrates are hydrophobic, although mycophenolic acid, which is hydrophilic, is also transported by P-gp, according to preliminary reports.[20] Several different classes of drugs are transported by P-gp (table I). These include antihistamines, anticancer agents, immunosuppressants, cardiac glycosides and steroids.[11] Several antihistamines bind to P-gp at physiological concentrations, including terfenadine,[10] fexofenadine[21] and astemizole.[22] Experimental evidence of P-gp interaction with some representative drugs is summarised in table II.
Interaction with P-gp does not occur with all antihistamines, however, since desloratadine, loratadine and cetirizine are not affected by changes in P-gp activity at physiological concentrations. In the case of desloratadine and loratadine, this has been demonstrated by a minimal effect at physiological concentrations on P-gp- mediated efflux of daunorubicin from fibroblasts expressing human P-gp.[23] In early reports of these in vitro studies, daunorubicin efflux was significantly suppressed by ketoconazole, a known potent P-gp inhibitor. However, under the same conditions, desloratadine failed to alter daunorubicin efflux.[23] These data suggest that desloratadine is minimally bound or transported by P-gp.
1.2 Central Nervous System Transport via P-Glycoprotein
A major site of P-gp expression is at the BBB, where P-gp functions to restrict access of drugs and other molecules into the brain.[15,24] In mdr1a
-/- mice (deficient for one of two P-gp genes), an increased brain-to-plasma ratio specifically characterises P-gp substrates.[15] For fexofenadine, the brain-to-plasma ratio in wild-type mice (mdr1a
+/+) was 0.17, whereas in mdr1a
-/- mice the ratio was doubled to 0.33. At 4 hours after administration of fexofenadine, these mdr1a
-/- mice also had a five-fold increase in plasma concentrations of fexofenadine compared with mdrla+/+
mice.[21] These data suggest that fexofenadine access to the brain is affected by P-gp transport.
Several other drugs are transported via the P-gp efflux pump associated with the BBB, thereby minimising their accumulation in brain tissue. Administration of ivermectin, vinblastine or cyclosporin to mdr1a
-/- mice resulted in increased concentrations of these drugs in brain tissue. Increased concentrations of the antidiarrhoeal agent loperamide were also observed in the brain tissue of mdr1a
-/- mice, causing an opioid-like effect in these mice.[15,17] In addition, the concentration of intravenously administered HIV protease inhibitors (indinavir, nelfinavir and saquinavir) increased seven-to 36-fold in the brain tissue of mdr1a
-/- mice compared with wild-type mice.[24] This same effect occurred in wild-type mice when they were administered the P-gp inhibitor valspodar (PSC-833); digoxin concentrations significantly increased in the brain tissue of wild-type mice treated with valspodar and digoxin.[17] Thus, it is apparent that P-gp plays a significant role in the transport of different classes of drugs, decreasing their access to the central nervous system (CNS). The consequences of normal P-gp activity in the BBB may be desirable (e.g. selective reduction of CNS adverse effects) or undesirable (e.g. decreasing the activity of antiretrovirals within the brain).
1.3 Intestinal Transport via P-Glycoprotein
P-gp plays an important role in the transport and efflux of drugs from intestinal epithelium, as elucidated by studies with HIV protease inhibitors.[25,26] Using in vitro models of absorption, indinavir, saquinavir and ritonavir have been shown to bind P-gp-transfected cell membrane preparations in vitro
[25] and have shown P-gp transport through Caco-2 epithelial cell monolayers.[26] Likewise, increased absorption of orally administered HIV protease inhibitors[24] or paclitaxel[27] resulted in two-to six-fold elevations of plasma drug concentrations in mdr1a
-/- mice compared with wild-type mice.[24,27] In addition, in wild-type mice, elimination of digoxin into the gut lumen was inhibited by oral administration of the P-gp inhibitor valspodar, suggesting basolateral-to-apical transepithelial transport of digoxin by P-gp.[17,28] In humans, elevated intestinal P-gp concentrations in renal transplant patients receiving oral cyclosporin (a substrate for P-gp) correlated with increased oral clearance and decreased blood concentration of the drug. Conversely, oral clearance of cyclosporin was decreased and plasma concentrations increased in patients expressing low levels of intestinal P-gp.[29]
When the antihistamine fexofenadine was given orally or intravenously to mdr1a
-/- mice, the fexofenadine concentration increased five-fold in the plasma compared with wild-type mice.[21] Although the mdr1a
-/- phenotype suggests that this change is due to P-gp transport, the interpretation of these results in whole animal systems is complicated by the fact that fexofenadine is also a substrate for OATP transport, which mediates cellular uptake of specific drugs (see section 2). In vitro experiments in polarised epithelial cells have also shown that P-gp affects the rate of fexofenadine transport. The basolateral to apical (secretory) transport of fexofenadine was significantly greater in cells expressing P-gp than in epithelial cells devoid of P-gp, reinforcing a role for P-gp transport mechanisms in the biodistribution of this drug.[21]
Collectively, these data demonstrate that intestinal absorption of certain drugs is restricted by P-gp transport. P-gp substrates that enter intestinal mucosal cells from either the apical side or the basolateral side are transported by P-gp through the apical side of mucosal epithelium into the intestinal lumen. Changes in P-gp transport function may account for the unexpected differences in the bioavailability of various drugs affected by P-gp.
1.4 Regulation of P-Glycoprotein and Drug Interactions
As described above, P-gp transport represents one of several major mechanisms by which the distribution of numerous drugs is controlled. It follows that drugs that induce or inhibit P-gp may have a profound effect on the pharmacokinetics and disposition of drugs transported by P-gp within the body, possibly compromising their bioavailability. These P-gp-related mechanisms are thought to be in part responsible for known drug-drug inter-actions that can lead to altered bioavailability of specific drugs.[10] For example, coadministration of rifampicin (rifampin) [a P-gp inducer] and digoxin (a P-gp substrate) decreases the bioavailability of digoxin,[30] and coadministration of erythromycin (a P-gp inhibitor) and talinolol (a minimally metabolised P-gp substrate) increases the bioavailability of talinolol.[31]
1.5 Induction of P-Glycoprotein: Effects on Drug Disposition
A variety of drugs have been shown to increase expression of P-gp (table III). In human colon carcinoma cell lines, midazolam and nifedipine selectively induce P-gp, and rifampicin, phenobarbital, clotrimazole, reserpine and isosafrole induce both the expression of P-gp and CYP3A4.[32] Thus, although most of these drugs have the potential for drug interaction through the CYP pathway, recent studies suggest that modulation of P-gp activity may be equally important in this respect.
Coadministration of the herbal preparation St John’s wort has also been reported to decrease digoxin serum concentrations through increased P-gp activity.[33] Although the following reports are preliminary, there is accumulating evidence that St John’s wort can affect the bioavailability of fexofenadine. A recent study observed a 50% reduction in the area under the concentration-time curve (AUC) for fexofenadine in healthy volunteers following coadministration of St John’s wort for 12 days.[34] In another study, a single dose of St John’s wort increased the Cmax of fexofenadine by 37%, although in this study repeated intake of St John’s wort did not affect fexofenadine pharmacokinetics.[35] The reason for these conflicting results is not clear; more research is needed.
The effects of drug interactions with P-gp suggest that, independently of CYP, P-gp inducers may play a significant role in altering drug bioavailability by decreasing intestinal absorption and possibly by increasing clearance through the kidney.
1.6 Inhibition of P-Glycoprotein: Effects on Drug Disposition
Some drugs have also been shown to inhibit P-gp transport mechanisms (reviewed by Silverman[11] and summarised in table III). Tanigawara et al.[7] showed that the transport of digoxin occurred via a P-gp-dependent mechanism located on the apical side of kidney epithelial cell membranes. Treatment with spironolactone, cyclosporin, quinidine and verapamil inhibited the P-gp-mediated transport of digoxin. Because these drugs are known to interact with digoxin, causing an increase in plasma drug concentrations and toxicity, it appears that the P-gp excretory pathway within the kidney may be an important site of drug inter-action.
Inhibition of intestinal P-gp has been proposed as a mechanism to explain increases in the bioavailability of certain drugs. For example, the P-gp inhibitor erythromycin has been reported to increase the bioavailability of many drugs, including the H1-receptor antagonists terfenadine and astemizole, the immunosuppressant cyclosporin and the ß-adrenergic antagonist talinolol.[10,22,29,31] Nonetheless, the extent to which inhibition of P-gp is responsible for these interactions requires further study.
Inhibition of P-gp transport mechanisms can be exploited to enhance the efficacy of tumour cell therapy, permitting increased access of antitumour drugs to tumour cells. P-gp inhibitors may also compromise the safety of certain drugs by increasing their absorption in the gut and decreasing their clearance through the kidney. The potential seriousness of the compromise in safety has been suggested by the changes in disposition of terfenadine and astemizole that led to life-threatening cardiac complications when these drugs were coadministered with agents known to inhibit P-gp and CYP metabolism.
2. Organic Anion Transporting Polypeptide (OATP) Transporters
2.1 Characteristics of OATP Transporters
Considerably less is known about OATP transporters. The OATP family of transporters is involved in the transport of endogenous substances such as bile acids, where they have been most widely studied as mediators of bile acid enterohepatic circulation.[9,36] In addition, OATP plays a key role in the uptake of drugs into cells, in contrast to the role of P-gp, which mediates drug efflux.[21] These transporters have been studied in animal models (oatp1 and oatp2 in the rat) and have been evaluated in vivo and in cell lines derived from hepatic and renal tissues.[37-40] The first identified human OATP transporter has 67% amino acid identity to rat oatp1.[41] Recently, several human OATP proteins have been identified, and the tissue-specific expression of these proteins is being studied.[8]
2.2 Structure, Localisation and Function of OATP Transporters
Rat oatp1 and oatp2 are glycoproteins with 12 putative transmembrane domains. OATP transporters are expressed in the rat on the liver sinusoidal membrane,[42] the apical membrane of the kidney,[42] and the choroid plexus of the brain.[43] Oatp1-mediated transport requires ATP, is Na+ –independent, and is bidirectional.[40]
2.3 OATP Substrates
Substrates of OATP transporters include a variety of endogenous compounds of differing chemical structures and a wide variety of drug classes (table IV). Oatp1 and oatp2 primarily transport bile acids and derivatives, steroids, peptidomimetics,[44] glucuronides[37,44-46] and anionic estrogen conjugates.[37,39,45,46] Oatp1 is capable of transporting endogenous organic anions such as the bile acid taurocholate,[36,37,47] estradiol 17-glucuronide,[37,46] the steroid hormone estrone-3-sulfate[45] and enalapril.[39] Oatp2 is similar to oatp1 in that it transports many of the same substrates as oatp1,[37] but demonstrates different specificities for some drugs such as digoxin.[48]
2.4 Drugs Transported by OATP Transporters
The angiotensin converting enzyme (ACE) inhibitor enalapril has been shown to be taken up by oatp1, which also transports the anionic dye bromosulfophthalein,[9,39] the thrombin inhibitor CRC-220,[44] ouabain[45] and temocaprilat.[38] Oatp2 is known to transport pravastatin,[49] digoxin, and thyroxine.[37,48,50] In addition, oatp1 and oatp2 also mediate the cellular uptake of fexofenadine, and human OATP is capable of fexofenadine transport as demonstrated by in vitro systems.[21]
Cvetkovic et al.[21] examined the role of OATP and P-gp in the cellular uptake and excretion of fexofenadine. They showed that drugs that alter P-gp transport activity, such as HIV protease inhibitors, statins, quinidine, betocanazole and verapamil, also affect the function of OATP transporters, suggesting that combined inhibition of OATP and P-gp may be the mechanistic explanation to account for the observed drug interactions involving fexofenadine. A number of cloned transporters present on the basolateral membrane of the hepatocyte and also in the kidney were investigated using a heterologous expression system. Oatp1 and oatp2 were found to effectively mediate the uptake of fexofenadine, consistent with the anionic nature of this drug.[21]
Initial reports from other studies have indicated that grapefruit juice can affect the bioavailability of H1-receptor antagonists, most likely through mechanisms involving OATP.[16] Fexofenadine is not a substrate for CYP and is normally excreted from epithelial cells via P-gp transport mechanisms.[21] In a clinical study involving healthy male and female volunteers, reported in abstract form, plasma concentrations of orally administered fexofenadine were significantly diminished by coadministration of grapefruit juice; the rate and extent of fexofenadine absorption was reduced by 30%. However, the bioavailability of another H1-receptor antagonist, desloratadine, was not significantly altered following coadministration with grapefruit juice.[6] In addition, Dresser et al.[51] recently showed in a preliminary report that other citrus juices besides grapefruit juice are potent inhibitors of intestinal OATP. They found that orange juice induced an approximately two-fold greater percentage reduction in the AUC of fexofenadine than grapefruit juice (-43% vs -20% reduction compared with water, respectively) in mdr1a/mdr1b-/- mice.[51] In another preliminary study, grapefruit juice, orange juice and apple juice reduced the AUC of fexofenadine by 67%, 72% and 77%, respectively, in healthy volunteers.[52]
In summary, the OATP transporters can have a profound effect on the absorption and bioavailability of drugs. Even common foods such as fruit juices can affect their activity, causing a large decrease in drug bioavailability. Also, because this is such a new area of research, new discoveries in the biodistribution of drugs may show that drug inter-actions previously attributed to CYP and P-gp mechanisms may be attributed to the OATP transporters.
3. Discussion
Traditional concerns with drug interactions have focused on oxidative metabolism via CYP isoenzymes. Recent research has demonstrated that other important mechanisms for affecting drug disposition involving the P-gp and OATP transporters must also be considered when evaluating the potential for drug interactions. To date, these three systems are considered to have the most potential to alter normal drug concentrations.[7,16,53] Such changes are well documented for numerous drugs of different therapeutic classes. The CNS efficacy of HIV protease inhibitors is seriously compromised by P-gp transport efflux mechanisms, which prevent accumulation of these drugs within the brain.[24] The limited oral bioavailability of HIV protease inhibitors and paclitaxel can also be explained by increased efflux from intestinal mucosal epithelium through P-gp transport mechanisms.[25,27,54] Furthermore, drugs that modulate the activity of P-gp and CYP isoenzymes (rifampicin, erythromycin and ketoconazole) are known to affect the bioavailability of coadministered drugs.[4,12,30,55,56] Interference with OATP can result in abnormal plasma drug concentrations, and some alterations in bioavailability may be the result of a combination of changes in P-gp, OATP and CYP processes.
Fexofenadine, the active metabolite of terfenadine, is not significantly metabolised by CYP isoenzymes, but has been shown to interact with transporters such as OATP and/or P-gp.[16,21,34,35,51,52] Pharmacokinetic studies have demonstrated elevated blood concentrations of fexofenadine in volunteers coadministered fexofenadine and erythromycin, or fexofenadine and ketoconazole.[5] A preliminary human study showed that grapefruit juice significantly decreased the bioavailability of orally administered fexofenadine.[6] Thus, even drugs that do not undergo oxidative metabolism via CYP may have their plasma concentrations changed by other mechanisms such as those that act through P-gp or OATP. The clinical significance of these findings needs further investigation.
However, control of antihistamine bioavailability by P-gp and OATP is not a class effect for H1-receptor antagonists. For example, the bioavailability of desloratadine, a new H1-receptor antagonist,[57,58] was unaltered by coadministration of erythromycin,[59] ketoconazole[60] or grapefruit juice,[6] according to preliminary reports. These data are consistent with minimal interaction of desloratadine with P-gp or OATP.[6,23]
4. Conclusions
Improvements in pharmacotherapy in recent years have translated into the broader and more intense use of pharmacological agents by everwidening patient populations. With this increase in pharmacotherapy among increasingly heterogeneous patient populations, the potential for drug interactions also grows. Interactions of a variety of therapeutic agents with the P-gp or OATP transport systems or the CYP metabolic pathway can result in changes in drug bioavailability. Differences appear to exist among the newer H1-receptor antagonists in the potential for drug interactions via these mechanisms, and these differences need to be considered when evaluating the appropriate anti-histamine for individual patients. Further research is needed to clarify the contributing roles of the P-gp and OATP transporters to clinical outcomes.