The ABCG2 C421A polymorphism does not affect oral nitrofurantoin pharmacokinetics in healthy Chinese male subjects

Abstract

WHAT IS ALREADY KNOWN ABOUT THIS SUBJECT

• An increasing number of drugs on the market or under development have been identified as substrates of the ATP-binding cassette drug efflux transporter breast cancer resistance protein (BCRP; ABCG2), which can affect the pharmacokinetics of drugs by reducing absorption and/or increasing biliary elimination.

• ABCG2 C421A, a single nucleotide polymorphism associated with decreased protein expression/transport activity in vitro and higher anti-cancer drug concentrations in carriers of the C421A polymorphism, may contribute to the intersubject pharmacokinetic variability of BCRP substrates.

• Predicting the potential influence of BCRP on drug disposition or drug interactions is challenging because of the lack of a well-characterized, BCRP-selective clinical probe substrate.

• Nitrofurantoin is potentially a suitable clinical BCRP probe substrate based on preclinical and clinical information available (e.g. in vitro transport studies, Bcrp knockout mouse studies, inhibition studies in rats, and milk secretion studies in rats and humans).

WHAT THIS STUDY ADDS

• The ABCG2 C421A SNP had no effect on oral nitrofurantoin plasma and urine pharmacokinetic parameters in healthy male Chinese subjects.

• Nitrofurantoin does not appear to be a useful clinical BCRP probe.

AIMS

A number of drugs are substrates or inhibitors of the efflux transporter breast cancer resistance protein (BCRP; ABCG2), which can limit systemic exposure by reducing absorption and/or increasing biliary elimination. The identification of a BCRP-selective clinical probe drug would provide a useful tool to understand the effect of genetic polymorphisms and transporter-based drug interactions on drug pharmacokinetics. The aim of this study was to assess the utility of nitrofurantoin as a clinical probe substrate for BCRP activity by evaluating the impact of genetic variation on nitrofurantoin pharmacokinetics.

METHODS

Nitrofurantoin pharmacokinetics were studied in an open-label, single-oral dose (100 mg) study in 36 male Chinese subjects who were pre-screened for ABCG2 421 CC, CA and AA genotypes (n = 12 each). Plasma and urine concentrations of nitrofurantoin were determined by LC/MS/MS and LC/UV respectively. anova was used to compare pharmacokinetic parameters among genotypes.

RESULTS

There were no significant differences in nitrofurantoin pharmacokinetics among the genotypic cohorts. The geometric mean nitrofurantoin plasma AUC_(0-∞) (95% confidence interval) values were 2.21 (2.00, 2.45), 2.42 (2.11, 2.78) and 2.32 (1.99, 2.70) µg h ml⁻¹ and half-life values were 0.79 (0.59, 1.0), 0.76 (0.64, 0.89) and 0.72 (0.62, 0.84) h for ABCG2 421 genotypes CC, CA and AA, respectively. The percentage of dose excreted unchanged in the urine was 43, 44 and 39%, respectively.

CONCLUSIONS

The ABCG2 C421A polymorphism had no effect on nitrofurantoin plasma and urine pharmacokinetic parameters in healthy Chinese subjects. These results indicate that nitrofurantoin is not a suitable clinical probe substrate for assessing BCRP activity.

Introduction

The gene ABCG2 codes for breast cancer resistance protein (BCRP), an ATP-binding cassette transporter that was initially identified for its role in the development of anticancer drug resistance. Later, it was shown that BCRP is constitutively expressed in healthy tissues including the intestine, liver, kidney, blood–brain barrier, breast and placenta [1]. In vitro and animal pharmacokinetic studies have suggested that BCRP plays an important role in the oral absorption and elimination of a variety of drugs, including many in the oncology, antiviral and central nervous system therapeutic areas. BCRP in enterocytes can serve as a barrier to drug absorption by transporting substrates back into the intestinal lumen. BCRP at the hepatocyte canalicular membrane contributes to drug elimination by transporting substrates out of the hepatocyte and into the bile. The combined result of intestinal and liver BCRP activity is lower systemic exposures of substrate drugs [1–4].

Several nonsynonymous single nucleotide polymorphisms (SNPs) in the human ABCG2gene are associated with reduced BCRP transport activity, including C421A, C376T, G34A, T1291C and T623C [5–10]. C421A, which results in a substitution of lysine for glutamine (Q141K) in the BCRP protein, is one of the most studied polymorphisms. In vitro experiments in a variety of C421A-transfected cell lines have shown decreased BCRP protein expression and/or less resistance to anticancer agents compared with cells transfected with wild-type ABCG2[5–8, 10]. In vivo, there is evidence that carriers of the C421A polymorphism have altered drug pharmacokinetics. The plasma exposures of intravenous diflomotecan and oral topotecan were shown to be threefold and 1.4-fold higher, respectively, in patients heterozygous for the 421A variant than in subjects homozygous for wild-type 421C [11, 12]. In addition, oral rosuvastatin plasma concentrations were higher in subjects with the 421CA and 421AA genotypes compared with subjects with wild-type 421CC, when controlled for SLCO1B1and CYP2C9genotypic status [13]. Based on these observations, genetic variation in ABCG2 may contribute to the intersubject variability observed for drugs that are BCRP substrates. The C421A allele frequency has been shown to vary by ethnicity; recent findings suggest that up to 46% of East Asians, 16–19% of Whites and <2% of sub-Saharan Africans are homo- or heterozygous carriers of the variant 421A allele [5, 14, 15].

Predicting the potential influence of BCRP on drug disposition is challenging because of the lack of well-characterized, BCRP-selective clinical probe substrates that have a safety profile acceptable for use in healthy volunteers [16]. Nitrofurantoin is an antibiotic used for the treatment of urinary tract infections. In vitro transport studies using mouse mammary epithelial cell monolayers and murine Bcrp- and human BCRP-transfected MDCKII cells have indicated that nitrofurantoin is a BCRP substrate [17–19]. In addition, in vivo pharmacokinetic studies have shown that Bcrp affects the absorption, biliary excretion and elimination of nitrofurantoin in mice. The plasma AUC of oral nitrofurantoin was fourfold higher in Bcrp knockout than wild-type mice. Following intravenous dosing, the nitrofurantoin AUC was twofold higher and the biliary excretion 48-fold lower in Bcrp knockout mice compared with Bcrp wild-type mice [17]. Furthermore, co-administration of nitrofurantoin with the BCRP inhibitor chrysin to rats resulted in a 75% reduction in biliary excretion and significantly increased nitrofurantoin AUC following both oral and intravenous administration [19]. BCRP has also been shown to be the primary transporter responsible for the active secretion of nitrofurantoin into the milk of lactating mice, and to limit the fetal distribution of nitrofurantoin in pregnant mice [17, 20, 21]. Nitrofurantoin is not a substrate for either the P-glycoprotein (MDR1) or multidrug resistance-associated protein 1 and 2 (MRP1, MRP2) efflux transporters in vitro[17, 19].

The results of the above-mentioned studies suggest that BCRP activity should influence nitrofurantoin disposition in man. Therefore, the objective of this clinical study was to evaluate the pharmacokinetics of orally dosed nitrofurantoin as a potential probe for assessing BCRP transport activity in vivo. Specifically, the influence of the ABCG2 C421A polymorphism on nitrofurantoin pharmacokinetics was addressed. The study was conducted in healthy Chinese subjects because of the higher frequency of the ABCG2 C421A polymorphism in this ethnic group and was limited to male subjects because sex differences in BCRP activity and/or expression have been reported [22].

Materials and methods

Subjects

Thirty-six healthy male subjects (age 21–45 years) on no other medications were recruited for this study. Following written informed consent, all subjects underwent an initial screening assessment for collection of a blood sample for genotyping and a second screening assessment within 28 days of the first dose that included a medical history, physical examination, blood pressure, ECG and clinical laboratory tests. Exclusion criteria included a positive urine drug test, past or current history of excessive alcohol or illicit drug use, use of any prescription or nonprescription drugs, vitamins, herbal and dietary supplements or grapefruit-containing products within 7 days or 5 half-lives prior to the first dose of study medication or during the clinical phase of the study, regular use of tobacco- or nicotine-containing products within 3 months of screening visit, any pre-existing conditions that would interfere with normal gastrointestinal anatomy or motility, hepatic and/or renal function, or a history of bone marrow transplant, porphyria, anuria, oliguria, diabetes, glucose-6-phosphate dehydrogenase deficiency, anaemia, vitamin B deficiency, HIV or viral hepatitis.

Study design

The study was approved by the ethics review board at National University of Singapore. Subjects were genotyped for ABCG2 C421A polymorphism prior to entry into the study and assigned to one of three cohorts: 421CC, 421CA or 421AA. Following a 10-h overnight fast, subjects received a single oral dose of nitrofurantoin (100 mg Apo-Nitrofurantoin; Apotex, Toronto, Ontario, Canada) with 250 ml of water and remained fasted until 4 h after dosing, when lunch was provided. Blood samples (2 ml) for the determination of nitrofurantoin plasma concentrations were collected in ethylenediamine tetraaceticacid-containing tubes prior to dosing and at 0.25, 0.5, 0.75, 1, 1.5, 2, 2.5, 3, 4, 5, 6, 10, 12, 16, 24 and 30 h post dose. A predose urine sample was collected; subjects were asked to empty their bladders completely immediately prior to dosing. Thereafter, all urine passed over the next 30 h was collected into separate pre-weighed containers over the following collection intervals: 0–12, 12–24 and 24–30 h. The blood samples were centrifuged to obtain plasma; all plasma and urine samples were stored frozen at −70°C until assayed. All blood, plasma and urine samples were protected from light during collection, storage and analysis.

Genotyping of ABCG2 polymorphisms

Whole blood samples (5 ml) for ABCG2genotyping were obtained from each subject. The genomic DNA extracted from peripheral leucocytes by standard methods was used as a template in amplification of the fragments encompassing the following SNPs of the ABCG2 gene: G34A, C376T, C421A, T623C and T1291C residing in exons 2, 4, 5, 6 and 11, respectively. With the exception of the primer pair for amplifying exon 5 that was designed by Kobayashi and colleagues [6], primer pairs published by Backstrom et al. [23] were used to amplify the other above-mentioned exons of the gene. The amplifications were performed in a total volume of 50 µl containing 1× Master Mix (Promega, Madison, WI, USA), 0.2 µM of each primer (Sigma-Proligo, St Louis, MO, USA) and 100 ng of DNA. Following an initial pre-denaturation step at 95°C for 5 min, the reactions were cycled (exon 5, 25 times; exons 2, 4, 6 and 11, 35 times) through denaturation at 95°C for 1 min, variable annealing temperatures (exon 5, 54°C; exons 2, 4, 6 and 11, 56°C) for 45 s to 1 min and extension at 72°C for 45 s to 1 min. The reactions were terminated by an additional extension step at 72°C for 10 min. The polymerase chain reactions (PCRs) were run on the Peltier Thermal Cycler (DNA Engine Dyad; MJ Research Inc, Waltham, MA, USA).

Prior to sequencing, the PCR products were purified using Exonuclease 1 (New England Biolabs, Beverly, MA, USA) and Shrimp Alkaline Phosphatase (Promega). The sequencing reactions were carried out using BigDye Terminator v3.1 Cycle Sequencing Kit (Applied Biosystems, Foster City, CA, USA). The sequences of genomic fragments were analyzed on the automated ABI Prism Model 3100 Avant Genetic Analyser (Applied Biosystems).

Bioanalytical methods

The concentration of nitrofurantoin in plasma samples and standards was determined by high-performance liquid chromatography (HPLC) with tandem mass spectrometry (MS/MS) using nitrofurazone as the internal standard. Nitrofurantoin was obtained from Sigma-Aldrich (St Louis, MO, USA) and nitrofurazone was obtained from Pfaltz & Bauer (Waterbury, CT, USA). Plasma proteins from a 50-µl plasma sample were precipitated by the addition of three volumes of acetonitrile containing nitrofurazone (500 ng ml⁻¹) and removed by centrifugation. The resulting supernatants were diluted with 300 µl of HPLC-grade water. HPLC was performed on a Shimadzu LC-10A HPLC system. Chromatography was performed on a Varian Polaris C18-A 5 µm, 2.1 × 50 mm column at a flow rate of 0.5 ml min⁻¹. The mobile phase for isocratic elution comprised 1 mM ammonium formate and 100% methanol (80 : 20 v/v). Samples were analysed by negative ion turbo ionspray LC/MS/MS with a PE/Sciex API 4000. The calibration range was 0.02–10 µg ml⁻¹. Raw data were analysed with PE/Sciex software Analyst 1.4.1. SMS 2000 (version 1.6; GlaxoSmithKline) was used to calculate peak area ratios and to construct the calibration curve from which concentrations of unknowns were interpolated. Performance of the method, as assessed by nitrofurantoin concentrations in quality control samples (60, 800 and 8000 ng ml⁻¹), showed that the average within-run precision [coefficient of variation (CV %)] was ≤10.8%. The between-run precision CV% was ≤4.8%.

The concentration of nitrofurantoin in urine samples and standards was determined by HPLC with ultraviolet detection (HLPC/UV) using furazolidone (Sigma Chemical Co., St Louis, MO, USA) as the internal standard. A 20-µl urine aliquot was mixed with 180 µl of HPLC-grade water and 20 µl of internal standard solution (10 µg ml⁻¹); then 50 µl of the resulting solution was injected onto the HPLC system. The HPLC system was a Shimadzu LC-10A system with a Thermo Hypersil BDS C18 5 µm, 4.6 × 250 mm column and an analytical BDS C18 5 µm guard column. The mobile phase consisted of 85% 0.05 M potassium phosphate monobasic (pH 3) and 15% acetonitrile and was pumped at a flow rate of 1 ml min⁻¹. The column temperature was maintained at 30°C and the UV detection wavelength was 370 nm. Calibration standards over the range of 0.1–50 µg ml⁻¹ and quality control samples (0.25, 2.5 and 20 µg ml⁻¹) were assayed along with each set of samples. The interday and intraday precision CV % ranged from 4 to 6% and 3 to 7%, respectively, and precision for the lowest calibration standard was 10%. Accuracy for the method was 99.6–106% (interday variability), 102–109% (intraday variability) and 99.3% for the lowest calibration standard.

Pharmacokinetic evaluation

Nitrofurantoin plasma pharmacokinetic parameters were estimated by noncompartmental methods with WinNonlin Version 4.1 (Pharsight, Mountain View, CA USA) using actual pharmacokinetic sampling times. The maximum observed drug concentration (C_max) and the first time of its occurrence (T_max) were taken directly from the concentration–time profile. The area under the concentration–time profile from zero time to infinity (AUC_(0−∞)) was calculated using the linear up/logarithmic down trapezoidal method and extrapolation to infinite time by the addition of C_last/λ_z, where λ_z is the apparent terminal phase elimination rate constant estimated by linear regression of the logarithmically transformed concentration data. A minimum of three terminal phase concentration values were used to estimate λ_z.

The total urinary recovery of nitrofurantoin (Ae_∞) was calculated as the sum of the amount of nitrofurantoin excreted in each urine collection interval. The percentage of the dose that was excreted as unchanged nitrofurantoin in urine (f_e) was calculated as f_e = Ae_∞/Dose × 100 and the renal clearance (CL_R) was calculated as CL_R = Ae_∞/AUC_(0−∞).

Statistical analyses

Statistical analyses were performed with SAS (version 8.2; SAS Inc., Cary, NC USA). The geometric mean and 95% confidence intervals (CIs) were calculated by genotypic cohort for each pharmacokinetic parameter, except T_max, for which median and range were calculated. One-way analysis of variance (anova) was used to compare log-transformed pharmacokinetic parameters among genotypic cohorts (CC, CA, AA). The ratio of geometric least-squares (GLS) means and associated 90% CIs were calculated for the comparison of cohort CA vs. cohort CC and cohort AA vs. cohort CC for each pharmacokinetic parameter, except T_max.

Results

Subject demographic characteristics

All 36 subjects completed the study and had evaluable pharmacokinetic parameters. The subject demographic characteristics are presented by ABCG2 421 genotypic cohort in Table 1. Similar age and body weight ranges were noted among cohorts. The single dose of nitrofurantoin was well tolerated and no adverse events were reported.

Table 1.

Demographic characteristics of study subjects

	ABCG2 421 genotype
	CC	CA	AA
N	12	12	12
Body weight (kg)	71.8 ± 7.8	68.9 ± 8.0	70.3 ± 6.6
	(58.9–86.7)	(57.5–81.9)	(56.5–79.3)
Age (years)	28.8 ± 7.1	36.3 ± 8.5	30.2 ± 6.6
	(23–43)	(24–45)	(21–42)

N, Sample size in each genotypic group. Data are presented as mean ± SD (range).

Nitrofurantoin pharmacokinetics

Following oral administration, nitrofurantoin was readily absorbed in subjects from all cohorts with a median T_max of 2.0–2.3 h. As shown in Figure 1, the mean nitrofurantoin plasma concentration–time profiles for each genotypic cohort are nearly superimposed on one another. The nitrofurantoin plasma and urine pharmacokinetic parameters are summarized by genotypic cohort in Table 2, along with the statistical comparisons for the ABCG2 421 AA and CA cohorts relative to the CC (wild-type) cohort. There were no significant differences in nitrofurantoin AUC_(0−∞), C_max, f_e or CL_R among carriers of the variant ABCG2 (AA or CA) and ABCG2 CC subjects.

Figure 1.

Arithmetic mean plasma concentration–time profiles of nitrofurantoin in ABCG2 421 CC, CA, and AA genotypic cohorts. AA, (•); CA, (▿); CC, (▪)

Table 2.

Summary of nitrofurantoin plasma and urine pharmacokinetic parameters and between genotype comparisons

	ABCG2 421 genotype			Ratio of geometric LS means (90% CI)
	CC (n = 12)	CA (n = 12)	AA (n = 12)	CA vs. CC	AA vs. CC
AUC(0−∞) (µg h ml−1)	2.21	2.42	2.32	1.09	1.05
	(2.00, 2.45)	(2.11, 2.78)	(1.99, 2.70)	(0.95, 1.26)	(0.91, 1.21)
Cmax (µg ml−1)	0.875	0.961	0.963	1.10	1.10
	(0.749, 1.02)	(0.780, 1.18)	(0.742, 1.25)	(0.87, 1.38)	(0.87, 1.39)
T1/2 (h)	0.78	0.76	0.72	0.97	0.93
	(0.59, 1.02)	(0.64, 0.89)	(0.62, 0.84)	(0.78, 1.21)	(0.74, 1.15)
% dose in urine	43.1	44.3	38.8	1.03	0.90
	(35.7, 51.9	(39.7, 49.3)	(33.6, 44.6)	(0.87, 1.21)	(0.76, 1.06)
CLR (l h−1)	19.4	18.3	16.7	0.94	0.86
	(16.2, 23.3)	(16.1, 20.7)	(14.2, 19.7)	(0.79, 1.12)	(0.72, 1.02)

N, Sample size in each genotypic cohort. Pharmacokinetic parameters shown as geometric mean and 95% confidence interval.

Results of additional genotypic analyses

Subjects were pre-screened and recruited into the study based on ABCG2 421 genotype status. Genotyping for additional ABCG2 SNPs was performed retrospectively. None of the subjects was found to be a carrier of the variant C376T, T623C or T1291C SNPs. Eleven subjects were found to be heterozygous carriers of the G34A SNP (seven in the 421 CA cohort and four in the 421 CC cohort). Three subjects were found to be homozygous carriers of the G34A SNP (one in the CA cohort and two in the CC cohort). A summary of the nitrofurantoin pharmacokinetic parameters by G34A polymorphism status did not show any significant differences among ABCG2 34 GG, GA or AA genotypes. The geometric mean AUC_0−∞ values were 2.30, 2.41 and 2.10 µg h ml⁻¹ and the C_max values were 0.957, 0.871 and 0.986 µg ml⁻¹ for the ABCG2 34 GG, GA and AA genotypes, respectively.

Discussion

A number of drugs in various therapeutic areas have been identified as substrates or modulators of the BCRP transporter [22]. The identification of a BCRP-selective probe drug would provide a useful tool to understand the clinical relevance of BCRP-mediated transport on drug absorption and elimination. One of the most widely studied BCRP substrates is the anticancer drug topotecan, where sevenfold and 2.2-fold higher systemic exposures have been observed in knockout mouse models and cancer patients, respectively, when co-administered with the dual BCRP/P-glycoprotein inhibitor GF120918 [24, 25]. However, topotecan is also a P-glycoprotein substrate, and this lack of specificity for BCRP coupled with its known cytotoxic effects makes topotecan a suboptimal probe substrate for use in healthy volunteer studies.

The antibiotic nitrofurantoin, which has a single dose safety profile acceptable for use in healthy subjects, has recently been shown to be a substrate of murine and human BCRP in vitro, but not P-glycoprotein, MRP1 or MRP2 [17, 19]. In both humans and rodents, oral nitrofurantoin is almost completely absorbed. Elimination occurs primarily by renal and biliary excretion, with some reductive metabolism by enzymes in human tissue and intestinal bacterial flora [26, 27]. The near complete absorption of nitrofurantoin suggests that BCRP transport does not substantially influence intestinal absorption of this permeable compound. However, BCRP is considered to play a role in the hepatobiliary first-pass effect and elimination of nitrofurantoin, consistent with the altered pharmacokinetics and biliary excretion observed in Bcrp knockout mice [17]. The aim of this study was to evaluate nitrofurantoin as a potential probe substrate for BCRP activity in vivo by comparing its pharmacokinetics in Chinese men who were carriers and noncarriers of the ABCG2 C421A variant polymorphism.

The pharmacokinetic parameters of nitrofurantoin in this study were similar to those previously reported for the macrocrystalline formulation [26, 28]. However, the nitrofurantoin plasma and urine pharmacokinetic parameters did not differ between carriers and noncarriers of the ABCG2 C421A polymorphism. The lack of effect of this polymorphism on nitrofurantoin AUC and half-life in humans was unexpected given the in vitro findings that nitrofurantoin is a substrate of human BCRP [17, 19] and the large differences observed in oral exposure and biliary elimination between Bcrp knockout and wild-type mice [17], as well as the effect of a Bcrp inhibitor on nitrofurantoin pharmacokinetics in rats [19].

Although the results were initially surprising, recent literature reports suggest that the contribution of the C421A polymorphism to the clinical pharmacokinetics of BCRP substrates may be substrate dependent. For example, 1.4-fold and fourfold differences have been noted in the pharmacokinetics of oral topotecan and intravenous diflomotecan, respectively, in heterozygous 421CA cancer patients compared with homozygous 421CC patients [11, 12]; no data were available for homozygous 421 AA patients. Similarly, oral rosuvastatin exposure was approximately twofold higher in a combined cohort of 421 AA and CA carriers compared with a cohort with the CC genotype, when controlled for potentially confounding genetic differences in CYP2C9 and OATP1B1 [13]. In contrast, there was no difference in the pharmacokinetics of intravenous irinotecan [29], oral lamivudine [30] and oral pravastatin between carriers and noncarriers of the C421A polymorphism [31].

The reason for the lack of influence of the ABCG2 C421A variant on the pharmacokinetics of nitrofurantoin remains to be elucidated. Potential explanations include the possibility that BCRP contributes less to the overall elimination of nitrofurantoin in humans in vivo than has been anticipated based on in vitro experiments in human BCRP-transfected cell lines and in vivo rodent studies. In humans, the BCRP pathway may represent a minor pathway of nitrofurantoin elimination, so that loss of the pathway in a C421A carrier results in no detectable change in the drug's pharmacokinetics. Alternatively, the existence of other non-BCRP elimination pathways may be able to compensate for the reduced BCRP activity in carriers of the C421A polymorphism. The availability of alternative elimination pathways may also explain the lack of effect of BCRP genotype on lamivudine and pravastatin pharmacokinetics described above, as both drugs are substrates for multiple transporters and/or enzymes [30, 31]. Although in vitro transport studies have shown that nitrofurantoin is not a substrate of human P-glycoprotein, MRP1 or MRP2 [17, 19], this does not preclude the involvement of some yet to be identified transporter(s) or enzyme(s) in the elimination of nitrofurantoin in humans. Of note, it is unlikely that an increase in renal elimination of nitrofurantoin is compensating for any reduced BCRP activity in 421 AA or 421 CA subjects in this study, as renal clearance values were similar across genotypes.

Another explanation for the lack of effect of the C421A polymorphism on nitrofurantoin pharmacokinetics in this study is that sufficient BCRP transport activity may be maintained in subjects of all genotypes. However, this seems unlikely given the number of reports that the C421A polymorphism influences BCRP expression or activity, including protein levels in human placental tissues samples [6] and reduced protein expression and/or transport activity in various transfected cell lines in vitro[5–8, 10]. Nevertheless, some studies have not observed a correlation between ABCG2 genotype and BCRP protein or mRNA expression in human tissue, including the intestine or heart [32, 33]. Further studies in this field will be of interest.

The marked difference in the apparent contribution of BCRP to nitrofurantoin pharmacokinetics in rodents and man is noteworthy. Potentially, rodent models may not provide reliable predictions for the contribution of BCRP to drug disposition in humans. Recent reports have suggested that there may be interspecies differences in the tissue expression of transporters, including BCRP [23, 34, 35]. However, most observations to date have been based on mRNA measurements, and it remains to be determined whether mRNA expression correlates with protein expression and activity.

In conclusion, although nitrofurantoin is a human BCRP substrate in vitro and rodent models suggest BCRP makes a significant contribution to the drug's pharmacokinetics in vivo, the results of this study indicate that oral nitrofurantoin is not a useful clinical probe for assessing BCRP activity in humans.