Multidrug Resistance–Associated Protein 4 Is Involved in the Urinary Excretion of Hydrochlorothiazide and Furosemide

Basic Science Articles

Maki Hasegawa,* Hiroyuki Kusuhara,* Masashi Adachi,† John D. Schuetz,† Kenji Takeuchi,* and Yuichi Sugiyama* *Graduate School of Pharmaceutical Sciences, University of Tokyo, Tokyo, Japan; and †Department of Pharmaceutical Sciences, St. Jude Children’s Research Hospital, Memphis, Tennessee

The role of ATP-binding cassette transporters in the urinary excretion of diuretics was investigated. Significant ATP- dependent uptake of hydrochlorothiazide (HCT) and furosemide was observed in membrane vesicles that expressed multidrug resistance–associated protein 4 (MRP4) and breast cancer resistance protein (BCRP). Unlike taurocholate uptake, S-methylglutathione had no effect on the ATP-dependent uptake of both compounds by MRP4. The functional importance of MRP4 and BCRP in the urinary excretion of HCT and furosemide was investigated using gene knockout mice. The renal clearance of HCT and furosemide was reduced significantly but not abolished in Mrp4 knockout mice compared with ,(wild-type mice (9.0 ؎ 0.9 versus 15 ؎ 2 ml/min per kg for HCT and 1.9 ؎ 0.3 versus 2.7 ؎ 0.1 ml/min per kg for furosemide ؎ and the amount of HCT that was associated with the kidney specimens was greater in Mrp4 knockout mice (21 ؎ 3 versus 13 1 nmol/g kidney). In contrast, Bcrp makes only a negligible contribution because the urinary excretion was unchanged in Bcrp knockout mice. Our results suggest that Mrp4, together with other unknown transporters, accounts for the luminal efflux of HCT and furosemide from proximal tubular epithelial cells. J Am Soc Nephrol 18: 37–45, 2007. doi: 10.1681/ASN.2005090966

he kidney plays important roles in the elimination of suggested that Oat1 mainly is responsible for the renal uptake endogenous waste and xenobiotics and their metabo- of hydrophilic and small organic anions, whereas Oat3 is re- T lites. Urinary excretion is the major detoxification sponsible for the uptake of more bulky organic anions (8–10). mechanism in the kidney, and this is governed by glomerular Diuretics interact with OAT1 and OAT3, and, particularly, loop filtration, tubular secretion across the proximal tubules, and diuretics such as furosemide and bumetanide are substrates of reabsorption. It is widely accepted that the renal secretion of OAT (11,12). Recently, Oat1 knockout mice were generated, drugs with anionic or cationic moieties across the renal tubules and it was demonstrated that Oat1 was responsible for the renal is achieved by cooperation between the uptake transporter at uptake and pharmacologic action of furosemide (13). However, the basolateral membrane and the efflux transporter at the an apical efflux mechanism for diuretics has not been reported brush border membrane (BBM) (1–4). yet. Loop diuretics and thiazides are actively secreted from blood Previous studies using BBM vesicles have suggested that the to the lumen by organic anion transport systems and exhibit luminal transport of organic anions is composed of two distinct diuretic effects by inhibiting the reabsorption of ions that are systems: (1) Electroneutral exchange of organic anions and (2) ϩ ϩ Ϫ mediated by Na -K -2Cl co-transporter in the loop of Henle voltage-driven facilitated transport (14–16). In addition, immu- ϩ Ϫ and Na -Cl co-transporter in the distal tubule from the lumi- nohistochemical staining has demonstrated the expression of nal side, respectively (5). In fact, co-administration of probene- ATP-binding cassette (ABC) transporters, such as multidrug cid, a nonspecific inhibitor of organic anion transport systems, resistance–associated protein 2 (MRP2/ABCC2) and 4 (MRP4/ inhibits the renal elimination and diuretic action in humans ABCC4), in the BBM of the proximal tubules of human and (6,7). rodent kidney (17,18), whereas breast cancer resistance protein Two organic anion transporters play important roles in the (BCRP/ABCG2) is found only in mouse kidney (19). Because basolateral uptake of organic anions. Organic anion transporter the substrate specificities of these ABC transporters are very 1 (OAT1/SLC22A6) and OAT3 (SLC22A8) have broad substrate broad (20–22), localization of ABC transporters along the BBM specificities, and kinetic analyses using rat kidney slices have of the proximal tubules has led to considerable interest in their role in the urinary excretion of drugs. Because BBM vesicles

Received September 19, 2005. Accepted October 10, 2006. that are prepared from the kidney predominantly are right- side-out oriented (23,24), they cannot be used for the detection Published online ahead of print. Publication date available at www.jasn.org. of ATP-dependent uptake. In vivo pharmacokinetic studies us- Address correspondence to: Dr. Hiroyuki Kusuhara, Department of Molecular ing gene-deficient or knockout animals are necessary to obtain Pharmacokinetics, Graduate School of Pharmaceutical Sciences, the University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo, 113-0033, Japan. Phone: ϩ81-3-5841- direct evidence regarding any involvement of ABC transporters 4774; Fax: ϩ81-3-5841-4766; E-mail: [email protected] in the luminal efflux of drugs.

In this study, we focused on MRP4 and BCRP. Both trans- to play important roles in tubular secretion (18,25,28), although porters accept a variety of organic anions as substrates, such as direct evidence of this still is lacking. For BCRP, an in vivo study p-aminohippurate (PAH), 17␤-estradiol-17␤-d-glucuronide, de- using Bcrp knockout mice has demonstrated that BCRP is re- hydroepiandrosterone sulfate (DHEAS), and methotrexate sponsible for the urinary excretion of some organic anions, such (18,25–27). van Aubel et al. (18) initially found that Mrp4 is as methotrexate and 6-hydroxy-5,7-dimethyl-2-methylamino-4- localized on the BBM, and, subsequently, Mrp4 was suggested (3-pyridylmethyl) benzothiazole (E3040) sulfate (29,30). In this

Figure 1. Expression of ATP-binding cassette (ABC) transporters in membrane vesicles and effect of diuretics on ATP-dependent uptake of typical substrates that are mediated by multidrug resistance–associated protein 4 (MRP4) and breast cancer resistance protein (BCRP). (A and B) Expression of ABC transporters in membrane vesicles was confirmed by Western blotting. For MRP4: Lane 1, MRP4-expressing vesicles; lane 2, green fluorescence protein (GFP)-expressing vesicles prepared from HEK293 cells (1 ␮g protein/lane). For BCRP: Lane 1, BCRP-expressing vesicles; lane 2, GFP-expressing vesicles prepared from HEK293 cells (0.1 ␮g protein/lane). (C) The uptake of typical substrates by membrane vesicles that expressed ABC transporters or GFP (3 ␮g of protein) for 5 min at 37°C was determined in the presence of 5 mM ATP (f) or AMP (Ⅺ). [3H]dehydroepiandrosterone sulfate (DHEAS; ␮ 3 ␮ 0.5 M) and [ H]Estrone-3-sulfate (E1S; 0.5 M) were selected as typical substrates of MRP4 and BCRP, respectively. (D) The effect of diuretics on the ATP-dependent uptake of the typical substrates by membrane vesicles that expressed ABC transporters was examined at the designated concentration. The bars represent MRP4-mediated uptake of [3H]DHEAS (f) and BCRP-mediated 3 Ⅺ uptake of [ H]E1S( ). Data were calculated by subtracting the ATP-dependent ligand uptake by the GFP-expressing membrane vesicles from that by the ABC transporter–expressing membrane vesicles. The unit of the inhibitor concentrations was ␮M. Each bar represents the mean Ϯ SE (n ϭ 3). *P Ͻ 0.05, **P Ͻ 0.01 versus control values; †P Ͻ 0.05, ††P Ͻ 0.01 versus the uptake in the presence of AMP. J Am Soc Nephrol 18: 37–45, 2007 Involvement of Mrp4 in Urinary Excretion of Diuretics 39 study, the functional involvement of Mrp4 and Bcrp in the (Netherlands Cancer Institute, Amsterdam, Netherlands), respectively. urinary excretion of hydrochlorothiazide (HCT) and furo- They are fertile and do not exhibit any physiologic abnormalities semide was examined by comparing their in vivo pharmacoki- (19,31). All experiments were carried out according to the guidelines netics using their corresponding gene knockout mice. provided by the Institutional Animal Care Committee (Graduate School of Pharmaceutical Sciences, University of Tokyo). Mice were housed and handled according to the Guide for Care and Use of Materials and Methods Laboratory Animals as adopted and promulgated by the National 3 3 [ H]Estrone-3-sulfate (E1S; 57.3 Ci/mmol) and [ H]DHEAS (60.0 Institutes of Health. Ci/mmol) were purchased from PerkinElmer Life Science, (Boston, MA). Benzthiazide was purchased from Wako Pure Chemical Indus- tries (Osaka, Japan). Furosemide, HCT, and other diuretics were pur- Membrane Vesicle Study chased from Sigma (St. Louis, MO). All other chemicals used were Membrane vesicles were prepared from recombinant adenovirus that commercially available and of analytical grade. contained human MRP4-infected (10 multiplicity of infection [MOI]) and BCRP-infected (2 MOI) HEK293 cells that were derived from Animals human embryonic kidney as described previously (26). As negative Female Mrp4 knockout and wild-type C57BL/6J mice (12 to 16 wk of controls, cells were infected with a virus that contained green fluores- age) and Bcrp knockout and wild-type FVB/NJcl mice (14 to 18 wk of cence protein (GFP) cDNA (10 MOI). All of these recombinant adeno- age) were used in this study. Mrp4 knockout mice and Bcrp knockout viruses had been established previously (26). Membrane vesicles were mice were constructed in the laboratory of J.D.S. and Dr. Schinkel isolated by the hypotonic method described previously in detail (26).

Figure 2. Transport of hydrochlorothiazide (HCT) and furosemide by MRP4 and BCRP. The uptake of HCT (50 ␮M; A) and furosemide (5 ␮M; B) by the membrane vesicles that were prepared from HEK293 cells that expressed MRP4, BCRP, and GFP was determined at 37°C for 5 min in medium that contained 5 mM ATP (f) or AMP (Ⅺ). Effect of S-methylglutathione (5 mM) was examined for the uptake of HCT (50 ␮M; C) and furosemide (5 ␮M; D) by membrane vesicles (20 ␮g of protein) at 37°C for 5 min. p, ATP and S-methylglutathione (5 mM); f, ATP alone; 2, AMP and S-methylglutathione (5 mM); Ⅺ, AMP alone. Each bar represents the mean Ϯ SE (n ϭ 3). *P Ͻ 0.05, **P Ͻ 0.01 versus control; ††P Ͻ 0.01 versus the uptake in the presence of AMP. S-methyl-GSH, S-methylglutathione. 40 Journal of the American Society of Nephrology J Am Soc Nephrol 18: 37–45, 2007

They finally were frozen in liquid nitrogen and stored at Ϫ80°C until Table 1. Renal function data in Mrp4 knockout and required. wild-type micea Transport studies were performed using a rapid filtration technique (26). For the inhibition study, the amount of protein of each membrane Parameter Wild-Type Mrp4 (Ϫ/Ϫ) vesicle that was used for each point was 3 ␮g. A 0.45-␮m nitrocellulose Ϯ Ϯ membrane filter (Millipore Corp., Bedford, MA) was used for filtration Body weight (g) 25 1250 of the stopped reaction mixture. Radioactivity that was retained on the Urine volume (␮l/h) 42 Ϯ 645Ϯ 7 filter was determined in a liquid scintillation counter (LS 6000SE; GFR (ml/min per kg) 12 Ϯ 010Ϯ 1 Beckman Instruments, Fullerton, CA). For the uptake of HCT and a Ϯ ϭ furosemide, 20 ␮g of protein was used for each point. Substrates that Data are means SEM (n 3). Mrp4, multidrug resistance–associated protein 4. were retained on the membrane filter (JH filter; Millipore) were dissolved in 1 ml of methanol that contained internal standard (benzthiazide) by sonication for 15 min. After centrifugation, supernatants were concentrated by centrifugal concentrator (CC-105; TOMY, Tokyo, Ja- 25% linear gradient of 10 mM ammonium acetate (pH 3.5 adjusted by pan) and dissolved in 75 ␮l of 10 mM ammonium acetate (pH 3.5 HCOOH) over 5 min at 0.4 ml/min. The eluate was introduced into the adjusted by HCOOH) plus acetonitrile at the ratio of 95 to 5. Then, 50-␮l MS via an electrospray interface. Detection was performed by selected aliquots were used for liquid chromatography/mass spectrometry ionization monitoring in negative ion mode (m/z 285, 296, and 430 for (LC/MS) quantification as described below. The uptake activity was furosemide, HCT, and benzthiazide, respectively). defined as the amount of ligand that was associated with the membrane vesicles divided by the ligand concentration in the transport buffer. Pharmacokinetic Analysis The renal clearance with respect to the plasma concentration

Western Blotting (CLrenal, p) of furosemide in Mrp4 knockout mice was calculated using Western blotting was performed as described previously (8). ECL the following equation because the plasma concentration did not reach advance Western blotting Detection reagent (Amersham Pharmacia steady-state: Biotech, Buckinghamshire, UK) was used for detection. Secondary an- ϭ CLrenal, p Xurine (30 to 90 min)/AUC (30 to 90 min) tibody was labeled horseradish peroxidase (HRP), and HRP that was present on a Western blot was detected using a luminescence analyzer where Xurine (30 to 90 min) and AUC (30 to 90 min) represent the cumu- (LAS-3000 mini; Fuji Film Corp., Tokyo, Japan) to evaluate the chemi- lative amount excreted into the urine from 30 to 90 min (nmol/kg) and the luminescence generated by the reaction of HRP with its substrate. The dilution of the primary antibodies were as follows: Polyclonal anti- MRP4 rabbit serum (Kamiya Biomedical Co., Seattle, WA), 1:1000; and monoclonal anti-BCRP (BXP-21; Kamiya Biomedical Co.), 1:200.

In Vivo Pharmacokinetic Analysis After anesthesia with intraperitoneal sodium pentobarbital (50 mg/kg), the bladder was catheterized for urine collection. After an intravenous bolus injection (furosemide 800 nmol/kg; HCT 600 nmol/kg), furosemide (20 nmol/min per kg) and HCT (83 nmol/min per kg) were infused via the jugular vein. Blood samples were collected via the jugular vein at 30, 60, and 90 min after administration and centrifuged. Urine was collected at 0 to 30, 30 to 60, and 60 to 90 min. At the end of the experiment, kidneys were removed. For determina- tion of the GFR, [3H]inulin (0.4 mg; 0.9 ␮Ci/min per kg) was infused via the jugular vein. Blood and urine samples were collected in the same way as for HCT and furosemide.

LC/MS Analysis The quantification of HCT and furosemide was performed using and HPLC (HPLC, Alliance 2690; Waters, Milford, MA) connected to a mass spectrometer (ZQ; Micromass, Manchester, UK). Ten microliters of plasma, 2 ␮l of urine, and 20 ␮l of 33% (wt/vol) kidney homogenate was precipitated with 50 ␮l (for plasma and urine sample) or 100 ␮l (for kidney samples) of methanol that contained an internal standard (ben- Figure 3. Time profiles of the plasma concentration and urinary zthiazide), mixed, and centrifuged, and then 45 ␮l (for plasma and excretion of HCT and furosemide in Mrp4 knockout and wild- urine) or 90 ␮l (for kidney) of the supernatants was concentrated by type mice. The plasma concentration (A and C) and urinary centrifugal concentrator (CC-105; TOMY, Tokyo, Japan). Concentrated excretion (B and D) of HCT (A and B) and furosemide (C and samples were dissolved in 75 ␮l of 10 mM ammonium acetate (pH 3.5 D) in Mrp4 knockout (E) and wild-type (F) mice were exam- adjusted by HCOOH) plus acetonitrile at the ratio of 95 to 5, and 50-␮l ined. HCT and furosemide were infused at 83 and 20 nmol/min aliquots were injected into the LC/MS. HPLC analysis was performed per kg after a loading dose of 600 and 800 nmol/kg, respec- on a CAPCELL PAK C18 column (MGII, 3 ␮m, 2.0 mm ID, 50 mm; tively. Each symbol represents the mean Ϯ SE (n ϭ 4). *P Ͻ 0.05 Shiseido, Tokyo, Japan) at 40°C. Elution was performed with a 95 to versus wild-type mice. J Am Soc Nephrol 18: 37–45, 2007 Involvement of Mrp4 in Urinary Excretion of Diuretics 41

Table 2. Pharmacokinetic parameters of HCT in Mrp4 knockout and wild-type micea

HCT Parameter Wild-Type Mrp4 (Ϫ/Ϫ) Ϯ Ϯ CLtotal (ml/min per kg) 35 3322 Ϯ Ϯ b CLrenal, p (ml/min per kg) 15 2 9.0 0.9 Ϯ Ϯ c CLrenal, k (ml/min per kg) 3.8 0.2 1.8 0.2 Ϯ Ϯ b Ckidney (nmol/g kidney) 13 1213 Ϯ Ϯ b Vurine (nmol/min per kg) 52 2365 Ϯ Ϯ b Furine (%) 43 2304 Ϯ Ϯ b Kp 3.9 0.4 5.4 0.5 ϫ Ϯ Ϯ CLrenal, p/(fp GFR) 3.3 0.4 2.3 0.2 aHydrochlorothiazide (HCT) was infused intravenously for 90 min. Each value was determined from the data shown in Ϯ ϭ Figure 3 and the equations described in Materials and Methods. Data are means SEM (n 4). CLtotal, total body clearance;

CLrenal, p, renal clearance with respect to the concentration in circulating plasma; CLrenal, k, renal clearance with respect to the kidney concentration; Ckidney, drug concentration in the kidney at 90 min; Vurine, urinary excretion rate; Furine, fractional urinary excretion; Kp, kidney to plasma concentration ratio; fp, plasma unbound fraction. bP Ͻ 0.05. cP Ͻ 0.01.

area under the plasma concentration curve from 30 to 90 min (nmol ϫ Millipore ultrafiltration membrane (10000 NMWL; Millipore). The param- min per ml), respectively. eter fp was calculated from following equation: Total body clearance of HCT (CLtotal), CLrenal, p, renal clearance with ϭ fp (Cfiltrate/Cplasma)/R respect to the kidney concentration (CLrenal, k), and fractional urinary where Cfiltrate,Cplasma, and R are the filtrate concentration, plasma excretion (Furine) were calculated from the following equations: ϭ concentration, and recovery rate, respectively. Drug recovery using CLtotal I/Cp ϭ PBS solution was determined to correct for membrane adsorption. The CLrenal, p Vurine/Cp ϭ Ϫ ϫ ϫ recovery rate of furosemide and HCT was 0.94 and 0.95, respectively. CLrenal, k (Vurine fp GFR Cp)/Ck F ϭ (V /I) ϫ 100 urine urine Statistical Analyses where I, Cp,Vurine,fp, and Ck represent the infusion rate (nmol/min per kg), plasma concentration at 90 min (␮M), urinary excretion rate from 60 Statistical analysis was performed by t test (in vivo study and uptake to 90 min (nmol/min per kg), and plasma unbound fraction and kidney study using membrane vesicles [Figures 1C and 2]) or ANOVA fol- ␮ lowed by Dunnett test (inhibition study using membrane vesicles [Fig- concentration at 90 min ( M), respectively. Ck is the drug concentrations in the kidney (␮M), estimated as the amount of diuretics in the kidney ure 1D]) to identify significant differences between two sets of data. (nmol/g kidney assuming that1gofkidney ϭ 1 ml). GFR was determined in a separate experiment and calculated from the urinary excretion rate of Results [3H]inulin from 60 to 90 min divided by the plasma concentration of Characterization of Membrane Vesicles and Effects of 3 [ H]inulin at 90 min. The fp was determined by ultrafiltration. Briefly, Diuretics on MRP4- and BCRP-Mediated Transport plasma that was obtained from C57BL/6J mice was incubated with furo- The expression of MRP4 and BCRP proteins in membrane semide and HCT for 30 min at 37°C, followed by ultrafiltration using a vesicles were confirmed by Western blotting (Figure 1, A and

Table 3. Pharmacokinetic parameters of furosemide in Mrp4 knockout and wild-type micea

Furosemide Parameter Wild-Type Mrp4 (Ϫ/Ϫ) Ϯ CLtotal (ml/min per kg) 3.0 0.1 — Ϯ Ϯ b CLrenal, p (ml/min per kg) 2.7 0.1 1.9 0.3 AUC 30 to 90 min (nmol ϫ min/ml) 560 Ϯ 25 610 Ϯ 45 Ϯ Ϯ Xurine 30 to 90 min (mmol/kg) 1.5 0.1 1.2 0.2 ϫ Ϯ Ϯ CLrenal, p/(fp GFR) 42 2356 aFurosemide was infused intravenously for 90 min. Each value was determined from the data shown in Figure 3 and the equations described in Materials and Methods. CLtotal of Mrp4 knockout mice was not determined because the plasma Ϯ ϭ concentration did not reach steady state as shown in Figure 3C. Data are means SEM (n 4). CLtotal, total body clearance; AUC 30 to 90 min, area under the plasma concentration curve from 30 to 90 min; Xurine 30 to 90 min, urinary excretion mass from 30 to 90 min. bP Ͻ 0.05. 42 Journal of the American Society of Nephrology J Am Soc Nephrol 18: 37–45, 2007

B). Transport activity of each set of vesicles was confirmed by Renal Excretion of HCT and Furosemide in Bcrp Knockout measuring the ATP-dependent uptake of typical substrates and Wild-Type Mice 3 3 The plasma concentrations and urinary excretion rates after (MRP4, [ H]DHEAS and BCRP, and [ H]E1S) by membrane vesicles (Figure 1C). The effects of diuretics on ATP-dependent intravenous infusion of HCT and furosemide in Bcrp knockout uptake by MRP4- and BCRP-expressing membrane vesicles mice and wild-type mice are shown in Figure 4, and the phar- were examined (Figure 1D). All of the tested diuretics inhibited macokinetic parameters are summarized in Table 4. Knockout MRP4-mediated transport of [3H]DHEAS, whereas those other of Bcrp did not affect the pharmacokinetic profiles of HCT and 3 furosemide (Figure 4, Table 4). than HCT inhibited BCRP-mediated transport of [ H]E1S. HCT rather enhanced BCRP-mediated transport. Furosemide is a ␮ potent inhibitor of MRP4 (IC50 10 M), whereas bumetanide Discussion ␮ In this study, we have demonstrated that HCT and furo- and ethacrynic acid are moderate inhibitors (IC50 10 to 100 M), Ͼ ␮ semide are substrates of MRP4 and BCRP using membrane and the others are only weak inhibitors (IC50 500 M). Furo- semide, ethacrynic acid, and trichlormethiazide inhibited BCRP vesicles, and in vivo studies using gene knockout mice were with a similar potency as that of MRP4. Bumetanide is a weaker carried out to examine whether Mrp4 and Bcrp play significant ␮ roles in the urinary excretion of furosemide and HCT. In the inhibitor of BCRP than MRP4 (IC50 100 to 1000 M), whereas acetazolamide is a more potent inhibitor of BCRP than MRP4. transport studies using membrane vesicles, we found that diuretics, such as loop diuretics, thiazides, and carbonic anhy- drase inhibitors, are inhibitors of MRP4 and BCRP, whereas ATP-Dependent Transport of HCT and Furosemide by HCT stimulated BCRP-mediated transport (Figure 1D). Furo- MRP4- and BCRP-Expressing Membrane Vesicles semide, ethacrynic acid, and trichloromethiazide had a similar The uptake of HCT and furosemide by MRP4- and BCRP- inhibitory effect on MRP4 and BCRP. In particular, loop diuret- expressing membrane vesicles was determined (Figure 2, A and ics were potent inhibitors of MRP4 and BCRP. In further anal- B). ATP-dependent uptake of HCT was observed only in ysis, it was found that both HCT and furosemide were sub- MRP4- and BCRP-expressing membrane vesicles but not in strates of MRP4 and BCRP, and furosemide exhibited markedly GFP-expressing vesicles (Figure 2A). Although the uptake of greater transport activities than HCT (Figure 2, A and B). The furosemide also was stimulated in the presence of ATP both in transport activities of HCT and furosemide were similar for MRP4- and BCRP-expressing and GFP-expressing vesicles, the MRP4 and BCRP. Rius et al. (32) demonstrated that the ATP- level and the degree of stimulation were greater in MRP4- and BCRP-expressing membrane vesicles than that in GFP-expressing vesicles (Figure 2B). Glutathione has been reported to stim- ulate the ATP-dependent uptake of taurocholate by MRP4 (32). However, the stimulatory effect was not observed for HCT and furosemide (Figure 2, C and D).

Renal Excretion of HCT and Furosemide in Mrp4 Knockout and Wild-Type Mice As summarized in Table 1, the physiologic parameters, such as body weight, urine volume, and GFR, were comparable in Mrp4 knockout and wild-type mice. HCT was given by intravenous infusion. Although the plasma concentration was similar in Mrp4 knockout and wild-type mice (Figure 3A), the urinary excretion rate was significantly reduced at 90 min in Mrp4 knockout mice compared with wild-type mice (Figure 3B). The HCT concentration in the kidney was greater in Mrp4 knockout mice than that in wild-type mice. The pharmacoki-

netic parameters are summarized in Table 2. CLrenal, p and

CLrenal, k were reduced significantly in Mrp4 knockout mice

(Table 2), although CLtotal was similar in the two strains. The fp of HCT in BL6J mice was 0.38 Ϯ 0.03. Furosemide also was given by intravenous infusion. The Figure 4. Time profiles of the plasma concentration and urinary concentrations of furosemide in plasma and urine were deter- excretion of HCT and furosemide in Bcrp knockout and wild- mined in Mrp4 knockout and wild-type mice. The urinary type mice. The plasma concentration (A and C) and urinary excretion rate was lower in Mrp4 knockout mice than in wild- excretion (B and D) of HCT (A and B) and furosemide (C and type mice (Figure 3D). The pharmacokinetic parameters are D) in Bcrp knockout (E) and wild-type (F) mice. HCT and summarized in Table 3. The CLrenal, p was reduced significantly furosemide were infused at 83 and 20 nmol/min per kg after a in Mrp4 knockout mice compared with wild-type mice (Table loading dose of 600 and 800 nmol/kg, respectively. Each sym- Ϯ Ϯ ϭ 3). The fp of furosemide in C57BL/6J mice was 0.0054 0.0011. bol represents the mean SE (n 4). J Am Soc Nephrol 18: 37–45, 2007 Involvement of Mrp4 in Urinary Excretion of Diuretics 43

Table 4. Pharmacokinetic parameters of HCT and furosemide in Bcrp knockout and wild-type micea

HCT Furosemide Parameter Wild-Type Bcrp (Ϫ/Ϫ) Wild-Type Bcrp (Ϫ/Ϫ) Ϯ Ϯ Ϯ Ϯ CLtotal (ml/min per kg) 31 3363 3.4 0.5 2.9 0.4 Ϯ Ϯ Ϯ Ϯ CLrenal, p (ml/min per kg) 13 2162 3.7 0.5 2.7 0.6 Ϯ Ϯ Ϯ Ϯ CLrenal, k (ml/min per kg) 3.6 0.3 3.9 0.3 4.8 0.4 3.6 0.4 Ϯ Ϯ Ϯ Ϯ Ckidney (nmol/g kidney) 16 3142 7.4 1.4 7.9 1.1 Ϯ Ϯ Ϯ Ϯ Vurine (nmol/min per kg) 52 2525322274 Ϯ Ϯ Ϯ Ϯ Furine (%) 43 2444 106 88914 Ϯ Ϯ Ϯ Ϯ Kp 3.9 0.4 4.1 0.4 0.78 0.06 0.75 0.08 ϫ Ϯ Ϯ Ϯ Ϯ CLrenal, p/(fp GFR) 3.0 0.4 4.2 0.6 56 85011 aHCT and furosemide were infused intravenously for 90 min. Each value was determined from the data shown in Figure 4 and the equations described in Materials and Methods. Each value represents the mean Ϯ SEM (n ϭ 4). Bcrp, breast cancer resistance protein.

dependent uptake of taurocholate by MRP4 is stimulated by co-transport of reduced glutathione or its S-methyl derivative. However, this is not the case with HCT and furosemide (Figure 2, C and D). In vitro transport studies suggest that MRP4 and BCRP are candidate transporters involved in the luminal efflux of furosemide and/or HCT in the kidney. This possibility was investigated by comparing pharmacokinetic profiles in knockout mice and wild-type mice. The renal clearance of HCT and furosemide was determined in wild-type and Mrp4 knockout mice. Taking their plasma unbound fractions and GFR into consideration, the renal clearance of HCT and furosemide is accounted for mainly by tubular secretion. Although the total clearance exhibited no change, the . renal clearance with respect to the plasma concentration of Figure 5. Schematic diagram illustrating the tubular secretion HCT was lower in Mrp4 knockout mice than that in wild-type mechanisms of HCT and furosemide. After furosemide and HCT are taken up by organic anion transporter 1 (OAT1) mice, approximately 60% of the control value (Table 2). Because and/or OAT3 into the renal tubular cells, MRP4 and unknown the K value was increased 1.7-fold in Mrp4 knockout mice p transporters are responsible for the luminal efflux. BCRP makes (Table 2), the intrinsic clearance of the luminal efflux was only a minor contribution. It is likely that luminal efflux clear- reduced to 45% of the control value in Mrp4 knockout mice. ance of HCT and furosemide is greater than the efflux clearance The discrepancy between the total body clearance and renal at the basolateral side. Under this condition, the uptake process clearance can be explained by introduction of the concept of a at the basolateral side becomes a rate-limiting step for the rate-limiting process. In the case in which the luminal efflux elimination of drugs from the plasma, and reduction of luminal clearance is significantly greater than the efflux clearance of the efflux as a result of knockout of Mrp4 hardly affects the plasma basolateral side, the uptake process becomes a rate-limiting one elimination clearance. NaDC, sodium-dicarboxylate co-trans- for the elimination of drugs from the plasma, and a reduction in porter. the luminal efflux clearance hardly affects the plasma concentration-time profile (illustrated in Figure 5). The increased kidney concentration that is produced by knockout of Mrp4 indi- part of the tubular secretion of HCT and furosemide at the BBM cates that the efflux that is mediated by Mrp4 makes a side in concert with OAT (Figure 5). significant contribution to the total efflux from inside the cells The renal clearance of HCT and furosemide in Mrp4 knock- ϫ and supports this speculation. For furosemide, the renal clear- out mice still was greater than their GFR (fp GFR; Tables 2 ance with respect to the plasma concentration was reduced in and 3), suggesting the involvement of other efflux transporters. Mrp4 knockout mice to approximately 70% of that in wild-type In contrast to Mrp4 knockout mice, there was no significant mice (Table 3). Because the plasma concentration did not reach difference in the urinary excretion of HCT and furosemide steady state, even after a 90-min infusion, the intrinsic param- between Bcrp knockout and wild-type mice (Table 4). This eter governing the luminal efflux could not be obtained. Pro- excludes the involvement of Bcrp, although the transport ac- longation of the plasma half-life in Mrp4 knockout mice as a tivities of HCT and furosemide are similar for MRP4 and BCRP result of reduced total body clearance may affect the duration in membrane vesicles (Figure 2, A and B). time in the body. These results suggest that Mrp4 accounts for Reduced MRP4 activity may be associated with a prolonga- 44 Journal of the American Society of Nephrology J Am Soc Nephrol 18: 37–45, 2007 tion of the plasma elimination half-life and may increase con- 6. Honari J, Blair AD, Cutler RE: Effects of probenecid on comitantly the kidney concentration more than expected from furosemide kinetics and natriuresis in man. Clin Pharmacol the increased plasma concentration. Generally, accumulation of Ther 22: 395–401, 1977 drugs in the kidney can be associated with nephrotoxicity, and, 7. Vree TB, van den Biggelaar-Martea M, Verwey-van Wissen particularly, the plasma concentration cannot be a predictor of CP: Probenecid inhibits the renal clearance of frusemide and its acyl glucuronide. Br J Clin Pharmacol 39: 692–695, adverse effects when the uptake process is rate limiting. MRP4 1995 confers resistance to adefovir and cidofovir (antiviral drugs) 8. Hasegawa M, Kusuhara H, Sugiyama D, Ito K, Ueda S, (33). These drugs are actively secreted into the urine, and one of Endou H, Sugiyama Y: Functional involvement of rat or- their adverse effects is severe nephrotoxicity (34). Because the ganic anion transporter 3 (rOat3; Slc22a8) in the renal expression of MRP4 overcomes the cytotoxicity of adefovir in uptake of organic anions. J Pharmacol Exp Ther 300: 746– HEK293 cells (35), it is possible that MRP4 activity is related to 753, 2002 their renal toxicity through the regulation of their tissue con- 9. Hasegawa M, Kusuhara H, Endou H, Sugiyama Y: Contri- centration. The factors that potentially cause a functional bution of organic anion transporters to the renal uptake of change in MRP4 include pathophysiologic conditions, aging, anionic compounds and nucleoside derivatives in rat. and genetic polymorphisms or mutations. Of these factors, J Pharmacol Exp Ther 305: 1087–1097, 2003 Denk et al. (36) found that the Mrp4 protein in the kidneys fell 10. Deguchi T, Kusuhara H, Takadate A, Endou H, Otagiri M, markedly 14 d after common bile duct ligation. Cholestasis may Sugiyama Y: Characterization of uremic toxin transport by organic anion transporters in the kidney. Kidney Int 65: increase the risk for drug-induced nephrotoxicity. 162–174, 2004 11. Uwai Y, Saito H, Hashimoto Y, Inui KI: Interaction and Conclusion transport of thiazide diuretics, loop diuretics, and acetazol- We have demonstrated that Mrp4 accounts, at least in part, amide via rat renal organic anion transporter rOAT1. for the urinary excretion of HCT and furosemide. Our findings J Pharmacol Exp Ther 295: 261–265, 2000 illustrate, for the first time, the role of MRP4 in the urinary 12. Hasannejad H, Takeda M, Taki K, Shin HJ, Babu E, Jutabha excretion of drugs in concert with basolateral organic anion P, Khamdang S, Aleboyeh M, Onozato ML, Tojo A, Eno- transporters. This finding will contribute to a better under- moto A, Anzai N, Narikawa S, Huang XL, Niwa T, Endou standing of the urinary excretion mechanisms of drugs and the H: Interactions of human organic anion transporters with factors that are associated with their nephrotoxicity. diuretics. J Pharmacol Exp Ther 308: 1021–1029, 2004 13. Eraly SA, Vallon V, Vaughn DA, Gangoiti JA, Richter K, Nagle M, Monte JC, Rieg T, Truong DM, Long JM, Barshop Acknowledgments BA, Kaler G, Nigam SK: Decreased renal organic anion This work was performed in the context of the Advanced and Inno- secretion and plasma accumulation of endogenous organic vational Research Program in Life Sciences from the Ministry of Edu- anions in OAT1 knock-out mice. J Biol Chem 281: 5072– cation, Culture, Sports, Science and Technology, Japan; by National 5083, 2006 Institutes of Health research grants GM60904, ES058571, and CA23099 14. Kahn AM, Branham S, Weinman EJ: Mechanism of urate (to J.D.S.); by Cancer Center support grant P30CA21745 (to J.D.S.); and and p-aminohippurate transport in rat renal microvillus by the American Lebanese-Syrian Associated Charities (to J.D.S.). membrane vesicles. Am J Physiol 245: F151–F158, 1983 We thank Yoshitane Nozaki and Tomoki Imaoka for technical sup- 15. Ohoka K, Takano M, Okano T, Maeda S, Inui K, Hori R: port. p-Aminohippurate transport in rat renal brush-border membranes: A potential-sensitive transport system and an Disclosures anion exchanger. Biol Pharm Bull 16: 395–401, 1993 None. 16. Roch-Ramel F, Guisan B, Schild L: Indirect coupling of urate and p-aminohippurate transport to sodium in human brush-border membrane vesicles. Am J Physiol 270: F61– References F68, 1996 1. Lee W, Kim RB: Transporters and renal drug elimination. 17. Schaub TP, Kartenbeck J, Konig J, Spring H, Dorsam J, Annu Rev Pharmacol Toxicol 44: 137–166, 2004 Staehler G, Storkel S, Thon WF, Keppler D: Expression of 2. Wright SH, Dantzler WH: Molecular and cellular physiol- the MRP2 gene-encoded conjugate export pump in human ogy of renal organic cation and anion transport. Physiol Rev kidney proximal tubules and in renal cell carcinoma. JAm 84: 987–1049, 2004 Soc Nephrol 10: 1159–1169, 1999 3. van de Water FM, Masereeuw R, Russel FG: Function and 18. van Aubel RA, Smeets PH, Peters JG, Bindels RJ, Russel regulation of multidrug resistance proteins (MRPs) in the FG: The MRP4/ABCC4 gene encodes a novel apical or- renal elimination of organic anions. Drug Metab Rev 37: ganic anion transporter in human kidney proximal tu- 443–471, 2005 bules: Putative efflux pump for urinary cAMP and cGMP. 4. Sekine T, Miyazaki H, Endou H: Molecular physiology of J Am Soc Nephrol 13: 595–603, 2002 renal organic anion transporters. Am J Physiol Renal Physiol 19. Jonker JW, Buitelaar M, Wagenaar E, Van Der Valk MA, 290: F251–F261, 2006 Scheffer GL, Scheper RJ, Plosch T, Kuipers F, Elferink RP, 5. Jackson E: Diuretics. In: Goodman & Gilaman’s the Pharma- Rosing H, Beijnen JH, Schinkel AH: The breast cancer cological of Basis of Therapeutics, 10th Ed., edited by Hard- resistance protein protects against a major chlorophyll- man J, Limberd LE, New York, McGraw Hill, 2001, pp derived dietary phototoxin and protoporphyria. Proc Natl 757–787 Acad Sci U S A 99: 15649–15654, 2002 J Am Soc Nephrol 18: 37–45, 2007 Involvement of Mrp4 in Urinary Excretion of Diuretics 45

20. Fardel O, Jigorel E, Le Vee M, Payen L: Physiological, Niwa T, Jonker JW, Sugiyama Y: Impaired renal excretion pharmacological and clinical features of the multidrug re- of 6-hydroxy-5,7-dimethyl-2-methylamino-4-(3-pyridylm- sistance protein 2. Biomed Pharmacother 59: 104–114, 2005 ethyl) benzothiazole (E3040) sulfate in breast cancer resis- 21. van Herwaarden AE, Schinkel AH: The function of breast tance protein (BCRP1/ABCG2) knockout mice. Drug Metab cancer resistance protein in epithelial barriers, stem cells Dispos 32: 898–901, 2004 and milk secretion of drugs and xenotoxins. Trends Phar- 30. Breedveld P, Zelcer N, Pluim D, Sonmezer O, Tibben MM, macol Sci 27: 10–16, 2006 Beijnen JH, Schinkel AH, van Tellingen O, Borst P, Schel- 22. Ritter CA, Jedlitschky G, Meyer zu Schwabedissen H, lens JH: Mechanism of the pharmacokinetic interaction Grube M, Kock K, Kroemer HK: Cellular export of drugs between methotrexate and benzimidazoles: Potential role and signaling molecules by the ATP-binding cassette trans- for breast cancer resistance protein in clinical drug-drug porters MRP4 (ABCC4) and MRP5 (ABCC5). Drug Metab interactions. Cancer Res 64: 5804–5811, 2004 Rev 37: 253–278, 2005 31. Leggas M, Adachi M, Scheffer GL, Sun D, Wielinga P, Du 23. Haase W, Schafer A, Murer H, Kinne R: Studies on the G, Mercer KE, Zhuang Y, Panetta JC, Johnston B, Scheper orientation of brush-border membrane vesicles. Biochem J RJ, Stewart CF, Schuetz JD: Mrp4 confers resistance to 172: 57–62, 1978 topotecan and protects the brain from chemotherapy. Mol 24. Venien C, Aubry M, Crine P, Le Grimellec C: Determina- Cell Biol 24: 7612–7621, 2004 tion of brush border membrane vesicle orientation using 32. Rius M, Nies AT, Hummel-Eisenbeiss J, Jedlitschky G, monoclonal antibodies recognizing extracytoplasmic and Keppler D: Cotransport of reduced glutathione with bile cytoplasmic domains of neutral endopeptidase-24.11. Anal salts by MRP4 (ABCC4) localized to the basolateral hepa- Biochem 174: 325–330, 1988 tocyte membrane. Hepatology 38: 374–384, 2003 25. Smeets PH, van Aubel RA, Wouterse AC, van den Heuvel 33. Schuetz JD, Connelly MC, Sun D, Paibir SG, Flynn PM, JJ, Russel FG: Contribution of multidrug resistance protein Srinivas RV, Kumar A, Fridland A: MRP4: A previously 2 (MRP2/ABCC2) to the renal excretion of p-aminohippu- unidentified factor in resistance to nucleoside-based anti- rate (PAH) and identification of MRP4 (ABCC4) as a novel viral drugs. Nat Med 5: 1048–1051, 1999 PAH transporter. J Am Soc Nephrol 15: 2828–2835, 2004 34. Ho ES, Lin DC, Mendel DB, Cihlar T: Cytotoxicity of 26. Kondo C, Suzuki H, Itoda M, Ozawa S, Sawada J, Koba- antiviral nucleotides adefovir and cidofovir is induced by yashi D, Ieiri I, Mine K, Ohtsubo K, Sugiyama Y: Func- the expression of human renal organic anion transporter 1. tional analysis of SNPs variants of BCRP/ABCG2. Pharm J Am Soc Nephrol 11: 383–393, 2000 Res 21: 1895–1903, 2004 35. Reid G, Wielinga P, Zelcer N, De Haas M, Van Deemter L, 27. Suzuki M, Suzuki H, Sugimoto Y, Sugiyama Y: ABCG2 Wijnholds J, Balzarini J, Borst P: Characterization of the transports sulfated conjugates of steroids and xenobiotics. transport of nucleoside analog drugs by the human mul- J Biol Chem 278: 22644–22649, 2003 tidrug resistance proteins MRP4 and MRP5. Mol Pharmacol 28. Masereeuw R, Notenboom S, Smeets PH, Wouterse AC, 63: 1094–1103, 2003 Russel FG: Impaired renal secretion of substrates for the 36. Denk GU, Soroka CJ, Takeyama Y, Chen WS, Schuetz JD, multidrug resistance protein 2 in mutant transport-defi- Boyer JL: Multidrug resistance-associated protein 4 is up- cient (TR-) rats. J Am Soc Nephrol 14: 2741–2749, 2003 regulated in liver but down-regulated in kidney in obstruc- 29. Mizuno N, Suzuki M, Kusuhara H, Suzuki H, Takeuchi K, tive cholestasis in the rat. J Hepatol 40: 585–591, 2004