Deficiency in the Organic Cation Transporters 1 and 2 (Oct1/Oct2 [Slc22a1/Slc22a2]) in Mice Abolishes Renal Secretion of Organic Cations

Abstract

The polyspecific organic cation transporters 1 and 2 (Oct1 and -2) transport a broad range of substrates, including drugs, toxins, and endogenous compounds. Their strategic localization in the basolateral membrane of epithelial cells in the liver, intestine (Oct1), and kidney (Oct1 and Oct2) suggests that they play an essential role in removing noxious compounds from the body. We previously showed that in Oct1^−/− mice, the hepatic uptake and intestinal excretion of organic cations are greatly reduced. Since Oct1 and Oct2 have extensively overlapping substrate specificities, they might be functionally redundant. To investigate the pharmacologic and physiologic roles of these proteins, we generated Oct2 single-knockout and Oct1/2 double-knockout mice. Oct2^−/− and Oct1/2^−/− mice are viable and fertile and display no obvious phenotypic abnormalities. Absence of Oct2 in itself had little effect on the pharmacokinetics of tetraethylammonium (TEA), but in Oct1/2^−/− mice, renal secretion of this compound was completely abolished, leaving only glomerular filtration as a TEA clearance mechanism. As a consequence, levels of TEA were substantially increased in the plasma of Oct1/2^−/− mice. This study shows that Oct1 and Oct2 together are essential for renal secretion of (small) organic cations. A deficiency in these proteins may thus result in increased drug sensitivity and toxicity.

The elimination of drugs, xenobiotics, and endogenous compounds is mediated by a variety of transporters primarily expressed in the liver, kidney, and intestine. Among these transporters, polyspecific organic cation transporters 1 and 2 (OCT1 and -2) have been shown in vitro to mediate the electrogenic transport of a broad range of structurally diverse cationic compounds. These include the prototypic organic cation tetraethylammonium (TEA), antidiabetics, neurotoxins, and a variety of endogenous compounds such as choline and monoamine neurotransmitters (1, 2, 11, 34). OCT1 and -2 (SLC22A1 and -2) belong to solute carrier family 22 (SLC22) of organic ion transporters, which consists of 12 members and includes the extraneuronal monoamine transporter (EMT/OCT3/SLC22A3) (8, 36, 37), the carnitine transporter (OCTN2/SLC22A5) (35), the urate anion exchanger (URAT1/SLC22A12) (5), and several polyspecific organic anion transporters. The members of this family are characterized by a predicted 12-transmembrane-domain (TMD) structure and are generally localized in the plasma membrane of epithelial cells (3, 18).

In rodents, Oct1 (Slc22a1) is expressed in the liver, kidney, and small intestine (11, 13, 27), whereas in humans, OCT1 is primarily expressed in the liver (6). Oct2 (Slc22a2) has a substrate specificity similar to that of Oct1 but is predominantly expressed in the kidney in both rodents and humans (10). Immunohistochemical studies with rats have demonstrated that Oct1 is localized at the sinusoidal (basolateral) membrane of hepatocytes in the liver, whereas in the kidney, Oct1 and Oct2 are both localized at the basolateral membrane of epithelial cells of the proximal tubules (15, 21, 31). The broad substrate specificity and strategic localization of Oct1 and Oct2 in the major excretory organs suggest that these proteins are essential in the removal of cationic toxins and waste products from the body via the liver, kidney, and intestine.

Previously, we generated Oct1^−/− mice and showed that absence of Oct1 resulted in greatly reduced hepatic uptake and direct intestinal excretion of substrate organic cations, indicating that Oct1 plays an essential role in the disposition of organic cations by the liver and intestine (13, 34). Despite its high expression in the kidney, the exact role of Oct1 in the renal elimination of organic cations remained unclear. In Oct1 knockout mice, loss of Oct1 from the liver and intestine resulted in increased excretion of drugs via the kidney. Loss of Oct1 from the kidney did not appear to affect the renal elimination of organic cations. This might be due to compensation by redundant transporters such as Oct2 (13).

In this study, we generated Oct2 knockout mice and Oct1/2 double-knockout mice to study the respective and possibly overlapping roles of Oct2 and Oct1 in physiology and pharmacology. We show here that the renal secretion of the model organic cation TEA was completely abolished in Oct1/2^−/− mice, resulting in substantially increased levels in plasma. These findings indicate that Oct1 and Oct2 together play an essential role in the renal secretion of organic cations and that a deficiency in these transporters may result in increased drug sensitivity and/or toxicity.

MATERIALS AND METHODS

Animals.

Mice were housed and handled in accordance with institutional guidelines complying with Dutch legislation. The animals used were Oct1^−/− (13), Oct2^−/−, Oct1/2^−/−, and wild-type mice. All mice were of a comparable mixed genetic background (on average, 50% 129/OLA, 50% FVB), and they were 9 to 14 weeks of age. Animals were kept in a temperature-controlled environment with a 12-h light, 12-h dark cycle. They received a standard diet (AM-II; Hope Farms, Woerden, The Netherlands) and acidified water ad libitum.

Materials.

[¹⁴C]TEA (55 Ci/mol) was from American Radiolabeled Chemicals, Inc. (St. Louis, Mo.); TEA (tetraethylammonium chloride) was from Fluka Chemie AG (Buchs, Switzerland); [³H]MPP⁺ (N-[methyl-³H]-1-methyl-4-phenylpyridinium acetate; 82 Ci/mmol) was from NEN Life Science Products, Inc. (Boston, Mass.); MPP⁺ iodide was from Research Biochemicals International (Natick, Mass.); methoxyflurane (Metofane) was from Medical Developments Australia Pty. Ltd. (Springvale, Victoria, Australia); [¹⁴C]inulin (6.7 Ci/mol) was from Amersham Life Science; anti-rat cytochrome P450 3a1 (monoclonal) was from Oxford Biomedical Research, Inc. (Oxford, Mich.); donkey anti-rabbit immunoglobulin (Ig), F(ab′)₂ fragment, was from Amersham Pharmacia Biotech; goat anti-mouse Ig was from DAKO (Glostrup, Denmark). All other compounds were reagent grade.

Cloning of 129/OLA Oct2 genomic DNA and construction of the targeting vector.

Mouse Oct2 genomic DNA sequences were cloned from a 129/OLA-derived genomic library constructed in bacteriophage λGEM12. A genomic sequence containing exon 1 was identified and cloned into the pGEM5 vector (Promega). From this construct, fragments were subcloned into the pGEM3 vector (Promega), resulting in replacement of a 2.3-kb StuI-Asp718 fragment containing exon 1 with a 1.9-kb pgk-neo cassette in reverse transcriptional orientation (Fig. 1A). Deletion of exon 1 resulted in removal of the start codon and of sequences encoding putative TMD1 and the large extracellular loop located between putative TMD1 and -2.

FIG. 1.

Targeted disruption of the Oct2 gene by homologous recombination. (A) In structures of the wild-type and mutant alleles, exons are indicated by closed boxes (exons are not drawn to scale). In the targeting construct, exon 1 was replaced with an inverted (as indicated with an arrow) pgk-neo cassette. Only relevant restriction sites are indicated. For Southern analysis, 5′ and 3′ probes were used on ScaI (5′)- and HindII (3′)-digested genomic DNA. Sizes of diagnostic restriction fragments for wild-type and targeted alleles are indicated by double-headed arrows (drawn to scale). (B) Schematic representation (drawn to scale) of the Oct1-3 (Slc22a1-3) gene cluster localized on mouse chromosome 17 based on the physical map of the mouse genome (7).

Electroporation and selection for recombinant ES cells.

129/OLA-derived E14 ES cells were cultured as previously described (13). For the generation of Oct1/2 double-knockout mice, ES cells were used in which Oct1 had previously been disrupted (13). For electroporation, 4 × 10⁷ cells were mixed with 100 μg of HindIII-linearized targeting DNA in 600 μl of phosphate-buffered saline. Electroporation was done in a 0.4-cm cuvette with a Bio-Rad gene pulser (model 1652078) at 3 μF and 0.8 kV/0.4 cm. The cells were then seeded on 10-cm-diameter tissue culture dishes without feeder cells. After 1 day, selection was started with 200 μg of G418 sulfate (Gibco) per ml. G418-resistant clones were picked and seeded onto feeder cells.

Southern analysis of ES cells and generation of chimeric mice.

Correct targeting in G418-resistant clones was verified by Southern analysis with 5′ and 3′ Oct2-specific probes. Hybridization of HindII-digested genomic DNA with the 3′ probe resulted in a wild-type band of 9.9 kb and a mutant band of 6.6 kb (probes and fragment sizes are indicated in Fig. 1A). Hybridization of ScaI-digested genomic DNA with a 5′ probe resulted in a wild-type band of 16.1 kb and a mutant band of 5.9 kb. Absence of additional pgk-neo cassettes inserted elsewhere in the genome was confirmed by hybridization with a neo-specific probe (data not shown). Chimeric mice were generated by microinjection of two independently targeted ES cell clones into blastocysts. By this approach, two independent Oct2^−/− and Oct1/2^−/− mouse lines were established.

Clinical-chemical analysis of plasma.

Standard clinical chemistry analyses of plasma were performed on a Hitachi 911 analyzer to determine levels of bilirubin, alkaline phosphatase, aspartate aminotransferase, alanine aminotransferase, lactate dehydrogenase, creatinine, urea, Na⁺, K⁺, Ca²⁺, Cl⁻, phosphate, total protein, and albumin.

Histopathological analysis.

Complete necropsies were performed on adult and aged mice of both sexes. For microscopic examination, tissues were fixed in 4% phosphate-buffered formalin, embedded in paraffin, sectioned at 4 μm, and stained with hematoxylin and eosin in accordance with standard procedures.

RNase protection analysis.

Total RNA was isolated from mouse tissues by use of Trizol reagent (Life Technologies, Inc. [GIBCO BRL], Rockville, Md.), in accordance with the manufacturer's instructions. RNase protection assays were performed as described previously (13, 26), with 10 μg of total RNA per sample. A mouse probe for Oct2 was made by cloning a 1,147-nucleotide (nt) PCR fragment (positions 457 to 1603 relative to the translation start) into the pGEM-T vector. After linearization with EcoRI, a 246-nt antisense RNA probe was generated by transcription with SP6 RNA polymerase, yielding a protected probe fragment of 197 nt. The mouse Gapdh probe was described previously (26).

Northern analysis.

Northern blot assays were performed in accordance with standard procedures. Blots were hybridized with a 617-nt probe for mouse Oct2 (positions 989 to 1605 relative to the translation start). The same blot was rehybridized with an Igf2r probe (cDNA covering exons 3 to 6) to check the amount and integrity of the RNA loaded.

Western analysis.

Crude membrane fractions were prepared as previously described (23). Protein concentrations were determined with the Bradford protein assay (Bio-Rad Laboratories, Munich, Germany). Proteins were subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis and transferred to nitrocellulose (HybondECL; Amersham Pharmacia Biotech). The filters were blocked for 1 h at room temperature with TBS-T (100 mM Tris [pH 7.6], 150 mM NaCl, 0.1% [wt/vol] Tween 20) with 5% skim milk powder. Incubation with the affinity-purified polyclonal antibody (ab1) (21) against rat Oct1 (dilution of 1:5,000) or with an anti-Rat cytochrome P450 3a1 (Cyp3a1) monoclonal antibody (dilution of 1:1,000) was performed at 4°C overnight (in TBS-T containing 5% skim milk powder). Antibodies were detected by incubation of the blot with horseradish peroxidase-conjugated donkey anti-rabbit (for Oct1 detection) or goat anti-mouse (for Cyp3a detection) IgG for 1 h at room temperature in TBS-T containing 5% skim milk powder. Antibody binding was visualized with the ECL Western blotting detection system (Amersham Pharmacia Biotech).

Pharmacokinetic experiments.

For intravenous drug administration, 5 μl of drug solution per g of body weight was injected into the tail vein of mice lightly anesthetized with methoxyflurane. Animals were sacrificed at the indicated time points by axillary bleeding after anesthesia with methoxyflurane. Urine was collected from the bladder, and organs and tissues were removed and homogenized in a 4% (wt/vol) bovine serum albumin solution. Where applicable, intestinal content was separated from intestinal tissue before homogenization. Levels of radioactivity in homogenates were determined by liquid scintillation counting. Continuous-infusion experiments were performed with intraperitoneally implanted micro-osmotic pumps with a pumping rate of 1 μl/h, a capacity of 100 μl, and a duration of 3 days (Alzet 1003D; Alza Corporation, Palo Alto, Calif.). For implantation, mice were first anesthetized with methoxyflurane, a median incision was made in the abdomen, and pumps were intraperitoneally inserted. Subsequently, the musculoperitoneal layer and skin were closed with a silk suture (size 5/0; Perma-hand silk; Ethicon, Norderstedt, Germany). Finally, mice received an intravenous bolus injection (5 μl/g of body weight) of the same drug solution in order to accelerate the point at which the distribution equilibrium was achieved. In all cases, a steady state was achieved within 1 day as determined by repeated sampling of plasma radioactivity. Two days after implantation, mice were sacrificed and steady-state drug levels were determined in plasma, tissues, and urine.

Determination of CL_renal of TEA.

Mice, anesthetized with a combination of ketamine (100 mg/kg) and xylazine (6.7 mg/kg), received intravenous [¹⁴C]TEA (0.2 mg/kg), and blood samples (of about 40 μl) were taken from the tail at 2.5, 5, 10, 20, 30, 40, 50, and 60 min. Urine was collected from the bladder after 60 min. Renal clearance (CL_renal) was determined with the following equation: CL_renal = TEA_{urine (0-60)}/AUC_(0-60). Where TEA_{urine (0-60)} and AUC_(0-60) are the cumulative urinary excretion of TEA up to 60 min and the area under the plasma concentration-time curve from 0 to 60 min as calculated by use of the linear trapezoidal rule, respectively.

Determination of GFR.

The glomerular filtration rate (GFR) was determined by measuring the clearance of inulin (CL_inulin). Mice, anesthetized with a combination of ketamine (100 mg/kg) and xylazine (6.7 mg/kg), received intravenous [¹⁴C]inulin (25 mg/kg), and blood samples (of about 40 μl) were taken from the tail at 5, 10, 20, 30, 40, 50, and 60 min. The GFR was determined with the following equation: GFR = CL_inulin = Dose/AUC_(0-∞). Where AUC_(0-∞) is the area under the plasma concentration-time curve extrapolated to infinity with the MW/Pharm software package (24). For calculation of the estimated GFR (in milliliters per hour), the following equation was used: GFR = 0.036 · BW^{0.74 ± 0.15}, where BW is body weight (12).

Statistical analysis.

All values are given as means ± standard deviations (SD). The two-tailed unpaired Student t test was used to assess the significance of differences between two sets of data. Differences were considered to be statistically significant when P was <0.05.

RESULTS

Generation and analysis of Oct2^−/− and Oct1/2^−/− mice.

The mouse Oct2 gene was disrupted by replacing exon 1 with an inverted pgk-neo cassette via homologous recombination in ES cells (Fig. 1A). By deletion of exon 1 and upstream sequences, the start codon and sequences encoding putative TMD1 and the large extracellular loop located between putative TMD1 and -2, and presumably the transcription start site as well, were removed. Because the Oct1 and Oct2 genes are closely linked within a cluster on mouse chromosome 17 (Fig. 1B), we had to sequentially inactivate both genes in one chromosome in order to generate Oct1/2 double-knockout mice. To do this, we disrupted the Oct2 gene in Oct1^+/− ES cells in which Oct1 had previously been disrupted (13). Correct targeting of the Oct2 allele in ES cell clones was verified by Southern analysis (data not shown). With standard blastocyst injection techniques, mice heterozygous and homozygous for the Oct2 and Oct1/2 disruptions were generated from two independent ES clones each. Absence of Oct2 mRNA in Oct2^−/− and Oct1/2^−/− mice was confirmed by Northern blotting and RNase protection analysis (Fig. 2A and data not shown). Western analysis with a polyclonal antibody (21) recognizing the C-terminal part of rat and mouse Oct1 confirmed that Oct1 protein was undetectable in the livers and kidneys of Oct1/2^−/− mice (Fig. 2B). Oct2^−/− and Oct1/2^−/− mice were fertile, had a normal life span, and were born at the expected Mendelian ratio, indicating that there was no reduction of embryonic viability. Standard plasma clinical chemical analysis and histological analysis with an emphasis on the liver, kidney, and intestine revealed no abnormalities. Absence of Oct2 and of Oct1 and Oct2 together therefore appears to be compatible with normal physiologic functioning of mice.

FIG. 2.

Oct1 and Oct2 RNA and protein analysis. (A) Northern analysis of total RNA from kidneys of wild-type and Oct2^−/− mice. Oct2 and Igf2r bands originate from the same gel, and their sizes are indicated. (B) Immunodetection of Oct1 in the livers and kidneys of wild-type and Oct1/2^−/− mice. A polyclonal antibody raised against rat Oct1 (which cross-reacts with mouse Oct1) was used on crude membrane fractions of liver (20 μg per lane) and kidney (10 μg per lane). The same blot was incubated with a monoclonal antibody raised against rat cytochrome P450 3a (Cyp3a; expressed only in the liver), which was used as a protein loading control. Molecular size markers are indicated in kilodaltons on the right. Part of these data was previously published (13).

Pharmacokinetics of [¹⁴C]TEA in Oct2^−/− and wild-type mice.

To study the pharmacologic role of Oct2, we compared the pharmacokinetics of the prototypic organic cation TEA in Oct2^−/− and wild-type mice. TEA is an excellent transported substrate for OCT1 and OCT2 but not for OCT3 (6, 9, 11). The CL_renal of TEA is substantially greater than the GFR, indicating that this CL_renal occurs mainly via tubular secretion (4, 22). TEA is not significantly bound to plasma proteins, and it is not substantially metabolized in mice, so radioactivity values give a good representation of unchanged TEA levels (4, 29).

Mice received intravenous [¹⁴C]TEA (0.2 mg/kg), and after 20 min, levels of radioactivity were measured in plasma, organs, feces, and urine (Table 1). No significant differences were observed in the distribution and excretion of TEA between Oct2^−/− and wild-type mice, except for the brain, in which levels were reduced with borderline significance in the Oct2^−/− mice (P = 0.046). The excretion of TEA in both the Oct2^−/− and wild-type mice was mainly via CL_renal (nearly 40% of the dose in 20 min), and a substantial amount accumulated in the liver (about 10% of the dose in 20 min), which corresponded well to our previous findings obtained with wild-type mice (13). From these results, it can be concluded that absence of Oct2 in itself has little effect on the pharmacokinetics of TEA, suggesting that its absence can be compensated for by other transporters in the kidney.

TABLE 1.

Levels of radioactivity in female wild-type and Oct2^−/− mice at 20 min after intravenous injection of 0.2 mg of [¹⁴C]TEA per kg^a

Parameter and tissue or fluid	TEA level		Mutant/wild-type ratio
	Wild type	Oct2−/−
Concn in:
Plasma	34.2 ± 3.9	38.3 ± 7.6	1.1
Brain	1.8 ± 0.4	1.3 ± 0.1c	0.7
Spleen	33.1 ± 3.3	32.7 ± 4.5	1.0
Kidney	552 ± 150	548 ± 230	1.0
Liver	482 ± 52	435 ± 69	0.9
Liver (%)	11.8 ± 0.5	10.3 ± 1.8	0.9
% Excretedb in:
Small intestine	5.1 ± 1.3	3.2 ± 1.6	0.6
Cecum	0.2 ± 0.1	0.2 ± 0.06	0.8
Colon	0.2 ± 0.1	0.12 ± 0.08	0.5
Urine	37 ± 22	39 ± 28	1.1

^aResults are expressed as mean [¹⁴C]TEA concentrations (nanogram equivalents per gram or milliliter), or as a percentage of the administered dose excreted, ± SD (n = 3 or 4).

^bExcretion represents total [¹⁴C]TEA found in the contents of the small intestine, cecum, and colon. Urine was collected from the bladder.

^cP < 0.05.

Decreased renal excretion of TEA in Oct1/2^−/− mice versus Oct1^−/− mice.

To further investigate the roles of Oct1 and Oct2 in renal secretion, we compared the pharmacokinetics of TEA in Oct1^−/− and Oct1/2^−/− mice. By direct comparison of Oct1^−/− mice with Oct1/2^−/− mice, the pharmacologic effect of Oct1 on the hepatic accumulation and intestinal secretion was eliminated (13). Oct1^−/− and Oct1/2^−/− mice received [¹⁴C]TEA intravenously at 0.2 mg/kg, and after 60 min, levels of radioactivity were measured in plasma, various organs, and urine (Fig. 3). The concentration of TEA in plasma was increased fourfold in Oct1/2^−/− mice, whereas urinary excretion was significantly decreased compared with that of Oct1^−/− mice (79.5% ± 8.4% of the dose in Oct1^−/− mice versus 56.3% ± 13.3% in Oct1/2^−/− mice; P < 0.05). These results show that when both Oct1 and Oct2 are absent, renal elimination of TEA is impaired and that this results in substantially increased levels of TEA in plasma.

FIG. 3.

Concentrations of TEA in plasma and its renal excretion in Oct1^−/− versus Oct1/2^−/− mice. Levels of radioactivity in Oct1^−/− and Oct1/2^−/− mice at 60 min after intravenous injection of [¹⁴C]TEA (0.2 mg/kg). (A) Levels of [¹⁴C]TEA in plasma. (B) Percentage of dose of [¹⁴C]TEA excreted in urine. Urine was collected from the bladder. *, P < 0.05; **, P < 0.01.

Steady-state pharmacokinetics of TEA and MPP⁺.

To better study the effect of the absence of Oct1 and Oct2 on renal elimination, we determined the pharmacokinetics of TEA under steady-state infusion conditions. This has the advantage that the concentration of a drug in plasma at steady state is determined only by the rate of infusion and the rate of elimination. For constant infusion of TEA, we intraperitoneally implanted micro-osmotic pumps into wild-type, Oct1^−/−, Oct2^−/−, and Oct1/2^−/− mice. After implantation, an intravenous bolus of the same drug solution was given to accelerate the point at which the distribution equilibrium was achieved. A steady state was reached within 1 day and maintained for at least 3 days, as determined by repeated sampling of plasma radioactivity. Two days after implantation, levels of radioactivity were measured in plasma, organs, and urine (Fig. 4). The steady-state concentration of TEA in plasma was about sixfold higher in Oct1/2^−/− mice compared to that in wild-type, Oct1^−/−, and Oct2^−/− mice, indicating that under these conditions, elimination of TEA was substantially impaired only in Oct1/2^−/− mice (Fig. 4A). Despite the sixfold increased levels of TEA in the plasma of Oct1/2^−/− mice, the urinary steady-state TEA concentration was not different from that of wild-type mice. The liver/plasma ratios of TEA in Oct1^−/− and Oct1/2^−/− mice were significantly reduced, confirming that absence of Oct1 results in decreased hepatic accumulation (Fig. 4C and reference 13). Surprisingly, absence of either Oct1 or Oct2 also resulted in an about twofold reduction in the accumulation of TEA in the kidney, whereas at the same time, levels of TEA in the plasma and urine of these mice were not different from those in wild-type mice. When both transporters were absent, this ratio was drastically decreased to about 10% of that in wild-type mice (Fig. 4D). Apparently, whereas both Oct1 and Oct2 contribute to TEA accumulation in the kidney, only when both transporters are absent does this have a pronounced effect on the renal secretion of this compound.

FIG. 4.

Steady-state pharmacokinetics of TEA. Steady-state levels of [¹⁴C]TEA in wild-type, Oct1^−/−, Oct2^−/−, and Oct1/2^−/− mice are shown. [¹⁴C]TEA was continuously infused at a rate of 37 ng/h with intraperitoneally implanted micro-osmotic pumps. (A) Steady-state levels of [¹⁴C]TEA in plasma. (B) Ratios of [¹⁴C]TEA concentrations in urine and plasma. (C) Ratios of [¹⁴C]TEA concentrations in liver and plasma. (D) Ratios of [¹⁴C]TEA concentrations in the kidneys and plasma. Results are means ± SD (n = 4). *, P < 0.05; **, P < 0.01 (compared to wild-type values).

We also investigated the effect of the absence of Oct1 and Oct2 on levels of the neurotoxin MPP⁺ in plasma and its renal secretion by using the same steady-state infusion approach. Previously, it has been shown that MPP⁺ is transported in vitro and in vivo by Oct1 (13, 20) and in vitro by Oct2 (6). However, in contrast to TEA, we did not find significantly different steady-state levels of MPP⁺ in plasma when we compared Oct1/2^−/− and wild-type mice (3.4 ± 0.4 ng/ml in wild-type mice versus 4.7 ± 1.1 ng/ml in Oct1/2^−/− mice at a continuous infusion rate of 37 ng/h). The hepatic and renal tissue-to-plasma ratios of MPP⁺ at steady state were also not significantly different between Oct1/2^−/− and wild-type mice (4.83 ± 1.09 versus 3.45 ± 1.09 in the liver and 4.03 ± 0.82 versus 7.5 ± 2.57 in the kidney, respectively). Possibly, MPP⁺ is not efficiently secreted by Oct1 and Oct2 or, alternatively, there may be other functionally redundant transporters, such as Oct3, which is expressed at low-to-moderate levels in the kidney and which has been shown to transport MPP⁺ (8, 16).

Renal secretion of TEA is abolished in Oct1/2^−/− mice.

The CL_renal of a drug is the result of glomerular filtration and tubular secretion and reabsorption. This means that when there is net tubular secretion of a drug, the CL_renal of that drug is greater than the GFR. As the CL_renal of TEA is primarily mediated via secretion, it is an excellent substrate with which to study this process (4, 22, 30). To establish the contribution of Oct1 and Oct2 to the secretion of TEA, we determined the CL_renal of TEA in Oct1/2^−/− and wild-type mice and compared it with the GFR. First, we determined the GFR in Oct1/2^−/− and wild-type mice by measuring the CL_inulin. [¹⁴C]inulin was administered intravenously at a dose of 25 mg/kg, and plasma concentration profiles were determined (Fig. 5A). The GFR (see Materials and Methods) was not significantly different between Oct1/2^−/− and wild-type mice and corresponded well to the estimated GFR calculated from the mean body weights (12) (Table 2). Next, we determined the CL_renal of TEA after intravenous administration of 0.2 mg of [¹⁴C]TEA per kg. The elimination of TEA from the plasma and the CL_renal of TEA (see Materials and Methods) were substantially decreased in Oct1/2^−/− mice compared to those in wild-type mice (Fig. 5B and C). The ratio of the CL_renal of TEA to the GFR was about 2.4 for wild-type mice, which is comparable to what has been found by others (4, 22, 30). In Oct1/2^−/− mice, the ratio of the CL_renal of TEA to the GFR was reduced to about 1, indicating that the net tubular secretion of TEA was completely abolished in these mice.

FIG. 5.

CL_inulin and clearance of TEA in wild-type versus Oct1/2^−/− mice. (A) Plasma [¹⁴C]inulin concentration-versus-time curves of wild-type and Oct1/2^−/− mice. Levels of radioactivity in plasma were determined at 5, 10, 20, 30, 40, 50, and 60 min after intravenous administration of 25 mg of [¹⁴C]inulin per kg. Results are means ± SD (n = 5). (B) Plasma [¹⁴C]TEA concentration-versus-time curves of wild-type and Oct1/2^−/− mice. Levels of radioactivity in plasma were determined at 2.5, 5, 10, 20, 30, 40, 50, and 60 min after intravenous administration of 0.2 mg of [¹⁴C]TEA per kg. Results are means ± SD (n = 5 to 7). (C) CL_renal of TEA in wild-type and Oct1/2^−/− mice. CL_renal was calculated by dividing the amount of TEA excreted in the urine over 60 min by the plasma AUC_(0-60). The estimated GFR was approximately 21 ml/h for both genotypes and is indicated with a dashed line. **, P < 0.01.

TABLE 2.

Pharmacokinetic parameters for the clearance of [¹⁴C]inulin and [¹⁴C]TEA in male wild-type and Oct1/2^−/− mice

Mice	Inulin clearancea			TEA clearanced
	AUC0-∞ (h · μg/ml)	GFR (ml/h)		AUC0-60 (h · μg/ml)	CLrenal (ml/h)	Estimated GFR (ml/h)	CLrenal/GFR ratio
		Measuredb	Estimatedc
Wild type	25.0 ± 3.2	27.7 ± 2.8	25.0 ± 3.6	51.6 ± 4.6	51.8 ± 5.3	21.2 ± 3.4	2.4
Oct1/2−/−	25.6 ± 3.3	27.7 ± 3.9	25.3 ± 3.6	134 ± 8.2e	19.2 ± 1.0e	20.7 ± 3.4	0.93

^aPharmacokinetic parameters for the clearance of insulin after intravenous administration of 25 mg of [¹⁴C]inulin per kg. Data are shown as mean ± SD (n = 5).

^bGFR measured was calculated from the following formula: dose/AUC_(0-∞).

^cEstimated GFR was calculated from the mean body weights as previously described (12).

^dPharmacokinetic parameters for the clearance of TEA after intravenous administration of 0.2 mg of [¹⁴C]TEA per kg. Data are shown as mean ± SD (n = 5 to 7).

^eP < 0.01.

DISCUSSION

In this study, we examined the in vivo functions of the polyspecific organic cation transporters Oct1 and Oct2 by analyzing Oct1, Oct2, and Oct1/2 knockout mouse models. We show here that Oct1 and Oct2 together are essential for the renal tubular secretion of the prototypic organic cation TEA. Moreover, we found a pronounced mutual redundancy between Oct1 and Oct2 in this function. The decreased renal secretion in Oct1/2^−/− mice resulted in substantially increased levels of TEA in plasma, indicating that these proteins can potentially reduce the systemic and tissue toxicity of noxious compounds and drugs that are Oct1 and Oct2 substrates. This study provides, to our knowledge, the first direct in vivo demonstration that molecularly defined basolateral drug transporters are essential for the renal tubular secretion of a substrate.

Many of the polyspecific drug transporters that have been identified and characterized to date are expressed in the liver and kidney. In the past decade, knockout mouse models for several of these transporters have been generated to elucidate their respective physiologic and pharmacologic functions. Whereas a clear pharmacologic function of these transporters could often be demonstrated in the liver, their exact in vivo role in the renal excretion of compounds remained essentially unclear (13, 14, 26, 28). The complexity of renal drug elimination, which involves passive filtration at the glomerulus, tubular secretion, and reabsorption, makes it intrinsically difficult to study this process in vivo. In contrast, hepatobiliary elimination is a relatively simple process, just involving the uptake of a compound from the bloodstream at the sinusoidal membrane of hepatocytes and its subsequent excretion at the canalicular membrane into the bile.

In our previous study, we showed that mice with a deficiency in Oct1 displayed greatly reduced hepatic uptake and direct intestinal excretion of substrate organic cations, demonstrating that Oct1 plays an essential role in the disposition of organic cations by the liver and intestine (13, 34). On the basis of our present study, it appears that the absence of Oct2 in itself has no immediately apparent physiologic or pharmacologic consequences. This is consistent with the possibility that Oct1 and Oct2 are mutually redundant in the kidney, the primary organ where both transporters are coexpressed. However, although Oct1 and Oct2 have similar substrate specificities, they are not identical (31, 32), and a deficiency in either Oct1 or Oct2 may already be sufficient to impair the renal secretion of some substrates. Moreover, in contrast to rodents, which express both Oct1 and Oct2 in the kidney, humans only express OCT2 (10). Therefore, a deficiency in OCT2 in humans might have effects on renal elimination similar to those of a deficiency in both Oct1 and Oct2 in mice.

Unexpectedly, we found that at steady-state levels in plasma, the relative accumulation of TEA in the kidney was decreased about twofold in both Oct1 and Oct2 knockout mice compared with wild-type mice, whereas at the same time the plasma and urinary concentrations of TEA were not different (Fig. 4). This can be explained when the transcellular transport rate of TEA through proximal tubular cells is primarily limited by the efflux processes across the apical membrane in wild-type, Oct1^−/−, and Oct2^−/− mice. In that case, loss of either Oct1 or Oct2 from the basolateral membrane of proximal tubular cells will reduce the accumulation of TEA in tubular cells but will not necessarily reduce the net transcellular transport rate and secretion. This observation is consistent with the data of Schäli et al. (25) that suggested a rate-limiting role for the apical membrane in the secretion of TEA by renal proximal tubules. That the effect is almost exactly twofold is probably determined by a fortuitous combination of the infusion rate and transport characteristics of Oct1 and Oct2. The similar effects of either Oct1 or Oct2 deficiency indicate that under the conditions applied, the TEA transport capacities of renal tubular Oct1 and Oct2 are virtually identical. In Oct1/2^−/− mice, the relative accumulation of TEA in the kidney was reduced even more, to about 10% of that in wild-type mice, and TEA secretion was (virtually) abolished. Apparently, when both Oct1 and -2 are absent, the transcellular transport of TEA across the proximal tubular cells is primarily limited by passage over the basolateral membrane, resulting in drastically decreased accumulation and net secretion.

The organic cation transporters may also be important for the transport of natural toxins or endogenous compounds. Many studies in vitro have shown that OCT1 and OCT2 can transport physiologically relevant endogenous compounds such as monoamine neurotransmitters (e.g., adrenaline, noradrenaline, dopamine), choline, and guanidine, suggesting that absence of these transporters might have physiologic consequences (1, 9, 10). The fact that our knockout mice are apparently healthy and show no signs of abnormal physiology indicates that these transporters are not absolutely essential and may possibly be compensated for by redundant transporters. It should be noted, however, that many physiologic aberrations only become apparent under specific or extreme conditions that may not occur under the relatively controlled conditions of our animal facility. For example, we have shown previously that mice with a deficiency in another polyspecific drug transporter, the breast cancer resistance protein (Bcrp1/Abcg2), are hypersensitive to a dietary phototoxin that is only sporadically present at toxic levels in the mouse diet (14).

Recently, several groups have reported the occurrence of polymorphisms in the human OCT1 and OCT2 genes, some of which have been shown to result in severely reduced transport activity (17, 19). Genetic deficiencies for these genes may have both positive and negative consequences for drug therapy. In the case of an OCT1 deficiency, reduced uptake of drugs into the liver may result in a decreased efficacy of drugs that have their therapeutic action in the liver. On the other hand, reduced uptake of drugs into the liver could be beneficial for drugs that have adverse effects in the liver or need hepatic metabolic activation. The latter is exemplified by the antidiabetic drug metformin, which has been shown to have reduced toxicity in Oct1^−/− mice (33, 34). Unlike rodents that express both Oct1 and Oct2 in the kidney, humans express only OCT2 (10). Therefore, it is likely that the Oct1/2^−/− mouse model better reflects the effect on renal function of an OCT2 deficiency in humans than the Oct2^−/− mouse model. On the basis of our findings, we expect that humans with a deficiency in OCT2 will have impaired renal elimination of some drugs and that this may result in increased exposure to these drugs. It will therefore be of interest to determine whether polymorphisms in the human OCT1 and OCT2 genes also correlate with altered drug pharmacokinetics in patients. If our findings can indeed be extrapolated to humans, these knockout mouse models will provide powerful tools for predicting and explaining drug sensitivity and toxicity, which may ultimately result in improved drug therapy.