Correlation between Epithelial Cell Permeability of Cephalexin and Expression of Intestinal Oligopeptide Transporter

Experimental Procedures

Materials

Cephalexin and other chemicals were purchased from Sigma Chemical (St. Louis, MO) with either analytical or HPLC grade. [³H]Glycyl-sarcosine (Gly-Sar) (400 mCi/mmol) was purchased from Moravek Biochemicals (Brea, CA). [¹⁴C]PEG 4000 (61.2 mCi/mmol) was provided by Amersham Pharmacia Biotech (Arlington Heights, IL). Cell culture reagents were purchased from Invitrogen (Carlsbad, CA) and culture supplies were from Corning (Corning, NY) and Falcon (Lincoln Park, NJ).

The rabbit polyclonal anti-hPEPT1 serum was kindly provided by Dr. Wolfgang Sadée (University of California, San Francisco, San Francisco, CA). The mouse monoclonal antibody raised against chicken villin also cross-reacts with human villin, and was purchased from Chemicon International (Temecula, CA). All the equipment and chemicals used in Western blot analysis were obtained from Bio-Rad (Hercules, CA) unless otherwise specified.

Animals

Male Sprague-Dawley rats weighing 250 to 350 g were used in this study. Rats 1 to 3, 4 to 7, 8 to 11, and 12 to 15 were from four different batches of rats purchased from Charles River (Raleigh, NC), respectively (Table1). Animals were treated humanely, and the animal experiments were performed in accordance with the Guide for the Care and Use of Laboratory Animals as adopted and promulgated by the National Institutes of Health.

Cells

Caco-2 Cells obtained from American Type Culture Collection (Manassas, VA) (ATCC HTB37, passage no. 44–52) were used in this study. The cells were routinely maintained in Dulbecco's modified Eagle's medium containing 10% fetal bovine serum, 1% nonessential amino acids, 1 mM sodium pyruvate, 1% l-glutamine, and penicillin/streptomycin (100 U/100 μg/ml). The cells grown in 100-mm tissue culture Petri dishes were passaged every 5 days at a split ratio of 1:5, and all cells were maintained in an atmosphere of 5% CO₂ and 90% relative humidity at 37°C.

Preparation of Caco-2/hPEPT1 Cells

Caco-2/hPEPT1 cells were prepared as described by Hsu et al. (1998). The overexpression of hPEPT1, cell viability, and transfection efficiency were characterized (Hsu et al., 1998). Briefly, Caco-2 cells were seeded in six-well culture plates (9.5-cm²growth area) at the density of 10⁴cells/cm². Three days after seeding, the medium was removed and cells were infected with 0.25 ml of recombinant replication-deficient adenovirus solution carrying hPEPT1 gene (Ad.RSV.hPEPT1). The multiplicity of infection (m.o.i.) of adenovirus solution was 1.5, 5, and 10 (pfu/cell) (3 pfu/cell is equal to 200 particles/cell) for Gly-Sar uptake and 0, 7.5, 15, 30, and 45 (pfu/cell) for cephalexin uptake. After a 2-h incubation at 37°C with shaking wells every 15 min, 2.25 ml of medium was added to each well and kept in CO₂ incubator. The uptake study was performed 3 days postinfection.

Uptake Study in Caco-2/hPEPT1 Cells

Three days postinfection, the medium was removed and cells were washed twice with 2 ml of uptake buffer containing 1 mM CaCl₂, 1 mM MgCl₂, 150 mM NaCl, 3 mM KCl, 1 mM NaH₂PO₄, 5 mMd-glucose, and 5 mM MES (pH 6.0). The osmolarity of the uptake medium was 300 ± 5 mmol/kg as measured by vapor pressure osmometer (Wescor Corp., Logan, UT). One milliliter of uptake buffer containing 1 mM [³H]Gly-Sar (0.5 μCi/ml) or 0.1 mM cephalexin was added to each well and the plates were agitating on the plate shaker (Maxi rotator; Lab-line Instrument Inc., Parkridge, IL) at 25°C. After incubation for a specific time, the buffer was removed and cells were washed with ice-cold uptake buffer three times to stop the uptake process. For the assay of [³H]Gly-Sar, 1 ml of ice-cold Milli-Q water containing 1% Triton X-100 was added to each well and the uptake of [³H]Gly-Sar was measured by liquid scintillation spectrometry (model LS-9000; Beckman Coulter, Inc., Fullerton, CA). For the assay of cephalexin, 1 ml of Milli-Q water was added to each well. The cells were then scraped off and sonicated for 10 min. Ice-cold trifluoroacetic acid was added into 700 μl of cell homogenate to make a final concentration of 4%. After centrifugation at 3000 rpm for 5 min, the supernatant was filtered through a membrane filter (0.45 μm; Gelman Instrument Co., Ann Arbor, MI) and injected into the HPLC system.

For the quantitation of PEPT1 in cell lysates, 1 ml of ice-cold solution D (50 mM Tris HCl, 20% glycerol, 2 mM ethylenediaminetetraacetic acid, and 1 mM phenylmethylsulfonyl fluoride, pH 7.5) was directly added into each well. Cells were scraped off, homogenized in a micro-ultrasonic cell disrupter (Kontes Glass, Vineland, NJ), and frozen at −80°C until assay. The protein concentration of each sample was determined with a Bio-Rad protein assay kit.

In Situ Rat Perfusion Study

The perfusion solution consisted of pH 6.5 MES buffer, 0.01% PEG 4000, and a trace amount of [¹⁴C]PEG 4000 as water flux marker. The osmolarity of perfusion solution was adjusted to 300 ± 5 mmol/kg. The concentration of cephalexin in perfusate was 0.5 mM, which is considerably lower than reportedK_m of cephalexin (K_m = 7.2 mM) measured by rat jejunum in situ single pass perfusion study (Sinko and Amidon, 1988).

Rats were fasted overnight with free access to water before the experiment. After anesthetizing with i.m. injection of ketamine (87 mg/kg) and i.p. injection of pentobarbital (40 mg/kg), the abdomen was shaved and an abdominal incision was carefully made to expose the intestinal segments. A 10-cm segment of jejunum was selected for the cannulation of both ends. After cannulation, the selected segment was rinsed with isotonic saline and covered with an isotonic saline-wet gauze. Drug solution was then perfused at a flow rate of 0.192 ml/min. After a steady state was reached (approximately 30 min), the perfusates were collected every 10 min for 90 min. At the end of the experiment, the rats were sacrificed by direct injection of air into the heart. The intestinal segments were quickly excised to measure the length, and mucosal samples were scraped off onto an ice-cold glass plate and transferred directly into ice-cold solution D and homogenized. After centrifugation at 12,000 rpm, 4°C for 5 min, the supernatant was collected and frozen at −80°C until assay. The protein amount of each sample was measured by Bio-Rad protein assay kit. The amount of [¹⁴C]PEG 4000 in perfusates was analyzed by scintillation counter. For the quantitation of cephalexin in perfusates, the samples were treated with 0.1 N ice-cold HCl to give a final pH of 2.3. After vortexed, the samples were centrifuged. The supernatant was filtered through a 0.45-μm syringe filter and injected into HPLC. During the experiment, the rats were kept on the heating slide to maintain their normal temperature.

HPLC Assay

Cephalexin in cell lysates and intestinal perfusates were analyzed by HPLC as described previously (Kovach et al., 1991) with some modification. The HPLC system included a pump (model 515; Waters, Milford, MA), an automatic injector (WISP model 712; Waters), a C₁₈ Supelco column (5 μm, 4.6 × 250 mm), and a UV detector (Waters 996 photodiode array detector) at 250 nm. The system was controlled by Millennium software (Waters, Milford, MA).

The mobile phase consisted of methanol/tetrahydrofuran/aqueous buffer (22:4:74). The aqueous buffer consisted of 5 mM 1-heptanesulfonic acid sodium salt and 1.5% (v/v) triethylamine, adjusted to pH 2.3 with concentrated phosphoric acid. It was eluted at a flow rate of 1.5 ml/min.

Quantitative Western Blot Analysis

PEPT1 was expressed principally in enterocytes of the intestine. It has been reported that there is a significant intersample variation in the content of enterocytes in mucosal samples (Lown et al., 1994). For example, the amount of enterocytes may vary significantly even in the same animal with a deep or superficial mucosal sampling. To normalize such variability, villin, an enterocyte-specific, constitutively expressed protein, was measured in each sample as an internal standard (Lown et al., 1994). It has also been used as an internal standard for quantitation of P-glycoprotein and CYP3A in mucosal biopsies in humans (Lown et al., 1997). Because purified PEPT1 or villin is not available currently, the level of PEPT1 in cell lysates and mucosal samples, expressed as the ratio of the optical density units of PEPT1 to villin, was a relative value. A serial dilution of the purified human α₁-antitrypsin was used to define the linear range of the optical density in Hyperfilm enhanced chemiluminescence (ECL) used for the detection.

Cell lysates or rat mucosal homogenates were boiled for 5 min at 100°C in sample buffer containing 5% 2-mercaptoethanol and 2.5% SDS. The samples were separated by SDS-PAGE with 8% polyacrylamide gel and transferred onto nitrocellulose membranes.

Because PEPT1 and villin have similar molecular masses (∼98 kDa for PEPT1 and ∼95 kDa for villin, respectively), the bands were too close to be separated simultaneously. Therefore, the amount of PEPT1 and villin on the same blot was quantitated separately, as described below.

For the assay of PEPT1, the blots were blocked with 10% nonfat dry milk in Tris buffer (pH 7.5) for 1 h. After washing with Tris buffer, the blots were incubated with a 1:1000 dilution of a rabbit anti-hPEPT1 polyclonal antibody at 25°C for 2 h and further incubated with an anti-rabbit IgG, peroxidase-linked specific whole antibody (ECL; Amersham Pharmacia Biotech) diluted 1:1000 for 1 h. The blots were then developed with the ECL kit (Amersham Pharmacia Biotech) and exposed to ECL Hyperfilm (Amersham Pharmacia Biotech). After film developing, the membranes were washed with Tris buffer and kept at 4°C until the assay of villin.

To quantitate villin on the same membranes, the blots were stripped with 30% H₂O₂ for 30 min. After washing with Tris buffer, the membrane were blocked with 5% nonfat dry milk for 1 h. The following procedures were same as that for PEPT1, except for using a 1:1500 dilution of a mouse anti-villin monoclonal antibody as a primary antibody and an anti-mouse IgG, peroxidase-linked specific whole antibody (ECL; Amersham Pharmacia Biotech) diluted 1:1500 as a secondary antibody.

Exposed films were scanned into binary images with Scan Jet IIc color scanner (Hewlett Packard, Palo Alto, CA) at a resolution of 300 dpi. The amount of the immunoblot protein was determined by computer-aided densitometry by using the public domain NIH Image program (http://rsb.info.nih.gov/nihimage/). Each protein band was converted to a peak where a quantitative number (area under the curve) was calculated.

Data Analysis

Uptake Permeability in Caco-2/hPEPT1 Cells.

The uptake permeability of Gly-Sar or cephalexin in Caco-2/hPEPT1 cells can be calculated by the following equation (Amidon et al., 1988):Equation 1dQ/dt is the initial rate of uptake, which was obtained from the slope of the linear portion of cephalexin uptake − time profile. The initial rates for the uptake of Gly-Sar were estimated by determining the uptake for 15 min, because our preliminary studies suggested that the uptake of Gly-Sar was linear up to 30 min. A is the surface area of the plate, andC_o is the substrate concentration. The values are normalized by the protein amounts (Pr) and expressed as centimeters per second per milligram of protein.

Intestinal Permeability in Rats.

A macroscopic mass balance approach and a mixing tank model were used to estimate the effective intestinal permeability (P_eff) of cephalexin at steady state (Amidon et al., 1980; Johnson and Amidon, 1988). P_eff was obtained by the following equations:Equation 2whereEquation 3is the fractional concentration of the drug absorbed adjusted for the water transport; C_in andC_out is inlet and outlet drug concentration, respectively; Q is the flow rate of the perfusate in the intestine; L is the length of the intestine; and R is the radius of the intestine.

Statistical Method

The results are shown as means ± S.D. for the number of determinations. Student's t test was used to evaluate statistical significance, with P < 0.001 andP < 0.1 as the minimum levels of significance.

Results

Uptake Permeability of Cephalexin and Its Correlation with hPEPT1 Expression in Caco-2/hPEPT1 Cells.

As shown in Fig. 1, uptake permeability of [³H]Gly-Sar was linearly correlated with the expression of hPEPT1 (hPEPT1/villin) (R² = 0.998) in Caco-2/hPEPT1 cells under various levels of Ad.RSV.hPEPT1 transfection (Fig. 1C). The expression of hPEPT1 and villin was also shown in Fig. 1, A and B, respectively. The intensity of the hPEPT1 band increased with increasing Ad.RSV.hPEPT1 infection level (Fig. 1A). Cephalexin uptake into Caco-2/hPEPT1 cells under different levels of Ad.RSV.hPEPT1 infection was shown in Fig. 2. Compared with normal Caco-2 cells (Fig. 2A), the cephalexin uptake was significantly enhanced in Caco-2/hPEPT1 cells (Fig. 2, B–E). The expression of hPEPT1 and villin in these cell lysates was also quantitated by Western blotting. The immunoblotting images of hPEPT1 and villin under different adenoviral infection levels were shown in Fig. 3. The endogenous hPEPT1 in normal Caco-2 cells after 6 days in culture was very low and undetectable by Western blotting assay (Fig. 3). On the other hand, the intensity of hPEPT1 band significantly increased when Caco-2 cells were infected with Ad.RSV.hPEPT1 at an increasing level from 7.5 to 45 pfu/cell (Fig.3). A good correlation was observed between the expression of hPEPT1 (hPEPT1/Villin) and uptake permeability of cephalexin withR² = 0.96 (P < 0.005) (Fig. 4A). In addition, Ad.RSV.hPEPT1 infection level in Caco-2/hPEPT1 cells was also well correlated with uptake permeability of cephalexin (R²= 0.84, P < 0.05) (Fig. 4B) and the expression of hPEPT1 (R² = 0.77, P< 0.1) (Fig. 4C), respectively. Moreover, our recent studies indicated that hPEPT1 expression in Caco-2/hPEPT1 cells also was well correlated with hPEPT1 mRNA level as determined by Northern blot analysis (K. Higaki, D.-M. Oh, X.-Y. Chu, H.-K. Han, Y.-Y. Chin, and G. L. Amidon, unpublished observation).

P_eff of Cephalexin and Its Correlation with PEPT1 Expression in Rat Jejunum.

In the rat perfusion study, the mean P_eff ± S.D. of cephalexin in jejunum of 15 rats was 3.89 ± 1.63 × 10⁻⁵ cm/s (Table1), exhibiting a 4-fold variation (Table 1). The amount of PEPT1 and villin in each jejunum mucosal sample was analyzed by Western blotting. PEPT1, an ∼98-kDa protein, and villin, an ∼95-kDa protein, were detected throughout all samples (Fig. 5). The PEPT1 level in jejunum after normalized by villin content in each mucosal sample exhibited a 5-fold variability (Table 1). The PEPT1 amount in ileum was also measured in the same study. It was comparable or somewhat higher than that in rat jejunum (Table 1). A significant correlation between PEPT1 expression (PEPT1/villin) and cephalexin permeability was observed in rat jejunum withR² = 0.63 (P < 0.001) (Fig. 6).

Discussion

Membrane transporters, in addition to metabolizing enzymes, are important factors that have been identified to be associated with the absorption, disposition, and detoxification of many endogenous substances and xenobiotics. Recently, there have been significant advances in the study of membrane transport, particularly in the molecular understanding of drug transporters, although such progress is still at an early stage. Recent studies have provided extensive reports on the cloning of transporters, their functional analysis, substrate specificity, as well as tissue distribution. However, there is limited knowledge about the contribution of the transporters to the in vivo absorption of drugs. For example, PEPT1 transporter is responsible for mucosal cell transport of many di- and tripeptide and peptide-like compounds. Targeting of PEPT1 transporter is, therefore, considered to be a useful strategy for oral drug delivery. However, there is still no evidence at present indicating that the absorption or bioavailability of these peptide-like compounds is correlated with PEPT1 expression in the intestine in vivo.

In this study, we measured the PEPT1 expression in hPEPT1 overexpressed Caco-2 cells and rat jejunum and determined its correlation with intestinal absorption phenotype and intestinal permeability, which is one of the most important parameters affecting rate and extent of drug absorption. Cephalexin, a substrate of PEPT1 (Dantzig and Bergin, 1990;Inui et al., 1992; Bretschneider et al., 1999), was used as a model compound. To our knowledge, this study provides the first demonstration of the correlation between transport expression and absorption permeability for peptide-type compounds both in vitro and in vivo and serves as a basis for genetic/molecular understanding of human variation in absorption.

Quantitation of PEPT1 is important for studying the mechanism of oral peptide transport. The immunoquantification of PEPT1 has been reported by Walker et al. (1998) by using immunohistochemical techniques along with pixel integration with confocal laser scanning microscopy. However, this technique has some limitations due to the inter- and intrasample variations during staining. In contrast, immunoblotting after SDS-PAGE provides qualitative and quantitative evaluation in protein detection. With the application of the ECL technique, a distinctive and sensitive nonradioactive system, the quantitative Western blotting gives a detection limit as low as 50 pg (Venembre et al., 1994). We adopted this approach and have subsequently developed a quantitative assay to measure PEPT1 amounts in Caco-2 cells and rat jejunum. Villin, an enterocyte-specific protein, was chosen as an internal standard to normalize the enterocyte content in each sample (Lown et al., 1994). Because it is difficult to separate the PEPT1 and villin band in SDS-PAGE, the amount of PEPT1 and villin in the same blot has to be quantitated separately. To do this, we analyzed PEPT1 or villin first and then used an excess amount of hydrogen peroxide, a substrate of horseradish peroxidase, to consume the activity of horseradish peroxidase-labeled secondary antibody for the first protein measured before the quantitation of the second protein. Our preliminary study has confirmed that no interference band of PEPT1 or villin was detected after hydrogen peroxide treatment, and the amount of protein was reproducible with or without hydrogen peroxide treatment. This allows us to quantitate PEPT1 and villin in the same blot and therefore eliminate the variation during electrophoresis and blotting. According to our results, this method was sensitive and reproducible for the quantitation of PEPT1.

To understand the possible contribution of PEPT1 to intestinal cephalexin absorption, we investigated the uptake permeability of cephalexin in vitro by using Caco-2/hPEPT1 cells and in an intestinal permeability study with the in situ rat jejunum perfusion method. In Caco-2/hPEPT1 cells, uptake of [³H]Gly-Sar, a representative substrate of PEPT1, was performed as a positive control. As a result, the uptake permeability of Gly-Sar was linearly correlated with the expression of hPEPT1 (hPEPT1/villin) withR² = 0.998 (Fig. 1). This suggested that hPEPT1 is the major transporter responsible for Gly-Sar uptake in Caco-2/hPEPT1 cells. This result also validated the overexpression of hPEPT1 in Caco-2/hPEPT1 cells and the PEPT1 quantitative method. Similar to that of Gly-Sar, a linear correlation was observed between uptake permeability of cephalexin and hPEPT1 expression withR² = 0.96 (P < 0.005) (Fig. 4A). This demonstrates that cephalexin permeability is directly proportional to the hPEPT1 expression in Caco-2/hPEPT1 cells. The hPEPT1-mediated uptake is therefore the dominant pathway for permeability of cephalexin in Caco-2/hPEPT1 cells. In addition, our recent observations indicate a good correlation between hPEPT1 and hPEPT1 mRNA level (K. Higaki, D.-M. Oh, X.-Y. Chu, H.-K. Han, Y.-Y. Chiu, and G. L. Amidon, unpublished observation), which further supports our hypothesis that hPEPT1 is the major transporter responsible for uptake of cephalexin in Caco-2/hPEPT1 cells. To obtain further evidence for the possible involvement of PEPT1 on intestinal absorption of cephalexin in vivo, the P_eff of cephalexin was determined in a rat in situ single pass perfusion study. There was a significant correlation between PEPT1 expression in jejunum mucosa and cephalexin intestinal permeability with anR² = 0.63 (P < 0.001) (Fig. 6). This suggests that PEPT1-mediated uptake is one of the important pathways for intestinal absorption of cephalexin in vivo. Compared with jejunum, considerable or even higher PEPT1 expression was observed in rat ileum (Table 1). The PEPT1 expression in ileum was also correlated with the PEPT1 expression in jejunum in the same rats withR² = 0.73 (P < 0.001). This suggests that the ileum is also an important segment that contributes to the overall intestinal absorption of cephalexin mediated by PEPT1.

Thus, our results indicate that PEPT1-mediated uptake is one of the major pathways for intestinal absorption of cephalexin in Caco-2/hPEPT1 cells and in situ in rats. However, there are still some concerns with the true correlation between PEPT1 expression and cephalexin permeability. For example, because PEPT1 antibody specific for apical or basolateral membrane is not available, it is difficult to distinguish PEPT1 expressed in apical membrane from that of total PEPT1 measured by our Western blotting assay. In addition, it is still unknown whether all PEPT1 expressed on the apical membrane is functionally active and related to cephalexin uptake. Finally, it is not known to what extent other transporters may be involved in peptide-like compound uptake or modulate or regulate PEPT1 activity. However, our results support PEPT1 as a major contributor to cephalexin uptake, and hence di- and tripeptide uptake. Further studies are needed to quantitate these additional factors in vitro and in vivo.

For the in vivo rat perfusion study, although villin has become an accepted marker to normalize the enterocyte content in intestinal mucosal samples, some other membrane marker proteins, such as sucrase or alkaline phosphatase, might be alternative markers with which to normalize the sample variation. This should be evaluated in a future study. Moreover, as shown in Table 1, the variation of PEPT1 in the rat jejunum in the studied animals was low (about 5-fold). Future studies should include animals expected to have a wider range of PEPT1 expression level and permeability variability, e.g., diet, disease. For cephalexin permeability, the surface area variation between animals in the perfusion study or paracellular transport may also be considered as the possible sources of variation. In addition to carrier-mediated process (Dantzig and Bergin, 1990; Inui et al., 1992; Bretschneider et al., 1999), passive diffusion has also been identified in intestinal absorption of cephalexin to limited extent (Yasuhara et al., 1977;Tsuji et al., 1979). Our recent human intestinal perfusion study indicated that 38% (P = 0.03) of the variation in cephalexin permeability was accounted by the variation from the permeability of propranolol, a marker for passive transport (K. Higaki, D.-M. Oh, X.-Y. Chu, H.-K. Han, Y.-Y. Chiu, and G. L. Amidon, unpublished observation). This further suggests that passive diffusion might also contribute to intestinal absorption of cephalexin.

Recent investigations have demonstrated that multiple transporters, other than PEPT1, are also expressed in brush-border membrane of the small intestine. For example, the studies with various dipeptides as inhibitors showed inconsistent results in uptake and/or transport of β-lactam antibiotics into Caco-2 cells (Delie and Rubas, 1997), which suggested the possible involvement of multiple transporters for transport of peptide-like compounds. In another study, a 127-kDa protein was identified as a component of the H⁺/oligopeptide transport system in brush-border membrane vesicles from rabbit small intestine by photoaffinity labeling with [³H]cephalexin and further photoreactive β-lactam antibiotics and dipeptides (Kramer et al., 1998). Antibodies prepared against the purified protein were able to inhibit the transport of cephalexin photoaffinity labeling of the 127-kDa protein in intact brush-border membrane vesicles (Kramer et al., 1992). This protein has been identified in the brush-border membrane of the epithelial cells in the jejunum and proximal tubule of the kidney (Kramer et al., 1995). In addition, Dantzig et al. (1994) reported the identification of another peptide transporter, termed the human intestinal peptide transporter-1 (HPT-1). HPT-1 and PEPT1 share only 16% identity and 41% similarity in their amino acid sequence (Liang et al., 1995). Chinese hamster ovary cells transfected with the HPT-1 gene showed 2- to 3-fold higher uptake of cephalexin than that in control (Dantzig et al., 1994). Furthermore, many peptide transporter families, including ATP-binding cassette-type peptide transporters, which contain ATP-binding cassette, have also been identified in small intestine (Yang et al., 1999). It is possible that these peptide transporters may also be involved in the intestinal absorption of cephalexin and other peptide-like compounds. Further study is needed to clarify the possible contribution of these transporters.

In conclusion, we have demonstrated a linear correlation between hPEPT1 level and uptake permeability of cephalexin in Caco-2/hPEPT1 cells, indicating that cephalexin uptake in Caco-2 cells is directly proportional to hPEPT1 expression. Furthermore, the significant correlation between cephalexin permeability and PEPT1 expression in rat jejunum is also demonstrating that a large part of absorption variability of cephalexin is due to the variation of PEPT1 transporter expression in vivo, thus providing the molecular bases for understanding absorption variability.