In vitro permeation of β-lactam antibiotics across rat jejunum and its correlation with oral bioavailability in humans

Abstract

Aims

To investigate the correlation between in vitro permeation of 11 β-lactam antibiotics across rat jejunum and their oral bioavailability in humans.

Methods

The absorptive and secretory permeation across rat jejunum was evaluated and apparent permeability coefficients (P_app) were determined.

Results

A steep, sigmoid-type curve was obtained for the relationship between P_app in the absorptive permeation and human oral bioavailability. When the ratios of P_app in the absorptive direction to P_app in the secretory direction were plotted against human oral bioavailability, a much improved correlation was obtained (r = 0.98, P < 0.001). The addition of glycylglycine to both mucosal and serosal media modified the permeation of ceftibuten and cephalexin from the absorptive to the secretory direction.

Conclusions

For 11 β-lactam antibiotics rat intestinal permeation correlated well with human oral bioavailability, especially when corrected for secretory transport.

Introduction

Previously, we reported that an energy-dependent efflux system was potentially involved in the poor absorption of orally inactive β-lactam antibiotics such as cefazolin [1, 2], giving a new insight into understanding the intestinal absorption of β-lactam antibiotics.

We have used excised rat small intestine and an Ussing chamber technique to elucidate the secretory-orientated transport of β-lactam antibiotics. However, the usefulness of an in vitro experimental model using rat intestine has been questioned. For example, using similar types of chambers, it was reported that there was no significant correlation between in vitro and in vivo drug permeability in rat jejunum [3]. Lennernas et al. [4] suggested that the in vitro rat intestinal model was useful for screening passive drug absorption in humans but that there were several concerns over the suitability of rat intestine for predicting in vivo absorption of drugs transported by carrier-mediated systems.

This study was undertaken to re-evaluate the correlation between in vitro rat intestinal permeation parameters and in vivo human oral bioavailability, focusing especially on the importance of the rat intestinal efflux system.

Methods

Materials

Cefditoren sodium was kindly supplied from Meiji Seika Kaisha (Tokyo, Japan). Cefdinir and cefixime were from Fujisawa Pharmaceutical Co. (Osaka, Japan). Ceftibuten was from Shionogi & Co. (Osaka, Japan). Ampicillin anhydrous was purchased from Wako Pure Chem. Ind. (Osaka, Japan). Other β-lactam antibiotics and glycylglycine (GlyGly) were from Sigma Chemical Co. (St Louis, MO, USA).

Permeation experiments across rat jejunum

Tyrode's solution containing 6 mm d-glucose was used as an experimental medium. Male Sprague-Dawley and Wistar rats, weighing 300-350 g, were used. Permeation experiments were performed as described previously [1] using diffusion chambers [5]. Permeation experiments were normally done for 2 h using 1 mm drug solution (pH 7.4). The cumulative amount of drug permeating in the mucosal-to-serosal (M to S, absorptive) or serosal-to-mucosal (S to M, secretory) direction was calculated and plotted against time. The slope of the linear portion of this plot was corrected for membrane surface area (1.78 cm²) and donor drug concentration to obtain the apparent permeability coefficient (P_app).

In this study, principles of good laboratory care were followed and animal experimentation was done in accordance with The Guidelines for the Care & Use of Laboratory Animals in Health Sciences University of Hokkaido.

H.p.l.c. analysis

All β-lactam antibiotics were assayed using a Shimadzu LC-10A h.p.l.c. system (Kyoto, Japan). A Nova-Pak C-18 (3.9 × 150 mm, 5 µm, Waters, Milford, MA, USA) column was used at 50 °C. Mobile phases were H₂O: CH₃CN: TFA 85:15:0.1 (v/v/v) for cefaclor, cefoperazone, cephaloridine and cephradine; 10 mm CH₃COONa: CH₃OH 8:2 for cefdinir, cefixime and cephalexin; 50 mm KH₂PO₄: CH₃OH 8:2 for ampicillin, cefazolin, cefditoren and ceftibuten. The flow rate was 0.6-1 ml min^-1. The wavelength was 215 nm for ampicillin, 230 nm for cefoperazone, 240 nm for cephaloridine, and 260 nm for other β-lactam antibiotics. The injection volume was 50 µl. A rectilinear relationship was obtained between peak heights and concentrations over the range of 0-10 µm. For all assays, the coefficients of variation were less than 6%, with intra- and inter assay precision of at least 98%. The limits of quantification were from 0.02 to 0.1 µm.

Data analysis

All results are expressed as mean ± s.e. mean. Statistical analysis was performed by using Student's t-test and P < 0.05 was considered significant. The relationship between in vitro and in vivo data was evaluated by means of linear regression analysis.

Results

Correlation between rat jejunal permeability and human oral bioavailability

Figure 1 demonstrates the relationship between P_app of 11 β-lactam antibiotics across rat jejunum and their urinary excretion ratios after oral administration in humans. Since most β-lactam antibiotics are poorly metabolized and excreted unchanged into the urine after absorption, urinary excretion ratios correspond to their oral bioavailability. Cefoperazone, which is excreted extensively into bile rather than urine [6], is an exception. However, its intestinal absorption has been regarded as negligible. When the P_{app(M to S)} was plotted against urinary excretion ratio, a correlation was obtained with a regression coefficient (r) of 0.89 (P < 0.05, Figure 1a, dotted line). There was a much improved correlation (r = 0.98, P < 0.001) when urinary excretion ratios were plotted vs permeability ratios expressed as P_{app(M to S)}/P_{app(S to M)} (Figure 1b).

Figure 1.

Relationship between the permeation parameters of 11 β-lactam antibiotics across rat jejunum and their urinary excretion ratios after oral administration in humans. The drugs were tested at a concentration of 1 mm. The data for urinary excretion were obtained from pharmaceutical manufacturers and from the work of Bergan [12]. The straight lines were obtained by fitting the data using linear regression analysis. 1, cephradine; 2, cephalexin; 3, cefaclor; 4, ampicillin; 5, cephaloridine; 6, cefazolin; 7, cefoperazone; 8, cefditoren; 9, ceftibuten; 10, cefixime; 11, cefdinir.

Effect of GlyGly on the permeation of four β-lactam antibiotics

Figure 2 demonstrates the effect of 20 mm GlyGly on the permeation of cefazolin, cefixime, ceftibuten and cephalexin. The dipeptide was added simultaneously to both the mucosal and serosal media (pH 7.4). Cefazolin permeation in both directions was unchanged in the presence of GlyGly. On the other hand, the absorptive permeation of the other three drugs was significantly decreased, whereas their S to M permeation increased. As a result, the secretory-orientated permeation of cefixime was promoted markedly, and the permeation of ceftibuten and cephalexin changed from the absorptive to the secretory direction.

Figure 2.

Effect of GlyGly on the permeation of four β-lactam antibiotics across rat jejunum. GlyGly was added to both the mucosal and serosal media at a concentration of 20 mm. The concentration of drugs was 1 mm. Results represent the mean ± s.e. of four experiments with jejunal segments from different rats. *P < 0.05, when compared with the control data.

Discussion

It has generally been accepted that the parameters of in vitro mucosal to serosal permeation could be an index of in vivo absorption. A linear regression analysis of the relationship between P_{app(M to S)} and urinary excretion ratios in humans after oral dosing for a series of β-lactam antibiotics gave a relatively high regression coefficient. However, the relationship was better described by a steep sigmoid-type curve. It was likely that such a curve would overestimate or underestimate the oral bioavailability of β-lactam antibiotics with a P_app of around 5–6 × 10⁻⁶ cm s^-1. When the permeability ratios were plotted against urinary excretion ratios, a much-improved correlation was obtained. For comparison, we calculated the correlation between urinary excretion ratios and the parameters of Caco-2 apical-to-basolateral permeation reported by Bretschneider et al. [7]. Although these workers emphasized that the extent of oral bioavailability of β-lactam antibiotics was determined mainly by their affinity for a peptide transporter (PEPT1), the r value from their study was 0.79. Accordingly, the present results suggested that the human oral bioavailability of various β-lactam antibiotics could be predicted more precisely by fitting their permeability ratios to equation 2 in Figure 1b. Our recent study has demonstrated that cefazolin does not exhibit secretory-orientated permeation across Caco-2 cell monolayers, indicating that Caco-2 cells have a much lower expression of the transporter responsible for the efflux of β-lactam antibiotics than the rat jejunum (unpublished data).

A new finding from the present study was that the permeability ratios of cefdinir and cefixime were secretory-orientated, in contrast to the absorptive-orientated permeation of ceftibuten, cephalexin and cephradine. It has been reported that cefdinir and cefixime were absorbed via PEPT1 [8–10]. The present results suggested the possibility that the efflux system for an orally inactive β-lactam such as cefazolin was also involved in the jejunal permeation of cefixime and cefdinir. To assess further the involvement of the efflux system in the jejunal transport of orally active β-lactam antibiotics, the permeation of four drugs was re-evaluated in the presence of GlyGly (Figure 2). Since peptide transporters exist not only on the brush-border but also on the basolateral membrane of intestinal epithelial cells [11], this dipeptide was added to both sides of intestinal segments. A finding that GlyGly had no effect on cefazolin permeation implied no interaction between the dipeptide and the efflux system. Furthermore, the permeation clearly changed to secretory-orientated manner when the interaction between orally active β-lactam antibiotics and peptide transporters was fully inhibited by GlyGly. These observations suggested that cefixime is partly pumped out into the lumen via an efflux system on the jejunal membrane after it is transported into the enterocytes via PEPT1. It is likely that this bi-directional movement of cefixime in the intestine leads to its low oral bioavailability. This efflux system appeared less involved in the absorption of highly absorbable β-lactam antibiotics such as ceftibuten and cephalexin.

In conclusion, the results from this study suggested (1) the oral bioavailability of β-lactam antibiotics is determined by differences in their affinities for peptide transporters and energy-dependent efflux system and (2) the rat intestine is a suitable tool for estimating the in vivo human oral bioavailability of a range of β-lactam antibiotics by using in vitro permeation parameters corrected for secretory transport.