Advertisement
Research ArticleArticle

Effect of Molecular Charge on Intestinal Epithelial Drug Transport: pH-Dependent Transport of Cationic Drugs

Katrin Palm, Kristina Luthman, Jenny Ros, Johan Gråsjö and Per Artursson
Journal of Pharmacology and Experimental Therapeutics November 1999, 291 (2) 435-443;

Abstract

The aim of this study was to investigate the effect of ionization on drug transport across the intestinal epithelium in order to include this effect in structure-absorption relationships. The pH-dependent permeation of one rapidly (alfentanil) and one slowly (cimetidine) transported basic model drug across Caco-2 cell monolayers was investigated. Both drugs had pKavalues in the physiological pH range. The permeability coefficients (Pc) of the model drugs were obtained at varying apical buffer pHs, thus varying the degree of drug ionization (from 5 to 95%). The relationship between Pc and the fraction of the drug in un-ionized form (fu) was analyzed to delineate the permeability coefficients of the un-ionized (Pc,u) and ionized (Pc,i) forms of the drugs. Theoretical estimates of the pKa values were also calculated from ionization energies for each model compound. For both drugs, a linear increase in Pc was observed with increasing fu. Transport of the un-ionized form was 150- and 30-fold more rapid than transport of the ionized form for alfentanil and cimetidine, respectively. However, when fu < 0.1, the contribution of the ionized form was significant. Because fu is <0.1 over the entire physiological pH range for a large number of drugs, these results will have implications on predictions of in vivo intestinal drug absorption both from in vitro studies in cell cultures and from computed structural properties of drug molecules.

The pH-partition theory states that only the un-ionized form of ionizable drug molecules can diffuse across cell membranes. This assumption is often extended to drug transport across epithelial barriers, such as the intestinal epithelium (Hogben et al., 1959). However, it has become clear from recent studies that the ionized form of drug molecules can also partition into phosphatidylcholine bilayers (Iseki et al., 1992;Ottiger and Wunderli-Allenspach, 1997; Avdeef et al., 1998), although to a much lesser extent than the un-ionized form. In addition, the paracellular pathway through intestinal epithelia has been shown to be more permeable to cationic than to either neutral or anionic drug molecules (e.g., Adson et al., 1994; Karlsson et al., 1994). Therefore, it cannot be ruled out that the intestinal epithelium is also permeable to the ionized form of drug molecules at high degrees of ionization. If this is the case, transport of the ionized form of the drugs should be considered in studies of epithelial permeability to drugs, and the pH-partition theory should be modified accordingly. To enable a rational approach to predictions of the intestinal absorption of proteolytic drugs, information on epithelial permeability to the un-ionized form of the drug, the pKa of the drug and the effect of molecular charge on drug transport; i.e., the selectivity of the intestinal epithelium for the un-ionized over the ionized form, are needed. Thus, a modification of the pH-partition theory would be particularly important for predictions of intestinal drug absorption from in vitro studies and studies of structure-absorption relationships.

Most experimental investigations on the relationship between pH, drug ionization and intestinal epithelial permeability have been made using animal models. Because of the complexity of these models, it has been difficult to determine the extent of the epithelial selectivity for the un-ionized over the ionized form. Deviations from the pH-partition theory have frequently been observed for both acidic and basic compounds. These deviations have been attributed to the presence of an acidic microclimate close to the intestinal wall and to aqueous boundary effects (Hogben et al., 1959; Högerle and Winne, 1983). Intestinal cell culture models, such as the Caco-2 cell monolayers, offer two major advantages in studies of the relationship between drug ionization and epithelial permeability. First, the absence of villous structures and a mucus layer makes it possible to minimize the influence of an acidic microclimate by using effective stirring and buffers (Shiau et al., 1985). Second, it is possible to determine the cell monolayer permeability unbiased by the aqueous boundary layer (Karlsson and Artursson, 1991).

The aim of this study was to quantify the effect of molecular charge on drug transport across the intestinal epithelium to include this effect in structure-absorption relationships. Therefore, the pH-dependent transport of two model drugs, both with pKavalues in the pH range found in vivo, across Caco-2 cell monolayers was investigated. Alfentanil, which is rapidly transported across Caco-2 cell monolayers, and cimetidine, which is more slowly transported, were selected as model drugs. The cell monolayer permeability coefficients (Pc) of the two basic drugs were determined in the pH range 5.0 to 8.0. For this purpose, a method for determining Pc values under nonsink conditions was developed. The relationship between Pc and the fraction of the drug in un-ionized form (fu) was established to delineate the Pcs of the un-ionized and ionized forms of the drugs. Finally, the contribution of the transport of the ionized form to the total transport in the studied pH range was investigated.

Materials and Methods

Selection of Model Drugs

A comprehensive drug compendium (Craig, 1990) was examined to identify drugs with pKa values of approximately 6.5. Such drugs were considered as suitable model drugs because they display the largest possible variation in the degree of ionization within the physiological pH-interval of the intestine. In addition, both rapidly and slowly transported drugs were considered because the effect of solute charge on the epithelial permeability to transcellularly and paracellularly transported drugs could be different.

Thirty compounds with pKa values between 6.2 and 6.8 were identified. These compounds were then evaluated based on levels of toxicity, chemical stability, solubility, active transport mechanisms, intestinal metabolism, and availability. Alfentanil (logarithm of the octanol/water partition coefficient = 2.16), which was confirmed to be stable and sufficiently soluble for the purposes of the study, was selected as an example of a rapidly transported drug. The second model drug, cimetidine (logarithm of the octanol/water partition coefficient = 0.40), has a low transport rate across Caco-2 monolayers (Pade and Stavchansky, 1997) and is believed to be transported across the intestinal epithelium predominantly by a passive diffusion process (Mummaneni and Dressman, 1994). Although affinity for active efflux systems, such as P-glycoprotein and the organic cation exchanger (Miyamoto et al., 1988;Dutt et al., 1992; Pan et al., 1994), has been observed for cimetidine, active efflux of the molecule can be significantly reduced by the addition of verapamil (Pan et al., 1994).

Drug Transport Studies

Drugs and Radiolabeled Markers.

Cimetidine and verapamil were purchased from Sigma (St. Louis, MO). Alfentanil was obtained as Rapifen IV solution from Janssen-Cilag through Apoteket AB, Stockholm, Sweden and [14C]-mannitol (specific radioactivity, 52 mCi/mmol) was purchased from New England Nuclear through Du Pont Scandinavian AB, Norrköping, Sweden.

Cell Culture

Caco-2 cells were obtained from American Tissue Culture Collection (Rockville, MD). The cells were maintained in Dulbecco’s modified Eagle’s medium, containing 10% heat-inactivated fetal calf serum and 1% nonessential amino acids, in a humidified atmosphere of 90% air and 10% CO2, as described elsewhere (Artursson, 1990). Cells (5 × 105) (passage number 90–105) were seeded on polycarbonate filter inserts (Transwell Costar, Badhoevedorp, the Netherlands; mean pore size 0.45 μm; diameters 6 and 12 mm) and cultivated in Dulbecco’s Modified Eagle’s Medium supplemented with 10% fetal calf serum, 1% nonessential amino acids, penicillin (100 units/ml), and streptomycin (100 μg/ml). All tissue culture media were obtained from Gibco through Laboratorie Design AB (Lidingö, Sweden). The cells were allowed to grow and differentiate for 22 to 32 days before the monolayers were used in drug transport experiments.

Drug Transport Experiments

The transport experiments were performed in Hanks balanced salt solution (HBSS). The HBSS was buffered with 25 mM acetate (pH 4.0–5.0), 20 mM 2-(N-morpholino)ethanesulfonic acid (MES) (pH 5.0–6.5), or 25 mM HEPES (pH 6.8–8.0), depending on the desired pH on the apical side. These buffers maintain the adjusted pH values during the transport experiments (data not shown; Ellens et al., 1997). The presence of an acidic microclimate at the mucosal surface is dependent on the presence of a mucus layer and is abolished by effective stirring and well-buffered apical solutions (Shiau et al., 1985). Consequently, we assumed that any acidic microclimate produced by the Caco-2 cells would be disrupted under the experimental conditions used in this study. HBSS of pH 7.4 buffered with 25 mM HEPES was always used on the basolateral side of the monolayers.

Transport rates were determined in both the apical-to-basolateral (a-b) and basolateral-to-apical (b-a) directions to detect polarized transport in the absorptive or secretory direction. Drugs were dissolved in the transport buffer to give a final concentration of 0.5 to 10 mM. All solutions were prewarmed to 37°C. The drug solutions were added to the donor side of the monolayers, and buffer without drug was added to the receiver side. The monolayers were incubated in a humidified atmosphere at 37°C. At regular time intervals (4–10 min for alfentanil and 20–30 min for cimetidine), samples were withdrawn from the receiving chamber and frozen (−20°C), pending HPLC analysis. All transport experiments were carried out in less than 120 min. The transport of the model compounds was followed at both low (100 rpm) and high (500 rpm) stirring rates using a calibrated plate shaker (IKA-Schüttler MTS4) to obtain Pc values unbiased by the aqueous boundary layer (see below). Monolayer permeability to the paracellular marker [14C]mannitol was used to investigate the integrity of the monolayers under the experimental conditions.

Analytical Methods

The samples from the drug transport experiments were analyzed using reversed phase HPLC. The system consisted of a Perkin-Elmer Isocratic LC Pump 250, a Perkin-Elmer Advanced LC Sample Processor ISS-200, a Spectra Physics UV100 detector, and the integration software program Chromatography Station for Windows. A Beckman Ultrasphere ODS (250 × 5.6 mm) with a particle size of 5 μm was used as analytical column. Mobile phases consisting of phosphate buffer (pH 3.0; 60 mM KH2PO4, 8 mM H3PO4), and 20% (cimetidine) or 30% (alfentanil) acetonitrile, resulted in retention times of 6 to 8 min at a flow rate of 1 ml/min. Cimetidine was detected at a wavelength of 225 nm and alfentanil at 232 nm. Samples containing [14C]mannitol were counted in a liquid scintillation counter (Tricarb 1900 CA; Packard Instruments).

Determination of Pc

At “sink” conditions, the apparent permeability coefficients (Papp, cm/s) were calculated from:Papp=(k·VR)/(A·60) Equation 1where k is the transport rate (min−1) defined as the slope obtained by linear regression of the cumulative fraction absorbed as a function of time (min), VRis the volume in the receiver chamber (ml), and A is the area of the filter (cm2). The fraction absorbed was defined as the ratio between the concentration on the receiver side (CR) at the end of an interval and the concentration on the donor side (CD,0) at the beginning of that interval. Alfentanil was transported so rapidly at higher pH values that sink conditions could not be maintained (Fig.1a). It was therefore necessary to develop a more general method for calculating Papp, one also applicable to nonsink conditions. Taking into account the effect of back flux from the receiver compartment, Fick’s law gives:dCR(t)/dt=[Papp·A·(CD(t)−CR(t))]/VR Equation 2where CD(t) and CR(t) is the concentration in the donor and receiver compartment, respectively, as a function of time. If the amount of drug in the system (M) is assumed to be constant within each sampling interval, CD(t) can be expressed as a function of CR(t). Thus, replacing CD(t) by this expression in eq. 2 and solving the differential equation for CR(t) gives: CR(t)=[M/(VD+VR)]+[CR,0−M/(VD+VR)]·e−Papp·A·(1/VD+1/VR)·t Equation 3where VD is the volume in the donor compartment, CR,0 is the drug concentration in the receiver compartment at the beginning of the interval and t is the time from the start of the interval. The sampling procedure necessitates the recalculation of M and CR,0 for each succeeding interval according to mass balance. Papp values for alfentanil were then determined by nonlinear regression, minimizing the sum of squared residuals (Σ(CR,i,obs − CR,i,calc)2), where CR,i,obs is the observed receiver concentration at the end of interval i and CR,i,calc is the corresponding concentration calculated according to eq. 3. The importance of the correction for nonsink conditions in calculations of Papp values of alfentanil is shown in Fig. 1a. Sink conditions were maintained in the transport experiments of cimetidine and Papp values of cimetidine calculated according to eqs. 1 and 3 were not different (Fig. 1b).

Figure 1
Figure 1

a, Amount alfentanil (1.2 mM) transported across Caco-2 cell monolayers at pH 8.0 as a function of time. At a low stirring rate (100 rpm; ○), sink conditions (which are maintained when the ratio between receiver and donor concentration, CR/CD, < 0.1) were marginally exceeded in the end of the last time interval (CR/CD = 0.11). There was not a significant difference (p > .05) between Papp of alfentanil at 100 rpm calculated according to eq. 1 (93.5 ± 5.4 · 10−6 cm/s), which assumes sink conditions, and eq. 3 (105 ± 8 · 10−6 cm/s), which takes into account the effect of back flux from the receiver compartment. However, at the higher stirring rate (500 rpm; ▪), sink conditions were exceeded already during the second time interval (CR/CD > 0.12) and the importance of applying a method applicable to nonsink conditions was demonstrated: Calculation of Papp according to eq. 1 (167 ± 6 · 10−6 cm/s) significantly (p < .01) underestimated the Papp of alfentanil at 500 rpm (197 ± 8 · 10−6 cm/s) calculated according to eq. 3. b, amount cimetidine (1 mM in the presence of verapamil) transported across Caco-2 cell monolayers at pH 8.0 as a function of time. Sink conditions were maintained both at low (100 rpm; ○) and high (500 rpm; ▪) stirring rates. Thus, Papp values calculated according to eqs. 1 and 3 were not significantly different at 100 rpm (1.68 ± 0.08 · 10−6 cm/s, according to both equations) or at 500 rpm (1.65 ± 0.17 · 10−6 cm/s and 1.67 ± 0.17 · 10−6 cm/s, according to eqs. 1 and 3, respectively). Values are presented as mean ± S.D.,n = 3 to 4.

Subsequently, the permeability coefficients of the cell monolayer (Pc,cm/s) for both alfentanil and cimetidine were calculated as described previously (Karlsson and Artursson, 1991):V/Papp=1/K+(1/Pc+1/Pf)·V Equation 4where V is the stirring rate, Pf is the calculated Pc of the filter support (Karlsson and Artursson, 1991) and K is a constant.

Theory

Calculation of pKa Values.

Conformational analysis. The conformational preferences of alfentanil and cimetidine were analyzed using molecular mechanics calculations. The MM3 force field, as implemented in the MacroModel program (version 5.5), was used in the analyses of alfentanil. Because parameters for cimetidine were missing in the MM3 force field, the conformational analyses of cimetidine were performed using the Merck Molecular Force Field, also as implemented in MacroModel. Systematic pseudo-Monte Carlo conformational search procedures (BatchMin v 5.5) were used for both compounds in vacuum and simulated water environment.

For alfentanil, a 10,000-step conformational search was carried out in each environment. Four separate 10,000-step searches were performed for cimetidine: two searches with different configurations of the amino function for both the cis- and trans- forms of the guanidine moiety. Energy minimizations were made using the truncated Newton algorithm with a maximum of 5000 iterations.

Calculations of pKa Using MolSurf.

The global minimum conformations found in the conformational analysis of each environment were first geometry optimized using the semiempirical AM1 method as implemented in the Spartan program (version 5.3). Then, a single-point quantum mechanical ab initio calculation using the 3–21G* basis set was performed on the AM1 optimized conformation. The wave function generated by the Spartan program was used by MolSurf to calculate the pKa values (pKa,calc) of the compounds (Norinder et al., 1997; Sjöberg, 1997).

The pKa values calculated using this method were found to be highly sensitive to the conformational properties of the molecules. Thus, pKa,calc was defined as the pKa value calculated for the global minimum conformations identified in a simulated water environment.

Determination of the Pc of Un-ionized and Ionized Drug Forms.

In the unmodified pH-partition theory, the following relationship is suggested:Pc=fu·Pc,u Equation 5where fu is the fraction of the drug in the un-ionized form (fu = 1 when the drug is completely un-ionized). fu is calculated from previously reported pKa values (Table1):fu=1/(1+10pKa­pH) Equation 6and Pc,u is the Pcof the un-ionized species. Thus, according to eq. 5, if the epithelial cell monolayer is impermeable to the ionized form, the ratio Pc/fu will equal Pc,u and will be independent of pH. However, if we assume that the intestinal epithelium is also permeable to the ionized species, eq. 5 must be expanded to:Pc=fu·Pc,u+fi·Pc,i Equation 7where fi is the fraction of drug in ionized form and Pc,i the Pc of the ionized species. Because fi + fu = 1, eq. 7 can be rewritten as:Pc=(Pc,u−Pc,i)·fu+Pc,i. Equation 8Thus, if an intercept significantly different from zero is obtained when Pc is plotted as a function of fu, this intercept equals the Pc of the ionized form of the drug and Pc,u can be calculated from the slope of the linear regression line. In this analysis, no assumptions are made with respect to whether transport of the un-ionized or ionized forms occurs via the transcellular or paracellular routes.

View this table:
Table 1

Structure and physicochemical properties of the model drugs  

Statistics

Values are generally expressed as mean ± one S.D. The differences between mean values were analyzed with Student’st test or one-way ANOVA, followed by Scheffe’s Ftest. When equal variances could not be obtained by logarithmic transformation, the analysis was made using Student’s ttests for unequal variances or the Bonferroni correction for multiple comparisons. A level of α = .05 was considered to be statistically significant. In the least-square linear regression analyses on Pc as a function of fu, a weighting scheme of 1/S.D.2 was used.

Results

pH-Dependent Integrity of Caco-2 Cell Monolayers.

The pH-dependent integrity of the Caco-2 cell monolayers was investigated by studying their permeability to the integrity marker mannitol (Fig.2). The Papp of mannitol remained constant in the apical pH interval 4.8 to 8.0 (1.51 ± 0.26 · 10−7 cm/s; n = 32;p > .05). Thus, the integrity of the Caco-2 monolayers was unaffected in this interval, which covers the variability in pH found in the human intestine in vivo (pH 5.5–7.5;Evans et al., 1988; Dressman et al., 1990). However, a significant increase in cell monolayer permeability to mannitol was observed at pH values of 4 (p < .001) and 4.5 (p < .05), indicating an effect on the Caco-2 cell monolayer integrity at pH < 4.8. Based on these results, an apical pH interval of 5.0 to 8.0 was selected for additional studies to include as wide range as possible for fu of the model drugs (fu of alfentanil varies from 0.03 to 0.97 and fu of cimetidine from 0.02 to 0.95 in this pH interval).

Figure 2
Figure 2

Permeability of Caco-2 cell monolayers to the hydrophilic marker mannitol was used to study the effect of pH on the integrity of the monolayers. Papp of mannitol was constant in the pH interval 5.0 to 8.0; i.e., throughout the physiological pH range in the human intestine (pH 5.5–7.5). A significant increase in monolayer permeability was observed at pHs of 4.5 and lower. Values are presented as mean ± S.D., n = 3 to 5. (*p < .05, and ***p < .001)

pH-Dependent Transport of Alfentanil, a Rapidly Transported Drug.

To confirm that the transport of alfentanil was predominantly passive, two types of experiments were performed. First, the contribution of polarized transport of alfentanil was investigated by comparing the Pc values of the drug in the a-b (absorption) and b-a (secretion) directions at pH 7.4 in both chambers. A b-a/a-b ratio of 1.0 indicates transport entirely by passive diffusion in Caco-2 monolayers (Karlsson et al., 1993; Burton et al., 1996). The rate of transport of alfentanil (1.2 mM) in the b-a direction was slightly higher than that in the a-b direction (p < .05), giving a b-a/a-b ratio of 1.1, Fig. 3a. However, because this ratio was only marginally higher than 1.0, it was concluded that active secretion is not likely to affect the passive transport of alfentanil in the a-b direction across the Caco-2 monolayers at this pH.

Figure 3
Figure 3

a, Caco-2 cell monolayer permeability to alfentanil (1.2 mM) at pH 7.4 in the a-b and b-a direction. The b-a/a-b ratio at pH 7.4 was low (1.1) and active secretion (b-a) was not considered to influence the passive transport of the drug in the a-b direction. b, b-a Pc of alfentanil determined at pH 5.0 and 8.0 on the apical side of the monolayers. The b-a permeability was found to be independent of apical pH. Values are presented as mean ± S.D.,n = 7 to 12. (comparison between a-b and b-a Pc values: *p < .05)

Second, the effect of the apical pH on b-a transport of alfentanil was investigated by comparing the b-a Pc values at the two extreme apical pH values, 5.0 and 8.0, with that at pH 7.4. The rate of transport of alfentanil in the b-a direction was significantly greater when the apical pH was reduced to 5.0 (p < .05; data not shown). However, alfentanil is ionized to a higher degree in the apical receiver chamber at pH 5.0 than at the more basic pH values of 7.4 and 8.0. This results in reduced back flux to the basolateral donor chamber and increased net transport in the b-a direction. Thus, when Pc in the b-a direction at pH 5.0 was corrected for the higher ionization of alfentanil in the apical chamber, no difference was observed between this and b-a Pc values for alfentanil at apical pHs of 7.4 and 8.0 (p > .05; Fig. 3b). These data indicate that transport of alfentanil through the monolayers in the a-b direction is not affected by polarized transport under the conditions used in this study. Therefore, it was concluded that alfentanil was suitable as a rapidly transported model drug for studies of pH-dependent passive transport.

The effects of pH on transport of alfentanil through the intestinal epithelium in the a-b direction were thus investigated. The Pc of alfentanil increased from 11.8 ± 0.21 · 10−6 cm/s at pH 5.0 to 327 ± 19 · 10−6 cm/s at pH 8.0. When the Pc values were plotted as a function of fu, a linear relationship was obtained (Fig.4a). The equation for this relationship was:Pc=2.10+321·fu (r2=0.983). Equation 9As can be seen from the general expression of this relationship in eq. 8, the intercept in Fig. 4a, 2.10 ± 0.54 · 10−6 cm/s (p < .001), is equal to Pc,i. A Pc,u value of 323 ± 7 · 10−6 cm/s was obtained from the slope of the graph. Thus, Pc,u was 154-fold higher than Pc,i, supporting the common assumption that permeation of the ionized form of the drug contributes only marginally to the total Pc. In the pH interval 5.0 to 8.0, the maximal relative contribution of the ionized form to Pc is obtained at pH 5.0, where Pc,i corresponds to 17% of the overall transport, Fig. 4b. In conclusion, the effect of the apical pH on the passive transport of alfentanil across Caco-2 cell monolayers can be attributed to changes in the fu of the drug in the donor solution.

Figure 4
Figure 4

pH-dependent transport of the rapidly transported model drug alfentanil across Caco-2 cell monolayers. a, cell monolayer permeability to alfentanil as a function of the fraction of the drug in un-ionized form on the apical side of the monolayers. A linear relationship was obtained (r2 =. 983); the Pc of the ionized form of alfentanil (2.10 ± 0.54 · 10−6 cm/s; p < .001) was obtained from the intercept. The monolayer permeability to the un-ionized form (323 ± 7 · 10−6 cm/s) was obtained from the slope of the regression line. b, calculated contribution of un-ionized (dotted line) and ionized forms (dashed line) to the calculated overall permeability of the monolayers to alfentanil (solid line) as a function of pH. (Observed Pcvalues are represented as ▪.) The contribution of the ionized form to the total transport was small. In the studied pH range, the largest relative contribution of the ionized form to the total transport (17%) was observed at pH 5.0. Experimentally determined values are presented as mean ± S.D., n = 6 to 12.

pH-Dependent Transport of Cimetidine, a Slowly Transported Drug.

As with alfentanil, the influence of active secretory mechanisms on the pH-dependent passive transport of cimetidine across the Caco-2 monolayers was studied initially. The Caco-2 monolayers were significantly more permeable to cimetidine (1 mM) in the b-a direction than they were in the a-b direction (p < .001; Fig. 5a). The b-a/a-b ratio was 4.4, indicating that active secretion could have an effect on the a-b transport rate (Karlsson et al., 1993; Burton et al., 1996). When the donor concentration of cimetidine was increased to 10 mM, the b-a/a-b ratio decreased to 2.9. Addition of 0.5 mM verapamil, a compound previously used to inhibit the active efflux of cimetidine (Pan et al., 1994), decreased the b-a/a-b ratio to 1.8. Therefore, subsequent transport experiments were carried out in the presence of verapamil. Interestingly, the decrease in the b-a/a-b ratio at pH 7.4 after addition of verapamil was caused entirely by decreased drug transport in the b-a direction, because there was no effect on a-b transport, Fig. 5a. There were no differences in the rate of b-a transport of cimetidine at apical pHs of 5.0, 7.4, and 8.0 (p > .05; Fig. 5b). These data indicate that the effects of polarized secretion are minimal with this drug, and that reliable estimates of the passive transport of cimetidine in the a-b direction can be obtained. Therefore, cimetidine was chosen as a model drug for studies of the pH-dependent passive transport of slowly transported compounds.

Figure 5
Figure 5

a, Caco-2 cell monolayer permeability to cimetidine in the a-b and b-a direction at pH 7.4. Pc values obtained at 1 and 10 mM cimetidine and at 1 mM cimetidine in the presence of verapamil are presented. At an apical pH of 7.4, the Pcvalues determined in the b-a direction was significantly higher than in the a-b direction. The b-a/a-b ratio decreased in the order 1 mM cimetidine > 10 mM cimetidine > 1 mM cimetidine + 0.5 mM verapamil. Therefore, 0.5 mM verapamil was added to limit the influence of active secretion in all subsequent transport experiments. The a-b transport was, however, found to be independent of the b-a/a-b ratio (p > .5). b, cell monolayer permeability to cimetidine in the b-a direction at pH 5.0 and 8.0 on the apical side of the monolayers. The b-a transport was not affected by the pH on the apical side (p > .5). Therefore, the a-b Pc values (obtained in the presence of verapamil) were used as estimates of passive transport in the subsequent analysis of the pH-dependent transport of cimetidine. Values are presented as mean ± S.D., n = 6 to 32. (comparison between a-b and b-a Pc values: **p < .01 and ***p < .001)

The pH-dependent transport of cimetidine in the a-b direction was thus investigated. The Pc of cimetidine increased from 0.072 ± 0.011 · 10−6 cm/s at pH 5.0 to 1.67 ± 0.13 · 10−6 cm/s at pH 8.0. A nearly linear relationship between Pc and fu was obtained in the pH interval 5.0 to 8.0 (Fig. 6a):Pc=0.045+1.44·fu (r2=0.946). Equation 10Substitution in the general expression of this relationship in eq. 8 resulted in a Pc,i of 0.045 ± 0.016 · 10−6 cm/s (p < .001) and a Pc,u of 1.49 ± 0.11 · 10−6 cm/s (33-fold higher than Pc,i). Thus, the ionized form of cimetidine contributed relatively more to the total transport of cimetidine than the ionized form of alfentanil contributed to the total transport of alfentanil (Figs. 4b and 6b). At lower pH values where fu decreases, the contribution of the ionized form to the total transport increases. Thus, at pH 5.0, more than 60% of the overall transport of cimetidine is attributable to Pc,i, Fig. 6b. In conclusion, the pH-dependent passive permeation of cimetidine is a result of changes in fu with pH in the donor solution.

Figure 6
Figure 6

pH-dependent transport of the slowly transported model drug cimetidine across Caco-2 cell monolayers. a, cell monolayer permeability to cimetidine as a function of the fraction of the drug in un-ionized form on the apical side of the monolayers. A nearly linear relationship was obtained (r2=. 946); the Pc of the ionized form of cimetidine (0.045 ± 0.016 × 10−6 cm/s;p < .001) was obtained from the intercept. The monolayer permeability to the un-ionized form (1.49 ± 0.11 × 10−6 cm/s) was obtained from the slope of the regression line. b, calculated contribution of un-ionized (dotted line) and ionized forms (dashed line) to the calculated overall permeability of the monolayers to cimetidine (solid line) as a function of pH. (Observed Pc values are represented as ▪.) The contribution of the ionized form to the total transport increased at lower pH values: at pH 5.0, 66% of the total transport was attributable to the ionized form. Experimentally determined values are presented as mean ± S.D., n = 6 to 12.

Calculations of Pc of Un-ionized Form of Drugs.

The pH-partition theory suggests that the transport of the ionized form of drugs is negligible. This implies that Pc,u can be calculated from Pc values observed at a single pH value, because the ratio Pc/fu would be constant and equal Pc,u (eq. 5). The significance of the transport of the ionized form when fu is small (as it is for the basic drugs in this study at lower pH values) was therefore investigated by plotting Pc/fu as a function of pH (Fig.7). At pHs ≤ 5.5, the ratio Pc/fu is significantly higher than Pc,u values calculated according to eq. 8 for both alfentanil (Fig. 7a) and cimetidine (Fig. 7b). Thus, because of the significant contribution of the transport of the ionized form, Pc,u, calculated as Pc/fu, will be overestimated if the transport experiments are performed when fu < 0.1. However, the Pc,u of alfentanil calculated as Pc/fu from Pc values determined at pH 5.0 and 5.5 are only 1.2-fold higher than Pc,u calculated according to eq. 8. For cimetidine, the greater contribution of the ionized form to the total transport results in a 3.1- and 1.4-fold overestimation of Pc,u from Pc/fu at pH 5.0 and 5.5, respectively. Thus, neglecting the transport of the ionized form in calculations of Pc,u from Pc values obtained at pH values corresponding to fu < 0.1 leads to overestimation of Pc,u, especially for slowly transported compounds such as cimetidine.

Figure 7
Figure 7

Observed monolayer permeability to: a, alfentanil and b, cimetidine, corrected for the fraction of the drug in un-ionized form, Pc/fu (●), as a function of pH. If the epithelium was impermeable to the ionized form, the Pc/fu ratio would equal Pc, u(dotted line) and be independent of pH. However, the simultaneous increase in the contribution of the ionized form of the drug to the total transport and decrease in fu at lower pH values, results in a significant contribution of the ionized form to the Pc/fu ratio (dashed line). Thus, the Pc/fu ratio is increased at lower pH values (<5.5). Therefore, Pc, u is overestimated from Pc values obtained at fu < 0.1 if the transport of ionized drug molecules is not considered. Experimentally determined values are presented as mean ± S.D.,n = 6 to 12. (comparison with determined values of Pc, u: *p < .05 and ***p < .001).

Prediction of the Pc of the Un-ionized Form of Drugs from Theoretically Calculated pKaValues.

In the early stages of the drug development process, the absorption of candidate drugs through the intestinal epithelium is determined at single pH values in vitro. Experimental values of pKa (required to calculate fu) are not always available at this early stage. We therefore investigated the accuracy of predicting the Pc,u of the model drugs from Pc determined at a single pH value and theoretical estimates of pKa and fu. Experimental pKa values and the pKa values calculated by the MolSurf program (pKa,calc) are presented in Table 1. The pKa values of alfentanil and cimetidine were underestimated by 0.7 to 0.8 U. Pc values determined at pH 6.5 were selected because pH 6.5 is in the middle of the studied pH interval, the acidic microclimate in the small intestine is close to pH 6.5 (Lucas and Blair, 1978), and fu > 0.1 at this pH. This procedure resulted in a predicted Pc,ufor alfentanil of 179 · 10−6 cm/s, or 1.8-fold lower than the experimentally determined Pc,u. A Pc,ufor cimetidine of 0.61 · 10−6 cm/s was predicted from pKa,calc. This value is 2.5-fold lower than the experimentally determined Pc,u.

Discussion

The results of this study support modification of the pH-partition theory to incorporate transport of both un-ionized and ionized forms of drugs across the intestinal epithelium. Transport of the ionized form contributed significantly to the total drug transport at pH values where the fu of the model drugs was < 0.1. Because many drugs have an fu < 0.1 in the entire physiological pH range in the intestine, the transport of the ionized forms of drugs should be considered to a greater extent than suggested by the traditional pH-partition theory.

The Caco-2 cell monolayer permeability to the paracellular marker mannitol was unaffected in the apical pH range 5.0 to 8.0. Because the pKa values of the phosphate groups in the cell membrane phospholipids are approximately 2 (Bangham, 1968), net charge of the cell membrane should be independent of pH in the present study. Only minor deviations from the theoretical relationships were observed in the pH-permeability curves of the model drugs, indicating that pH had no significant effect on the integrity of the cell membrane. Thus, the pH-dependent changes in drug permeation observed in this study were the result of changes in the fuof the drugs and not of pH-dependent changes in cell membrane properties.

If the cell membrane is assumed to be impermeable to the ionized form, the obtained Pc,i values represent the paracellular transport of the ionized species. Pc,i for alfentanil was considerably higher than that for cimetidine. However, in theory, the larger molecular size of alfentanil (mw = 416) compared with cimetidine (mw = 252), should result in lower paracellular transport of the ionized form of alfentanil (according to a size-restricted diffusion of the paracellular pathway).

One possible explanation for this unexpected result is that the values of pKa obtained from the literature are not applicable to the experimental conditions used in this study. Because Pc,i is determined by extrapolation of the relationship between fu and Pc, the method of calculating Pc,i will be sensitive to the values of pKa used in the analysis, especially for alfentanil. However, the pKa of alfentanil (6.5) used in the determination of Pc,i is in good agreement with the inflection point of the experimentally obtained pH-permeability curve for alfentanil (6.4; Fig. 4b). Moreover, although the pH-permeability curve for cimetidine (Fig. 6b) indicates a pKa of 7.1 under these experimental conditions, calculations based on this pKavalue do not result in a markedly different value for Pc,i (0.072 ± 0.029 · 10−6 cm/s) than that obtained using a pKa value of 6.8 (Pc,i = 0.045 ± 0.016 · 10−6 cm/s). Thus, the use of the literature to obtain values of pKa for use in the calculations does not explain the difference in Pc,i between alfentanil and cimetidine.

Another explanation could be that the ionized form of alfentanil can be transported across the epithelium transcellularly, although at a much slower rate than the un-ionized form. However, the pKa values of the drug molecules can be different at the amphiphilic membrane interface compared with the aqueous solution pKa values (Retzinger et al., 1986). Such a shift in pKa will result in an apparent weaker base at the membrane interface and therefore in an increased transcellular transport. Theoretically, the pKa-shift is likely to be associated with different conformational preferences in the aqueous solution and at the amphiphilic interface. A less polar environment will promote conformations more prone to form intramolecular hydrogen bonds. Indeed, calculations of pKa from ionization energies indicated that intramolecular hydrogen bonding significantly influenced the pKa of the identified low-energy conformations of alfentanil and cimetidine.

The rate of transport of the un-ionized forms of the drugs was 150- and 30-fold higher than the rate of transport of the ionized forms of alfentanil and cimetidine, respectively. In early in vitro studies, the transport of the un-ionized form of weak acids across rat small intestinal tissue was found to be 10,000-fold more rapid than the transport of the ionized form (Tai and Jackson, 1981). The selectivity of the rat gastric mucosa for the un-ionized form of weak bases was, however, found to be considerably smaller, approximately 500-fold (Tai and Jackson, 1982). The lower epithelial selectivity observed for the un-ionized form of bases compared with acids may be the result of the higher paracellular permeability of intestinal epithelia for cationic compared with anionic drug molecules (e.g., Adson et al., 1994;Karlsson et al., 1994).

For the basic model drugs investigated in this study, the intestinal epithelial Caco-2 cell monolayers displayed an even lower selectivity for the un-ionized form than previously observed for rat gastrointestinal epithelia. Because ionized molecules will presumably permeate the epithelium mainly via the paracellular route, the selectivity of the epithelium for the un-ionized form of drugs (Pc,u/Pc,i) is dependent on the paracellular permeability of the epithelium. The paracellular route in Caco-2 monolayers is more permeable than that in gastric mucosa, which might explain the lower selectivity observed in Caco-2 monolayers compared with rat mucosa. However, apart from the level of paracellular permeability, the selectivity of the tissue will also be highly dependent on the level of transcellular permeability to the un-ionized form of drugs (Jackson, 1987). This was also observed in this study, where the Caco-2 monolayers were less selective for the un-ionized form of the hydrophilic drug cimetidine than for the un-ionized form of the more lipophilic drug alfentanil.

An important implication of these results is that a more generally applicable model for predictions of human intestinal permeability to drugs from permeability studies in Caco-2 monolayers or from the molecular structure of the drug must account for the effect of drug ionization. This can be done in at least two ways. One approach is to compensate for the degree of ionization by determination of the Pc,u of the compounds. Because the contribution of the ionized form to the total transport is small, Pc,u can be calculated by correcting the observed Pc for the fraction of the drug in the un-ionized form (Pc/fu) when fu > 0.1. However, many drugs are bases with pKa values > 9 or acids with pKa values < 4 (Jackson, 1987). For this category of drugs, calculation of Pc,ufrom Pc determinations at a single pH value poses a problem because fu of the drugs is < 0.1 throughout the entire pH range of 5.0 to 8.0. For example, in the structurally diversified set of compounds used to investigate the relationship between the fraction absorbed after oral administration to humans and structural descriptors (Palm et al., 1997), 12 of 20 compounds fall into this category. Thus, Pc,uwould be difficult to determine for a large number of drugs. Therefore, a relationship between Pc,u and structural descriptors will be difficult to establish for compounds of greater structural diversity.

The second approach to accounting for the degree of ionization is to include a structural descriptor of molecular charge in the structure-permeability models. In the traditional structure-permeability relationship between octanol/water partition coefficients and epithelial permeability to drugs, the charge effect is accounted for by including the apparent distribution coefficients (logD) at the relevant pH. Because organic solvents are more selective for the un-ionized form of the solutes (of the order of 106; see Jackson, 1987) than is the intestinal epithelium (see discussion above), logD values will frequently exaggerate the effect of ionization. The use of an alternative structural descriptor of molecular charge, such as pKa or fu, is therefore more attractive. Predictions of intestinal drug absorption from calculated structural descriptors would then depend on computational methods to estimate the pKavalues of the drugs. In this study, the pKavalues calculated from ionization energies gave promising results. However, further validation of this method and other computational methods of calculating pKa are needed (Leahy et al., 1997).

In summary, the results of this study clearly show that the degree of ionization of the drugs under investigation has a pronounced influence on their permeation through the intestinal epithelium. The implications of this for predictions of the in vivo absorption of proteolytic drugs after oral administration remain to be investigated.

Footnotes

  • Send reprint requests to: Per Artursson, Department of Pharmacy, Box 580, Uppsala University, SE-751 23 Uppsala, Sweden. E-mail: per.artursson{at}galenik.uu.se

  • ↵1 This work was supported by Grant 95–98 from Centrala Försöksdjursnämnden, Grant 9478 from The Swedish Medical Research Council, Grant K11163-302 from The Swedish Natural Science Research Council, and The Swedish Fund for Scientific Research without Animals.

  • ↵2 Present address: Department of Internal Medicine, University Hospital Groningen, Groningen, the Netherlands.

  • Abbreviations:
    Pc
    permeability coefficient of the cell monolayer
    fu
    fraction of drug in un-ionized form
    Pc,u
    cell monolayer permeability coefficient of the un-ionized form
    Pc,i
    cell monolayer permeability coefficient of the ionized form
    HBSS
    Hanks’ balanced salt solution
    a-b
    apical-to-basolateral
    b-a
    basolateral-to-apical
    Papp
    apparent permeability coefficient
    CD
    concentration in the donor compartment
    CR
    concentration in the receiver compartment
    fi
    fraction of drug in ionized form
    MES
    2-(N-morpholino)ethanesulfonic acid
    • Received March 8, 1999.
    • Accepted June 2, 1999.

References

  1. ↵
    1. Adson A,
    2. Raub TJ,
    3. Burton PS,
    4. Barsuhn CL,
    5. Hilgers AR,
    6. Audus KL,
    7. Ho NF
    (1994) Quantitative approaches to delineate paracellular diffusion in cultured epithelial cell monolayers. J Pharm Sci 83:1529–1536.
    OpenUrlCrossRefPubMed
  2. ↵
    1. Artursson P
    (1990) Epithelial transport of drugs in cell culture. I: A model for studying the passive diffusion of drugs over intestinal absorbtive (Caco-2) cells. J Pharm Sci 79:476–482.
    OpenUrlCrossRefPubMed
  3. ↵
    1. Avdeef A,
    2. Box KJ,
    3. Comer JEA,
    4. Hibbert C,
    5. Tam KY
    (1998) pH-metric logP 10. Determination of liposomal membrane-water partition coefficients of ionizable drugs. Pharm Res 15:209–215.
    OpenUrlCrossRefPubMed
  4. ↵
    1. Bangham AD
    (1968) Membrane models with phospholipids. Prog Biophys Mol Biol 18:29–95.
    OpenUrlCrossRef
  5. ↵
    1. Burton PS,
    2. Conradi RA,
    3. Ho NFH,
    4. Hilgers AR,
    5. Borchardt RT
    (1996) How structural features influence the biomembrane permeability of peptides. J Pharm Sci 85:1336–1340.
    OpenUrlCrossRefPubMed
  6. ↵
    1. Drayton CJ
    1. Craig PN
    (1990) Drug Compendium. in Cumulative Subject Index & Drug Compendium: Vol. 6, ed Drayton CJ (Pergamon Press, Oxford), pp 237–991.
  7. ↵
    1. Dressman JB,
    2. Berardi RR,
    3. Dermentzoglou LC,
    4. Russell TL,
    5. Schmaltz SP,
    6. Barnett JL,
    7. Jarvenpaa KM
    (1990) Upper gastrointestinal (GI) pH in young, healthy men and women. Pharm Res 7:756–761.
    OpenUrlCrossRefPubMed
  8. ↵
    1. Dutt A,
    2. Priebe TS,
    3. Teeter LD,
    4. Kuo MT,
    5. Nelson JA
    (1992) Postnatal development of organic cation transport and mdr gene expression in mouse kidney. J Pharmacol Exp Ther 261:1222–1230.
    OpenUrlAbstract/FREE Full Text
  9. ↵
    1. Ellens H,
    2. Eddy EP,
    3. Lee CP,
    4. Dougherty P,
    5. Largo A,
    6. Xiang JN,
    7. Elliott JD,
    8. Cheng HY,
    9. Ohlstein E,
    10. Smith PL
    (1997) In vitro permeability screening for identification of orally bioavailable endothelin receptor antagonists. Adv Drug Deliv Rev 23:99–109.
  10. ↵
    1. Evans DF,
    2. Pye G,
    3. Bramley R,
    4. Clark AG,
    5. Dyson TJ,
    6. Hardcastle JD
    (1988) Measurements of gastrointestinal pH profiles in normal ambulant human subjects. Gut 29:1035–1041.
    OpenUrlAbstract/FREE Full Text
  11. ↵
    1. Hogben CAM,
    2. Tocco DJ,
    3. Brodie BB,
    4. Schanker LS
    (1959) On the mechanism of intestinal absorption of drugs. J Pharmacol Exp Ther 125:275–282.
    OpenUrlAbstract/FREE Full Text
  12. ↵
    1. Högerle ML,
    2. Winne D
    (1983) Drug absorption by the rat jejunum perfused in situ: Dissociation from the pH-partition theory and role of microclimate-pH and unstirred water layer. Naunyn- Scmiedeberg’s Arch Pharmacol 322:249–255.
  13. ↵
    1. Iseki K,
    2. Hirano T,
    3. Fukushi Y,
    4. Kitamura Y,
    5. Miyazaki S,
    6. Takada M,
    7. Sugawara M,
    8. Saitoh H,
    9. Miyazaki K
    (1992) The pH-dependent uptake of enoxazin by rat intestinal brush-border membrane vesicles. J Pharm Pharmacol 44:722–726.
    OpenUrlPubMed
  14. ↵
    1. Johnson LR
    1. Jackson MJ
    (1987) Drug transport across gastrointestinal epithelia. in Physiology of the Gastrointestinal Tract, ed Johnson LR (Raven Press, New York), pp 1597–1621.
  15. ↵
    1. Karlsson J,
    2. Artursson P
    (1991) A method for the determination of cellular Pcs and aqueous boundary layer thickness in monolayers of intestinal epithelial (Caco-2) cells grown in permeable filter chambers. Int J Pharm 71:55–64.
  16. ↵
    1. Karlsson J,
    2. Kuo SM,
    3. Ziemniak J,
    4. Artursson P
    (1993) Transport of celiprolol across human intestinal epithelial (Caco-2) cells: Mediation of secretion by multiple transporters including P-glycoprotein. Br J Pharmacol 110:1009–1016.
    OpenUrlPubMed
  17. ↵
    1. Karlsson J,
    2. Ungell AL,
    3. Artursson P
    (1994) Effect of an oral rehydration solution on paracellular drug transport in intestinal epithelial cells and tissues: Assessment of charge and tissue selectivity. Pharm Res 11:S248.
    OpenUrl
  18. ↵
    1. Leahy DE,
    2. Duncan R,
    3. Ahr HJ,
    4. Bayliss MK,
    5. de Boer ABG,
    6. Darvas F,
    7. Fentem JH,
    8. Fry JR,
    9. Hopkins R,
    10. Houston JB,
    11. Karlsson J,
    12. Keddris GL,
    13. Pratten MK,
    14. Priteo P,
    15. Smith DA,
    16. Straughan DW
    (1997) Pharmacokinetics in early drug research. Alternatives to Laboratory Animals 25:17–31.
    OpenUrl
  19. ↵
    1. Lucas ML,
    2. Blair JA
    (1978) The magnitude and distribution of the acid microclimate in proximal jejunum and its relation to luminal acidification. Proc R Soc Lond B Biol Sci 200:27–41.
    OpenUrlPubMed
  20. ↵
    1. Miyamoto Y,
    2. Ganapathy V,
    3. Leibach FH
    (1988) Transport of guanidine in rabbit intestinal brush-border membrane vesicles. Am J Physiol 255:G85–G92.
    OpenUrlAbstract/FREE Full Text
  21. ↵
    1. Mummaneni V,
    2. Dressman JB
    (1994) Intestinal uptake of cimetidine and ranitidine in rats. Pharm Res 11:1599–1604.
    OpenUrlCrossRefPubMed
  22. ↵
    1. Norinder U,
    2. Österberg T,
    3. Artursson P
    (1997) Theoretical calculation and prediction of Caco-2 cell permeability using MolSurf parametrization and PLS statistics. Pharm Res 14:1786–1791.
    OpenUrlCrossRefPubMed
  23. ↵
    1. Ottiger C,
    2. Wunderli-Allenspach H
    (1997) Partition behaviour of acids and bases in a phosphatidylcholine liposome-buffer equilibirum dialysis system. Eur J Pharm Sci 5:223–231.
    OpenUrlCrossRef
  24. ↵
    1. Pade V,
    2. Stavchansky S
    (1997) Estimation of the relative contribution of the transcellular and paracellular pathway to the transport of passively absorbed drugs in the Caco-2 cell culture model. Pharm Res 14:1210–1215.
    OpenUrlCrossRefPubMed
  25. ↵
    1. Palm K,
    2. Stenberg P,
    3. Luthman K,
    4. Artursson P
    (1997) Polar molecular surface properties predict the intestinal absorption of drugs in humans. Pharm Res 14:568–571.
    OpenUrlCrossRefPubMed
  26. ↵
    1. Pan BF,
    2. Dutt A,
    3. Nelson JA
    (1994) Enhanced transepithelial flux of cimetidine by Madin-Darby canine kidney cells overexpressing human P-glycoprotein. J Pharmacol Exp Ther 270:1–7.
    OpenUrlAbstract/FREE Full Text
  27. ↵
    1. Retzinger GS,
    2. Cohen L,
    3. Lau SH,
    4. Kézdy FJ
    (1986) Ionization and surface properties of verapamil and several verapamil analogues. J Pharm Sci 75:976–982.
    OpenUrlCrossRefPubMed
  28. ↵
    1. Shiau YF,
    2. Fernandez P,
    3. Jackson MJ,
    4. McMongale S
    (1985) Mechanisms maintaining a low-pH microclimate in the intestine. Am J Physiol 248:G608–G617.
    OpenUrl
  29. ↵
    1. van der Waterbeemd H,
    2. Testa B,
    3. Folkers G
    1. Sjöberg P
    (1997) MOLSURF—A generator of chemical descriptors for QSAR. in Computer-Assisted Lead Finding and Optimization: Current Tools for Medicinal Chemistry, eds van der Waterbeemd H, Testa B, Folkers G (Verlag Helvetica Chimica Acta, Basel, Switzerland), pp 83–92.
  30. ↵
    1. Tai CY,
    2. Jackson MJ
    (1981) Weak-acid transport in the small intestine: Discrimination in the lamina propria. J Membr Biol 59:35–43.
    OpenUrlCrossRefPubMed
  31. ↵
    1. Tai CY,
    2. Jackson MJ
    (1982) Transport of weak bases across rat gastric mucosa in vivo and in vitro. J Pharmacol Exp Ther 222:372–378.
    OpenUrlAbstract/FREE Full Text
View Abstract
Back to top

In this issue

Download PDF
Article Alerts
Email Article
Citation Tools
Research ArticleArticle

Effect of Molecular Charge on Intestinal Epithelial Drug Transport: pH-Dependent Transport of Cationic Drugs

Katrin Palm, Kristina Luthman, Jenny Ros, Johan Gråsjö and Per Artursson
Journal of Pharmacology and Experimental Therapeutics November 1, 1999, 291 (2) 435-443;
del.icio.us logo Digg logo Reddit logo Twitter logo CiteULike logo Facebook logo Google logo Mendeley logo

Jump to section