- Email: [email protected]
A new system for the prediction of drug absorption using a pH‐controlled Caco‐2 model: Evaluation of pH‐dependent soluble drug absorption and pH‐related changes in absorption
A New System for the Prediction of Drug Absorption Using a pH-Controlled Caco-2 Model: Evaluation of pH-Dependent Soluble Drug Absorption and pH-Related Changes in Absorption XIN HE, SHOTA KADOMURA, YOH TAKEKUMA, MITSURU SUGAWARA, KATSUMI MIYAZAKI Department of Pharmacy, Hokkaido University Hospital, School of Medicine, Kita 14 jo, Nishi 5 chome, Kita-ku, Sapporo 060-8648, Japan
Received 1 April 2003; revised 28 May 2003; accepted 11 June 2003
ABSTRACT: One purpose of this study was to develop a new system for the prediction of pH-dependent soluble drug absorption that takes into account the physiological condition of the gastrointestinal tract. Another purpose was to establish several models of different gastric acidities: a normal gastric acidity model, a low gastric acidity model (a model of achlorhydria), a temporarily elevated gastric acidity model (a model of a case in which an acidic drug was coadministered to temporarily elevate gastric acidity in the case of low gastric acidity), a weak antacid model (a model of a case in which a weak antacid drug, such as an H2 receptor antagonist, was coadministered to temporarily elevate pH up to 6), and a strong antacid model (a model of a case in which a strong antacid drug, such as magnesium hydroxide, was coadministered to temporarily elevate pH up to 8.0). These models were used to evaluate variation in pH-related absorption in humans. Dipyridamole preparation (Persantin1 tablets) and glibenclamide preparation (Euglucon1 tablets), both poorly water-soluble and pH-dependent soluble drugs, were chosen as model drugs to determine whether absorption is altered by changes in levels of gastric acid. The extent of absorption of dipyridamole was remarkably lower when gastric pH was continuously elevated to 6.0, whereas it was increased when gastric pH temporarily decreased to 1.8. The extent of absorption of glibenclamide increased dramatically when gastric pH temporarily increased to 8.0, but did not change when gastric pH temporarily increased to 6.0. These results are consistent with reported results obtained in clinical studies. The results suggest that pH-related variations in absorption in humans can be accurately predicted using our new system. ß 2004 Wiley-Liss, Inc. and the American Pharmacists Association J Pharm Sci 93:71–77, 2004
Keywords: prediction; Caco-2 cells; in vitro models; oral absorption; permeability; drug interactions; solubility
INTRODUCTION Various factors cause variation in pH in the gastrointestinal tract. It is well known that oral bioavailability is influenced greatly by solubility
Correspondence to: Katsumi Miyazaki (Telephone: 81-117065680; Fax: 81-11- 7561505; E-mail: [email protected]) Journal of Pharmaceutical Sciences, Vol. 93, 71–77 (2004) ß 2004 Wiley-Liss, Inc. and the American Pharmacists Association
and stability of drugs, which depend on pH. Gastric acid in humans has pH of 1–2. Variations in gastric acidity results in changes in the absorption profiles of drugs with pH-sensitive dissolution. We have developed an in vitro system for prediction of drug absorption that takes into account dissolution of solid drugs and pH changes in the gastrointestinal tract.1 This system enables prediction of the oral absorption of a relatively water-soluble drug in humans based on the cumulative permeation of the drug across a
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 93, NO. 1, JANUARY 2004
71
72
HE ET AL.
Caco-2 monolayer.2 The primary aim of this study was to develop a new system for predicting pHdependent soluble drug absorption that takes into account physiological conditions in the gastrointestinal tract. The second aim was to evaluate variation in pH-related absorption in humans. We established several models with different gastric acidities: a normal gastric acidity model, a low gastric acidity model (a model of achlorhydria), a temporarily elevated gastric acidity model (a model of a case in which an acidic drug was coadministered to temporarily elevate gastric acidity), a weak antacid model (a model of a case in which a weak antacid drug, such as an H2 receptor antagonist, was coadministered to temporarily elevate pH up to 6), and a strong antacid model (a model in which a strong antacid drug, such as magnesium hydroxide, was coadministered to temporarily elevate pH up to 8.0). Dipyridamole preparation (Persantin1 tablets) and glibenclamide preparation (Euglucon1 tablets) were chosen as model drugs to determine whether absorption is altered by variations in levels of gastric acidity in these models. The extent of absorption of dipyridamole, a poorly soluble weak base with a reported pKa of 6.4,3 was reported to be altered to a considerable extent by the pH of different digestive fluids;4 that is, dipyridamole dissolves readily in the stomach, but incompletely in the intestine. The absorption of dipyridamole is dissolution rate limited over at least a part of the usual physiological range of gastrointestinal pH. Because achlorhydric subjects have high pH values in both the stomach and the intestine regardless of the prandial phase,5 and because the dissolution rate of dipyridamole is very low in that pH range, achlorhydrics exhibit a slow rate and poor extent of absorption compared with subjects who have a low gastric pH. Furthermore, it would be beneficial for achlorhydrics to take glutamic acid hydrochloride prior to taking dipyridamole for complete absorption.6 Glibenclamide, a weakly acidic drug with a reported pKa of 6.8, does not dissolve at pH 1–6. However, its solubility increases at pH > 7.7 Comparable enhancement of glibenclamide absorption can be achieved by elevating gastric pH.8 The oral absorption of glibenclamide, therefore, did not change when ranitidine, a weak antacid drug that increases the pH value up to 6.0, was coadministered.9 On the other hand, coadministration of magnesium hydroxide, a strong antacid reagent that increases pH in the stomach up to 8.0,10 resulted in a dramatic increases in the extent of JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 93, NO. 1, JANUARY 2004
absorption of glibenclamide.8 This study predicted changes that occur in the extent of absorption of poorly water-soluble and pH-dependent soluble drugs in people with abnormal levels of gastric acidity and in people with normal levels of gastric acidity using the models just described.
EXPERIMENTAL Materials Glibenclamide preparation (Euglucon1 tablets, each 2.5 mg) was purchased from Yamanouchi Pharmaceutical Company, Ltd. (Tokyo, Japan). Dipyridamole preparation (Persantin1 tablets, each 12.5 mg) was purchased from Nippon Boehringer Ingelheim Company, Ltd. (Hyogo, Japan). Fluorescein isothiocyanate (FITC)-dextran (molecular weight, 4400) was purchased from Sigma Chemical Company (St. Louis, MO). System for Predicting Drug Absorption As shown in Figure 1, our in vitro system for predicting drug absorption takes into account drug dissolution and change in pH in the gastrointestinal tract. In this system, a solid drug is added to a drug-dissolving vessel (DDV; modeled stomach, 10 mL), and the dissolved drug is transferred to a pH adjustment vessel (PAV; modeled intestine, 10 mL). Each of these vessels is a plastic vial. The compositions of the drug-dissolving solution, pH adjustment solution, and receiver solution are shown in Table 1. The flow rate (0.5 mL/min) of
Figure 1. Scheme of the drug absorption prediction system.
NEW MODEL SYSTEM TO PREDICT DRUG ABSORPTION
73
Table 1. Compositions of Flowing Solutions (mM)
Component KCl KH2PO4 NaCl Na2HPO4 D-glucose CaCl2 MgSO4 MES HEPES HCl NaOH
Drug-Dissolving Solution (pH 1.8)
pH Adjustment Solutiona
Receiver Solutionb (pH 7.4)
— — 61.4 — 50 2.52 0.81 — — 13.6 —
10.7 0.88 199 0.68 — — — 40 — — 13.6
5.37 0.44 137 0.34 25 1.26 0.41 — 10 — —
a pH was adjusted to 6.0 (models A, B, and C) or 7.5 (models D, E, and F) with Tris before addition of NaOH. b pH was adjusted to 7.4 with Tris.
each solution is controlled by a peristaltic pump. The drug solution is transferred to the donor compartment of side-by-side diffusion chamber (Corning Costar Company). To prevent the inflow of undissolved powder from the PAV, glass wool is set between the PAV and the donor compartment of side-by-side diffusion chamber. Mounted between the donor and receiver compartments is a Caco-2 monolayer grown on a Snapwell (0.4 mm in pore size, 1 cm2 in growth area; Corning Costar Company). Samples of dipyridamole or glibenclamide are collected every 10 min over a period of 300 min. A silicon tube (i.d., 0.5 mm) is used to connect each vessel and the compartment. All of the solutions and both vessels and both compartments are preheated to 378C and maintained at that temperature.
absorption of dipyridamole due to changes in pH in the gastrointestinal tract was evaluated by using these models. Glibenclamide is sparingly soluble in water at gastric pH, whereas its solubility increases at pH >7.7 Also, glibenclamid is absorbed not only in
Models As shown in Figure 2, the normal gastric acidity model (model A) was used for prediction of pHdependent soluble drug absorption that takes into account the physiological condition of the gastrointestinal tract. The DDV modeled stomach, the PAV modeled intestine, and the drug-dissolving solution modeled gastric juice. Dipyridamole is a pH-dependent soluble drug that dissolves readily in acidic conditions, whereas its solubility decreases dramatically at pH of 3.0. So, we established a low gastric acidity model (model B, model of achlorhydria) and a temporarily elevated gastric acidity model (model C), in which gastric acidity was elevated temporarily by coadministration of an acidic drug. The variation in the oral
Figure 2. Scheme for models of different gastric acidities: (A) normal gastric acidity model; (B) low gastric acidity model; (C) temporarily elevated gastric acidity model; (D) normal gastric acidity model (PAVpH7.5); (E) weak-antacid model; and (F) strong-antacid model. JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 93, NO. 1, JANUARY 2004
74
HE ET AL.
the stomach and small intestine but also in the large intestine (pH 7–8).11 Furthermore, the integrity of the Caco-2 monolayers was unaffected in the apical pH interval 4.8–8.0.12 Considering these facts, the pH value of the solution in the PAV was adjusted to 7.5. The weak antacid model (model E) is a model of a case in which a weak antacid drug, such as ranitidine, was coadministered to temporarily elevate pH up to 6.0. The strong antacid model (model F) is a model of a case in which a strong antacid drug, such as magnesium hydroxide, was coadministered to temporarily elevate pH up to 8.0. As a control experiment, a model of normal gastric acidity (model D) was established using a PAV solution with pH of 7.5. Cell Culture Caco-2 cells were purchased from American Type Culture Collection (Rockville, MD) and maintained in plastic culture flasks (Falcon, Becton Dickinson and Company, Lincoln Park, NJ). The stock cells were subcultivated before reaching confluence. The growth medium used was Dulbecco’s modified Eagle’s medium (DMEM; Sigma) supplemented with 10% fetal bovine serum (ICN Biomedicals, Inc., Aurora, OH), 1% nonessential amino acids (Gibco), and 4 mM glutamine without antibiotics. The monolayer cultures were grown in a CO2 incubator (5% CO2) at 378C. The cells were harvested with 0.25% trypsin and 0.2% EDTA (0.5–1 min at 378C), resuspended, and seeded into a new flask. Cells between the 35th and 50th passages were used in this study. For the transport studies, Caco-2 cells were seeded on Snapwell at a cell density of 8 104 cells/ filter. The cell monolayers were fed a fresh growth medium every 2 days and were used on days 20– 28 for the transport experiments. To evaluate the integrity of the monolayer, transepithelia electrical resistance (TEER) was measured with a Millicell-ERS (Millipore Company, Bedford, MA). The TEER of the filter was subtracted from the measured total TEER of Caco-2 cell epithelia. Caco-2 monolayers were used when their TEERs has reached >600 O cm2. When a Caco-2 monolayer was used, the permeation rate of 100 mM FITC-dextran was measured after completion of the permeation experiment to check that the barrier function had been maintained during the experiment. The permeation rate of FITC-dextran to the receiver compartment over a period of 1 h was <0.1%. JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 93, NO. 1, JANUARY 2004
Analysis The concentrations of drugs were determined by high-performance liquid chromatography (HPLC; L-6000, Hitachi Company, Ltd., Tokyo, Japan) using an L-4200H UV-VIS detector. A Asahipak ODP-50 (250 4.6 mm; dipyridamole) or a ERCODS-1161 (100 4.6 mm; glibenclamide) was used as analytical column. Mobile phases consisting of phosphate buffer (10 mM Na2HPO4) and 50% acetonitrile, adjusted to pH 8.0 (dipyridamole) or pH 6.5 (glibenclamide) with sodium hydroxide. Dipyridamole was detected at wavelength of 280 nm at 558C and glibenclamide at 230 nm at room temperature. The measurement of FITC-dextran was carried out in a 650-60 fluorescence spectrophotometer (Hitachi Company) at an excitation wavelength of 495 nm and emission wavelength of 514 nm.
RESULTS AND DISCUSSION Changes of pH Values of the Outflow from the PAV We determined pH values of the outflow from the PAV every 10 min over a period of 300 min in each model. The time courses of pH values are shown in Figure 3. The results suggest a suitable pH value of the solution in the PAV (modeled intestine, pH 6.0) in the normal gastric acidity model (model A) and indicate that the temporarily decreased pH values in temporarily elevated gastric acidity model (model C) occurred. The results also suggest that pH values of the solution in the PAV in weak antacid model (model E) did not change compared with that in normal gastric acidity model (PAV-pH 7.5; model D), whereas a temporary elevation in pH values in strong antacid model (model F) occurred. Evaluation of pH-Related Changes in Absorption of Dipyridamole The amounts of dipyridamole preparation (Persantin1 tablets, each 12.5 mg) that eluted into the donor compartment were measured in the normal gastric acidity model (model A), low gastric acidity model (model B), and temporarily elevated gastric acidity model (model C). The time course of the efflux of dipyridamole to the donor compartment of the side-by-side diffusion chamber in each model is shown in Figure 4a. As shown in Table 2, the cumulative efflux in model C
NEW MODEL SYSTEM TO PREDICT DRUG ABSORPTION
Figure 3. pH values of the outflow from the PAV in each model. Key: (&) normal gastric acidity model (model A); (*) low gastric acidity model (model B); (~) temporarily elevated gastric acidity model (model C); (&) normal gastric acidity model (PAV-pH7.5) (model D); (~) weak-antacid model (model E); (*) strongantacid model (model F). Each value is the mean SEM of three independent experiments.
(45.3%) was almost the same as that in model A (44.6%). However, the cumulative efflux in model B (17.9%) was significantly (p < 0.001) lower than that in model A.
75
The amounts of dipyridamole preparation (Persantin1 tablets, each 12.5 mg) that permeated across a Caco-2 monolayer were also investigated using models A, B, and C. The purpose of this experiment was to compare the amounts of dipyridamole absorbed and degrees of absorption variation due to changes in pH in different models. The results are shown in Table 2. The cumulative permeation in model B (0.017%) was significantly (p < 0.001) decreased compared with that in model A (0.047%), whereas there was no significant difference between the cumulative permeations in model C (0.047%) and model A. The results suggest that the extent of absorption of dipyridamole, which needs a low gastric pH for complete absorption, is remarkably lower when gastric pH is elevated to 6.0, whereas it is increased when gastric pH is temporarily decreased to 1.8. Our results agree with those obtained in studies that show that the oral absorption in subjects with low levels of gastric acidity is much lower than that in subjects with normal levels and that the oral absorption can be improved to some degree by coadministration of acidic drugs, which can temporarily increase gastric acidity.5,6 Evaluation of pH-Related Changes in Absorption of Glibenclamide The time courses of the elution of the glibenclamide preparation (Euglucon1 tablets, each 2.5 mg) into the donor compartment in the normal gastric acidity model (PAV-pH 7.5; model D),
Figure 4. Time courses of the efflux of (a) dipyridamole or (b) glibenclamide to the donor compartment of a side-by-side chamber in each model. Each data point is the mean SEM of 4–5 independent experiments. JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 93, NO. 1, JANUARY 2004
76
HE ET AL.
Table 2. Cumulative Efflux in Donor Compartment and Cumulative Permeation in Receiver Compartment in Each Modela
Drug Dipyridamole
Glibenclamide
Donor Compartment
Receiver Compartment
Model
Cumulative Efflux (% of Dose)
Cumulative Permeation (% of Dose)
A B C D E F
44.6 3.3 17.9 2.6b 45.3 3.7 64.4 3.8 69.5 10.0 78.5 2.5c
0.047 0.006 0.017 0.005b 0.047 0.010 0.059 0.013 0.074 0.002 0.135 0.020c
a
Each value represents the mean SEM of the results of 3–5 independent experiments. Significantly different from model A ( p < 0.001). c Significantly different from model D ( p < 0.001). b
weak antacid model (model E), and strong antacid model (model F) are shown in Figure 4b. As shown in Table 2, the cumulative efflux in model E (69.5%) was almost the same as that in model D (64.4%). However, there was a significant (p < 0.001) difference between the cumulative efflux in model F (78.5%) and model D. Also shown in Table 2 are the amounts of glibenclamide preparation (Euglucon1 tablets, each 2.5 mg) that permeated across the Caco-2 monolayer in models D, E, and F. The purpose of this experiment was to compare the amounts of glibenclamide absorbed and the degrees of variation in absorption due to coadministration of various antacids in different models. The cumulative permeation in model F (0.135%) was significantly ( p < 0.001) higher than that in model D (0.059%), but the cumulative permeation in model E (0.074%) was almost the same as that in model D. These results are consistent with those reported by Neuvonen and Kivisto8 and Kubacka et al.9
CONCLUSIONS A new system for predicting pH-dependent soluble drug absorption using a pH-controlled Caco-2 model was developed. We established several models for different gastric acidities: a normal gastric acidity model, a low gastric acidity model, a temporarily elevated gastric acidity model, and two antacid models. Dipyridamole and glibenclamide, poorly water-soluble and pH-dependent soluble drugs, were chosen as model drugs. The pH-related changes in the absorption of these drugs could be accurately predicted using our JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 93, NO. 1, JANUARY 2004
system. It is useful to know variations in the absorption of medications that require a special gastric pH for complete absorption due to changes in pH in the gastrointestinal tract because the extent of absorption of these medications may be unexpectedly increased or decreased by the concomitant intake of drugs that change the gastrointestinal pH.
REFERENCES 1. Kobayashi M, Sada N, Sugawara M, Iseki K, Miyazaki K. 2001. Development of a new system for prediction of drug absorption that takes into account drug dissolution and pH change in the gastro-intestinal tract. Int J Pharm 221:87–94. 2. He X, Sugawara M, Kobayashi M, Takekuma Y, Miyazaki K. 2003. An in vitro system for prediction of oral absorption of relatively water-soluble drugs and ester prodrugs. Int J Pharm 263:35–44. 3. Foye WO. 1981. Principles of Medicinal Chemistry. Philadelphia: Lea and Febiger. 4. Kohri N, Miyata N, Takahashi M, Endo H, Iseki K, Miyazaki K, Takechi S, Nomura A. 1992. Evaluation of pH-independent sustained-release granules of dipyridamole by using gastric-acidity-controlled rabbits and human subject. Int J Pharm 81:49– 58. 5. Russell TL, Berardi RR, Barnett JL, Dermentzoglou LC, Jarvenpaa KM, Schmaltz SP, Dressman JB. 1993. Upper gastrointestinal pH in seventynine healthy, elderly, North American men and women. Pharm Res 10:187–196. 6. Russel TL, Berardi RR, Barnett JL, O’Sullivan TL, Wagner JG, Dressman JB. 1994. pH-related changes in the absorption of dipyridamole in the elderly. Pharm Res 11:136–143.
NEW MODEL SYSTEM TO PREDICT DRUG ABSORPTION
7. Kaali RN, Rye RM, Wiseman D, York P. 1987. Dissolution of glibenclamide: The role of particle size. J Pharm Pharmacol 39:44. 8. Neuvonen PJ, Kivisto KT. 1991. The effects of magnesium hydroxide on the absorption and efficacy of two glibenclamide preparations. Br J Clin Pharmacol 32:215–220. 9. Kubacka RT, Antal EJ, Juhl RP. 1987. The paradoxical effect of cimetidine and ranitidine on glibenclamide pharmacokinetics and pharmacodynamics. Br J Clin Pharmacol 23:743– 751.
77
10. Tuderman V, Klinge E, Lind H. 1979. In vitro evaluation of antacid suspensions marketed in Finland. Acta Pharm Fenn 88:63–72. 11. Brockmeier D, Grigoleit HG, Leonhardt H. 1985. Absorption of glibenclamide from different sites of the gastro-intestinal tract. Eur J Clin Pharmacol 29:193–197. 12. Palm K, Luthman K, Ros J, Grasjo T, Artursson P. 1999. Effect of molecular charge on intestinal epithelial drug transport: pH-dependent transport of cationic drug. J Pharmacol Exp Ther 291:435– 443.
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 93, NO. 1, JANUARY 2004
Received 1 April 2003; revised 28 May 2003; accepted 11 June 2003
ABSTRACT: One purpose of this study was to develop a new system for the prediction of pH-dependent soluble drug absorption that takes into account the physiological condition of the gastrointestinal tract. Another purpose was to establish several models of different gastric acidities: a normal gastric acidity model, a low gastric acidity model (a model of achlorhydria), a temporarily elevated gastric acidity model (a model of a case in which an acidic drug was coadministered to temporarily elevate gastric acidity in the case of low gastric acidity), a weak antacid model (a model of a case in which a weak antacid drug, such as an H2 receptor antagonist, was coadministered to temporarily elevate pH up to 6), and a strong antacid model (a model of a case in which a strong antacid drug, such as magnesium hydroxide, was coadministered to temporarily elevate pH up to 8.0). These models were used to evaluate variation in pH-related absorption in humans. Dipyridamole preparation (Persantin1 tablets) and glibenclamide preparation (Euglucon1 tablets), both poorly water-soluble and pH-dependent soluble drugs, were chosen as model drugs to determine whether absorption is altered by changes in levels of gastric acid. The extent of absorption of dipyridamole was remarkably lower when gastric pH was continuously elevated to 6.0, whereas it was increased when gastric pH temporarily decreased to 1.8. The extent of absorption of glibenclamide increased dramatically when gastric pH temporarily increased to 8.0, but did not change when gastric pH temporarily increased to 6.0. These results are consistent with reported results obtained in clinical studies. The results suggest that pH-related variations in absorption in humans can be accurately predicted using our new system. ß 2004 Wiley-Liss, Inc. and the American Pharmacists Association J Pharm Sci 93:71–77, 2004
Keywords: prediction; Caco-2 cells; in vitro models; oral absorption; permeability; drug interactions; solubility
INTRODUCTION Various factors cause variation in pH in the gastrointestinal tract. It is well known that oral bioavailability is influenced greatly by solubility
Correspondence to: Katsumi Miyazaki (Telephone: 81-117065680; Fax: 81-11- 7561505; E-mail: [email protected]) Journal of Pharmaceutical Sciences, Vol. 93, 71–77 (2004) ß 2004 Wiley-Liss, Inc. and the American Pharmacists Association
and stability of drugs, which depend on pH. Gastric acid in humans has pH of 1–2. Variations in gastric acidity results in changes in the absorption profiles of drugs with pH-sensitive dissolution. We have developed an in vitro system for prediction of drug absorption that takes into account dissolution of solid drugs and pH changes in the gastrointestinal tract.1 This system enables prediction of the oral absorption of a relatively water-soluble drug in humans based on the cumulative permeation of the drug across a
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 93, NO. 1, JANUARY 2004
71
72
HE ET AL.
Caco-2 monolayer.2 The primary aim of this study was to develop a new system for predicting pHdependent soluble drug absorption that takes into account physiological conditions in the gastrointestinal tract. The second aim was to evaluate variation in pH-related absorption in humans. We established several models with different gastric acidities: a normal gastric acidity model, a low gastric acidity model (a model of achlorhydria), a temporarily elevated gastric acidity model (a model of a case in which an acidic drug was coadministered to temporarily elevate gastric acidity), a weak antacid model (a model of a case in which a weak antacid drug, such as an H2 receptor antagonist, was coadministered to temporarily elevate pH up to 6), and a strong antacid model (a model in which a strong antacid drug, such as magnesium hydroxide, was coadministered to temporarily elevate pH up to 8.0). Dipyridamole preparation (Persantin1 tablets) and glibenclamide preparation (Euglucon1 tablets) were chosen as model drugs to determine whether absorption is altered by variations in levels of gastric acidity in these models. The extent of absorption of dipyridamole, a poorly soluble weak base with a reported pKa of 6.4,3 was reported to be altered to a considerable extent by the pH of different digestive fluids;4 that is, dipyridamole dissolves readily in the stomach, but incompletely in the intestine. The absorption of dipyridamole is dissolution rate limited over at least a part of the usual physiological range of gastrointestinal pH. Because achlorhydric subjects have high pH values in both the stomach and the intestine regardless of the prandial phase,5 and because the dissolution rate of dipyridamole is very low in that pH range, achlorhydrics exhibit a slow rate and poor extent of absorption compared with subjects who have a low gastric pH. Furthermore, it would be beneficial for achlorhydrics to take glutamic acid hydrochloride prior to taking dipyridamole for complete absorption.6 Glibenclamide, a weakly acidic drug with a reported pKa of 6.8, does not dissolve at pH 1–6. However, its solubility increases at pH > 7.7 Comparable enhancement of glibenclamide absorption can be achieved by elevating gastric pH.8 The oral absorption of glibenclamide, therefore, did not change when ranitidine, a weak antacid drug that increases the pH value up to 6.0, was coadministered.9 On the other hand, coadministration of magnesium hydroxide, a strong antacid reagent that increases pH in the stomach up to 8.0,10 resulted in a dramatic increases in the extent of JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 93, NO. 1, JANUARY 2004
absorption of glibenclamide.8 This study predicted changes that occur in the extent of absorption of poorly water-soluble and pH-dependent soluble drugs in people with abnormal levels of gastric acidity and in people with normal levels of gastric acidity using the models just described.
EXPERIMENTAL Materials Glibenclamide preparation (Euglucon1 tablets, each 2.5 mg) was purchased from Yamanouchi Pharmaceutical Company, Ltd. (Tokyo, Japan). Dipyridamole preparation (Persantin1 tablets, each 12.5 mg) was purchased from Nippon Boehringer Ingelheim Company, Ltd. (Hyogo, Japan). Fluorescein isothiocyanate (FITC)-dextran (molecular weight, 4400) was purchased from Sigma Chemical Company (St. Louis, MO). System for Predicting Drug Absorption As shown in Figure 1, our in vitro system for predicting drug absorption takes into account drug dissolution and change in pH in the gastrointestinal tract. In this system, a solid drug is added to a drug-dissolving vessel (DDV; modeled stomach, 10 mL), and the dissolved drug is transferred to a pH adjustment vessel (PAV; modeled intestine, 10 mL). Each of these vessels is a plastic vial. The compositions of the drug-dissolving solution, pH adjustment solution, and receiver solution are shown in Table 1. The flow rate (0.5 mL/min) of
Figure 1. Scheme of the drug absorption prediction system.
NEW MODEL SYSTEM TO PREDICT DRUG ABSORPTION
73
Table 1. Compositions of Flowing Solutions (mM)
Component KCl KH2PO4 NaCl Na2HPO4 D-glucose CaCl2 MgSO4 MES HEPES HCl NaOH
Drug-Dissolving Solution (pH 1.8)
pH Adjustment Solutiona
Receiver Solutionb (pH 7.4)
— — 61.4 — 50 2.52 0.81 — — 13.6 —
10.7 0.88 199 0.68 — — — 40 — — 13.6
5.37 0.44 137 0.34 25 1.26 0.41 — 10 — —
a pH was adjusted to 6.0 (models A, B, and C) or 7.5 (models D, E, and F) with Tris before addition of NaOH. b pH was adjusted to 7.4 with Tris.
each solution is controlled by a peristaltic pump. The drug solution is transferred to the donor compartment of side-by-side diffusion chamber (Corning Costar Company). To prevent the inflow of undissolved powder from the PAV, glass wool is set between the PAV and the donor compartment of side-by-side diffusion chamber. Mounted between the donor and receiver compartments is a Caco-2 monolayer grown on a Snapwell (0.4 mm in pore size, 1 cm2 in growth area; Corning Costar Company). Samples of dipyridamole or glibenclamide are collected every 10 min over a period of 300 min. A silicon tube (i.d., 0.5 mm) is used to connect each vessel and the compartment. All of the solutions and both vessels and both compartments are preheated to 378C and maintained at that temperature.
absorption of dipyridamole due to changes in pH in the gastrointestinal tract was evaluated by using these models. Glibenclamide is sparingly soluble in water at gastric pH, whereas its solubility increases at pH >7.7 Also, glibenclamid is absorbed not only in
Models As shown in Figure 2, the normal gastric acidity model (model A) was used for prediction of pHdependent soluble drug absorption that takes into account the physiological condition of the gastrointestinal tract. The DDV modeled stomach, the PAV modeled intestine, and the drug-dissolving solution modeled gastric juice. Dipyridamole is a pH-dependent soluble drug that dissolves readily in acidic conditions, whereas its solubility decreases dramatically at pH of 3.0. So, we established a low gastric acidity model (model B, model of achlorhydria) and a temporarily elevated gastric acidity model (model C), in which gastric acidity was elevated temporarily by coadministration of an acidic drug. The variation in the oral
Figure 2. Scheme for models of different gastric acidities: (A) normal gastric acidity model; (B) low gastric acidity model; (C) temporarily elevated gastric acidity model; (D) normal gastric acidity model (PAVpH7.5); (E) weak-antacid model; and (F) strong-antacid model. JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 93, NO. 1, JANUARY 2004
74
HE ET AL.
the stomach and small intestine but also in the large intestine (pH 7–8).11 Furthermore, the integrity of the Caco-2 monolayers was unaffected in the apical pH interval 4.8–8.0.12 Considering these facts, the pH value of the solution in the PAV was adjusted to 7.5. The weak antacid model (model E) is a model of a case in which a weak antacid drug, such as ranitidine, was coadministered to temporarily elevate pH up to 6.0. The strong antacid model (model F) is a model of a case in which a strong antacid drug, such as magnesium hydroxide, was coadministered to temporarily elevate pH up to 8.0. As a control experiment, a model of normal gastric acidity (model D) was established using a PAV solution with pH of 7.5. Cell Culture Caco-2 cells were purchased from American Type Culture Collection (Rockville, MD) and maintained in plastic culture flasks (Falcon, Becton Dickinson and Company, Lincoln Park, NJ). The stock cells were subcultivated before reaching confluence. The growth medium used was Dulbecco’s modified Eagle’s medium (DMEM; Sigma) supplemented with 10% fetal bovine serum (ICN Biomedicals, Inc., Aurora, OH), 1% nonessential amino acids (Gibco), and 4 mM glutamine without antibiotics. The monolayer cultures were grown in a CO2 incubator (5% CO2) at 378C. The cells were harvested with 0.25% trypsin and 0.2% EDTA (0.5–1 min at 378C), resuspended, and seeded into a new flask. Cells between the 35th and 50th passages were used in this study. For the transport studies, Caco-2 cells were seeded on Snapwell at a cell density of 8 104 cells/ filter. The cell monolayers were fed a fresh growth medium every 2 days and were used on days 20– 28 for the transport experiments. To evaluate the integrity of the monolayer, transepithelia electrical resistance (TEER) was measured with a Millicell-ERS (Millipore Company, Bedford, MA). The TEER of the filter was subtracted from the measured total TEER of Caco-2 cell epithelia. Caco-2 monolayers were used when their TEERs has reached >600 O cm2. When a Caco-2 monolayer was used, the permeation rate of 100 mM FITC-dextran was measured after completion of the permeation experiment to check that the barrier function had been maintained during the experiment. The permeation rate of FITC-dextran to the receiver compartment over a period of 1 h was <0.1%. JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 93, NO. 1, JANUARY 2004
Analysis The concentrations of drugs were determined by high-performance liquid chromatography (HPLC; L-6000, Hitachi Company, Ltd., Tokyo, Japan) using an L-4200H UV-VIS detector. A Asahipak ODP-50 (250 4.6 mm; dipyridamole) or a ERCODS-1161 (100 4.6 mm; glibenclamide) was used as analytical column. Mobile phases consisting of phosphate buffer (10 mM Na2HPO4) and 50% acetonitrile, adjusted to pH 8.0 (dipyridamole) or pH 6.5 (glibenclamide) with sodium hydroxide. Dipyridamole was detected at wavelength of 280 nm at 558C and glibenclamide at 230 nm at room temperature. The measurement of FITC-dextran was carried out in a 650-60 fluorescence spectrophotometer (Hitachi Company) at an excitation wavelength of 495 nm and emission wavelength of 514 nm.
RESULTS AND DISCUSSION Changes of pH Values of the Outflow from the PAV We determined pH values of the outflow from the PAV every 10 min over a period of 300 min in each model. The time courses of pH values are shown in Figure 3. The results suggest a suitable pH value of the solution in the PAV (modeled intestine, pH 6.0) in the normal gastric acidity model (model A) and indicate that the temporarily decreased pH values in temporarily elevated gastric acidity model (model C) occurred. The results also suggest that pH values of the solution in the PAV in weak antacid model (model E) did not change compared with that in normal gastric acidity model (PAV-pH 7.5; model D), whereas a temporary elevation in pH values in strong antacid model (model F) occurred. Evaluation of pH-Related Changes in Absorption of Dipyridamole The amounts of dipyridamole preparation (Persantin1 tablets, each 12.5 mg) that eluted into the donor compartment were measured in the normal gastric acidity model (model A), low gastric acidity model (model B), and temporarily elevated gastric acidity model (model C). The time course of the efflux of dipyridamole to the donor compartment of the side-by-side diffusion chamber in each model is shown in Figure 4a. As shown in Table 2, the cumulative efflux in model C
NEW MODEL SYSTEM TO PREDICT DRUG ABSORPTION
Figure 3. pH values of the outflow from the PAV in each model. Key: (&) normal gastric acidity model (model A); (*) low gastric acidity model (model B); (~) temporarily elevated gastric acidity model (model C); (&) normal gastric acidity model (PAV-pH7.5) (model D); (~) weak-antacid model (model E); (*) strongantacid model (model F). Each value is the mean SEM of three independent experiments.
(45.3%) was almost the same as that in model A (44.6%). However, the cumulative efflux in model B (17.9%) was significantly (p < 0.001) lower than that in model A.
75
The amounts of dipyridamole preparation (Persantin1 tablets, each 12.5 mg) that permeated across a Caco-2 monolayer were also investigated using models A, B, and C. The purpose of this experiment was to compare the amounts of dipyridamole absorbed and degrees of absorption variation due to changes in pH in different models. The results are shown in Table 2. The cumulative permeation in model B (0.017%) was significantly (p < 0.001) decreased compared with that in model A (0.047%), whereas there was no significant difference between the cumulative permeations in model C (0.047%) and model A. The results suggest that the extent of absorption of dipyridamole, which needs a low gastric pH for complete absorption, is remarkably lower when gastric pH is elevated to 6.0, whereas it is increased when gastric pH is temporarily decreased to 1.8. Our results agree with those obtained in studies that show that the oral absorption in subjects with low levels of gastric acidity is much lower than that in subjects with normal levels and that the oral absorption can be improved to some degree by coadministration of acidic drugs, which can temporarily increase gastric acidity.5,6 Evaluation of pH-Related Changes in Absorption of Glibenclamide The time courses of the elution of the glibenclamide preparation (Euglucon1 tablets, each 2.5 mg) into the donor compartment in the normal gastric acidity model (PAV-pH 7.5; model D),
Figure 4. Time courses of the efflux of (a) dipyridamole or (b) glibenclamide to the donor compartment of a side-by-side chamber in each model. Each data point is the mean SEM of 4–5 independent experiments. JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 93, NO. 1, JANUARY 2004
76
HE ET AL.
Table 2. Cumulative Efflux in Donor Compartment and Cumulative Permeation in Receiver Compartment in Each Modela
Drug Dipyridamole
Glibenclamide
Donor Compartment
Receiver Compartment
Model
Cumulative Efflux (% of Dose)
Cumulative Permeation (% of Dose)
A B C D E F
44.6 3.3 17.9 2.6b 45.3 3.7 64.4 3.8 69.5 10.0 78.5 2.5c
0.047 0.006 0.017 0.005b 0.047 0.010 0.059 0.013 0.074 0.002 0.135 0.020c
a
Each value represents the mean SEM of the results of 3–5 independent experiments. Significantly different from model A ( p < 0.001). c Significantly different from model D ( p < 0.001). b
weak antacid model (model E), and strong antacid model (model F) are shown in Figure 4b. As shown in Table 2, the cumulative efflux in model E (69.5%) was almost the same as that in model D (64.4%). However, there was a significant (p < 0.001) difference between the cumulative efflux in model F (78.5%) and model D. Also shown in Table 2 are the amounts of glibenclamide preparation (Euglucon1 tablets, each 2.5 mg) that permeated across the Caco-2 monolayer in models D, E, and F. The purpose of this experiment was to compare the amounts of glibenclamide absorbed and the degrees of variation in absorption due to coadministration of various antacids in different models. The cumulative permeation in model F (0.135%) was significantly ( p < 0.001) higher than that in model D (0.059%), but the cumulative permeation in model E (0.074%) was almost the same as that in model D. These results are consistent with those reported by Neuvonen and Kivisto8 and Kubacka et al.9
CONCLUSIONS A new system for predicting pH-dependent soluble drug absorption using a pH-controlled Caco-2 model was developed. We established several models for different gastric acidities: a normal gastric acidity model, a low gastric acidity model, a temporarily elevated gastric acidity model, and two antacid models. Dipyridamole and glibenclamide, poorly water-soluble and pH-dependent soluble drugs, were chosen as model drugs. The pH-related changes in the absorption of these drugs could be accurately predicted using our JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 93, NO. 1, JANUARY 2004
system. It is useful to know variations in the absorption of medications that require a special gastric pH for complete absorption due to changes in pH in the gastrointestinal tract because the extent of absorption of these medications may be unexpectedly increased or decreased by the concomitant intake of drugs that change the gastrointestinal pH.
REFERENCES 1. Kobayashi M, Sada N, Sugawara M, Iseki K, Miyazaki K. 2001. Development of a new system for prediction of drug absorption that takes into account drug dissolution and pH change in the gastro-intestinal tract. Int J Pharm 221:87–94. 2. He X, Sugawara M, Kobayashi M, Takekuma Y, Miyazaki K. 2003. An in vitro system for prediction of oral absorption of relatively water-soluble drugs and ester prodrugs. Int J Pharm 263:35–44. 3. Foye WO. 1981. Principles of Medicinal Chemistry. Philadelphia: Lea and Febiger. 4. Kohri N, Miyata N, Takahashi M, Endo H, Iseki K, Miyazaki K, Takechi S, Nomura A. 1992. Evaluation of pH-independent sustained-release granules of dipyridamole by using gastric-acidity-controlled rabbits and human subject. Int J Pharm 81:49– 58. 5. Russell TL, Berardi RR, Barnett JL, Dermentzoglou LC, Jarvenpaa KM, Schmaltz SP, Dressman JB. 1993. Upper gastrointestinal pH in seventynine healthy, elderly, North American men and women. Pharm Res 10:187–196. 6. Russel TL, Berardi RR, Barnett JL, O’Sullivan TL, Wagner JG, Dressman JB. 1994. pH-related changes in the absorption of dipyridamole in the elderly. Pharm Res 11:136–143.
NEW MODEL SYSTEM TO PREDICT DRUG ABSORPTION
7. Kaali RN, Rye RM, Wiseman D, York P. 1987. Dissolution of glibenclamide: The role of particle size. J Pharm Pharmacol 39:44. 8. Neuvonen PJ, Kivisto KT. 1991. The effects of magnesium hydroxide on the absorption and efficacy of two glibenclamide preparations. Br J Clin Pharmacol 32:215–220. 9. Kubacka RT, Antal EJ, Juhl RP. 1987. The paradoxical effect of cimetidine and ranitidine on glibenclamide pharmacokinetics and pharmacodynamics. Br J Clin Pharmacol 23:743– 751.
77
10. Tuderman V, Klinge E, Lind H. 1979. In vitro evaluation of antacid suspensions marketed in Finland. Acta Pharm Fenn 88:63–72. 11. Brockmeier D, Grigoleit HG, Leonhardt H. 1985. Absorption of glibenclamide from different sites of the gastro-intestinal tract. Eur J Clin Pharmacol 29:193–197. 12. Palm K, Luthman K, Ros J, Grasjo T, Artursson P. 1999. Effect of molecular charge on intestinal epithelial drug transport: pH-dependent transport of cationic drug. J Pharmacol Exp Ther 291:435– 443.
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 93, NO. 1, JANUARY 2004