- Email: [email protected]
Effect of Food Intake on the Oral Absorption of Poorly Water-Soluble Drugs: In Vitro Assessment of Drug Dissolution and Permeation Assay System
Effect of Food Intake on the Oral Absorption of Poorly Water-Soluble Drugs: In Vitro Assessment of Drug Dissolution and Permeation Assay System MAKOTO KATAOKA, YOSHIE MASAOKA, SHINJI SAKUMA, SHINJI YAMASHITA Faculty of Pharmaceutical Sciences, Setsunan University, Nagaotoge-cho 45-1, Hirakata, Osaka 573-0101, Japan
Received 21 November 2005; revised 25 April 2006; accepted 6 May 2006 Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/jps.20691
ABSTRACT: The aim of the present work was to establish appropriate conditions for the dissolution/permeation system (D/P system) to estimate the effect of food intake on oral drug absorption. The D/P system is an in vitro assay system to evaluate the drug dissolution and permeation processes after oral administration. Caco-2 monolayer was used as a model membrane of the intestinal epithelium. In this study, two types of simulated intestinal fluid reflecting the fasted and the fed state conditions of the human gastrointestinal tract were used. Drugs were applied to the D/P system as a powder, then, permeated amounts of drugs into the basal side were monitored. A sigmoidal correlation was obtained between in vivo oral absorption (% absorbed of dose) and in vitro permeated amount (% of dose/2 h) under both states. From the D/P system, the estimated absorption of albendazole in both states was found to correspond well with in vivo observation. Moreover, the D/P system could estimate the effect of self-emulsifying formulation on the oral absorption of danazol, quantitatively. In conclusion, the D/P system was proved to be a useful assay system not only for the oral absorption of drugs, but also for the food effect on the absorption. ß 2006 Wiley-Liss, Inc. and the American Pharmacists Association J Pharm Sci 95:2051–2061, 2006
Keywords: absorption; biopharmaceutics classification system (BCS); Caco-2 cells; dissolution; food effects; in vitro/in vivo correlations (IVIVC); permeability; solubility
INTRODUCTION The oral absorption of poorly water-soluble drugs, such as danazol and griseofulvin, is known to increase when they are administered after food intake.1,2 As a main cause of the effect of food, the secretion of bile juice into the GI tract is accelerated by food intake, which enhances the solubility and dissolution rate and therefore the absorption of these drugs after oral administration. Within pharmaceutical companies, in vivo animal studies have been carried out to detect the effect of food on oral absorption of new drug Correspondence to: Makoto Kataoka (Telephone: þ81-72866-3126; Fax: þ81-72-866-3126; E-mail: [email protected]) Journal of Pharmaceutical Sciences, Vol. 95, 2051–2061 (2006) ß 2006 Wiley-Liss, Inc. and the American Pharmacists Association
candidates before advancing to clinical studies. However, animal studies often show large variations depending on the species and might result in an uncertain estimation on drug absorption in humans. Therefore, an in vitro screening system that can evaluate the effect of food intake on the oral absorption of drugs is strongly desired. Galia et al.3 have proposed the use of fasted and/ or fed state simulated intestinal fluids (FaSSIF and FeSSIF) in the drug dissolution study to simulate the effect of food on drug dissolution, then on the bioavailability/bioequivalence of commercial drug products. FaSSIF and FeSSIF contain taurocholate (NaTC) and lecithin as bile acid and lipid to mimic the composition of human intestinal fluid. Dressman et al. have reported that FaSSIF and FeSSIF can be used in the drug dissolution test and are useful in formulation
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 95, NO. 9, SEPTEMBER 2006
2051
2052
KATAOKA ET AL.
development.4,5 On the other hand, Ingels et al. have reported that the use of FaSSIF as the apical medium in the Caco-2 permeation study may improve the biorelevance of the model to evaluate the oral absorption in humans.6–8 Both approaches with simulated intestinal fluids may not be sufficient to evaluate the effect of food intake on the total processes of oral drug absorption, because they consist of two processes, the dissolution and permeation of drugs, which occur sequentially in the GI tract. In order to evaluate the oral absorption of poorly soluble drugs, we have developed an in vitro system with which both dissolution and permeation processes can be analyzed simultaneously. In this system (dissolution/permeation system, D/P system), a Caco-2 monolayer is mounted in the side-by-side chambers. Both the apical and basal sides of the monolayer are filled with media and are constantly stirred. Drugs are applied to the apical medium as a solid form (powder or granule); then, dissolution in the apical side and permeation to the basal side of the drug are simultaneously monitored. In the previous report, it was demonstrated that the amount of drug that permeated to the basal side of the Caco-2 monolayer showed a sigmoidal correlation to their absorption in vivo in humans, regardless of the class of drugs in the Biopharmaceutics Classification System (BCS). Consequently, the D/P system is a useful tool to evaluate the oral absorption of poorly soluble drugs after oral administration as the solid dosage form.9 Ginski et al. have reported the use of the continuous dissolution/Caco-2 system to investigate drug absorption.10,11 They observed the dissolution-absorption relationships and determined the rate limiting process of absorption in rapid and sustained dissolving formulations. Kobayashi et al.12 have reported a system in which the effect of pH change in the GI tract on drug dissolution and permeation can be evaluated. Furthermore, Motz et al. have developed a flow through permeation cell system to simultaneously monitor the dissolution and permeability of drugs.13,14 Those studies and devices have generated the useful information in considering the dissolution-permeation relationship in oral drug absorption. However, in order to predict the in vivo oral absorption of drugs quantitatively from the in vitro study, physiological conditions in the human GI tract, such as the fluid volume, fluid composition, and fluid pH must be taken into consideration, especially for the absorption of poorly JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 95, NO. 9, SEPTEMBER 2006
water-soluble drugs. In our D/P system, the assay conditions were fixed carefully, based on the in vivo physiological conditions of the human GI tract to successfully obtain the IVIVC in oral drug absorption.9 In this study, we have tried to predict the effect of food intake on oral drug absorption in humans from in vitro experiments with the D/P system. Various kinds of simulated intestinal fluids were prepared and applied to the apical side of the D/P system to reflect in vivo fasted and fed state conditions in the GI tract. Then, we compared the results for the D/P system with in vivo observations in humans.
EXPERIMENTAL Materials The Caco-2 cell line was purchased from American Type Culture Collection (Rockville, MD) at passage 17. Dulbecco’s modified Eagle medium (D-MEM) was purchased from Sigma-Aldrich (St. Louis, MO). Nonessential amino acids (NEAA), fetal bovine serum (FBS), L-glutamate, trypsin (0.25%)EDTA (1 mM) and antibiotic-antimycotic mixture (10000 U/mL penicillin G, 10000 mg/mL streptomycin sulfate and 25 mg/mL amphotericin B in 0.85% saline) were purchased from Gibco Laboratories (Lenexa, KS). Danazol was purchased from Sigma-Aldrich. Albendazole, egg-phosphatidylcholine (lecithin), sodium taurocholate (NaTC) and bovine serum albumin (BSA) were purchased from WAKO Pure Chemical Industries, Ltd. (Japan). Gelucire1 44/14 was obtained from ´ (France). All other reagents used GATTEFOSSE were of the highest purity.
Preparation of Caco-2 Monolayer Caco-2 cells were grown in D-MEM supplemented with 10% FBS, 1% L-glutamate, 1% NEAA and 5% antibiotic-antimycotic solution as culture medium at 38C in culture flasks (Nippon Becton Dickinson Co., Ltd., Tokyo, Japan) in humidified air with a 5% CO2 atmosphere. Cells were harvested with typsin-EDTA and seeded on polycarbonate filters (0.3 mm pores, 4.20 cm2 growth area) inside a cell culture insert (Nippon Becton Dickinson Co., Ltd., Tokyo, Japan) at a density of 3 105 cells/filter. The culture medium (1.5 mL in the insert and 2.6 mL in the well) was replaced every 48 h for the first 6 days and every 24 h thereafter. After DOI 10.1002/jps
ESTIMATION OF FOOD EFFECTS ON ORAL DRUG ABSORPTION
18–21 days in culture, the Caco-2 monolayer was utilized for the following experiments.
permeation processes of drugs can be analyzed simultaneously.9 In this system, the Caco-2 monolayer is mounted in side-by-side chambers. The effective surface area of the Caco-2 monolayer is 1.77 cm2. Both sides of the Caco-2 monolayer are consistently stirred at 200 rpm with magnetic stirrers. The volume of apical and basal sides is set to 8 and 5.5 mL, respectively. As an apical medium, various fluids shown in Table 1 were used according to the purpose of the study. TM containing a 4.5 w/v% concentration of BSA (pH 7.4) was used as a basal medium in all studies. All experiments were performed at 378C.
Preparation of Various Fluids As a buffer solution in this study, Hank’s balanced salts solution (HBSS), containing KCl 5.36 mM, NaCl 136.89 mM, Na2HPO4 0.34 mM, KH2PO4 0.44 mM, NaHCO3 4.17 mM, CaCl2 1.26 mM, MgCl2 0.49 mM, MgSO4 0.41 mM and glucose 19.45 mM, was used (transport medium, TM). The pH of the TM was adjusted to 5.0, 6.5 or 7.4 with HEPES. Simulated intestinal fluids (FaSSIF and FeSSIF) were prepared according to the report of Galia et al.3 as FaSSIF; KH2PO4 3.9 g/L, NaTC 3 mM, Lecithin 0.75 mM, KCl 7.7 g/L, adjusted pH to 6.5 by NaOH, and FeSSIF; Acetic acid 8.65 g/L, NaTC 15 mM, Lecithin 3.75 mM, KCl 15.2 g/L, adjusted pH to 5.0 by NaOH. Modified simulated intestinal fluids, FaSSIFmod and FeSSIFmod, were prepared based on TM by adding NaTC and lecithin as the same concentration with those in FaSSIF and FeSSIF, respectively (Tab. 1). The pH of FaSSIFmod and FeSSIFmod was adjusted to 6.5 and 5.0 with HEPES, respectively. FeSSIFmod6.5 has a same composition with FeSSIFmod, but pH was adjusted to 6.5.
Effect of Various Fluids on the Transepithelial Electric Resistance (TEER) of Caco-2 Monolayers The effects of various fluids (applied to the apical side) on the TEER of the Caco-2 monolayer were monitored in the D/P system. Before applied to the D/P system, the Caco-2 monolayer was preincubated in the culture well with TM (pH 6.5 for the apical and 7.4 for the basal side). After preincubation for 20 min, the Caco-2 monolayer with support filter was taken from the insert and was mounted between the chambers of the D/P system. Then, the apical side of the monolayer was filled with various fluids in Table 1 (TM, TM þ NaTC, FaSSIF, FeSSIF, FaSSIFmod, and FeSSIFmod). The basal side was filled with TM with 4.5 w/v% BSA (pH 7.4) in all experiments. Then, the TEER of the Caco-2 monolayer was routinely measured at appropriate intervals over
Chambers for the Dissolution/Permeation System (D/P System) The D/P system is an in vitro assay system of oral drug absorption with which both dissolution and
Table 1.
2053
Various Simulated Intestinal Fluids Applied to D/P System
TMb TM þ NaTCc Simulated intestinal fluidd FaSSIF FeSSIF Simulated intestinal fluid (modified) FaSSIFmod FeSSIFmod FeSSIFmod6.5
pH
Taurocholate
Lecithin
Osmolality (mOsm)a
6.5 6.5
— 5 mM
— —
306.3 0.6 313.7 0.6
6.5 5.0
3 mM 15 mM
0.75 mM 3.75 mM
270 10 635 10
6.5 5.0 6.5
3 mM 15 mM 15 mM
0.75 mM 3.75 mM 3.75 mM
311.7 0.6 327.0 1.0 325.7 0.6
a
Osmolality was measured by OSMOSTAT (OM-6040, ARKARY). TM; Transport medium (HBSS supplemented with 25 mM glucose). c NaTC; Sodium taurocholate. d Galia et al.3 b
DOI 10.1002/jps
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 95, NO. 9, SEPTEMBER 2006
2054
KATAOKA ET AL.
2 h by Millcell-ERS (MILLIPORE, MA). In order to apply the electrodes to the D/P system, the shape of the electrodes were slightly modified.
Dissolution Profile of Danazol in the D/P System The dissolution profile of danazol was observed by using only the apical chamber without mounting Caco-2 monolayers. Instead of Caco-2 monolayers, a flat sheet of aluminum foil was mounted between the chambers to prevent the leakage of apical solution to the basal compartment. As an apical medium, FaSSIFmod or FeSSIFmod was used. One milligram of danazol was applied to the apical chamber as a powder and aliquots of samples were routinely taken from the apical solution over 2 h. Samples were filtrated through cellulose acetate filter to determine the dissolved amount of danazol.
Permeation Profiles of Various Drugs in the D/P System Permeation profiles of various drugs (listed in Tab. 2) were observed in the D/P system with FaSSIFmod, FeSSIFmod, or FeSSIFmod6.5 as an apical medium. The Caco-2 monolayer was applied to the D/P system using the same method as described above. The apical side of the monolayer was filled with FaSSIFmod, FeSSIFmod or FeSSIFmod6.5, and the basal side was filled with TM with 4.5 w/v% BSA (pH 7.4). Drugs were applied to the apical side as a powder or as a suspension with Gelucire1 44/14. An applied amount of drugs were set to 1% of their clinical dose. Then, aliquots of samples were taken from the basal solution at appropriate intervals over 2 h. The volume of the basal solution was maintained by adding fresh medium. Samples were
pretreated with methanol and centrifugation and the deproteinized supernatant was used to determine the amount of drugs permeated through the Caco-2 monolayer. After finishing the experiment, the apical solution was immediately collected and filtered through a cellulose acetate filter to determine the final concentration of the dissolved drug. The TEER of the Caco-2 monolayer was checked before and after the experiment. In most of the experiments, no significant decreases in the TEER value during the experiments were observed (data not shown). Preparation of Self-Emulsifying Formulation Gelucire1 44/14 was used to prepare the selfemulsifying formulation of danazol. Danazol was suspended into Gelucire1 44/14 and a concentration of 1 mg/50 mL. Then, 1 mg of danazol (thus 50 mL of suspension) was applied to the apical side of the D/P system, followed by the same procedures of sampling of both the apical and basal solution for drug concentration measurement. Analytical Methods All samples were analyzed with a reversed-phase HPLC system (LC-10A Shimadzu Co., Kyoto, Japan) equipped with a variable wavelength ultraviolet detector (SPD-10A, Shimadzu Co., Kyoto Japan). The column (J’sphere ODS-H80 75 4.6 mm I.D., YMC, Japan) was used with a mobile phase consisting of 50 mM phosphate buffer (pH 2.5) and acetonitrile. Acyclovir, albendazole, atenolol, carbamazepine, chlorpheniramine, danazol, griseofulvin, ketoprofen, metoprolol, nadolol, piroxicam, pranlukast, propranolol, and warfarin were quantified with the variable ultraviolet detector at 255, 310, 226, 285, 225, 286, 325, 260, 225, 215, 326, 260, 228 and 311 nm, respectively.
Table 2. Effect of Simulated Intestinal Fluids on Dissolved and Permeated Amount of Various Drugs in D/P System Dissolved Amount (% of dose/2 h)
Propranolol Danazol Albendazole
Applied Amount (mg)
FaSSIFmod
0.1 1.0 4.0
3.698 0.242
FeSSIFmod
FeSSIFmod6.5
Completely dissolved 13.307 14.288 2.300 1.928
Permeated Amount (% of dose/2 h) FaSSIFmod
FeSSIFmod
FeSSIFmod6.5
4.430 0.125 0.041
0.905 0.256 0.088
3.401 0.250 0.122
Values are expressed as the mean of at least three independent experiments. JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 95, NO. 9, SEPTEMBER 2006
DOI 10.1002/jps
ESTIMATION OF FOOD EFFECTS ON ORAL DRUG ABSORPTION
RESULTS Effect of Simulated Intestinal Fluids on the TEER of Caco-2 Monolayers Figure 1 shows the effect of various fluids on the TEER of Caco-2 monolayers. When FeSSIF was applied to the apical side of Caco-2 monolayers, the TEER value decreased rapidly to almost zero after 1 h incubation. Since the osmolarity of FeSSIF is twofold higher than that of isotonic fluid, this hyper-osomolarity would cause the destructive damage to the monolayers. Then, isotonic fluid having the same concentration of NaTC and lecithin with FeSSIF was prepared (FeSSIFmod). After applying FeSSIFmod, the TEER value decreased only slightly over the 2 h. Since this pattern of TEER change was almost same that observed with TM, it was apparent that FeSSIFmod did not cause any significant damage to the Caco-2 monolayers. The other fluids listed in Table 1 were also verified to have no significant effect on the TEER of Caco-2 monolayers during the experimental period. In the following study, therefore, FaSSIFmod (having the same concentration of NaTC and lecithin with FaSSIF, but with the osmolarity was finely adjusted to 312 mOsm) and FeSSIFmod were used as the fasted and fed state simulated intestinal fluids, instead of FaSSIF and FeSSIF.
Figure 1. Effect of various simulated intestinal fluids on the TEER of Caco-2 monolayers in the D/P system. TM (&), TM þ NaTC (&), FaSSIF (~), FeSSIF(~), FaSSIFmod (*), and FeSSIFmod (*) were used as the simulated intestinal fluids in the apical side of the D/P system then the time-courses of the TEER were measured for 2 h. The data are expressed as the mean SE of at least three independent experiments. DOI 10.1002/jps
2055
Effect of Simulated Intestinal Fluids on Drug Dissolution and Permeation The time courses of the dissolution and permeation of danazol (applied amount ¼ 1 mg) in the D/P system were measured with FaSSIFmod and FeSSIFmod as apical media. When FeSSIFmod was used, danazol was dissolved rapidly and the dissolved amount reached 13.3% of the dose for 2 h. In contrast, the dissolution of danazol was apparently slow and reached only 3.7% of the dose in FaSSIFmod (Fig. 2a). The permeated amount from the FeSSIFmod medium was also
Figure 2. Effect of simulated intestinal fluids on dissolution (a) and permeation (b) of danazol in the D/P system. The time-courses of danazol dissolution and permeation were observed for 2 h in the D/P system when FaSSIFmod (*) and FeSSIFmod (*) were used as the simulated intestinal fluids in the apical side of the D/P system. The data are expressed as the mean SE of at least three independent experiments. JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 95, NO. 9, SEPTEMBER 2006
2056
KATAOKA ET AL.
in vitro permeated amount of drugs in the D/P system (% of dose/2 h) is summarized in Table 3, with the fraction dose absorbed in humans in the fasted and fed states. Among the drugs listed in Table 3, the oral absorption of danazol and griseofulvin is known to be affected by food intake. Charman et al. have reported that the oral absorption of danazol in the fed state was approximately threefold higher than that in the fasted state in humans. The oral bioavailability of griseofulvin was reported to be dose-dependent and increased almost twofold (70%–120%) by food intake.2 The oral absorptions of danazol and griseofulvin in the fasted and fed states were calculated based on these reports in Table 3. In the case of danazol, the AUC after oral administration as the emulsion formulation under the fed state was reported to show the linear relation with dose up to 200 mg. Since the oral absorption of danazol is limited by its solubility, this linearity in AUC means that the absorption was almost complete. Then, we have assumed that the oral absorption of danazol in human (applied as 100 mg emulsion formulation under the fed state) was 100%. In contrast, the oral absorption of drugs classified to class 1 in BCS, such as propranolol, metoprolol, piroxicam, and ketoprofen was not affected by food intake.15–18 The oral bioavailability of
greater than that from FaSSIFmod, although the difference was rather small (about twofold difference) compared to that observed for dissolution (Fig. 2b). The same experiments were performed for propranolol and albendazole. In the FeSSIFmod medium, the dissolved amount of albendazole was much greater than that in FaSSIFmod, while the permeated amount increased only 2-fold in FeSSIFmod. In the case of propranolol, the permeated amount was markedly reduced in FeSSIFmod despite of the complete dissolution in both media. Then, one more fed state simulated fluid was prepared by adjusting the pH of FeSSIFmod to 6.5 (FeSSIFmod6.5). As summarized in Table 2, FeSSIFmod6.5 dramatically increased the permeated amount of propranolol and albendazole, but not that of danazol. Correlation Between Human Absorption and In Vitro Permeation in the D/P System In order to obtain the correlation between human oral absorption and the in vitro permeated amount in D/P system, FaSSIFmod and FeSSIF mod6.5 were used as apical media. Thirteen drugs were used as references and were applied to the apical side as a powder. The amount of each drug applied was set to 1% of its clinical dose. The
Table 3. Parameters of 13 Drugs for Dissolution and Permeation in D/P System and their Absorption in Humans
Acyclovir Atenolol Carbamazepine Chlorpheniramine Danazol Griseofulvin (low dose) Griseofulvin (high dose) Ketoprofen Metoprolol Nadolol Piroxicam Pranlukast Propranolol Warfarin
Clinical Dosea (mg)
Applied Amountb (mg)
FaSSIFmod Permeated Amount (% of dose/2 h)
200 50 100 6 100 250 500 50 40 80 10 225 10 5
2.0 0.5 1.0 0.06 1.0 2.5 5.0 0.5 0.4 0.8 0.1 2.25 0.1 0.05
0.049 0.097 8.393 5.364 0.125 0.640 0.244 6.295 4.268 0.124 15.659 0.072 4.430 10.793
FeSSIFmod6.5 Permeated Amount (% of dose/2 h)
Human Fasted (%)
Absc Fed (%)
0.029 0.103 — — 0.250 — 0.295 1.742 3.119 — 14.695 0.040 3.401 9.751
20 50 83 80 24 80 40 85 95 34 95 15 95 94
20 50 — — 76 — 78 85 95 — 95 15 95 94
The Pharmacological Basis of Therapeutics (10th edition). Values were expressed as the mean of at least three experiments. a Oral dose. b Applied amount to D/P system (1% of clinical dose). c Represents mean values for % absorption obtained from individual d rug references or the Goodman & Gilman’s. JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 95, NO. 9, SEPTEMBER 2006
DOI 10.1002/jps
ESTIMATION OF FOOD EFFECTS ON ORAL DRUG ABSORPTION
atenolol (a moderately absorbed drug classified to class 3) was reported to be reduced slightly by food intake, but not a significant extent.19 Concerning other class 1 and 3 drugs, warfarin and acyclovir, therefore, we have assumed that their absorption will not be affected by food intake. Then, the following equation was applied to the data sets in Table 3 to obtain the correlation between the in vivo human absorption and the in vitro permeated amount (% of dose/2 h) in the D/P system under the fasted and fed conditions, individually. Human absorption ð%Þ ¼
Absmax PAg PAg50 þ PAg
ð1Þ
where Absmax is the maximum absorption (defined as 100%), PA is the in vitro permeated amount in the D/P system (% of dose/2 h), PA50 is the permeated amount, which corresponds to 50% of the absorption in vivo, and g is a Hill’s coefficient. PA50 and g were obtained by fitting the permeated amount (PA) of drugs in the D/P system and their oral absorption in human (Tab. 3) by using MULTI program developed by Yamaoka et al.20 The correlation curves are described in Figure 3 by incorporating the obtained value of PA50 and g to Eq. 1. PA50 and g in the fasted state were 0.334 0.096 (SD) and 0.883 0.178, and in the fed state were 0.121 0.021 and 1.334 0.198, respectively. In Figure 3, the correlation curve for the fed state-absorption shifted to the left of that for the fasted state. This was due to the fact that
2057
the permeated amount of most drugs in the D/P system was lower in FeSSIFmod6.5 than that in FaSSIFmod. Estimation of In Vivo Human Absorption from In Vitro Data in D/P System By using the in vivo and in vitro correlation (IVIVC) of oral drug absorption in Figure 3, we have tried to predict the effect of food intake on the absorption of albendazole from the in vitro data in the D/P system. When 4 mg of albendazole (clinical dose: 400 mg) was applied to the D/P system as a powder, the permeated amount (% of dose/2 h) was 0.041 % with FaSSIFmod and 0.122 % with FeSSIFmod6.5, respectively (Tab. 2). These values were plotted on the standard curves in Figure 3 to predict the in vivo absorption under the fasted and fed states. The estimated absorption of albendazole was low in the fasted state and increased 3.5-fold in the fed state (Tab. 4). Danazol was suspended in Gelucire1 44/14 (1 mg/50 mL) and was applied to the apical side of the D/P system. The dissolved amount of danazol with Gelucire1 44/14 in FaSSIFmod reached 52.3% of the dose/2 h; in FeSSIFmod6.5 danazol was completely dissolved. From the permeated amount, the in vivo absorption of danazol with Gelucire1 44/14 was estimated as 74% in the fasted and 97% in the fed state, respectively (Tab. 5), suggesting that Gelucire1 44/14 can enhance the absorption of danazol while minimizing the effect of food on it.
DISSCUSSION In the previous report, in order to simulate the in vivo oral absorption of poorly water-soluble drugs with an in vitro D/P system, the assay conditions were carefully fixed, based on the in vivo physiological conditions of the human gastrointestinal (GI) tract. One of the key factors Table 4. In Vivo and In Vitro Estimated Absorption of Albendazole Under Fasted and Fed States in Humans Figure 3. Correlation between the in vivo oral absorption in humans and the in vitro permeated amount in the D/P system. The symbols were the amounts permeated over 2 h in the D/P system when FaSSIFmod (*) and FeSSIFmod6.5 (*) were used as the simulated intestinal fluids in the apical side of the D/P system. Correlation curves in the fasted (dashed line) and the fed (solid line) states were obtained from Eq. 1. DOI 10.1002/jps
Food State
In Vivo Absorption AUCa (mmolh/L)
In Vitro Estimated Absorption (%)
5.0 3.0 19.9 8.2
14 49
Fasted Fed a
Data from Ref. 24.
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 95, NO. 9, SEPTEMBER 2006
2058
KATAOKA ET AL.
Table 5. In Vivo and In Vitro Estimated Absorption of Danazol as Solid Dosage Form and Lipid-Based Formulation In Vivo Absorption (Human) Dosage Form and Food State Solid dosage form Fasted state Fed state Lipid-based formulation Fasted state Fed state
In Vitro Estimated Absorption (D/P System)
AUCa (ngh/mL)
Oral Abs (%)
Dissolution (% of Dose)
Permeation (% of Dose)
Oral Abs (%)
204 125 639 259
24 76
3.698 14.288
0.125 0.250
30 71
779 189 844 194
92 100
52.335 109.262
1.101 1.787
74 97
Values are expressed as the mean of at least three independent experiments. a Data from Ref. 1.
to determine the absorption of poorly soluble drugs should be the balance of dose and water volume in the GI tract. In the D/P system, the applied amount of drugs were set to 1% of their clinical dose, because the volume of the apical chamber, 8 mL, could be regarded as about 1/100 of the physiological volume of the human GI tract that was reported to be approximately 500–1000 mL.4 Bile acids are secreted into the duodenum from the gall bladder. Their average concentration in the human intestinal tract (of healthy subjects) was reported to be 5–15 mM21,22 and highly affected by the food intake. In the previous study, a 5 mM concentration of NaTC was added to the apical solution, which corresponds to the concentration under fasted conditions in humans. In this study, to establish the in vitro conditions that can predict the effect of food intake on drug absorption quantitatively, simulated intestinal fluids were applied to the apical side of the D/P system. Simulated intestinal fluids, FaSSIF and FeSSIF, were proposed by Galia et al.3 and are now widely used as dissolution media to simulate the drug dissolution in the human GI tract under the fasted and fed states.4,5,23,24 Ingels et al. have reported the effect of simulated intestinal fluids on drug permeation across Caco-2 monolayers.6–8 Although FaSSIF did not alter the morphology and integrity of the Caco-2 monolayer, FeSSIF was found not to be suitable for use in Caco-2 assay due to the serious damage to monolayers. The same results were obtained in our study in the D/P system. Since only FeSSIF but not FaSSIF caused the rapid decrease in TEER of Caco-2 monolayers (Fig. 1), the high osmotic pressure of FeSSIF (635 10 mOsm) was considered to cause the JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 95, NO. 9, SEPTEMBER 2006
damage to Caco-2 monolayers. Also, the high concentration of acetic acid in FeSSIF (8.65 g/L: 144 mM) was suspected to decrease the TEER since Mariadason et al. have reported the pronounced effect of acetic acid on the integrity of tight-junction of Caco-2 monolayer.25 Then, we have prepared FaSSIFmod and FeSSIFmod as isotonic media without acetic acid. Both media were confirmed not to cause any significant changes in the TEER value. In the case of intestinal membrane in vivo, the mucus layer on the surface of the epithelium prevents the direct contact of luminal fluid to the epithelial cells that may diminish the effect of osmotic pressure. Therefore, intestinal membrane is considered to have a high tolerance to the osmotic difference and therefore, not injured by the food intake in vivo. The dissolution of danazol in FeSSIFmod in the D/P system was 3.5-fold higher than in FaSSIFmod (Fig. 2a), which corresponds well with the dissolution study with FaSSIF and FeSSIF reported by Dressman et al.5 In addition, dissolution patterns of griseofulvin in FaSSIFmod and FeSSIFmod were not significantly different from those observed in FaSSIF and FeSSIF (data not shown). The in vitro permeation of propranolol, danazol, and albendazole was examined with FaSSIFmod and FeSSIFmod as the apical media in the D/P system. As shown in Table 2, the permeated amount of danazol and albendazole with FeSSIFmod was higher than that with FaSSIFmod. This result clearly indicates that the food effect on the absorption of danazol and albendazole can be observed in vitro by applying simulated intestinal fluid to the D/P system. However, although the dissolution of albendazole was much faster in FeSSIFmod, the permeated amount was only DOI 10.1002/jps
ESTIMATION OF FOOD EFFECTS ON ORAL DRUG ABSORPTION
slightly increased by FeSSIFmod, suggesting the lowered permeability to Caco-2 monolayers in this condition. Furthermore, the permeated amount of propranolol was dramatically decreased when FeSSIFmod was applied. As one of the reasons, the lower pH of FeSSIFmod (pH 5.0) might facilitate the ionization of the basic drugs (propranolol and albendazole) and lower the permeability to Caco-2 monolayer. The regional pH of the gastrointestinal fluid was varied by the intake of food. In the upper intestine, the pH value was higher (around 6.5) in the fasted state than that in the fed state (around 5.0). These pH values were adapted to FaSSIF and FeSSIF. However, in the middle and lower portion of the small intestine, the fluid pH is rather stable and not influenced by food intake.4 Assuming that the drug was absorbed from the whole regions of the small intestine, the pH value of the simulated intestinal fluid in the fed state should be adjusted to 6.5 in the D/P system. When FeSSIFmod6.5 was applied to the D/P system, the permeated amount of propranolol became almost the same as that obtained with FaSSIFmod. Moreover, the permeated amount of albendazole with FeSSIFmod6.5, became approximately 1.5-fold higher than that with FeSSIFmod. In contrast, no significant differences were observed in the permeated amount of danazol when FeSSIFmod6.5 was applied instead of FeSSIFmod. In the case of poorly water-soluble drugs such as albendazole, danazol, and griseofulvin, D/P system gave the higher permeated amount in FeSSIFmod6.5 condition than in FaSSIFmod one (Tabs. 2 and 3, that reflected correctly the in vivo food effect on oral absorption. However, the permeated amount of other drugs from FeSSIFmod6.5 was lesser than that from FaSSIFmod. As a result, the correlation curve for the fed state-absorption shifted to the left of that for the fasted state in Figure 3. This could be due to (1) in the case of lipophilic drugs, such as ketoprofen or metoprolol, dissolved drugs are taken up into micelles formed by NaTC and lecithin that reduced the free concentration of drugs in the apical solution, (2) in the case of hydrophilic drugs, such as acyclovir, the higher viscosity of FeSSIFmod6.5 reduced the diffusivity of drugs in the apical side that lower the permeation into basal side. These situations should also occur in vivo, however, as a final absorbed fraction, in vivo absorption is rather insensitive to these factors because, once dissolved in the lumen, lipophilc drugs (with high permeability) will be absorbed almost completely in vivo regardless of the presence of micelles. Also, strong DOI 10.1002/jps
2059
contraction of the GI tract will mask the effect of reduced diffusivity of drugs. In contrast, permeated amount in the D/P system is considered to be sensitive to these factors. In order to minimize these differences between in vivo and in vitro parameters and to enable the accurate simulation of the effect of food, we have described the standard curve for each fasted and fed state, individually (Fig. 3). In order to obtain the parameters for in vivo and in vitro correlation (IVIVC) in Figure 3, we have used Hill type equation because this equation has often been used to describe the correlation having a maximum value of the response. The rationale of the obtained IVIVC in oral drug absorption in the fasted and fed states was confirmed by the following studies in the D/P system, (1) prediction of the effect of food intake on the in vivo absorption of albendazole, (2) estimation of the effect of lipidbased formulation on the in vivo absorption of danazol under the fasted and fed states. Lange et al. have observed the in vivo oral absorption of albendazole under the fasted and the fed states in humans26 and reported that the area under the plasma concentration-time curve (AUC) of albendazole was enhanced approximately fourfold by food intake (Tab. 4). From in vitro data in the D/P system, the absorption of albendazole in the fed state was predicted to be approximately 3.5-fold higher than that in the fasted state (14% and 49% of dose, respectively, Table 4), corresponding well with in vivo observation. Charman et al. have reported that the emulsion formulation (with glycerol mono-oleate) enhanced the oral absorption of danazol as well as minimized the effect of food on it.1 In this study, Gelucire1 44/ 14 was used as the emulsifying reagent in the D/P system. Gelucire1 44/14 is known to spontaneously form a microemulsion in the GI tract after oral administration by mixing with bile juice, resulting in the improved oral absorption of poorly soluble drugs.27,28 Gelucire1 44/14 markedly enhanced the permeated amount of danazol in the D/P system both from FaSSIFmod and FeSSIFmod6.5 (Tab. 5). The predicted human oral absorption of danazol from in vitro results with Gelucire1 44/14 was 74% under the fasted state and 97% under the fed state. These results coincided well with the in vivo observation of danazol absorption with emulsion formulation, although the estimated absorption in the fasted state was still lower than that in the fed state. In the report of Charman et al., 30 mL of emulsion formulation including 100 mg of danazol was used JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 95, NO. 9, SEPTEMBER 2006
2060
KATAOKA ET AL.
for clinical study. In our study, 50 mL of the lipidbase formulation including 1 mg of danazol was applied. Relative amount of lipids content in our study is, therefore, much smaller than that in the clinical study. This could be a reason of lower estimation of danazol absorption in the D/P system especially under the fasted state. The permeated amount in the D/P system could be expressed as Z2 Pemeated amountð=2 hÞ ¼ Peff S Cdissolved dt 0
¼ Peff S AUCdissolved ð2Þ where Peff is an effective permeability, S is a surface area of Caco-2 monolayer and Cdissolved is a concentration of dissolved drug in the apical solution. From this equation, it is obvious that the permeated amount of drugs in the D/P system was affected not only by the solubility and permeability, but also by dissolution rate in the apical side. For instance, if the dissolution was much faster than the permeation, the permeated amount became the function of permeability and solubility, whereas if the dissolution was slower than the permeation, the permeated amount was profoundly affected by the dissolution rate, which corresponds to the dissolution rate-limited absorption. Since this is also true for in vivo, the D/P system enables the evaluation of oral drug absorption in all cases of permeability, solubility and dissolution rate limited. Consequently, this system could be a useful tool to screen the effect of food on the oral absorption of new drug candidates at the early stage of drug development in the pharmaceutical industry. Recently, the oral absorption of some drugs, such as alendronate sodium hydrate or nadolol,29–32 has been reported to be inhibited by food intake in humans. This negative effect of food on oral drug absorption is now under investigation in the D/P system and is a subject for future report.
CONCLUSIONS In this study, we have established the in vitro system to evaluate the effect of food on the oral absorption of poorly soluble drugs. In addition, this system was found to be able to detect the effect of formulations on drug absorption. In vitro D/P system may be a useful tool, not only to JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 95, NO. 9, SEPTEMBER 2006
predict the oral drug absorption under fasted and fed states, but also to consider the effectiveness of drug formulations to improve the oral absorption of poorly soluble drugs.
REFERENCES 1. Charman WN, Rogge MC, Boddy AW, Berger BM. 1993. Effect of food and a monoglyceride emulsion formulation on danazol bioavailability. J Clin Pharmacol 33:381–386. 2. Crounse RG. 1963. Effective use of griseofulvin. Arch Dermatol 87:176–178. 3. Galia E, Nicolaides E, Horter D, Lobenberg R, Reppas C, Dressman JB. 1998. Evaluation of various dissolution media for predicting in vivo performance of class I and II drugs. Pharm Res 15: 698–705. 4. Dressman JB, Amidon GL, Reppas C, Shah VP. 1998. Dissolution testing as a prognostic tool for oral drug absorption: Immediate release dosage forms. Pharm Res 15:11–22. 5. Dressman JB, Reppas C. 2000. In vitro-in vivo correlations for lipophilic, poorly water-soluble drugs. Eur J Pharm Sci 11:S73–S80. 6. Ingels F, Deferme S, Destexhe E, Oth M, van den Mooter G, Augustijns P. 2002. Simulated intestinal fluid as transport medium in the Caco-2 cell culture model. Int J Pharm 232:183–192. 7. Ingels FM, Augustijns PF. 2003. Biological, pharmaceutical, and analytical considerations with respect to the transport media used in the absorption screening system, Caco-2. J Pharm Sci 92: 1545–1558. 8. Ingels F, Beck B, Oth M, Augustijns P. 2004. Effect of simulated intestinal fluid on drug permeability estimation across Caco-2 monolayers. Int J Pharm 274:221–232. 9. Kataoka M, Masaoka Y, Yamazaki Y, Sakane T, Sezaki H, Yamashita S. 2003. In vitro system to evaluate oral absorption of poorly water-soluble drugs: Simultaneous analysis on dissolution and permeation of drugs. Pharm Res 20:1674–1680. 10. Ginski MJ, Taneja R, Polli JE. 1999. Prediction of dissolution-absorption relationships from a continuous dissolution/Caco-2 System. AAPS PharmSci 1: E3. 11. Ginski MJ, Polli JE. 1999. Prediction of dissolutionabsorption relationships from a dissolution/Caco-2 system. Int J Pharm 177:117–125. 12. 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. 13. Motz S, Schaefer UF, Eichinger T, Lehr CM. 2004. Development of a flow through system with DOI 10.1002/jps
ESTIMATION OF FOOD EFFECTS ON ORAL DRUG ABSORPTION
14.
15. 16.
17.
18.
19.
20.
21.
22.
23.
24.
combined dissolution and permeation devices. Globalization of Pharmaceutics Education Network. Motz S, Schaefer UF, Eichinger T, Lehr CM. 2004. Development of a flow through system with combined dissolution and permeation devices. Pharmaceutical Sciences World Congress. Welling PG. 1989. Effects of food on drug absorption. Pharmacol Ther 43:425–441. Olkkola KT, Brunetto AV, Mattila MJ. 1994 . Pharmacokinetics of oxicam nonsteroidal antiinflammatory agents. Clin Pharmacokinet 26: 107–120. Verbeeck RK, Richardson CJ, Blocka KL. 1986. Clinical pharmacokinetics of piroxicam. J Rheumatol 13:789–796. Bannwarth B, Lapicque F, Netter P, Monot C, Tamisier JN, Thomas P, Royer RJ. 1988. The effect of food on the systemic availability of ketoprofen. Eur J Clin Pharmacol 33:643–645. Melander A, Stenberg P, Liedholm H, Schersten B, Wahlin-Boll E. 1979. Food-induced reduction in bioavailability of atenolol. Eur J Clin Pharmacol 16:327–330. Yamaoka K, Tanigawara Y, Nakagawa T, Uno T. 1981. A pharmacokinetic analysis program (multi) for microcomputer. J Pharmacobiodyn 4:879–885. Peeters TL, Vantrappen G, Janssens J. 1980. Bile acid output and the interdigestive migrating motor complex in normal and in cholecystectomy patients. Gastroenterology 79:678–681. Tangerman A, van Schai A, van der Hoek EW. 1986. Analysis of conjugated and unconjugated bile acids in serum and jejunal fluid of normal subjects. Clin Chim Acta 159:123–132. Schulte-Lobbert S, Westerhoff K, Wilke A, Schubert-Zsilavecz M, Wurglics M. 2003. Development of a high-performance-liquid-chromatographic method for the determination of biapigenin in biorelevant media. J Pharm Biomed Anal 33: 53–60. Kostewicz ES, Wunderlich M, Brauns U. Becker R, Bock T, Dressman JB. 2004. Predicting the precipitation of poorly soluble weak bases upon entry
DOI 10.1002/jps
25.
26.
27.
28.
29.
30.
31.
32.
2061
in the small intestine. J Pharm Pharmacol 56: 43–51. Mariadason JM, Barkla DH, Gibson PR. 1997. Effect of short-chain fatty acids on paracellular permeability in Caco-2 intestinal epithelium model. Am J Physiol 272:G705–G712. Lange H, Eggers R, Bircher J. 1988. Increased systemic availability of albendazole when taken with a fatty meal. Eur J Clin Pharmacol 34:315– 317. Hauss DJ, Fogal SE, Ficorilli JV, Price CA, Roy T, Jayaraj AA, Keirns JJ. 1998. Lipid-based delivery systems for improving the bioavailability and lymphatic transport of a poorly water-soluble LTB4 inhibitor. J Pharm Sci 87:164–169. Joshi HN, Tejwani RW, Davidovich M, Sahasrabudhe VP, Jemal M, Bathala MS, Varia SA, Serajuddin AT. 2004. Bioavailability enhancement of a poorly water-soluble drug by solid dispersion in polyethylene glycol-polysorbate 80 mixture. Int J Pharm 269:251–258. Gertz BJ, Holland SD, Kline WF, Matuszewski BK, Freeman A, Quan H, Lasseter KC, Mucklow JC, Porras AG. 1995. Studies of the oral bioavailability of alendronate. Clin Pharmacol Ther 58:288– 298. Yamaguchi T, Ikeda C. Sekine Y. 1986. Intestinal absorption of a beta-adrenergic blocking agent nadolol. I. Comparison of absorption behavior of nadolol with those of other beta-blocking agents in rats. Chem Pharm Bull (Tokyo) 34:3362–3369. Yamaguchi T, Ikeda C, Sekine Y. 1986. Intestinal absorption of a beta-adrenergic blocking agents nadolol. II. Mechanism of the inhibitory effect on the intestinal absorption of nadolol by sodium cholate in rats. Chem Pharm Bull (Tokyo) 34: 3836–3843. Yamaguchi T, Oida T, Ikeda C, Sekine Y. 1986. Intestinal absorption of a beta-adrenergic blocking agent nadolol. III. Nuclear magnetic resonance spectroscopic study on nadolol-sodium cholate micellar complex and intestinal absorption of nadolol derivatives in rats. Chem Pharm Bull (Tokyo) 34:4259–4264.
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 95, NO. 9, SEPTEMBER 2006
Received 21 November 2005; revised 25 April 2006; accepted 6 May 2006 Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/jps.20691
ABSTRACT: The aim of the present work was to establish appropriate conditions for the dissolution/permeation system (D/P system) to estimate the effect of food intake on oral drug absorption. The D/P system is an in vitro assay system to evaluate the drug dissolution and permeation processes after oral administration. Caco-2 monolayer was used as a model membrane of the intestinal epithelium. In this study, two types of simulated intestinal fluid reflecting the fasted and the fed state conditions of the human gastrointestinal tract were used. Drugs were applied to the D/P system as a powder, then, permeated amounts of drugs into the basal side were monitored. A sigmoidal correlation was obtained between in vivo oral absorption (% absorbed of dose) and in vitro permeated amount (% of dose/2 h) under both states. From the D/P system, the estimated absorption of albendazole in both states was found to correspond well with in vivo observation. Moreover, the D/P system could estimate the effect of self-emulsifying formulation on the oral absorption of danazol, quantitatively. In conclusion, the D/P system was proved to be a useful assay system not only for the oral absorption of drugs, but also for the food effect on the absorption. ß 2006 Wiley-Liss, Inc. and the American Pharmacists Association J Pharm Sci 95:2051–2061, 2006
Keywords: absorption; biopharmaceutics classification system (BCS); Caco-2 cells; dissolution; food effects; in vitro/in vivo correlations (IVIVC); permeability; solubility
INTRODUCTION The oral absorption of poorly water-soluble drugs, such as danazol and griseofulvin, is known to increase when they are administered after food intake.1,2 As a main cause of the effect of food, the secretion of bile juice into the GI tract is accelerated by food intake, which enhances the solubility and dissolution rate and therefore the absorption of these drugs after oral administration. Within pharmaceutical companies, in vivo animal studies have been carried out to detect the effect of food on oral absorption of new drug Correspondence to: Makoto Kataoka (Telephone: þ81-72866-3126; Fax: þ81-72-866-3126; E-mail: [email protected]) Journal of Pharmaceutical Sciences, Vol. 95, 2051–2061 (2006) ß 2006 Wiley-Liss, Inc. and the American Pharmacists Association
candidates before advancing to clinical studies. However, animal studies often show large variations depending on the species and might result in an uncertain estimation on drug absorption in humans. Therefore, an in vitro screening system that can evaluate the effect of food intake on the oral absorption of drugs is strongly desired. Galia et al.3 have proposed the use of fasted and/ or fed state simulated intestinal fluids (FaSSIF and FeSSIF) in the drug dissolution study to simulate the effect of food on drug dissolution, then on the bioavailability/bioequivalence of commercial drug products. FaSSIF and FeSSIF contain taurocholate (NaTC) and lecithin as bile acid and lipid to mimic the composition of human intestinal fluid. Dressman et al. have reported that FaSSIF and FeSSIF can be used in the drug dissolution test and are useful in formulation
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 95, NO. 9, SEPTEMBER 2006
2051
2052
KATAOKA ET AL.
development.4,5 On the other hand, Ingels et al. have reported that the use of FaSSIF as the apical medium in the Caco-2 permeation study may improve the biorelevance of the model to evaluate the oral absorption in humans.6–8 Both approaches with simulated intestinal fluids may not be sufficient to evaluate the effect of food intake on the total processes of oral drug absorption, because they consist of two processes, the dissolution and permeation of drugs, which occur sequentially in the GI tract. In order to evaluate the oral absorption of poorly soluble drugs, we have developed an in vitro system with which both dissolution and permeation processes can be analyzed simultaneously. In this system (dissolution/permeation system, D/P system), a Caco-2 monolayer is mounted in the side-by-side chambers. Both the apical and basal sides of the monolayer are filled with media and are constantly stirred. Drugs are applied to the apical medium as a solid form (powder or granule); then, dissolution in the apical side and permeation to the basal side of the drug are simultaneously monitored. In the previous report, it was demonstrated that the amount of drug that permeated to the basal side of the Caco-2 monolayer showed a sigmoidal correlation to their absorption in vivo in humans, regardless of the class of drugs in the Biopharmaceutics Classification System (BCS). Consequently, the D/P system is a useful tool to evaluate the oral absorption of poorly soluble drugs after oral administration as the solid dosage form.9 Ginski et al. have reported the use of the continuous dissolution/Caco-2 system to investigate drug absorption.10,11 They observed the dissolution-absorption relationships and determined the rate limiting process of absorption in rapid and sustained dissolving formulations. Kobayashi et al.12 have reported a system in which the effect of pH change in the GI tract on drug dissolution and permeation can be evaluated. Furthermore, Motz et al. have developed a flow through permeation cell system to simultaneously monitor the dissolution and permeability of drugs.13,14 Those studies and devices have generated the useful information in considering the dissolution-permeation relationship in oral drug absorption. However, in order to predict the in vivo oral absorption of drugs quantitatively from the in vitro study, physiological conditions in the human GI tract, such as the fluid volume, fluid composition, and fluid pH must be taken into consideration, especially for the absorption of poorly JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 95, NO. 9, SEPTEMBER 2006
water-soluble drugs. In our D/P system, the assay conditions were fixed carefully, based on the in vivo physiological conditions of the human GI tract to successfully obtain the IVIVC in oral drug absorption.9 In this study, we have tried to predict the effect of food intake on oral drug absorption in humans from in vitro experiments with the D/P system. Various kinds of simulated intestinal fluids were prepared and applied to the apical side of the D/P system to reflect in vivo fasted and fed state conditions in the GI tract. Then, we compared the results for the D/P system with in vivo observations in humans.
EXPERIMENTAL Materials The Caco-2 cell line was purchased from American Type Culture Collection (Rockville, MD) at passage 17. Dulbecco’s modified Eagle medium (D-MEM) was purchased from Sigma-Aldrich (St. Louis, MO). Nonessential amino acids (NEAA), fetal bovine serum (FBS), L-glutamate, trypsin (0.25%)EDTA (1 mM) and antibiotic-antimycotic mixture (10000 U/mL penicillin G, 10000 mg/mL streptomycin sulfate and 25 mg/mL amphotericin B in 0.85% saline) were purchased from Gibco Laboratories (Lenexa, KS). Danazol was purchased from Sigma-Aldrich. Albendazole, egg-phosphatidylcholine (lecithin), sodium taurocholate (NaTC) and bovine serum albumin (BSA) were purchased from WAKO Pure Chemical Industries, Ltd. (Japan). Gelucire1 44/14 was obtained from ´ (France). All other reagents used GATTEFOSSE were of the highest purity.
Preparation of Caco-2 Monolayer Caco-2 cells were grown in D-MEM supplemented with 10% FBS, 1% L-glutamate, 1% NEAA and 5% antibiotic-antimycotic solution as culture medium at 38C in culture flasks (Nippon Becton Dickinson Co., Ltd., Tokyo, Japan) in humidified air with a 5% CO2 atmosphere. Cells were harvested with typsin-EDTA and seeded on polycarbonate filters (0.3 mm pores, 4.20 cm2 growth area) inside a cell culture insert (Nippon Becton Dickinson Co., Ltd., Tokyo, Japan) at a density of 3 105 cells/filter. The culture medium (1.5 mL in the insert and 2.6 mL in the well) was replaced every 48 h for the first 6 days and every 24 h thereafter. After DOI 10.1002/jps
ESTIMATION OF FOOD EFFECTS ON ORAL DRUG ABSORPTION
18–21 days in culture, the Caco-2 monolayer was utilized for the following experiments.
permeation processes of drugs can be analyzed simultaneously.9 In this system, the Caco-2 monolayer is mounted in side-by-side chambers. The effective surface area of the Caco-2 monolayer is 1.77 cm2. Both sides of the Caco-2 monolayer are consistently stirred at 200 rpm with magnetic stirrers. The volume of apical and basal sides is set to 8 and 5.5 mL, respectively. As an apical medium, various fluids shown in Table 1 were used according to the purpose of the study. TM containing a 4.5 w/v% concentration of BSA (pH 7.4) was used as a basal medium in all studies. All experiments were performed at 378C.
Preparation of Various Fluids As a buffer solution in this study, Hank’s balanced salts solution (HBSS), containing KCl 5.36 mM, NaCl 136.89 mM, Na2HPO4 0.34 mM, KH2PO4 0.44 mM, NaHCO3 4.17 mM, CaCl2 1.26 mM, MgCl2 0.49 mM, MgSO4 0.41 mM and glucose 19.45 mM, was used (transport medium, TM). The pH of the TM was adjusted to 5.0, 6.5 or 7.4 with HEPES. Simulated intestinal fluids (FaSSIF and FeSSIF) were prepared according to the report of Galia et al.3 as FaSSIF; KH2PO4 3.9 g/L, NaTC 3 mM, Lecithin 0.75 mM, KCl 7.7 g/L, adjusted pH to 6.5 by NaOH, and FeSSIF; Acetic acid 8.65 g/L, NaTC 15 mM, Lecithin 3.75 mM, KCl 15.2 g/L, adjusted pH to 5.0 by NaOH. Modified simulated intestinal fluids, FaSSIFmod and FeSSIFmod, were prepared based on TM by adding NaTC and lecithin as the same concentration with those in FaSSIF and FeSSIF, respectively (Tab. 1). The pH of FaSSIFmod and FeSSIFmod was adjusted to 6.5 and 5.0 with HEPES, respectively. FeSSIFmod6.5 has a same composition with FeSSIFmod, but pH was adjusted to 6.5.
Effect of Various Fluids on the Transepithelial Electric Resistance (TEER) of Caco-2 Monolayers The effects of various fluids (applied to the apical side) on the TEER of the Caco-2 monolayer were monitored in the D/P system. Before applied to the D/P system, the Caco-2 monolayer was preincubated in the culture well with TM (pH 6.5 for the apical and 7.4 for the basal side). After preincubation for 20 min, the Caco-2 monolayer with support filter was taken from the insert and was mounted between the chambers of the D/P system. Then, the apical side of the monolayer was filled with various fluids in Table 1 (TM, TM þ NaTC, FaSSIF, FeSSIF, FaSSIFmod, and FeSSIFmod). The basal side was filled with TM with 4.5 w/v% BSA (pH 7.4) in all experiments. Then, the TEER of the Caco-2 monolayer was routinely measured at appropriate intervals over
Chambers for the Dissolution/Permeation System (D/P System) The D/P system is an in vitro assay system of oral drug absorption with which both dissolution and
Table 1.
2053
Various Simulated Intestinal Fluids Applied to D/P System
TMb TM þ NaTCc Simulated intestinal fluidd FaSSIF FeSSIF Simulated intestinal fluid (modified) FaSSIFmod FeSSIFmod FeSSIFmod6.5
pH
Taurocholate
Lecithin
Osmolality (mOsm)a
6.5 6.5
— 5 mM
— —
306.3 0.6 313.7 0.6
6.5 5.0
3 mM 15 mM
0.75 mM 3.75 mM
270 10 635 10
6.5 5.0 6.5
3 mM 15 mM 15 mM
0.75 mM 3.75 mM 3.75 mM
311.7 0.6 327.0 1.0 325.7 0.6
a
Osmolality was measured by OSMOSTAT (OM-6040, ARKARY). TM; Transport medium (HBSS supplemented with 25 mM glucose). c NaTC; Sodium taurocholate. d Galia et al.3 b
DOI 10.1002/jps
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 95, NO. 9, SEPTEMBER 2006
2054
KATAOKA ET AL.
2 h by Millcell-ERS (MILLIPORE, MA). In order to apply the electrodes to the D/P system, the shape of the electrodes were slightly modified.
Dissolution Profile of Danazol in the D/P System The dissolution profile of danazol was observed by using only the apical chamber without mounting Caco-2 monolayers. Instead of Caco-2 monolayers, a flat sheet of aluminum foil was mounted between the chambers to prevent the leakage of apical solution to the basal compartment. As an apical medium, FaSSIFmod or FeSSIFmod was used. One milligram of danazol was applied to the apical chamber as a powder and aliquots of samples were routinely taken from the apical solution over 2 h. Samples were filtrated through cellulose acetate filter to determine the dissolved amount of danazol.
Permeation Profiles of Various Drugs in the D/P System Permeation profiles of various drugs (listed in Tab. 2) were observed in the D/P system with FaSSIFmod, FeSSIFmod, or FeSSIFmod6.5 as an apical medium. The Caco-2 monolayer was applied to the D/P system using the same method as described above. The apical side of the monolayer was filled with FaSSIFmod, FeSSIFmod or FeSSIFmod6.5, and the basal side was filled with TM with 4.5 w/v% BSA (pH 7.4). Drugs were applied to the apical side as a powder or as a suspension with Gelucire1 44/14. An applied amount of drugs were set to 1% of their clinical dose. Then, aliquots of samples were taken from the basal solution at appropriate intervals over 2 h. The volume of the basal solution was maintained by adding fresh medium. Samples were
pretreated with methanol and centrifugation and the deproteinized supernatant was used to determine the amount of drugs permeated through the Caco-2 monolayer. After finishing the experiment, the apical solution was immediately collected and filtered through a cellulose acetate filter to determine the final concentration of the dissolved drug. The TEER of the Caco-2 monolayer was checked before and after the experiment. In most of the experiments, no significant decreases in the TEER value during the experiments were observed (data not shown). Preparation of Self-Emulsifying Formulation Gelucire1 44/14 was used to prepare the selfemulsifying formulation of danazol. Danazol was suspended into Gelucire1 44/14 and a concentration of 1 mg/50 mL. Then, 1 mg of danazol (thus 50 mL of suspension) was applied to the apical side of the D/P system, followed by the same procedures of sampling of both the apical and basal solution for drug concentration measurement. Analytical Methods All samples were analyzed with a reversed-phase HPLC system (LC-10A Shimadzu Co., Kyoto, Japan) equipped with a variable wavelength ultraviolet detector (SPD-10A, Shimadzu Co., Kyoto Japan). The column (J’sphere ODS-H80 75 4.6 mm I.D., YMC, Japan) was used with a mobile phase consisting of 50 mM phosphate buffer (pH 2.5) and acetonitrile. Acyclovir, albendazole, atenolol, carbamazepine, chlorpheniramine, danazol, griseofulvin, ketoprofen, metoprolol, nadolol, piroxicam, pranlukast, propranolol, and warfarin were quantified with the variable ultraviolet detector at 255, 310, 226, 285, 225, 286, 325, 260, 225, 215, 326, 260, 228 and 311 nm, respectively.
Table 2. Effect of Simulated Intestinal Fluids on Dissolved and Permeated Amount of Various Drugs in D/P System Dissolved Amount (% of dose/2 h)
Propranolol Danazol Albendazole
Applied Amount (mg)
FaSSIFmod
0.1 1.0 4.0
3.698 0.242
FeSSIFmod
FeSSIFmod6.5
Completely dissolved 13.307 14.288 2.300 1.928
Permeated Amount (% of dose/2 h) FaSSIFmod
FeSSIFmod
FeSSIFmod6.5
4.430 0.125 0.041
0.905 0.256 0.088
3.401 0.250 0.122
Values are expressed as the mean of at least three independent experiments. JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 95, NO. 9, SEPTEMBER 2006
DOI 10.1002/jps
ESTIMATION OF FOOD EFFECTS ON ORAL DRUG ABSORPTION
RESULTS Effect of Simulated Intestinal Fluids on the TEER of Caco-2 Monolayers Figure 1 shows the effect of various fluids on the TEER of Caco-2 monolayers. When FeSSIF was applied to the apical side of Caco-2 monolayers, the TEER value decreased rapidly to almost zero after 1 h incubation. Since the osmolarity of FeSSIF is twofold higher than that of isotonic fluid, this hyper-osomolarity would cause the destructive damage to the monolayers. Then, isotonic fluid having the same concentration of NaTC and lecithin with FeSSIF was prepared (FeSSIFmod). After applying FeSSIFmod, the TEER value decreased only slightly over the 2 h. Since this pattern of TEER change was almost same that observed with TM, it was apparent that FeSSIFmod did not cause any significant damage to the Caco-2 monolayers. The other fluids listed in Table 1 were also verified to have no significant effect on the TEER of Caco-2 monolayers during the experimental period. In the following study, therefore, FaSSIFmod (having the same concentration of NaTC and lecithin with FaSSIF, but with the osmolarity was finely adjusted to 312 mOsm) and FeSSIFmod were used as the fasted and fed state simulated intestinal fluids, instead of FaSSIF and FeSSIF.
Figure 1. Effect of various simulated intestinal fluids on the TEER of Caco-2 monolayers in the D/P system. TM (&), TM þ NaTC (&), FaSSIF (~), FeSSIF(~), FaSSIFmod (*), and FeSSIFmod (*) were used as the simulated intestinal fluids in the apical side of the D/P system then the time-courses of the TEER were measured for 2 h. The data are expressed as the mean SE of at least three independent experiments. DOI 10.1002/jps
2055
Effect of Simulated Intestinal Fluids on Drug Dissolution and Permeation The time courses of the dissolution and permeation of danazol (applied amount ¼ 1 mg) in the D/P system were measured with FaSSIFmod and FeSSIFmod as apical media. When FeSSIFmod was used, danazol was dissolved rapidly and the dissolved amount reached 13.3% of the dose for 2 h. In contrast, the dissolution of danazol was apparently slow and reached only 3.7% of the dose in FaSSIFmod (Fig. 2a). The permeated amount from the FeSSIFmod medium was also
Figure 2. Effect of simulated intestinal fluids on dissolution (a) and permeation (b) of danazol in the D/P system. The time-courses of danazol dissolution and permeation were observed for 2 h in the D/P system when FaSSIFmod (*) and FeSSIFmod (*) were used as the simulated intestinal fluids in the apical side of the D/P system. The data are expressed as the mean SE of at least three independent experiments. JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 95, NO. 9, SEPTEMBER 2006
2056
KATAOKA ET AL.
in vitro permeated amount of drugs in the D/P system (% of dose/2 h) is summarized in Table 3, with the fraction dose absorbed in humans in the fasted and fed states. Among the drugs listed in Table 3, the oral absorption of danazol and griseofulvin is known to be affected by food intake. Charman et al. have reported that the oral absorption of danazol in the fed state was approximately threefold higher than that in the fasted state in humans. The oral bioavailability of griseofulvin was reported to be dose-dependent and increased almost twofold (70%–120%) by food intake.2 The oral absorptions of danazol and griseofulvin in the fasted and fed states were calculated based on these reports in Table 3. In the case of danazol, the AUC after oral administration as the emulsion formulation under the fed state was reported to show the linear relation with dose up to 200 mg. Since the oral absorption of danazol is limited by its solubility, this linearity in AUC means that the absorption was almost complete. Then, we have assumed that the oral absorption of danazol in human (applied as 100 mg emulsion formulation under the fed state) was 100%. In contrast, the oral absorption of drugs classified to class 1 in BCS, such as propranolol, metoprolol, piroxicam, and ketoprofen was not affected by food intake.15–18 The oral bioavailability of
greater than that from FaSSIFmod, although the difference was rather small (about twofold difference) compared to that observed for dissolution (Fig. 2b). The same experiments were performed for propranolol and albendazole. In the FeSSIFmod medium, the dissolved amount of albendazole was much greater than that in FaSSIFmod, while the permeated amount increased only 2-fold in FeSSIFmod. In the case of propranolol, the permeated amount was markedly reduced in FeSSIFmod despite of the complete dissolution in both media. Then, one more fed state simulated fluid was prepared by adjusting the pH of FeSSIFmod to 6.5 (FeSSIFmod6.5). As summarized in Table 2, FeSSIFmod6.5 dramatically increased the permeated amount of propranolol and albendazole, but not that of danazol. Correlation Between Human Absorption and In Vitro Permeation in the D/P System In order to obtain the correlation between human oral absorption and the in vitro permeated amount in D/P system, FaSSIFmod and FeSSIF mod6.5 were used as apical media. Thirteen drugs were used as references and were applied to the apical side as a powder. The amount of each drug applied was set to 1% of its clinical dose. The
Table 3. Parameters of 13 Drugs for Dissolution and Permeation in D/P System and their Absorption in Humans
Acyclovir Atenolol Carbamazepine Chlorpheniramine Danazol Griseofulvin (low dose) Griseofulvin (high dose) Ketoprofen Metoprolol Nadolol Piroxicam Pranlukast Propranolol Warfarin
Clinical Dosea (mg)
Applied Amountb (mg)
FaSSIFmod Permeated Amount (% of dose/2 h)
200 50 100 6 100 250 500 50 40 80 10 225 10 5
2.0 0.5 1.0 0.06 1.0 2.5 5.0 0.5 0.4 0.8 0.1 2.25 0.1 0.05
0.049 0.097 8.393 5.364 0.125 0.640 0.244 6.295 4.268 0.124 15.659 0.072 4.430 10.793
FeSSIFmod6.5 Permeated Amount (% of dose/2 h)
Human Fasted (%)
Absc Fed (%)
0.029 0.103 — — 0.250 — 0.295 1.742 3.119 — 14.695 0.040 3.401 9.751
20 50 83 80 24 80 40 85 95 34 95 15 95 94
20 50 — — 76 — 78 85 95 — 95 15 95 94
The Pharmacological Basis of Therapeutics (10th edition). Values were expressed as the mean of at least three experiments. a Oral dose. b Applied amount to D/P system (1% of clinical dose). c Represents mean values for % absorption obtained from individual d rug references or the Goodman & Gilman’s. JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 95, NO. 9, SEPTEMBER 2006
DOI 10.1002/jps
ESTIMATION OF FOOD EFFECTS ON ORAL DRUG ABSORPTION
atenolol (a moderately absorbed drug classified to class 3) was reported to be reduced slightly by food intake, but not a significant extent.19 Concerning other class 1 and 3 drugs, warfarin and acyclovir, therefore, we have assumed that their absorption will not be affected by food intake. Then, the following equation was applied to the data sets in Table 3 to obtain the correlation between the in vivo human absorption and the in vitro permeated amount (% of dose/2 h) in the D/P system under the fasted and fed conditions, individually. Human absorption ð%Þ ¼
Absmax PAg PAg50 þ PAg
ð1Þ
where Absmax is the maximum absorption (defined as 100%), PA is the in vitro permeated amount in the D/P system (% of dose/2 h), PA50 is the permeated amount, which corresponds to 50% of the absorption in vivo, and g is a Hill’s coefficient. PA50 and g were obtained by fitting the permeated amount (PA) of drugs in the D/P system and their oral absorption in human (Tab. 3) by using MULTI program developed by Yamaoka et al.20 The correlation curves are described in Figure 3 by incorporating the obtained value of PA50 and g to Eq. 1. PA50 and g in the fasted state were 0.334 0.096 (SD) and 0.883 0.178, and in the fed state were 0.121 0.021 and 1.334 0.198, respectively. In Figure 3, the correlation curve for the fed state-absorption shifted to the left of that for the fasted state. This was due to the fact that
2057
the permeated amount of most drugs in the D/P system was lower in FeSSIFmod6.5 than that in FaSSIFmod. Estimation of In Vivo Human Absorption from In Vitro Data in D/P System By using the in vivo and in vitro correlation (IVIVC) of oral drug absorption in Figure 3, we have tried to predict the effect of food intake on the absorption of albendazole from the in vitro data in the D/P system. When 4 mg of albendazole (clinical dose: 400 mg) was applied to the D/P system as a powder, the permeated amount (% of dose/2 h) was 0.041 % with FaSSIFmod and 0.122 % with FeSSIFmod6.5, respectively (Tab. 2). These values were plotted on the standard curves in Figure 3 to predict the in vivo absorption under the fasted and fed states. The estimated absorption of albendazole was low in the fasted state and increased 3.5-fold in the fed state (Tab. 4). Danazol was suspended in Gelucire1 44/14 (1 mg/50 mL) and was applied to the apical side of the D/P system. The dissolved amount of danazol with Gelucire1 44/14 in FaSSIFmod reached 52.3% of the dose/2 h; in FeSSIFmod6.5 danazol was completely dissolved. From the permeated amount, the in vivo absorption of danazol with Gelucire1 44/14 was estimated as 74% in the fasted and 97% in the fed state, respectively (Tab. 5), suggesting that Gelucire1 44/14 can enhance the absorption of danazol while minimizing the effect of food on it.
DISSCUSSION In the previous report, in order to simulate the in vivo oral absorption of poorly water-soluble drugs with an in vitro D/P system, the assay conditions were carefully fixed, based on the in vivo physiological conditions of the human gastrointestinal (GI) tract. One of the key factors Table 4. In Vivo and In Vitro Estimated Absorption of Albendazole Under Fasted and Fed States in Humans Figure 3. Correlation between the in vivo oral absorption in humans and the in vitro permeated amount in the D/P system. The symbols were the amounts permeated over 2 h in the D/P system when FaSSIFmod (*) and FeSSIFmod6.5 (*) were used as the simulated intestinal fluids in the apical side of the D/P system. Correlation curves in the fasted (dashed line) and the fed (solid line) states were obtained from Eq. 1. DOI 10.1002/jps
Food State
In Vivo Absorption AUCa (mmolh/L)
In Vitro Estimated Absorption (%)
5.0 3.0 19.9 8.2
14 49
Fasted Fed a
Data from Ref. 24.
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 95, NO. 9, SEPTEMBER 2006
2058
KATAOKA ET AL.
Table 5. In Vivo and In Vitro Estimated Absorption of Danazol as Solid Dosage Form and Lipid-Based Formulation In Vivo Absorption (Human) Dosage Form and Food State Solid dosage form Fasted state Fed state Lipid-based formulation Fasted state Fed state
In Vitro Estimated Absorption (D/P System)
AUCa (ngh/mL)
Oral Abs (%)
Dissolution (% of Dose)
Permeation (% of Dose)
Oral Abs (%)
204 125 639 259
24 76
3.698 14.288
0.125 0.250
30 71
779 189 844 194
92 100
52.335 109.262
1.101 1.787
74 97
Values are expressed as the mean of at least three independent experiments. a Data from Ref. 1.
to determine the absorption of poorly soluble drugs should be the balance of dose and water volume in the GI tract. In the D/P system, the applied amount of drugs were set to 1% of their clinical dose, because the volume of the apical chamber, 8 mL, could be regarded as about 1/100 of the physiological volume of the human GI tract that was reported to be approximately 500–1000 mL.4 Bile acids are secreted into the duodenum from the gall bladder. Their average concentration in the human intestinal tract (of healthy subjects) was reported to be 5–15 mM21,22 and highly affected by the food intake. In the previous study, a 5 mM concentration of NaTC was added to the apical solution, which corresponds to the concentration under fasted conditions in humans. In this study, to establish the in vitro conditions that can predict the effect of food intake on drug absorption quantitatively, simulated intestinal fluids were applied to the apical side of the D/P system. Simulated intestinal fluids, FaSSIF and FeSSIF, were proposed by Galia et al.3 and are now widely used as dissolution media to simulate the drug dissolution in the human GI tract under the fasted and fed states.4,5,23,24 Ingels et al. have reported the effect of simulated intestinal fluids on drug permeation across Caco-2 monolayers.6–8 Although FaSSIF did not alter the morphology and integrity of the Caco-2 monolayer, FeSSIF was found not to be suitable for use in Caco-2 assay due to the serious damage to monolayers. The same results were obtained in our study in the D/P system. Since only FeSSIF but not FaSSIF caused the rapid decrease in TEER of Caco-2 monolayers (Fig. 1), the high osmotic pressure of FeSSIF (635 10 mOsm) was considered to cause the JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 95, NO. 9, SEPTEMBER 2006
damage to Caco-2 monolayers. Also, the high concentration of acetic acid in FeSSIF (8.65 g/L: 144 mM) was suspected to decrease the TEER since Mariadason et al. have reported the pronounced effect of acetic acid on the integrity of tight-junction of Caco-2 monolayer.25 Then, we have prepared FaSSIFmod and FeSSIFmod as isotonic media without acetic acid. Both media were confirmed not to cause any significant changes in the TEER value. In the case of intestinal membrane in vivo, the mucus layer on the surface of the epithelium prevents the direct contact of luminal fluid to the epithelial cells that may diminish the effect of osmotic pressure. Therefore, intestinal membrane is considered to have a high tolerance to the osmotic difference and therefore, not injured by the food intake in vivo. The dissolution of danazol in FeSSIFmod in the D/P system was 3.5-fold higher than in FaSSIFmod (Fig. 2a), which corresponds well with the dissolution study with FaSSIF and FeSSIF reported by Dressman et al.5 In addition, dissolution patterns of griseofulvin in FaSSIFmod and FeSSIFmod were not significantly different from those observed in FaSSIF and FeSSIF (data not shown). The in vitro permeation of propranolol, danazol, and albendazole was examined with FaSSIFmod and FeSSIFmod as the apical media in the D/P system. As shown in Table 2, the permeated amount of danazol and albendazole with FeSSIFmod was higher than that with FaSSIFmod. This result clearly indicates that the food effect on the absorption of danazol and albendazole can be observed in vitro by applying simulated intestinal fluid to the D/P system. However, although the dissolution of albendazole was much faster in FeSSIFmod, the permeated amount was only DOI 10.1002/jps
ESTIMATION OF FOOD EFFECTS ON ORAL DRUG ABSORPTION
slightly increased by FeSSIFmod, suggesting the lowered permeability to Caco-2 monolayers in this condition. Furthermore, the permeated amount of propranolol was dramatically decreased when FeSSIFmod was applied. As one of the reasons, the lower pH of FeSSIFmod (pH 5.0) might facilitate the ionization of the basic drugs (propranolol and albendazole) and lower the permeability to Caco-2 monolayer. The regional pH of the gastrointestinal fluid was varied by the intake of food. In the upper intestine, the pH value was higher (around 6.5) in the fasted state than that in the fed state (around 5.0). These pH values were adapted to FaSSIF and FeSSIF. However, in the middle and lower portion of the small intestine, the fluid pH is rather stable and not influenced by food intake.4 Assuming that the drug was absorbed from the whole regions of the small intestine, the pH value of the simulated intestinal fluid in the fed state should be adjusted to 6.5 in the D/P system. When FeSSIFmod6.5 was applied to the D/P system, the permeated amount of propranolol became almost the same as that obtained with FaSSIFmod. Moreover, the permeated amount of albendazole with FeSSIFmod6.5, became approximately 1.5-fold higher than that with FeSSIFmod. In contrast, no significant differences were observed in the permeated amount of danazol when FeSSIFmod6.5 was applied instead of FeSSIFmod. In the case of poorly water-soluble drugs such as albendazole, danazol, and griseofulvin, D/P system gave the higher permeated amount in FeSSIFmod6.5 condition than in FaSSIFmod one (Tabs. 2 and 3, that reflected correctly the in vivo food effect on oral absorption. However, the permeated amount of other drugs from FeSSIFmod6.5 was lesser than that from FaSSIFmod. As a result, the correlation curve for the fed state-absorption shifted to the left of that for the fasted state in Figure 3. This could be due to (1) in the case of lipophilic drugs, such as ketoprofen or metoprolol, dissolved drugs are taken up into micelles formed by NaTC and lecithin that reduced the free concentration of drugs in the apical solution, (2) in the case of hydrophilic drugs, such as acyclovir, the higher viscosity of FeSSIFmod6.5 reduced the diffusivity of drugs in the apical side that lower the permeation into basal side. These situations should also occur in vivo, however, as a final absorbed fraction, in vivo absorption is rather insensitive to these factors because, once dissolved in the lumen, lipophilc drugs (with high permeability) will be absorbed almost completely in vivo regardless of the presence of micelles. Also, strong DOI 10.1002/jps
2059
contraction of the GI tract will mask the effect of reduced diffusivity of drugs. In contrast, permeated amount in the D/P system is considered to be sensitive to these factors. In order to minimize these differences between in vivo and in vitro parameters and to enable the accurate simulation of the effect of food, we have described the standard curve for each fasted and fed state, individually (Fig. 3). In order to obtain the parameters for in vivo and in vitro correlation (IVIVC) in Figure 3, we have used Hill type equation because this equation has often been used to describe the correlation having a maximum value of the response. The rationale of the obtained IVIVC in oral drug absorption in the fasted and fed states was confirmed by the following studies in the D/P system, (1) prediction of the effect of food intake on the in vivo absorption of albendazole, (2) estimation of the effect of lipidbased formulation on the in vivo absorption of danazol under the fasted and fed states. Lange et al. have observed the in vivo oral absorption of albendazole under the fasted and the fed states in humans26 and reported that the area under the plasma concentration-time curve (AUC) of albendazole was enhanced approximately fourfold by food intake (Tab. 4). From in vitro data in the D/P system, the absorption of albendazole in the fed state was predicted to be approximately 3.5-fold higher than that in the fasted state (14% and 49% of dose, respectively, Table 4), corresponding well with in vivo observation. Charman et al. have reported that the emulsion formulation (with glycerol mono-oleate) enhanced the oral absorption of danazol as well as minimized the effect of food on it.1 In this study, Gelucire1 44/ 14 was used as the emulsifying reagent in the D/P system. Gelucire1 44/14 is known to spontaneously form a microemulsion in the GI tract after oral administration by mixing with bile juice, resulting in the improved oral absorption of poorly soluble drugs.27,28 Gelucire1 44/14 markedly enhanced the permeated amount of danazol in the D/P system both from FaSSIFmod and FeSSIFmod6.5 (Tab. 5). The predicted human oral absorption of danazol from in vitro results with Gelucire1 44/14 was 74% under the fasted state and 97% under the fed state. These results coincided well with the in vivo observation of danazol absorption with emulsion formulation, although the estimated absorption in the fasted state was still lower than that in the fed state. In the report of Charman et al., 30 mL of emulsion formulation including 100 mg of danazol was used JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 95, NO. 9, SEPTEMBER 2006
2060
KATAOKA ET AL.
for clinical study. In our study, 50 mL of the lipidbase formulation including 1 mg of danazol was applied. Relative amount of lipids content in our study is, therefore, much smaller than that in the clinical study. This could be a reason of lower estimation of danazol absorption in the D/P system especially under the fasted state. The permeated amount in the D/P system could be expressed as Z2 Pemeated amountð=2 hÞ ¼ Peff S Cdissolved dt 0
¼ Peff S AUCdissolved ð2Þ where Peff is an effective permeability, S is a surface area of Caco-2 monolayer and Cdissolved is a concentration of dissolved drug in the apical solution. From this equation, it is obvious that the permeated amount of drugs in the D/P system was affected not only by the solubility and permeability, but also by dissolution rate in the apical side. For instance, if the dissolution was much faster than the permeation, the permeated amount became the function of permeability and solubility, whereas if the dissolution was slower than the permeation, the permeated amount was profoundly affected by the dissolution rate, which corresponds to the dissolution rate-limited absorption. Since this is also true for in vivo, the D/P system enables the evaluation of oral drug absorption in all cases of permeability, solubility and dissolution rate limited. Consequently, this system could be a useful tool to screen the effect of food on the oral absorption of new drug candidates at the early stage of drug development in the pharmaceutical industry. Recently, the oral absorption of some drugs, such as alendronate sodium hydrate or nadolol,29–32 has been reported to be inhibited by food intake in humans. This negative effect of food on oral drug absorption is now under investigation in the D/P system and is a subject for future report.
CONCLUSIONS In this study, we have established the in vitro system to evaluate the effect of food on the oral absorption of poorly soluble drugs. In addition, this system was found to be able to detect the effect of formulations on drug absorption. In vitro D/P system may be a useful tool, not only to JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 95, NO. 9, SEPTEMBER 2006
predict the oral drug absorption under fasted and fed states, but also to consider the effectiveness of drug formulations to improve the oral absorption of poorly soluble drugs.
REFERENCES 1. Charman WN, Rogge MC, Boddy AW, Berger BM. 1993. Effect of food and a monoglyceride emulsion formulation on danazol bioavailability. J Clin Pharmacol 33:381–386. 2. Crounse RG. 1963. Effective use of griseofulvin. Arch Dermatol 87:176–178. 3. Galia E, Nicolaides E, Horter D, Lobenberg R, Reppas C, Dressman JB. 1998. Evaluation of various dissolution media for predicting in vivo performance of class I and II drugs. Pharm Res 15: 698–705. 4. Dressman JB, Amidon GL, Reppas C, Shah VP. 1998. Dissolution testing as a prognostic tool for oral drug absorption: Immediate release dosage forms. Pharm Res 15:11–22. 5. Dressman JB, Reppas C. 2000. In vitro-in vivo correlations for lipophilic, poorly water-soluble drugs. Eur J Pharm Sci 11:S73–S80. 6. Ingels F, Deferme S, Destexhe E, Oth M, van den Mooter G, Augustijns P. 2002. Simulated intestinal fluid as transport medium in the Caco-2 cell culture model. Int J Pharm 232:183–192. 7. Ingels FM, Augustijns PF. 2003. Biological, pharmaceutical, and analytical considerations with respect to the transport media used in the absorption screening system, Caco-2. J Pharm Sci 92: 1545–1558. 8. Ingels F, Beck B, Oth M, Augustijns P. 2004. Effect of simulated intestinal fluid on drug permeability estimation across Caco-2 monolayers. Int J Pharm 274:221–232. 9. Kataoka M, Masaoka Y, Yamazaki Y, Sakane T, Sezaki H, Yamashita S. 2003. In vitro system to evaluate oral absorption of poorly water-soluble drugs: Simultaneous analysis on dissolution and permeation of drugs. Pharm Res 20:1674–1680. 10. Ginski MJ, Taneja R, Polli JE. 1999. Prediction of dissolution-absorption relationships from a continuous dissolution/Caco-2 System. AAPS PharmSci 1: E3. 11. Ginski MJ, Polli JE. 1999. Prediction of dissolutionabsorption relationships from a dissolution/Caco-2 system. Int J Pharm 177:117–125. 12. 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. 13. Motz S, Schaefer UF, Eichinger T, Lehr CM. 2004. Development of a flow through system with DOI 10.1002/jps
ESTIMATION OF FOOD EFFECTS ON ORAL DRUG ABSORPTION
14.
15. 16.
17.
18.
19.
20.
21.
22.
23.
24.
combined dissolution and permeation devices. Globalization of Pharmaceutics Education Network. Motz S, Schaefer UF, Eichinger T, Lehr CM. 2004. Development of a flow through system with combined dissolution and permeation devices. Pharmaceutical Sciences World Congress. Welling PG. 1989. Effects of food on drug absorption. Pharmacol Ther 43:425–441. Olkkola KT, Brunetto AV, Mattila MJ. 1994 . Pharmacokinetics of oxicam nonsteroidal antiinflammatory agents. Clin Pharmacokinet 26: 107–120. Verbeeck RK, Richardson CJ, Blocka KL. 1986. Clinical pharmacokinetics of piroxicam. J Rheumatol 13:789–796. Bannwarth B, Lapicque F, Netter P, Monot C, Tamisier JN, Thomas P, Royer RJ. 1988. The effect of food on the systemic availability of ketoprofen. Eur J Clin Pharmacol 33:643–645. Melander A, Stenberg P, Liedholm H, Schersten B, Wahlin-Boll E. 1979. Food-induced reduction in bioavailability of atenolol. Eur J Clin Pharmacol 16:327–330. Yamaoka K, Tanigawara Y, Nakagawa T, Uno T. 1981. A pharmacokinetic analysis program (multi) for microcomputer. J Pharmacobiodyn 4:879–885. Peeters TL, Vantrappen G, Janssens J. 1980. Bile acid output and the interdigestive migrating motor complex in normal and in cholecystectomy patients. Gastroenterology 79:678–681. Tangerman A, van Schai A, van der Hoek EW. 1986. Analysis of conjugated and unconjugated bile acids in serum and jejunal fluid of normal subjects. Clin Chim Acta 159:123–132. Schulte-Lobbert S, Westerhoff K, Wilke A, Schubert-Zsilavecz M, Wurglics M. 2003. Development of a high-performance-liquid-chromatographic method for the determination of biapigenin in biorelevant media. J Pharm Biomed Anal 33: 53–60. Kostewicz ES, Wunderlich M, Brauns U. Becker R, Bock T, Dressman JB. 2004. Predicting the precipitation of poorly soluble weak bases upon entry
DOI 10.1002/jps
25.
26.
27.
28.
29.
30.
31.
32.
2061
in the small intestine. J Pharm Pharmacol 56: 43–51. Mariadason JM, Barkla DH, Gibson PR. 1997. Effect of short-chain fatty acids on paracellular permeability in Caco-2 intestinal epithelium model. Am J Physiol 272:G705–G712. Lange H, Eggers R, Bircher J. 1988. Increased systemic availability of albendazole when taken with a fatty meal. Eur J Clin Pharmacol 34:315– 317. Hauss DJ, Fogal SE, Ficorilli JV, Price CA, Roy T, Jayaraj AA, Keirns JJ. 1998. Lipid-based delivery systems for improving the bioavailability and lymphatic transport of a poorly water-soluble LTB4 inhibitor. J Pharm Sci 87:164–169. Joshi HN, Tejwani RW, Davidovich M, Sahasrabudhe VP, Jemal M, Bathala MS, Varia SA, Serajuddin AT. 2004. Bioavailability enhancement of a poorly water-soluble drug by solid dispersion in polyethylene glycol-polysorbate 80 mixture. Int J Pharm 269:251–258. Gertz BJ, Holland SD, Kline WF, Matuszewski BK, Freeman A, Quan H, Lasseter KC, Mucklow JC, Porras AG. 1995. Studies of the oral bioavailability of alendronate. Clin Pharmacol Ther 58:288– 298. Yamaguchi T, Ikeda C. Sekine Y. 1986. Intestinal absorption of a beta-adrenergic blocking agent nadolol. I. Comparison of absorption behavior of nadolol with those of other beta-blocking agents in rats. Chem Pharm Bull (Tokyo) 34:3362–3369. Yamaguchi T, Ikeda C, Sekine Y. 1986. Intestinal absorption of a beta-adrenergic blocking agents nadolol. II. Mechanism of the inhibitory effect on the intestinal absorption of nadolol by sodium cholate in rats. Chem Pharm Bull (Tokyo) 34: 3836–3843. Yamaguchi T, Oida T, Ikeda C, Sekine Y. 1986. Intestinal absorption of a beta-adrenergic blocking agent nadolol. III. Nuclear magnetic resonance spectroscopic study on nadolol-sodium cholate micellar complex and intestinal absorption of nadolol derivatives in rats. Chem Pharm Bull (Tokyo) 34:4259–4264.
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 95, NO. 9, SEPTEMBER 2006