# Fileset

[manuscript_JPS_interplay.docx](https://mdr.nims.go.jp/filesets/6ff1f0b4-fcf1-4da6-818b-345f7a162064/download)

## Creator

Michinori Oikawa, Satoru Matsuura, Takeyuki Okudaira, Ryo Ito, Kanako Arima, Masahiro Fushimi, Takamasa Oda, Kaoru Ohyama, [Kohsaku Kawakami](https://orcid.org/0000-0002-3466-9365)

## Rights

[Creative Commons BY-NC-ND Attribution-NonCommercial-NoDerivs 4.0 International](https://creativecommons.org/licenses/by-nc-nd/4.0/)

## Other metadata

[Bridging the gap between in vitro and in vivo solubility-permeability interplay](https://mdr.nims.go.jp/datasets/29848b30-a949-45e5-add3-0e7013f3ee37)

## Fulltext

Bridging the Gap between in vitro and in vivo Solubility-Permeability InterplayMichinori Oikawa1, Satoru Matsuura2, Takeyuki Okudaira3, Ryo Ito4, Kanako Arima1, Masahiro Fushimi1, Takamasa Oda2, Kaoru Ohyama5, Kohsaku Kawakami5,6**To whom correspondence should be addressedE-mail: kawakami.kohsaku@nims.go.jp, Tel. +81-29-860-44241Sawai Pharmaceutical Co., Ltd., 5-2-30, Miyahara, Yodogawa-ku, Osaka, 532-0003, Japan2Nippon Shinyaku Co., Ltd, 14, Nishinosho-Monguchi-cho, Kisshoin, Minami-ku, Kyoto, 601-8550, Japan3Taiho Pharmaceutical Co., Ltd, 224-2, Ebisuno, Hiraishi, Kawauchi-cho, Tokushima, 771-0194, Japan4Towa Pharmaceutical Co., Ltd., Kyoto Research Park KISTIC #202, 134, Chudoji Minami-machi, Shimogyo-ku, Kyoto, 600-8813, Japan 5Research Center for Macromolecules and Biomaterials, National Institute for Materials Science, 1-1 Namiki, Tsukuba, Ibaraki 305-0044, Japan6Graduate School of Pure and Applied Sciences, University of Tsukuba, 1-1-1 Tennodai, Tsukuba, Ibaraki 305-8577, JapanManuscript pages: 32, Figures: 9, Tables: 7Submitted to J. Pharm. Sci.AbstractUse of solubilization carriers for poorly soluble drugs may disturb transmembrane absorption by lowering the activity of drug molecules, which is known as the solubility-permeability interplay. However, although many in vitro studies have indicated the negative impacts of use of solubilization carriers for oral absorption, in vivo studies that showed the interplay effect are limited. This study provides systematic in vitro, in situ, and in vivo investigation of the interplay effect of cyclodextrin using dexamethasone as a model drug. The evaluation methods included permeation through polymeric, artificial lipid, cell, and intestinal closed-loop membranes. Then, the results were compared with oral administration studies in mice and dogs. Although the interplay effect was clearly observed in the in vitro studies, no obvious interplay was found in the in vivo studies, suggesting that the interplay effect is more prominent in the in vitro permeation studies. Absence of in vivo interplay was attributed to the dilution effect in the gastrointestinal tract, interaction of the drug with living components, and clearance of the drug after membrane permeation. Overall, this investigation clearly demonstrated the applicability and limitations of in vitro permeation studies for predicting the interplay effects of solubilizers after the oral administration.Keywords: solubility-permeability interplay; solubilization; poorly soluble drug; oral absorption; cyclodextrin1. IntroductionPoorly soluble drugs often exhibit low oral bioavailability, for which various solubilization technologies, including amorphous solid dispersions, nanocrystals, and self-emulsifying drug delivery systems, are available1-3. However, use of solubilization technologies can disturb apparent membrane permeability due to the solubility-permeability interplay4. Many in vitro studies have revealed that the solubilization of poorly soluble drugs using carriers, such as surfactant micelles, cyclodextrins, and cosolvents, does not contribute to the enhancement of transmembrane flux due to the decrease in the activity of the drug5-8. However, in vitro permeation studies are simplified systems, where careful interpretation is required. In living bodies, many factors, including dilution by intestinal fluid, interference by intestinal components, large absorption area, and clearance of the drug after the absorption, can hinder the interplay effect. Traditional in vitro evaluation systems were not developed with intention to use with solubilization excipients9; therefore, the effects of excipients on the permeation behaviors of poorly soluble drugs and if it really reflects the in vivo absorption require further investigation9.          This study focused on the interplay effect caused by cyclodextrins, a class of molecules showing solubilization effect for poorly soluble drugs. Despite their successful inclusion in many commercial products, they can inhibit the absorption process owing to the interplay effect, which has clearly been observed in in vitro permeation studies4,5. Addition of a competitive substrate for cyclodextrins enhances the peroral absorption of the guest drug10, suggesting the existence of interplay effect in vivo. Understanding the relationship between in vitro permeation study results and actual in vivo absorption is important for the effective use of solubilization agents, such as cyclodextrins.          Various methods are available for the in vitro evaluation of membrane permeation. Simple methods have an advantage in throughput; however, as they ignore many biological factors, careful consideration is required in interpreting the results. Commonly used simple cell-free evaluation methods are phospholipid or liposome-based assays9. The cellular systems used to evaluate membrane permeation of drugs include Caco-211,12 and MDCK cells13. Excised intestines are also options for evaluation14. Although many examples of the solubility–permeability interplay have been reported in the literature, reports on this interplay in oral administration studies10,15,16 has been limited. In this study, we examined the solubility-permeability interplay using dexamethasone (DMS) and -cyclodextrin (CD) in vitro, in situ, and in vivo studies. Membrane permeation of DMS was hindered by the addition of CD in the in vitro permeation studies, whereas this interplay was not observed in the oral administration studies in mice and dogs. The reasons for these contradictory observations are discussed in this paper to provide guidance on the interpretation of results of in vitro permeation studies with solubilization agents.2. Material and methods2.1 MaterialsDMS and -cyclodextrin (CD) were obtained from FUJI FILM Wako Pure Chemicals (Tokyo, Japan). Sodium taurocholate (STC) was supplied by Nacalai Tesque (Kyoto, Japan). Caco-2 cells were purchased from KAC (Kyoto, Japan). GIT lipids and acceptor sink buffer (ASB) were supplied by Pion (Billerica, MA, USA). All reagents used in this study were of reagent grade and used as supplied.2.2 Solubility Measurement          Solubility of DMS in buffered solutions was determined at 37 °C using 50 mM Hank's buffer (pH 7.4). CD and STC were added to the buffer for evaluating their solubilization capacities for DMS. Excess DMS was loaded into a test tube (n = 3), followed by the addition of 5 mL of the test medium. Subsequently, the tubes were rotated in a temperature-controlled oven at 37 °C for 1 day. The solutions were filtered through a nylon syringe filter with a pore size of 0.45 m. The glassware and syringe filters used for filtration were incubated in the same oven prior to use for avoiding unexpected precipitation during processing. The filtrates were diluted with 50% ethanol aqueous solution and subjected to HPLC measurement on the Prominence system (Shimadzu, Kyoto, Japan). A YMC-Pack Pro C18 (150 mm × 2.0 mmID, YMC, Kyoto, Japan) was used as the separation column, to which a mixture of acetonitrile and water at a ratio of 35/65 (v/v) was introduced as the mobile phase at a flow rate of 0.2 mL/min. Injection volume and detection wavelength were 2 L and 241 nm, respectively. Linearity was confirmed in the concentration range from 0.2 to 200 g/mL. DMS concentrations in the following in vitro studies were evaluated using similar HPLC procedures except the closed-loop and Caco-2 studies.2.3 Polymeric Membrane Permeation Study          The permeability of DMS through a polymeric membrane in the presence or absence of the solubilization agents was investigated on the dissolution/permeation (D/P) system17 using dialysis membrane with a cut-off molecular weight of 1,000 Da. Donor and acceptor chambers were filled with 8 and 5.5 mL of the test solutions prepared by 100 mM phosphate buffer, respectively, after attaching the dialysis membrane between the chambers. The donor solution contained 100 M DMS with or without additives. Effective surface area of the membrane for permeation was 1.77 cm2. Temperature of the chambers was maintained at 25 °C. The solutions were mixed using a magnetic stirrer at 200 rpm. 100 L of the acceptor solution was collected at 15, 30, 60, 90, and 120 min and subjected to HPLC analysis to determine the DMS concentration. Apparent membrane permeability (Papp) was determined following the general procedure11,12.          For observing the effect of the acceptor volume, a microtube with semipermeable membrane (Pur-A-Lyzer™ Midi Dialysis Kit, Sigma-Aldrich, St. Louis, MO, USA, cut-off molecular weight: 1000 Da) was used. 800 L of 100 M DMS dissolved in 50 mM phosphate buffer solution with or without 6 mM CD were sealed in the microtube, and the tube was suspended in 16 mL or 800 mL of 50 mM phosphate buffer solution at 25 °C. The outer medium was mixed using a magnetic stirrer at 750 rpm. 100 L of the outer medium was collected at 15, 30, 60, 90, and 120 min and subjected to HPLC analysis to determine the DMS concentration. Papp was determined following the general procedure.2.4 Artificial Lipid Membrane Permeation Study           Permeability of DMS through an artificial lipid membrane was investigated on the FluxTM apparatus18 (Pion, Billerica, MA, USA). The membrane was prepared by dropping 25 L of 20% lecithin/dodecane solution (GIT Lipid) on a PVDF filter support with a 0.45 m pore size. After attaching the membrane between the donor and acceptor chambers, 20 mL of the solutions were added to both sides. 100 M DMS dissolved in 50 mM phosphate buffer with or without additives and ASB were used as donor and acceptor solutions, respectively. Surface area of the membrane for permeation was 1.54 cm2. Temperature of the chambers was maintained at 37 °C by circulating hot water. UV signal was monitored for both sides using fiber optic UV-vis probes equipped with tips with a 20-mm path length.2.5 Caco-2 Study          Caco-2 cells were seeded on collagen Vitrigel membranes and 0.33 cm2 of growth area on a cell culture insert (ad-MED Vitrigel 2, Kanto Chemical Co., Inc., Tokyo, Japan) at a density of 5 × 104 cells/insert. The culture medium was replaced every 2-3 d. The cells were subjected to the study after 2 weeks. Permeability was assessed using Hank’s buffer containing 10 mM HEPES in a 24-well plate. The insert and well volumes were 200 and 500 L, respectively. 4.5% bovine serum albumin was added to the acceptor (basal) side. After applying 100 M of DMS with or without CD/STC, the solution in the acceptor side was collected at 30, 60, 90, and 120 min. Membrane integrity was confirmed by measuring the transepithelial electrical resistance (TEER) before and after the experiments. The initial value was above 1000 Ωcm2. The reported value will be alerted when the value after the experiment was below 300 Ωcm2. The DMS concentrations were measured using LC/MS/MS on API 4000 (AB Sciex, Framingham, MA, USA). For the calibration standards, 200 μL of internal standard (IS) solution using Limaprost-d3 were added to 20 μL of standard working solution. For the collected samples, 200 μL of IS solution were added to 20 μL of the collected samples after 100-folds and 10-folds of dilution with 50% methanol for the donor and the acceptor side, respectively. The samples were vortexed and centrifuged at 3000× g at 4°C for 3 min, followed by filtration and addition of 200 μL of water to obtain the solution for analysis. The analytical column (YMC-Triart PFP, 3 μm, 3.0 × 50 mm, YMC, Kyoto, Japan) was kept at 40°C. The mobile phase was a mixture of an equal volume of 0.1 % formic acid solution and acetonitrile at a flow rate of 0.4 mL/min. Injection volume was 2 μL. Multiple reaction monitoring mode (negative) with electrospray ionization was used to detect the product ions of m/z 361 (DMS) and 302 (IS), which were produced from their parent ions of m/z 391 and 382, respectively. The range of quantification under this analytical conduction was from 0.5 to 5000 ng/mL. The peak area of the chromatogram was calculated using Analyst (ver. 1.6.3, AB Sciex, Framingham, MA, USA).2.6 In situ Absorption Study Using Intestinal Closed-loop of Rats          An intestinal closed-loop study was performed using 8-week-old male SD (Crl:CD) rats. The study was approved by the ethical committee of Hamri Co. (No. 24-018). The rats were housed in a temperature-controlled room at 24±3 °C with a relative humidity of 50±20% and a 12 h light/dark cycle. The rats were fasted for 16 h before the operation under isoflurane anesthesia. A 10-cm segment below the pylorus was used to form a closed-loop. The intestinal loops of half of the animals were flushed using 10 mL of prewarmed saline at 37 °C. 1 mL of 100 M DMS solutions with or without CD were applied into the closed-loop. 200 μL of blood samples were collected using pre-heparinized syringes 0.25, 0.5, 1, 2, and 4 h after application of the test solutions. Plasma samples were harvested by centrifugation of the blood at 2,300 × g for 10 min and kept at −80 °C until analysis. DMS concentrations in blood samples were measured with LC/MS using TripleQuad TM 4500 (AB Sciex). For the calibration standards, 10 μL of IS solution of triamcinolone acetonide (TACA), 10 μL of standard working solution, and 30 μL of acetonitrile were added to 10 μL of rat blank plasma. For the collected samples, 5 μL of IS solution and 20 μL of acetonitrile were added to 5 μL of the collected samples. The samples were vortexed and centrifuged at 13,000 × g at 4°C for 5 min. 40 μL of water was added to 10 μL of supernatants, and the mixture was subjected to the analysis. The analytical column (Imtakt Unison UK-C18, 3 μm, 2.0 × 50 mm, Imtakt, Kyoto, Japan) was kept at 40°C. Mobile phase A was 0.1 % formic acid solution, and mobile phase B was acetonitrile containing 0.1% formic acid. The following gradient condition was applied: 0-1 min, B 30%; 1-1.1 min, B 30-70%; 1.1-5 min, B 70%; 5-5.1 min, B 70-30%; 5.1-9.1 min, B 30%. The flow rate of the mobile phase and injection volume were 0.3 mL/min and 10 μL, respectively. Multiple reaction monitoring mode (positive) with electrospray ionization was used to detect the product ions of m/z 373.1 (DMS) and 415 (IS), which were produced from their parent ions of m/z 393.2 and 435, respectively. The range of quantification under this analytical conduction was from 1 to 1000 ng/mL. The peak area of the chromatogram was calculated by Analyst (ver. 1.6.3, AB Sciex). Pharmacokinetic (PK) parameters were calculated using the trapezium method. P-values for log(Cmax) and log(AUC0-4 hr) were calculated using JMP ver. 17  (JMP Statistical Discovery LLC, Cary, NC, USA).2.7 Oral Administration Study Using Mice and Dogs          Oral administration studies were conducted using 5-week-old male ICR (Crl:CD1) mice and 65-M-old male beagle dogs in both fed and fasted states. Each study was approved by the ethical committee of Hamri Co. (No. 23-102) and Taiho Pharmaceutical Co. Ltd. (No. A22A009), respectively.          Mice were housed in a temperature-controlled room at 24±3 °C with a relative humidity of 50±30% under a 12 h light/dark cycle. Half of the mice were fasted for 16 h before and 4 h after administration with free access to water. 5 mL/kg of 100 M DMS solutions, with or without CD, were orally administered to mice (n = 3). 45 μL of blood samples were withdrawn through the jugular vein using pre-heparinized syringes 0.25, 0.5, 1, 2, and 4 h after administration. Plasma samples were harvested by centrifugation of the blood at 2,300 × g at  5°C for 10 min and were kept at −80°C until analysis.          Dogs were housed in a temperature-controlled room at 20–28 °C with a relative humidity of 30-70% under a 12 h light/dark cycle. In the fasted group, dogs were fasted for 20 h before to 24 h after administration and allowed free access to water, except for 30 min before to 2 h after administration. 50 mL of 100 M DMS solutions, with or without CD, were administered using a catheter (n = 6). Pentagastrin was administered intramuscularly at a dose of 10 g/kg 30 min before and 30/90 min after administration. In the fed group, dogs were fasted for 20 h before to 30 min before administration. 100 g of canine feed were administered in the form of dumplings containing 70 g of water. One mL blood samples were withdrawn from the forelimb using pre-heparinized syringes pre, 0.25, 0.5, 1, 2, 4, 8, and 24 h after administration and handled following the same procedure as that for the mouse study.          DMS concentrations in the mouse/dog blood samples were measured with LC/MS using TripleQuad TM 4500, and API 4000 (AB Sciex), respectively. For the mouse/dog calibration standards, 10 μL or 50 μL of IS solution using TACA, 10 μL or 50 μL of standard working solution, and 30 μL or 150 μL of acetonitrile were added to 10 μL or 50 μL of the mouse or dog blank plasma samples, respectively. For the mouse/dog collected samples, 5 μL or 50 μL of IS solution and 20 μL or 200 μL of acetonitrile were added to 5 μL or 50 μL of the collected samples, respectively. The samples were vortexed and centrifuged at 13,000 × g at 4°C for 5 min. 40 μL or 50 μL of water was added to 10 μL or 50 μL of the supernatant, respectively, and the mixture was used for the analysis. Further details are the same as those for the closed-loop study.3. Results3.1 Solubilization Capacities of CD and STC for DMS          Figure 1 shows solubilities of DMS in CD and STC Hank's buffer solutions. Solubility in the presence of CD increased linearly with increasing DMS concentration. The stability constant between DMS and CD was determined to be 2.81 mM-1 assuming 1:1 complex formation using slope of the plot. The critical micellar concentration (cmc) of STC was determined to be ca. 6 mM using surface tension and fluorescent measurements, which agreed with a reported value19. A linear increase in the solubility of DMS was observed above cmc, which indicating the entrapment of DMS in STC micelles. Almost the same solubility values were obtained when the Hank's buffer was replaced with the phosphate buffer at the same pH (data not shown).Figure 1. Solubility of DMS in Hank's buffer in the presence of (a) CD and (b) STC. Solubility of DMS in the presence of STC micelles (i.e., when the STC concentration was higher than the cmc), is presented by closed symbols. The experiments were triplicated to obtain the mean values. Standard deviations are presented by error bars, but they are smaller than the symbols in most cases.3.2 In vitro Permeation Study through a Polymeric Membrane          Figure 2 shows the effects of CD and STC on apparent permeability of DMS through the polymeric membrane. Addition of STC did not influence the permeability of DMS despite the decrease in free DMS concentration, whereas significant decrease in the permeability was observed when CD was added. The applied STC concentrations, 3 and 20 mM, were below and above cmc of STC, respectively. Observation for STC is consistent with the general idea that encapsulation of drugs into STC micelles does not inhibit their membrane permeation20,21. Decrease in apparent permeability of DMS in the presence of CD was recovered by the coexistence of STC, regardless of formation of STC micelles, most likely due to competition between DMS and STC for interaction with CD, as bile acids can be trapped by cyclodextrins22,23. It was obvious from solubility measurement as well, where solubility of DMS in the presence of 6 mM of CD was 2.28 mM, but it decreased to 1.47 or 0.561 mM, respectively, in the coexistence of 3 mM or 20 mM STC. In the presence of 20 mM STC, which is a concentration level in the gastrointestinal tract, apparent permeability of DMS in the presence of 3 or 6 mM of CD was almost the same with that for the CD-free solutions.Figure 2. Effects of CD and STC on permeability of DMS through the polymeric membrane investigated using the D/P system. Each measurement was triplicated to present the mean values with standard deviations as error bars. Additives used are indicated in the figure.          Figure 3 shows the effect of the donor/acceptor volume ratio on permeation of DMS. The ratio in D/P study described above was 0.69. Thus, it was increased to 20 and 1000 to confirm the effects of the acceptor capacity. As we used different systems for these evaluations, we simply focused on ratios of the apparent permeation rates in the presence and absence of 6 mM CD. The ratio was 6.7 % in the D/P study. It increased to 30% and 56%, respectively, by increasing the volume ratio. Driving force of the permeation is difference in activity of DMS in both sides of the membrane. Difference in the DMS activity is maintained to be large during the study, if the volume ratio is sufficiently large, which allows fast dissociation of DMS from its complex with CD to diminish the apparent permeability ratios in the presence and absence of CD. Thus, the interplay effect was clearly demonstrated to depend on the acceptor capacity.Figure 3. Effect of donor/acceptor volume ratio on permeation of DMS.3.3 In vitro Permeation Study through an Artificial Lipid Membrane (Flux)          Unlike polymeric membranes, artificial lipid membranes can uptake hydrophobic drug molecules. Therefore, partitioning of drug molecules between the aqueous phases and the membrane influences the permeation process. Figure 4 shows the effects of CD and STC on apparent permeability of DMS through the artificial lipid membrane. Similar to that in the polymeric membrane, addition of STC did not significantly influence apparent permeability, and decrease in permeability was observed in the presence of CD, which was recovered by coexistence of STC. Interplay effect of CD was also observed in this membrane, but it was weaker than that observed for the polymeric membrane.Figure 4. Effects of CD and STC on apparent permeability of DMS through the artificial lipid membrane investigated using the Flux apparatus. Accepter concentration as a function of time is presented. Each measurement was repeated twice or thrice to present the most representative data. Black: DMS only (100 mM), blue: DMS + 1 mM CD, orange: DMS + 6 mM CD, green: DMS + 10 mM STC, red: DMS + 6 mM CD + 10 mM STC.3.4 In vitro Permeation Study through a Caco-2 Membrane          Figure 5 shows the effects of CD and STC on apparent permeability of DMS through the Caco-2 monolayer membrane. Integrity of Caco-2 membrane was reported to be unaffected by the addition of 5 w/v% of CD or 20 mM of STC24. However, in our study, addition of CD or STC at concentrations over 6 mM and 8 mM, respectively, significantly decreased the TEER values, indicating their unignorable influence on membrane integrity. Papp of DMS in the presence of CD was ca. 33% of that of the pure DMS solution. STC did not appear to influence apparent permeability of DMS (Figure 5(a)), although uncertainty on membrane integrity remains for addition of 8 mM STC. Decrease in apparent permeability in the presence of CD was recovered by the coexistence of STC (Figure 5(b)). Permeability of DMS increased with increasing concentration of STC at a constant CD concentration of 6 mM, which could be understood by the competition effect. Thus, these observations were qualitatively the same as those for the polymeric and lipid membrane studies. However, the interplay effect was reduced by employing cell membrane.Figure 5. Caco-2 cell permeability of DMS in the presence of CD and STC. The experiments were repeated at least for three times to show mean values with standard deviations as error bars. Addition of CD or STC at concentrations over 6 mM and 8 mM, respectively, decreased the TEER values, which indicated loss of integrity of the membrane.3.5 In situ Closed-loop Study          Figure 6 shows the closed-loop study of DMS, where the influence of flashing of intestine and addition of CD were investigated. PK parameters are summarized in Table 1. Flashing of intestine improved the absorption of DMS, which was likely to be mainly originated from difference in initial absorption rates, suggesting that some biological components, including proteins, lipids, and bile acids, played inhibitory roles in the absorption of DMS. Addition of CD decreased the AUC to approximately 75%, regardless of the flashing conditions, possibly due to the solubility-permeability interplay effect; however, CD was less influential than expected from the in vitro permeation studies. Table 2 shows p-values for log(Cmax) and log(AUC0-4 hr) calculated using JMP, indicating that the addition of CD resulted in a statistically significant reduction of the exposure to DMS, with p-values below 0.05. As the interplay effect of CD was not influenced by flashing, the interactions between CD and living components was not likely to influence absorption.Figure 6. Plasma concentration profiles of DMS in the closed-loop study with (black) or without (red) flashing of the intestine. 1 mL of 100 M DMS was applied in the presence (closed) or absence (open) of 6 mM CD. Experiments were triplicated to show the mean values with standard deviations as error bars.Table 2. P-values for log(Cmax) and log(AUC0-4 hr) calculated using JMP.  log(Cmax) log(AUC0-4 hr) Sample 0.00347 0.0155 Flashing 0.344 0.0666 Sample × Flashing 0.971 0.7623.6 Oral Administration StudyFigure 7 shows the plasma concentration profiles of DMS after oral administration in mice. Table 3 summarized the PK parameters. Addition of CD did not influence absorption of DMS in the fasted state, whereas absorption was slightly suppressed in the fed state. These results were statistically analyzed using JMP (Table 4). Secretion of bile salts in the fed state did not influence the absorption behavior in the presence of CD, as it should result in better absorption in the fed state. Despite absence of the interplay effect in the fasted state, the reduction in free water in the fed state possibly stressed the interplay effect. In the dog study (Figure 8, Table 5), no interplay effects were observed, as proved by p-values for log(Cmax) and log(AUC0-24 hr) over 0.05 (Table 5). Neither CD nor bile salts impacted the absorption of DMS. It should be noted that CD is not likely to be decomposed in the gastrointestinal tract25 and not absorbed with the drug26. Thus, this observation clearly demonstrated that the solubility-permeability interplay observed in in vitro studies may not represent fate of the drug after oral administration in living bodies.Figure 7. Plasma DMS concentrations after oral administration of DMS solutions to mice. CD Conc: 0 mM (black), 1 mM (blue), and 6 mM (red). Experiments were triplicated to show the mean values with standard deviations as error bars.Table 4. P-values for log(Cmax) and log(AUC0-4 hr) calculated using JMP.  Mouse Dog  log(Cmax) log(AUC0-4 hr) log(Cmax) log(AUC0-24 hr) Sample 0.343 0.0540 0.593 0.646 Fed condition 0.00533 0.000170 0.0576 0.609 Sample × Fed condition 0.166 0.0322 0.442 0.522Figure 8. Plasma DMS concentrations after oral administration of DMS solutions to dogs. Experiments were triplicated to show the mean values with standard deviations as error bars.4. Discussion4.1 In vivo and in vitro Comparison of the Solubility-Permeability Interplay Effects          The solubility-permeability interplay has been recognized for more than 30 years10. However, the detailed discussion was limited at that time, as candidate compounds had much higher solubility compared to recent ones, which did not require solubilization technologies for administration. The potential negative impact of the interplay effect was recalled relatively recently5,6 after increased demand for solubilization technologies for poorly soluble drugs. Active research has been performed using in vitro apparatuses to observe the interplay effects caused by various solubilization carriers. Even bile salt micelles have been anticipated to disturb oral absorption8, although they are generally thought to aid the absorption of poorly soluble drugs27. On the other hand, only limited numbers of in vivo interplay effect10,15,16 have been reported, which indicates that the cases where the interplay gives serious negative effects should be limited. Thus, systematic study on the interplay effect to bridge the gap between in vitro and in vivo studies is required.Using the stability constant for DMS and CD, the free fraction of DMS in the presence of CD is calculated using the following equation:   (1)Here [D], [CD], and [D・CD] are the concentrations of DMS, CD, and their complex, respectively. Subscript t means the total concentration. Examples of free DMS concentrations calculated using equation (1) are shown in Figure 9. If 6 mM CD is added to 100 M DMS, the free fraction is only 5.7%. In the coexistence of STC, the complexation efficiency of DMS and CD is expected to decrease, because STC can also form complex with CD. When the STC concentration is lower than cmc, the following equation can be obtained by focusing on the CD concentration under the assumption that DMS and STC do not interfere upon interaction with CD.  (2)where Here T represents STC. K1 and K2 are the binding constants of DMS/CD and STC/CD, respectively. K1 was determined to be 2.81 mM-1 in our study and K2 is reported to be 2.3 mM-1 in literature23. Using these values, free DMS and STC concentrations in the coexistence of CD and STC can also be calculated (Figure 9). Owing to the competing effect, free DMS concentration increases in the presence of STC at the same CD concentration. For instance, in the presence of 3 mM CD, free DMS concentration is 10.9 M. It increases to 14.5 and 28.3 M in the coexistence of 1 and 3 mM STC, respectively. However, as presented later, experimental activity of DMS in the coexistence of CD and STC was likely to be higher than this calculation, suggesting complexity of interference.Figure 9. Calculated free DMS and STC concentrations as a function of CD concentration when STC concentration is below the cmc. Black and red lines represent the DMS and STC concentrations, respectively. Total STC concentrations are 0 mM (solid), 1 mM (dotted), and 3 mM (break).          Table 6 summarizes the relative permeation rates of DMS in the presence of CD and STC in each evaluation system except for oral administration. The relative permeation rates across the polymeric membrane agreed well with the model calculation for the free fraction in the DMS/CD binary system. Thus, investigation using a polymer membrane reflected the activity of DMS in the presence of additives.          When a polymeric membrane is used, the distribution of the drug into the membrane may be ignored. If the same simple buffer of the same volume is used for both the donor and the acceptor, the increase in drug concentration in the acceptor, Ca, as a function of time, t, can be described as follows: , (3)where k and Cd are the rate constant and the drug concentration in the donor, respectively. The activity coefficient is approximated to unity in a simple buffer. If a solubilization agent is added to both sides, equation (5) is rewritten as , (4)where Sa and Sd are the solubilities of the drug, and a and d are the activity coefficients in the acceptor and donor, respectively. The activity may be approximated by the free concentration; however, it is also influenced by interaction with the solubilization agents.           When a lipid membrane is used, partitioning of the drug in the membrane can no longer be ignored. The equation is divided into two parts as follows.From donor to membrane:   (5)From membrane to acceptor:  (6)Here m, Sm, and Cm are the activity coefficient, solubility, and drug concentration in the lipid membrane, respectively. kd and ka are the permeation coefficients into and out of the lipid membrane, respectively. mSm must be sufficiently higher than dSd for enabling partitioning of the drug into the membrane. However, an excessively large mSm inhibits the partitioning of the drug from the membrane to the acceptor. As the difference in the solubility in the donor and the acceptor is no longer the direct factor to influence membrane permeation, the solubility-permeability interplay is weakened in the presence of the lipid membrane. As expected, although the relative permeation rates in the presence of 6 mM CD were 5.7 and 6.7, respectively, relative to the CD-free solution for the model calculation and polymeric membrane, respectively, it was 19 with the lipid membrane.           Cell and intestinal membranes are living systems which have more structural and compositional complexity. However, an interplay effect for 6 mM CD was similarly observed for Caco-2. A dramatic reduction of the interplay effect was found in the loop study as proved by the increased ratio to 60.          DMS can be trapped to the bile salt micelle. However, the interaction is not strong to leave 83% activity, even in the presence of 20 mM STC, as proved by the polymeric membrane study. Nevertheless, a weak interplay was likely to exist, as the flashing of the intestinal loop increased the permeation of DMS, although STC was not the only living components to influence the permeation process. Possible positive impact of the bile salt micelle in living body includes interaction with mucus and local condensation28. Presumably these contributions are sufficient to cancel the interplay effect in the oral administration studies.          Coexistence of CD and STC dramatically weakened the interplay effect caused by CD in the in vitro studies. For instance, in the presence of 10 mM STC, which is a much lower concentration than that in the living bodies, the relative permeation rate reached 83%, which was much larger than 6.7% observed for the STC free solution in the polymeric membrane study. Similar observations were made for other in vitro studies. Based on this observation, positive food effect was expected for oral administration studies in the presence of CD. However, this effect was not observed in the dog study, whereas the opposite effect was observed in the mouse study, suggesting that the interplay by CD was not active in the gastrointestinal tract.Table 6. Comparison of the permeation rates in different in vitro evaluation systems (DMS solution as 100%) Additives Model calculation* Polymer membrane (D/P) Lipid membrane(Flux) Cell membrane (Caco-2) Intestinal membrane (Rat loop) None 100 100 100 100 100 1 mM CD 28 38 32 30 − 3 mM CD 11 10 23 18 − 6 mM CD 5.7 6.7 19 33*** 60** 1 mM STC 100 − 94 103 − 3 mM STC 100 106 94 107 − 8 mM STC 93 − − 117*** − 10 mM STC 91 − 87 132*** − 20 mM STC 83 105 79 − − 6 mM CD1 mM STC 6.6 22 − 40*** − 6 mM CD3 mM STC 9.8 54 23 47*** − 6 mM CD8 mM STC 31 75 − 71*** − 6 mM CD10 mM STC 41 83 44 81*** −* Percentages of free fractions. **Calculated from the plasma concentration at 15 min. *** Membrane integrity was lost, as indicated by the significant decrease of TEER.4.2 Origin of the Absence of the Interplay Effect in the Oral Administration Study          The difference in the permeation rate of DMS in the absence and presence of 6 mM CD was estimated to be 18-folds in the model calculation. It decreased to 15, 5.3, 3.0, and 1.7-folds for polymeric, artificial lipid, Caco-2, and close-loop membranes, respectively. The solubility-permeability effect by CD was weakened by approaching real living systems and completely disappeared in the oral administration studies, possibly due to following reasons. As the intestinal fluid is available in the small intestine, dilution may lead to the release of captured DMS from CD due to the relatively low stability constant. This effect should be more prominent in large animals. Also, the interplay effect may be weakened by the much larger absorption area in the gastrointestinal tract compared to that in the in vitro model systems, which can facilitate fast absorption. Moreover, living components may mask the interplay effect by the solubilizers. As the flashing in the loop study slightly enhanced permeation, the interplay might also be caused by living components.           Probably the most important factor missing in in vitro systems is clearance of drug after permeation. In living bodies, the permeated drug is immediately transferred into the systemic circulation. If Ca ≒ 0, the driving force for permeation in equation (6) is always infinity. In fact, difference in solubilization capacity of the acceptor in the in vitro evaluation system is known to influence the permeation behavior29. Thus, we performed a simple study where the volume of the acceptor phase was varied to confirm whether permeation was influenced by the acceptor volume (Figure 3). The result could explain the observation that large animals show less interplay due to the large distribution volume. This impact should be much larger in humans. These findings indicate that the interplay effect is more prominent in the in vitro permeation studies. The discussion above is summarized in Table 7. Although interplay can occur in vivo, the possibility is much lower than expected in in vitro studies because of these reasons. The interplay effect in vivo should be found only when the interaction between the drug and solubilization carrier is sufficiently strong. The stability constant between CD and DMS was 2.81 mM-1, which is moderately strong. That for cinnarizine and CD, which was a pair to exhibit the interplay effect, had the stability constant of 6.2 mM-1 30. When the formulation is administered in a solid state, cyclodextrins may help the dissolution process, which is an additional advantage for improving absorption of poorly soluble drugs31. Nevertheless, too large binding constant (> 17 mM-1) was presented to fail in enhancement of the oral absorption32. This information should work as a guidance for the formulation study utilizing solubilization carrier.Table 7 Possible origin of difference between in vitro and in vivo solubility-permeability interplayAlthough the interplay may theoretically be found for solubilization by bile salt micelles, it may rarely occurs considering the morphology of the bile salt micelles. Unlike micelles formed by synthetic surfactants, they do not have a spherical structure to separate the hydrophobic region completely from the outer atmosphere but have a disk structure33 to allow the entrapped drug molecules to access to the outside media. Moreover, structure of micelles is not static but constituent molecules, including solubilized drugs, are dynamically exchanged. Molecular dynamics studies on bile salt micelles have revealed that entrapped drug molecules are frequently exposed to the outer atmosphere34. As they are generally believed to aid absorption of poorly soluble drugs, it seems to be a natural conclusion that bile salt micelles generally should not disturb membrane permeation by the interplay effect.5. ConclusionsThis study provided systematic investigation on the interplay effect of CD using DMS as a model compound. Model calculations revealed that addition of 6 mM CD to 100 M DMS decreased the activity of DMS to 5.7%. Significant interplay effect was observed in the permeation study using polymeric membrane to find that permeation was only 6.7% relative to that in the absence of CD, which was assumed to simply reflect the activity of DMS. The interplay effect was a little suppressed by employing the artificial lipid membrane and sink buffer to the acceptor to find 19% of the permeation, as this system considers distribution of the drug into the membrane phase and clearance of the drug after permeation. In the Caco-2 and closed-loop studies, the relative permeabilities increased to 33% and 60%, respectively. In the oral administration studies in mice and dogs, the interplay effect was absent. Thus, interplay effect was prominent in the in vitro permeation studies, which was explained by the dilution effect in the gastrointestinal tract, interaction of the drug with living components, large absorption area in the gastrointestinal tract, and rapid clearance of the drug after the absorption. Bile salts did not decrease the activity of DMS, even when trapped. In summary, the interplay effect is likely to be limited to systems where significantly strong interactions exist between the drug molecule and carrier.FundingThis work was funded by Materials Open Platform for Pharmaceutical Science and each author’s affiliation.AcknowledgmentsThe authors would like to thank Physio Mckina Co., Ltd. for use of the mFlux apparatus. The authors are grateful for discussion and technical support by Shizuka Ono, Ph.D. (Ono Pharmaceutical Co., Ltd.), Atsushi Nagayasu, Shinji Kizuki, Junko Chikamoto (Taiho Pharmaceutical Co., Ltd.), Airi Ito, Takumi Sumi, Mitsuki Uehara, Toshinori Tanaka (Nippon Shinyaku Co., Ltd), Sayaka Koda, Mayako Hori, and Motoki Onishi (Towa Pharmaceutical Co., Ltd.). This study was conducted as part of Materials Open Platform for Pharmaceutical Science (Center of Excellence for Pharmaceutical Materials Science) led by National Institute for Materials Science.Data AvailabilityAll the data used for preparing this manuscript are available upon request to the corresponding author.Conflicts of interestThe authors declare no conflict of interest.References1. Brouwers, J., Brewster, M. E., Augustijns, P. Supersaturating Drug Delivery Systems: The Answer to Solubility-Limited Oral Bioavailability? J. Pharm. Sci. 2009;98:2549-2572.2. Kawakami, K. Modification of Physicochemical Characteristics of Active Pharmaceutical Ingredients and Application of Supersaturatable Dosage Forms for Improving Bioavailability of Poorly Absorbed Drugs. Adv. Drug Deliv. Rev. 2012;64:480-495.3. Newman, A., Knipp, G., Zografi, G. Assessing the Performance of Amorphous Solid Dispersions. J. Pharm. Sci. 2012;101:1355-1377.4. Porat, D., Dahan, A. Active Intestinal Drug Absorption and the Solubility-Permeability Interplay. Int. J. Pharm. 2018;537:84-93.5. Dahan, A., Miller, J.M., Hoffman, A., Amidon, G.E., Amidon, G.L. The Solubility-Permeability Interplay in Using Cyclodextrins as Pharmaceutical Solubilizers: Mechanistic Modeling and Application to Progesterone. J. Pharm. Sci. 2010;99:2739-27496. Miller, J.M., Beig, A., Krieg, B.J., Carr, R.A., Borchardt, T.B., Amidon, G.E., Amidon, G.L., Dahan, A. The Solubility-Permeability Interplay: Mechanistic Modeling and Predictive Application of the Impact of Micellar Solubilization on Intestinal Permeation. Mol. Pharmaceutics 2011;8:1848-1856.7. Raina, S.A., Zhang, G.G.Z., Alonzo, D.E., Wu, J., Zhu, D., Catron, N.D., Gao, Y., Taylor, L.S. Impact of Solubilizing Additives on Supersaturation and Membrane Transport of Drugs. Pharm. Res. 2015;32:3350–3364.8. Elkhabaz, A., Moseson, D. E., Brouwers, J., Augustijns, P., Taylor, L. S. Interplay of Supersaturation and Solubilization: Lack of Correlation between Concentration-Based Supersaturation Measurements and Membrane Transport Rates in Simulated and Aspirated Human Fluids. Mol. Pharmaceutics 2019;16:5042-5053.9. Buckley, S.T, Fischer, S.M., Fricker, G., Brandl, M. In vitro Models to Evaluate the Permeability of Poorly Soluble Drug Entities: Challenges and Perspectives. Eur. J. Pharm. Sci. 2012;45:235-250.10. Tokumura, T., Nanba, M., Tsushima, Y., Tatsuishi, K., Kayano, M., Machida, Y., Nagai, T. Enhancement of Bioavailability of Cinnarizine from Its b-Cyclodextrin Complex on Oral Administration with DL-Phenylalanine as a Competing Agent. J. Pharm. Sci. 1986;75:391-394.11. Yamashita, S., Furubayashi, T., Kataoka, M., Sakane, T., Sezaki, H., Tokuda, H. Optimized Conditions for Prediction of Intestinal Drug Permeability Using Caco-2 Cells. Eur. J. Pharm. Sci. 2000;10:195-204.12. Takahashi, Y., Kondo, H., Yasuda, T., Watanabe, T., Kobayashi, S., Yokohama, S. Common Solubilizers to Estimate the Caco-2 Transport of Poorly Water Soluble Drugs. Int. J. Pharm. 2002;246:85-94.13. Taub, M.E., Kristensen, L., Frokjaer, S. Optimized Conditions for MDCK Permeability and Turbidimetric Solubility Studies Using Compounds Representative of BCS Classes I-IV. Eur. J. Pharm. Sci. 2002;15:331-340.14. Hillgren, K.M., Kato, A., Borchardt, R.T. In Vitro Systems for Studying Intestinal Drug Absorption. Med. Res. Rev. 1995; 15:83-109.15. K. Kawakami, T. Yoshikawa, T. Hayashi, Y. Nishihara, K. Masuda. Microemulsion Formulation for Enhanced Absorption of Poorly Soluble Drugs II. In vivo Study. J. Control. Release 2002;81:75-82.16. Hens, B., Brouwers, J., Corsetti, M., Augustijns, P. Gastrointestinal Behavior of Nano- and Microsized Fenofibrate: In Vitro Evaluation in Man and In Vitro Simulation by Assessment of the Permeation Potential. Eur. J. Pharm. Sci. 2015;77:40-47.17. Kataoka, M., Masaoka, Y., Yamaaki, Y., Sakane, T., Sezaki, H., Yamashita, S. In Vitro System to Evaluate Oral Absorption of Poorly Water-Soluble Drugs: Simultaneous Analysis on Dissolution and Permeation of Drugs. Pharm. Res. 2003;20:1674-1680.18. Tsinman, K., Tsinman, O., Lingamaneni, R., Zhu, S., Riebesehl, B., Grandeury, A., Juhnke, M., van Eerdenbrugh. Ranking Itraconazole Formulations Based on the Flux through Artificial Lipophilic Membrane. Pharm. Res. 2018;35:161.19. Natalini, B., Sardella, R., Gioiello, A., Ianni, F., Di Michele, A., Marinozzi, M. Determination of Bile Salt Critical Micellar Concentration on the Road to Drug Discovery. J. Pharm. Biopharm. Anal. 2014;87:62-81.20. Sugano, K. Estimation of Effective Intestinal Membrane Permeability Considering Bille Micelle Solubilisation. Int. J. Pharm. 2009;368:116-122.21. Lu, J., Ormes, J.D., Lowinger, M., Wu, W., Mitra, A., Mann, A.K.P., Litster, J.D., Taylor, L.S. Impact of Endogenous Bile Salts on the Thermodynamics of Supersaturated Active Pharmaceutical Ingredient Solutions. Cryst. Growth Des. 2017;17:1264-1275.22. Miyajima, K., Yokoi, M., Komatsu, H., Nakagaki, M. Interaction of b-Cyclodextrin with Bile Salts in Aqueous Solutions. Chem. Pharm. Bull. 1986;34:1395-1398.23. Holm, R., Shi, W., Hartvig, R.A., Askjaer, S., Madsen, C., Westh, P. Thermodynamics and Structure of Inclusion Compounds of Tauro- and Glyco-Conjugated Bile Salts and b-Cyclodextrin. Phys. Chem. Chem. Phys. 2009;11:5070-5078.24. Ingels, F.M., Augustijns, P.F. Biological, Pharmaceutical, and Analytical Considerations with Respect to the Transport Media Used in the Absorption Screening System, Caco-2. J. Pharm. Sci. 2003;92:1545-1558.25. De Bie, A.T.H.J., Van Ommen, B., Bär, A. Disposition of [14C]g-Cyclodextrin in Germ-Free and Conventional Rats. Reg. Toxicol. Pharmacol. 1998;27:150-158.26. Gerlóczy, A., Fónagy, A., Keresztes, P., Perlaky, L., Szejtli, J. Absorption, Distribution, Excretion and Metabolism of Orally Administered 14C-b-Cyclodextrin in Rat. Arzneim.-Forsch./Drug Res. 1985;35:1042-1047.27. Holm, R., Mullertz, A., Mu, H. Bile Salts and Their Importance for Drug Absorption. Int. J. Pharm. 2013;453:44-55.28. Kakemi, K., Sezaki, H., Konishi, R., Kimura, T., Murakami, M., Effect of Bile Salts on the Gastrointestinal Absorption of Drugs. I. Chem. Pharm. Bull. 1970;18:275-280.29. Narula, A., Sabra, R., Li, N. Mechanisms and Extent of Enhanced Passive Permeation by Colloidal Drug Particles. Mol. Pharmaceutics 2022;19:3085-3099.30. Tokumura, T., Ueda, H., Tsushima, Y., Kasai, M., Kayano, M., Amada, I., Nagai, T. Inclusion Complexes of Cinnarizine with b-Cyclodextrin in Aqueous Solution and in the Solid State. Chem. Pharm. Bull. 1984;32:4179-4184.31. Aihara R., Messerschmid, R., Mizoguchi, M., Wada, K., Minami, K., Higashino, H., Takagi, T., Kataoka, M., Yamashita, S. In vitro-in vivo correlation in the effect of cyclodextrin on oral absorption of poorly soluble drugs. Int. J. Pharm. 2021;600:120494.32. Carrier, R.L., Miller, L.A., Ahmed, I. The Utility of Cyclodextrins for Enhancing Oral Bioavailability. J. Control. Release 2007;123:78-99.33. Carey, M.C. Bile Salt Structure and Phase Equilibria in Aqueous Bile Salt and Bile Salt-Lecithin Systems. Hepatology 1984;4:138S-150S.34. Parrow, A., Larsson, P., Augustijns, P., Bergström, C.A.S. Molecular Dynamics Simulations of Self-Assembling Colloids in Fed-State Human Intestinal Fluids and Their Solubilization of Lipophilic Drugs. Mol. Pharmaceutics 2023;20:451-460.25image2.emfimage3.emfimage4.emfimage5.emfimage6.emfimage7.emfimage8.emfimage9.emfimage10.emfimage11.emfimage12.emfimage13.emfimage1.emf00.10.20.30.40.50 5 10 15 20 25 30DMS Solubility (mM)STC conc (mM)(b)0123450 2 4 6 8 10 12DMS Solubility (mM)CD conc (mM)(a)K1:1 = 2.81 mM-1cmc