1521-009X/43/4/590–602$25.00 http://dx.doi.org/10.1124/dmd.114.060038DRUG METABOLISM AND DISPOSITION Drug Metab Dispos 43:590–602, April 2015Copyright ª 2015 by The American Society for Pharmacology and Experimental Therapeutics

A New Physiologically Based Pharmacokinetic Model for thePrediction of Gastrointestinal Drug Absorption:

Translocation Model s

Hirotaka Ando,1 Akihiro Hisaka,2 and Hiroshi Suzuki

Department of Pharmacy (H.A., H.S.) and Pharmacology and Pharmacokinetics, The University of Tokyo Hospital, Faculty ofMedicine, The University of Tokyo, Tokyo, Japan (A.H.)

Received July 15, 2014; accepted January 23, 2015

ABSTRACT

This study aimed to construct a new local pharmacokinetic modelof gastrointestinal absorption, the translocation model (TLM),using an anatomically relevant, minimally segmented structure toexplain linear and nonlinear intestinal absorption, metabolism,and transport. The TLM was based on the concept of a singleabsorption site that flexibly moves, expands, and shrinks alongwith the length of the gastrointestinal tract after the intake of anoral dose. The structure of the small intestine is continuous, andvarious time- and location-dependent issues are freely incorpo-rated in the analysis. Since the model has only one absorptionsite, understanding and modification of factors affecting absorp-tion are simple. The absorption site is composed of four com-partments: solid drug in the lumen, solution drug in the lumen,

concentration in the enterocytes, and concentration in the laminapropria. The lamina propria includes the blood capillaries. Bloodflow in the absorption site of the lamina propria appropriatelyaccounts for the absorption. In the TLM, the permeability of theapical membrane and that of the basolateral membrane are dis-tinct. By considering plicate, villi, and microvilli expansions of thesurface area, the apparent permeability measured in Caco-2 ex-periments was converted to the effective permeability in vivo. Theintestinal availability, bioavailability, and dose product of intestinalavailability and absorption rate relationship of the model drugswere well explained using the TLM. The TLM would be a usefultool for the consideration of local pharmacokinetics in the gastro-intestinal tract in various situations.

Introduction

In drug discovery, the prediction and understanding of the ab-sorption profile of oral drug candidates are required because they arekey information to determine the exposure of the drug to therapeutictargets in the body. Accordingly, the development of a logical phys-iologically based mathematical model that can predict the absorptionprocess of drugs in the gastrointestinal (GI) tract is important. Sev-eral models for the prediction of local pharmacokinetics in the GItract have been reported so far, including the advanced compartmentalabsorption and transit (ACAT) model (Yu and Amidon, 1999; Agoramet al., 2001), segregated-flow model (Cong et al., 2000), QGut model(Yang et al., 2007), advanced dissolution, absorption, and metabolism

model (Jamei et al., 2009), and GI transit time model (Bergstrandet al., 2009, 2012; Hénin et al., 2012). Among these, the QGut modelis the simplest one in model description that predicts intestinalavailability (FG) by using Q, which is a virtual hybrid parameterof the membrane permeability clearance and blood flow rate in theintestinal villi (Gertz et al., 2010). Although the fundamental prin-cipal of the QGut model is similar to that of the segregated-flowmodel (Pang and Chow, 2012), the QGut model is not capable ofanalyzing nonlinear pharmacokinetics because the model does notconsider drug concentration in the enterocytes, where nonlinear pro-cesses, such as metabolism and transport, actually occur (Hisaka et al.,2010).The ACAT and advanced dissolution, absorption, and metabolism

models assume a compartmentalized structure of the intestine, andfirst-order kinetics is applied to the movement of drugs within it.These models can consider location-dependent longitudinal changes inthe physiologic conditions. It has been reported that the ACAT modelsuccessfully predicts the oral pharmacokinetics of various drugs(Heikkinen et al., 2012). However, even by incorporating the ACATmodel into a sophisticated whole body physiologically based pharma-cokinetic model, the oral pharmacokinetics of only 23% of drugs areaccurately predicted (Poulin et al., 2011).

This research was supported in part by a Grant-in-Aid for Scientific Researchon Innovative Areas HD Physiology Project from the Japanese Ministry ofEducation, Culture, Sports, Science, and Technology [Grant 22136015].

1Current affiliation: Discovery Research Laboratories, Kyorin PharmaceuticalCo., Ltd., Tochigi, Japan.

2Current affiliation: Geriatric Pharmacology and Therapeutics, Graduate Schoolof Pharmaceutical Sciences, Chiba University, Chiba, Japan.

dx.doi.org/10.1124/dmd.114.060038.s This article has supplemental material available at dmd.aspetjournals.org.

ABBREVIATIONS: A, protein amount; ACAT, advanced compartmental absorption and transit; AFE, average fold error; CLint,H, hepatic intrinsicclearance; CYP3A, cytochrome P450 3A; F, bioavailability; f, free fraction; FA, fraction absorbed; FAFG, product of FA and FG; FG, intestinalavailability; FH, hepatic availability; FaSSIF, fasted state simulated intestinal fluid; GI, gastrointestinal; Hvilli, height of villi; HLw, half-life fordisappearance of water; Ht, hematocrit; kGE, gastric emptying rate constant; Km, Michaelis constant; L, length; ME, microvilli expansion; P,permeability; Papp, apparent permeability; PE, plicate expansion; P-gp, P-glycoprotein; Q, blood flow rate; RMSE, root mean square prediction error;Sol, solubility; Tent, thickness of enterocyte; TLM, translocation model; V, volume; Vmax, maximum rate; VE, villi expansion; z, location.

590

http://dmd.aspetjournals.org/content/suppl/2015/01/23/dmd.114.060038.DC1Supplemental material to this article can be found at:

at ASPE

T Journals on M

ay 21, 2020dm

d.aspetjournals.orgD

ownloaded from

Furthermore, it should be noted that diagonal (moves from mucosato submucosa) changes are taken into consideration less than lon-gitudinal (moves along with the absorption site) changes in thesemodels. The concentration of drugs in the enterocytes is sometimesvague since the necessary distinction of permeability between theapical and basolateral membranes is lacking or only implicitly defined.

Furthermore, heterogeneous flow that is caused by the differences invilli abundance in different regions should be adequately considered(Kalampokis et al., 1999).In most of the past physiologic absorption models, the blood flow

rate to the whole small intestine is used for the analysis of absorption.In addition, these models often assume that the basolateral transfer is

Scheme 1. Differential equations of translocation model

CLmet, metabolic clearance in enterocytes; CLout, output clearance from enterocytes; FLlumen, flux for disappearance from lumen; kexcr, rate constant for excretion from lumen;kGE,solid, gastric emptying rate of solid form drug (Jamei et al., 2009); Labs, length of absorption site; Mtime,1, first-order moments for time of drug as solid and solutionin lumen; t, mean residence time in lumen; Xsolid,stomach, Xstomach, Xsolid,lumen, Xlumen, Xent, and Xpropria, amount of drug in stomach (as solid), stomach (as solution), lumen(as solid), lumen (as solution), enterocytes, and lamina propria, respectively; zabs, location of absorption site.

Prediction of Gastrointestinal Absorption by Translocation Model 591

at ASPE

T Journals on M

ay 21, 2020dm

d.aspetjournals.orgD

ownloaded from

mostly unidirectional; thus, the effect of the blood stream on theabsorption rate is ignored. The consequence of changes in the bloodflow on the absorption rate becomes important when the basolateraltransfer of drugs from the enterocytes to the blood stream is assumedto be reversible. The importance of blood flow is discussed moreappropriately in several articles (Winne, 1978; Schulz and Winne,1987; Chen and Pang, 1997; Pang and Chow, 2012). A physiologicmodel was reported by Cong et al. (2001), in which branching of theblood flow to the mucosa and submucosa was considered. Pang andChow mentioned the importance of intestinal blood flow branchingand proposed the adequate mucosal to intestinal blood flow ratio (fQ)(Pang and Chow, 2012). When the effect of the blood stream on theabsorption of drugs is considered based on the physiologic structureof the intestine, branching in both the diagonal and longitudinal di-rections should be considered. It is well known that blood flow to theduodenum and jejunum is greater than to the ileum, which is in accordwith the physiologic function of the intestine. To date, few absorptionmodels appropriately take into consideration both the diagonal andlongitudinal blood flow branching, except the segregated-flow model,which can factor for the longitudinal blood flow in animals by di-viding the lumen and enterocytes into three compartments (Tam et al.,2003).In this study, a new model, the translocation model (TLM), is

developed to solve these issues and describe drug absorption in theGI tract in a more physiologic manner (Figs. 1 and 2). Althoughcomplicated factors needed to be considered in the new model, wehoped to keep the whole structure as simple as possible. This isbecause the model should be comprehensive and completely open tousers. Considering all of these issues, we aimed to construct a newmodel that could accurately predict various absorption events withonly one absorption site.

Materials and Methods

Theory of Location and Distribution Functions. In TLM, the location andlength of the absorption site in the small intestine were calculated using l(t), the

location function, and s2(t), the variance function, at time t. Those functionswere obtained by using a fitting analysis of the observed location and itsvariance for a nonabsorbable index drug in the intestinal tract. We assume thata drug is being dissolved completely here. The dissolution process will beconsidered later. When a total drug concentration (density per length) ata distance z from an exit of the stomach is expressed as Clumen,total(z), nth ordermoment of distribution of a drug in the intestinal tract is defined by eq. 1, whereLsi is the length of the small intestine from the duodenum to the ileum

Mlocation;n ¼Z Lsi

0zn × Clumen;totalðzÞdz: ð1Þ

The mean and variance of location of the index drug in the intestine aregiven by l and s2, respectively

l ¼ Mlocation;1=Mlocation;0 ð2Þ

s2 ¼ Mlocation;2=Mlocation;0 2l2: ð3Þ

Fig. 1. The concept of the translocation model. (A) The anatomicstructure of the intestinal villus reported by Lin et al. (1999). (B)The anatomic structure of the gastrointestinal tract included in thetranslocation model. (C) State of the compartments that werecreated for the translocation model (translocation compartments)immediately after a solid drug reached the duodenum. (D) State ofthe translocation compartments after a dissolved drug reached thecolon.

Fig. 2. The structure of the translocation model. The translocation model includesfour movable compartments; solid and solution in lumen, enterocyte, and laminapropria containing the capillaries.

592 Ando et al.

at ASPE

T Journals on M

ay 21, 2020dm

d.aspetjournals.orgD

ownloaded from

Equations 2 and 3 correspond to the mean residence time and variance ofresidence time in the moment analysis of blood drug concentration timeprofiles. In the TLM, it was assumed that Clumen(z) at the observation time isconstant within the absorption site, and it is otherwise zero for simplification.Based on this hypothesis, the length of the absorption site Labs is given by eq. 4,which was derived from eqs. 2 and 3. In the following, subscript abs denotesthe absorption site:

Labs ¼ 2ffiffiffi3

ps: ð4Þ

In theory, the equations for l and s can be arbitrarily selected as capable ofexplaining the observed movements of a drug in the GI tract. In this study, eqs.5 and 6 were obtained using a fitting analysis of the reported data (Kimura andHigaki, 2002), which showed the time course of the percentage dose of radio-activity in the jejunum, ileum, and colon after administration of nonabsorbable99mTc–diethylenetriamine pentaacetic acid in humans (Fig. 3A)

lbolusðtÞ ¼ m1   t

m2 þ tð5Þ

Fig. 3. Time-dependent parameters defined for the translocation model. (A) Time course of 99mTc–diethylenetriamine pentaacetic acid amount in each site of thegastrointestinal tract reported by Kimura and Higaki (2002). Open circle, closed triangle, open square, and closed circle represent the drug in the stomach, jejunum, ileum,and colon, respectively. Time course of (B) location (solid line), with proximal and distal ends of the absorption site (dotted line), (C) drug in the lumen, and (D) meanresidence time are simulated using a nonabsorbable condition after a single input.

Fig. 4. Location-dependent parameters defined for the translocation model. Location-dependent expression of (A) radius of the intestinal lumen, (B) plasma flow rate,(C) CYP3A amount, (D) P-glycoprotein level, (E) apical surface area, (F) volume of the lumen, (G) volume of the enterocytes, and (H) volume of the lamina propria.

Prediction of Gastrointestinal Absorption by Translocation Model 593

at ASPE

T Journals on M

ay 21, 2020dm

d.aspetjournals.orgD

ownloaded from

sbolusðtÞ ¼ s1�e2 s2 t 2 e2 s3 t

�þ s4: ð6Þ

The subscript bolus means that the functions correspond to bolus inputs ofa drug into the system, and t is the observation time. As a result of the fittinganalysis, m1, m2, s1, s2, s3, and s4 were obtained as 415 (1.35%), 1.50 (4.71%),92.4 (4.68%), 0.605 (4.35%), 45.9 (36.8%), and 3.72 (5.03%), respectively(values in the parentheses are CV values calculated from the fitting analysis).Equations 5 and 6 were used to describe the time-dependent change of the drugabsorption site and are shown in Fig. 3.

When a drug enters the GI tract with an input function I(t) but not as a bolus,such as continuous or intermittent administration, the location function shouldbe given with a convolution of I(t) and eq. 5

lðtÞ ¼ �lbolus*I�ðtÞ ¼Z

lbolusðyÞIðt2 yÞdy: ð7Þ

To simplify the convolution analysis, another moment analysis is applied todrugs in the intestine to calculate the mean residence time t

Mtime;n ¼Z t

0sn × Xlumen;totalðsÞds ð8Þ

t ¼ Mtime;1=Mtime;0 ð9Þ

where Xlumen,total(t) is the total amount of drug in the lumen at time t. Cal-culation of differentials of Mtime,n is described in the Supplemental Material andapplied for calculation of Mtime,1 in Scheme 1. When the velocity of a drug in theabsorption site is constant, even in the absence or presence of continuous druginput/output, a location of the average of the drug distribution moves in accordancewith t. Therefore, the convolution of eq. 7 can be approximated as follows:

lðtÞ ¼ �lbolus*I�ðtÞ @  lbolusðtÞ: ð10Þ

Equation 10 means that if we calculate t from the time courses of a druginput/output appropriately, the drug movement is described apparently withlbolus(t). Equation 10 generates errors when velocity of the drug movement isvariable in the absorption site. Theoretically, it is possible to cover these errorsby calculating higher moments. However, the simplest model was adopted inthis study because it showed sufficient performance (Fig. 10). Similarly, s wasalso approximated using t for simplification

dðtÞ @  dbolusðtÞ: ð11Þ

In the TLM, parameters, such as permeability and clearance, are determineddepending on the location of the absorption site (Table 1 and Fig. 4). Strictlyspeaking, these parameters are possibly variable even within the absorption site.However, for simplification, it was assumed that representative values at theaverage location could be applied monotonously for the absorption site in theTLM. In the present model, the stomach and large intestine were modeledseparately as well stirred compartments. It was assumed that a drug in thestomach traverses to the entrance of the duodenum with a first-order rateconstant of kGE. The value for kGE for a solution was assumed as 5.46 hours21

from the fitting analysis using eqs. 5 and 6. Input into the large intestine isdetermined by movements of a drug in the small intestine. In the present study,it was assumed that absorption does not occur from the stomach and largeintestine, considering the characteristics of the drugs used for the analysis.

Radius and Adjusted Location of the Intestinal Tracts. The function forthe radius (r) of the lumina of the small intestine, consisting of the duodenum,jejunum, and ileum, was described by a linear function in accordance witha report by Willmann et al. (Fig. 4A; Willmann et al., 2003, 2004).

r ¼ rini 2 rgrad ×zcðcmÞ ð12Þ

where zc (cm) was the center of the absorption site. It needs to be noted that theintestine should be adequately filled with contents to agree to its true radius tothe value calculated by eq. 12. In the process of absorption, this is not alwaysrealized. For this reason, the internal volume of the intestine will be calculatedbased on a different assumption later. The values of rini and rgrad (1.56 and0.00265, respectively, with r =21.00, and Akaike’s information criterion =248.7)were obtained from a regression analysis to the radii that were used in GastroPlus(Table 2).

Surface Area and Volume at the Absorption Site. The surface areas of theabsorption sites of the basolateral and apical membranes of the small intestine(Sabs,bas and Sabs,api, respectively) were calculated using eqs. 13 and 14

Sabs;bas ¼ pðrin þ routÞffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiðrin 2 routÞ2þL2abs

q� PE� VEabs

�cm2

� ð13Þ

Sabs;api ¼ Sabs;bas �ME�cm2

� ð14Þ

where rin and rout are the radii at the proximal and distal end of theabsorption site; PE is plicate expansion; VEabs is villi expansion; and ME ismicrovilli expansion (Fig. 5) at the absorption site. Equation 13 was derivedfrom the formula of surface area of the frustum. VEabs was calculated usingeq. 15 because of a gradual decrease in the size of the villi, depending on thelocation in the small intestine

VEabs ¼ VEini 2 ðVEini 2 1Þ zcLsi

ð15Þ

where VEini is villi expansion at the entrance of the duodenum. We assumedthat PE and ME are constant in the small intestine. For PE, VEini, and ME, 3,10, and 20 were used for the small intestine, respectively (DeSesso and Jacobson,2001).

The surface area of the whole length of the small intestine (Ssi) was calculatedusing eq. 16 as a solution of integration of the cross-sectional area of the intestinallumen from 0 to Lsi

Ssi ¼ 2p × PE × ME

"�1þ VEini

�rini × Lsi

22

13þ Vini

6

!rgrad × L2si

#�cm2

�:

ð16Þ

Ssi will be used for calculation of the blood flow (eq. 20). The volume of theabsorption site for the intestinal lumen, Vlumen,abs, was not calculated from thesizes of the absorption site described by eqs. 4 and 12 because these equationsassume that the lumen is adequately filled with contents. Instead, it wascalculated using eq. 17 as the total of the water volume taken with the drug andphysiologic intestinal water volume in the absorption site

Vlumen;abs ¼ LabsLall

Vcont þ Vwater;inie2 lnð2Þt=HLw�ml

� ð17Þ

Fig. 5. Definitions of apical and basolateral permeability in the translocationmodel. In this model, different surface areas were defined for the apical andbasolateral membrane. P1, P2, P3, and P4 represent the passive permeability fromthe lumen to enterocyte, enterocyte to lumen, enterocyte to lamina propria, andlamina propria to enterocyte, respectively. PP-gp represents the active permeabilityby P-gp (eq. 29).

594 Ando et al.

at ASPE

T Journals on M

ay 21, 2020dm

d.aspetjournals.orgD

ownloaded from

where Vwater,ini is the volume of water that is taken with an orally administereddrug (200 ml), and Vcont represents the volume of contents in the intestine undernormal fasted conditions (105 ml) (Schiller et al., 2005). HLw is the half-life forthe disappearance of water taken with the drug. In this study, we assumed thatHLw is 1 hour. Several fold changes of HLw did not affect the absorptionprofile of drugs in preliminary simulations (data not shown).

The volume of the enterocytes at the absorption site, Vent,abs, was calculatedas the product of the thickness of the enterocytes (Tent = 0.0018 cm) (Sugano,2009) and Sabs,bas

Vent;abs ¼ Sabs;bas ×Tent�ml� ð18Þ

where volume of the lamina propria, Vpropria,abs, at the absorption site was cal-culated based on the assumption of almost complete coverage of the intestinallumen by villi as the product of the villi height (Hvilli = 0.07 cm) (Sugano, 2009)and the surface area of the intestinal lumen subtracted by Vent, abs

Vpropria;abs ¼ Sabs;bas ×Hvilli

VEabs2Vent;abs

�ml� ð19Þ

Plasma Flow at the Absorption Site. The plasma flow rate to enterocytes atthe absorption site, Qplasma,abs, was calculated using eq. 20 by multiplying theblood flow rate to all enterocytes, Qblood,enterocyte, and the ratio of the surfacearea of the absorption site (Sabs,api) to the whole small intestine (Ssi) and thenconverting to plasma flow considering hematocrit (Ht) (0.45)

Qplasma;abs ¼ ð12HtÞ Qblood;enterocyte  Sabs;apiSsi

ðml=hÞ ð20Þ

Qblood, enterocyte was calculated to be 18,000 ml/h by assuming that the bloodflow into the superior mesenteric artery was 37,200 ml/h, which accounts for10% of the cardiac output. Eighty percent of the blood flow of the superiormesenteric artery flows into the mucosa, and then 60% of the mucosal bloodflows into the epithelium cells of the villi (Jamei et al., 2009).

Dissolution of a Solid Formulation. Solubility in fasted state simulatedintestinal fluid (FaSSIF) (SolFaSSIF) was calculated using eq. 21, as reported byPoulin et al. (2011)

SolFaSSIF ¼ SolWater

þ SolWaterMWWater

MWDrug

�100:75logPþ2:27

�×MWDrug × CBSðmg=mlÞ

ð21Þ

where SolWater, MWWater, MWDrug, log P, and CBS represent water solubility,molecular weight of water and the drug, n-octanol:buffer partition coefficient ofthe drug, and the bile salt concentration of sodium taurocholate (4 nM),respectively. The dissolution constants for the small intestine were defined usingeqs. 22 and 23

kdis ¼ 3gSol2C

rrj   Tðh2 1Þ ð22Þ

g ¼ 9:9� 102 5 × MW2 0:453�cm2=s

� ð23Þ

where Sol, g, r, rj, T, and Ci represent solubility, diffusion coefficient, density(1.2 mg/ml), particle radius (0.0025 cm), diffusion layer thickness (0.003 cm), andthe concentration in the stomach or lumen, respectively (Agoram et al., 2001).

Expression Level of Metabolizing Enzymes and Transporters. It has beenreported that cytochrome P450 (CYP) 3A is most abundantly expressed in theduodenum and less expressed in the distal end of the ileum (Paine et al., 1997).Therefore, a linear gradient was assumed for the expression of CYP3A, de-creasing from the proximal end of the duodenum to the distal end of the ileum.The expression of CYP3A at the absorption site, ACYP3A,abs, was calculated asa product of Labs and the density of CYP3A expression (eq. 24; Table 1) for thesmall intestine (z , Lsi; Lsi = 306 cm)

ACYP3A;abs ¼ 2ACYP3A;total

�12

zcLsi

�LabsLsi

�pmol

� ð24Þ

where ACYP3A,total is the total amount of CYP3A expressed in the smallintestine (70,500 pmol). The factor 2 in eq. 24 indicated that the density is doublewhen zc is zero. Conversely, ACYP3A,abs decreases to zero as the location moves tothe large intestine since no expression of CYP3A was reported in the colon(Bruyere et al., 2010). Metabolic clearance of the substrates of CYP3A wascalculated from ACYP3A,abs and Michaelis constants for each substrate (Scheme 1).

It has been reported that the expression level of P-glycoprotein (P-gp)increases relatively from the proximal to distal end of small intestine (Stephenset al., 2001; Englund et al., 2006; Bruyere et al., 2010); however, the absoluteexpression level is rarely reported. Therefore, we assumed that the relativeexpression level of P-gp in interest (Arel,Pgp,abs) increases linearly in the lower partof the intestine. For the lower site below, Arel,Pgp,abs was calculated using eq. 25

Arel;Pgp;abs ¼Ainit;Pgp þ Agrad;Pgp

zcLsi

LabsLsi

�pmol

� ð25Þ

where Ainit,Pgp and Agrad,Pgp represent the amount of P-gp in the entrance ofduodenum, and the slope for the increase in the P-gp expression amount, re-spectively. In this study, 0 and 1 were used for Ainit,Pgp and Agrad,Pgp, respectively.To calculate the actual expression level of P-gp, a scaling factor was estimated fromin vivo observations, as will be described later.

Membrane Permeability. The components of the permeability constants ofthe enterocytes, i.e., passive intake of apical, passive efflux of apical, activeefflux of apical by P-gp, passive output from basal, and passive reverse intakefrom basal, were designated as P1, P2, PP-gp, P3, and P4, respectively (Fig. 5).Based on the structural symmetry and similarity of membranes, values for thepassive permeability (P1–P4) were assumed to be the same as the passivepermeability of a single membrane, which is denoted as Ps (when the pH of theenvironments is the same for both sides of the membrane). Ps for the Caco-2system was calculated from the apparent permeability (Papp) determined inCaco-2 assays (Gertz et al., 2010) using eq. 26

Ps;Caco2 2 ¼ 1þMECaco-2MECaco-2

Papp;Caco-2 ð26Þ

where MECaco-2 is the apical/basolateral relative surface area ratio for Caco-2cells. Equation 26 was derived from the equation reported by Tachibana et al.(2010). For the equation, PS1 = P1,Caco-2S·MECaco-2, PS2 = P2,Caco-2S·MECaco-2,PS3 = P3,Caco-2S, and PS4 = P4,Caco-2S were assumed. The terms Km and Vmax

were removed because we used the equation to calculate passive diffusion.In the present study, an MECaco-2 value of 4 was adopted as the average of the

values determined for propranolol and naproxen as nonsubstrates of P-gp (Ohuraet al., 2011). The values of Papp,Caco-2 for the substrates of P-gp (quinidine andverapamil) were calculated as the average values of Papp,Caco-2 at the three highestconcentrations (approximately 10–100 mM) because the activity of P-gp wassaturated at the three highest concentrations in the cited experiments (Tachibanaet al., 2010).

Ps for the in vivo system was determined by eq. 27

Ps;in vivo ¼ P2 ¼ P3 ¼ P4 ¼ Ps;Caco-2 � psfpassive ð27Þ

where psfpassive is the scaling factor, which is obtained from approximation ofthe simulated and reported FA for nine drugs (acyclovir, atenolol, cimetidine,ciprofloxacin, enalaprilat, gabapentin, methotrexate, ranitidine, and sulpiride).These drugs were selected based on the following criteria: 1) FA , 0.9; 2)Caco-2 permeability is available in the literature (Thomas et al., 2008); 3) FH. 0.7;and 4) CLH/CLtot , 0.3. First, each value of psf for the individual drug wasestimated by fitting the analysis to agree to the reported FA and predicted FA

as possible using TLM (Supplemental Table 1). Then, psfpassive was de-termined as the average of the psf for the nine drugs. Calculation of P1 will bedescribed later.

For the drugs that are substrates of P-gp, the active permeability by P-gp invitro (eq. 28) and in vivo (eq. 29) were described as follows:

PPgp;Caco-2MECaco-2S ¼ Vmax;Pgp;Caco-2Km;Pgp þ Ccell

ð28Þ

PPgp;absSabs;api ¼ psfPgp ×Vmax;Pgp ×Arel;Pgp

Km;Pgp þ CEnterocyte;uð29Þ

Prediction of Gastrointestinal Absorption by Translocation Model 595

at ASPE

T Journals on M

ay 21, 2020dm

d.aspetjournals.orgD

ownloaded from

where psfPgp, Ccell, and CEnterocyte,u represent the scaling factor for P-gptransport and the concentrations in the Caco-2 cell and enterocyte,respectively. Michaelis constants for the transport by P-gp (Km,Pgp andVmax,Pgp,Caco-2) were calculated using a fitting analysis with data fromCaco-2 cell assays (Tachibana et al., 2010), as previously described, witha slight modification using eq. 30

where Ca represents the drug concentration in the apical chamber. The Km,Pgp

values were 0.113 and 0.283 mg/ml for quinidine and verapamil, re-spectively. For Ps,Caco-2 values, it was confirmed that the values obtained byfitting the analysis with eq. 30 were practically the same as those obtainedfrom eq. 26.

For apical uptake permeability, P1 was calculated using eq. 31, consideringdifferences of pH between the Caco-2 assay and intestinal lumen. Caco-2assays are generally carried out at a pH of 7.4, but it has been reported that thesurface of the intestinal lumen is somewhat more acidic (Table 1; Bolger et al.,2009).

P1 ¼ Ps;in vivo � 1þ 10pHCaco‐2 2 pKaacid þ 10pKabase 2 pHCaco‐2

1þ 10pHlumen 2 pKaacid þ 10pKabase 2 pHlumenð31Þ

where pHCaco-2, pKaacid, and pKabase represent the pH for the Caco-2 assay (i.e.,7.4), acidic pKa, and basic pKa, respectively. A value for pHlumen represents thepH of the intestinal lumen of the small intestine that is described by thefollowing equation:

pHlumen ¼ pHini þ pHgrad × LsizcLsi

: ð32Þ

The values for pHini and pHgrad were 5.85 and 0.00455, respectively, withr = 0.979 and Akaike’s information criterion =213.7. The values for a product

of psfPgp and Vmax,Pgp in eq. 26 were obtained simultaneously by a fitting analysisusing the TLM. The values were estimated to explain the dose response of FAFGappropriately for verapamil and quinidine, which is a substrate of P-gp but a weak

TABLE 1

Properties of intestine in the translocation model

Category Parameters Equations

Physical pHlumen pHlumen ¼ 5:85þ 1:39zcLsi

r r ¼ 2 0:00265 × zc þ 1:56

Sabs,api Sabs;api ¼ pðrin þ routÞffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiðrin 2 routÞ2 þ L2abs

q� PE� VEabs �ME

Vlumen Vlumen;abs ¼ LabsLall

Vcont þ Vwater;ini × e2 lnð2Þt=HLw

Vent Vent;abs ¼ Sabs;bas � Tent

Vpropria Vpropria;abs ¼ Sabs;bas � Hvilli

VEabs2Vent;abs

QQplasma;abs ¼ ð12HtÞQblood;enterocyte

SabsSsi

Protein expression ACYP3AACYP3A;abs ¼ 2ACYP3A;total

�12

zcLsi

�LabsLsi

APgp Arel;Pgp;abs ¼Ainit;Pgp þ Agrad;Pgp

zcLsi

LabsLsi

Permeability P1P1 ¼ Ps;in vivo � 1þ 10pHCaco‐2 2 pKaacid þ 10pKabase 2 pHCaco‐2

1þ 10pHlumen 2 pKaacid þ 10pKabase 2 pHlumen

P2–P4 P2 ¼ P3 ¼ P4 ¼ Ps;in vivo

Papp;  Caco‐2 ¼ 1

2�1þMECaco‐2

�S × Ca

Ps;  Caco‐2S

hMECaco‐2

�Ca 2Km;Pgp

�2Km;Pgp

i2Vmax;Pgp;Caco‐2

þffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi�Ps;Caco‐2S

hMECaco‐2

�Ca 2Km;Pgp

�2Km;Pgp

i2Vmax;Pgp;Caco‐2

�2þ4MECaco‐2

�1þMECaco‐2

��Ps;Caco‐2S

�2Km;PgpCa

r ð30Þ

TABLE 2

Size of the gastrointestinal tract used for the translocation model

Radiusa Lengtha Volumea Cumulative Length Note

cm cm ml cm

Stomach 9.67 28.3b 47 — Stomach endDuodenum 1.53 14.1 42 14.1 Duodenum endJejunum1 1.45 58.4 154 72.5Jejunum2 1.29 58.4 122 130.9 Jejunum endIleum1 1.13 58.4 94 189.3Ileum2 0.980 58.4 71 247.7Ileum3 0.820 58.4 49 306.1 Ileum end

aUsed in GastroPlus (Bolger et al., 2009; Heikkinen et al., 2012).bLength of stomach was not incorporated into the translocation model.

596 Ando et al.

at ASPE

T Journals on M

ay 21, 2020dm

d.aspetjournals.orgD

ownloaded from

substrate of CYP3A (Supplemental Fig. 2). Based on the above theory,fundamental simultaneous ordinary differential equations for the TLM wereconstructed, as shown in Scheme 1.

Data Collection for Drug-Related Parameters. The drug-related param-eters for 20 drugs (alfentanil, alprazolam, buspirone, cisapride, cyclosporin,felodipine, lovastatin, midazolam, nifedipine, nisoldipine, repaglinide, rifabutin,saquinavir, sildenafil, simvastatin, trazodone, triazolam, zolpidem, quinidine,and verapamil) used in this model are summarized in Table 3. Km values forCYP3A were taken from the literature (Lavrijsen et al., 1988; Rotzinger et al.,1998; Ekins et al., 1999; Kajosaari et al., 2005; Ku et al., 2008; Polasek et al.,

2010; Gertz et al., 2011; Heikkinen et al., 2012;). When no informationwas available for the Km values of CYP3A, the Km values of CYP3A4were used. Metabolic intrinsic clearance, membrane permeability, and theblood-free fraction were taken from the literature (Gertz et al., 2011). Theobserved FG was estimated using the drug-drug interaction method re-ported by Hisaka et al. (2014) and taken from other sources (Varma et al.,2010). The other observed or reported values required for analysis in thisstudy are shown in Tables 3 and 4. Unbounded fractions in the intestinallumen (flumen), enterocyte (fent), and lamina propria (fpropria) were assumedto be 1.

TABLE 3

Drug-specific parameters used for the prediction of absorption

Drugs

Dose Km,u Vmax Papp,Caco-2

fB MWmmol

mM mmol/h per pmolcm/h

CYP3A P-gp CYP3A P-gp

Alfentanil 7220 22.8 NA 1.30 0 0.105 0.137 417Alprazolam 1290 265 NA 0.238 0 0.0918 0.341 309Buspirone 4150 7.98 NA 1.04 0 0.0914 0.062 386Cisapride 16,100 3.20 NA 0.526 0 0.108 0.0200 466Cyclosporin 474,000 1.40 0.0567 0.0333 1.38�106 0.0286 0.0140 1203Felodipine 26,000 5.31 NA 7.45 0 0.0151 0.068 384Lovastatin 49,400 7.80 NA 22.8 0 0.0522 0.03 405Midazolam 9200 3.31 NA 1.35 0 0.117 0.056 326Nifedipine 28,900 11.0 NA 1.45 0 0.0846 0.066 346Nisoldipine 12,900 2.10 NA 9.66 0 0.0720 0.003 388Repaglinide 552 15.6 NA 0.442 0 0.0868 0.025 453Rifabutin 177,000 10.9 NA 0.338 0 0.0342 0.483 847Saquinavir 1,490,000 0.300 0.863 1.09 2.47�106 0.0497 0.038 671Sildenafil 52,600 15.0 NA 1.01 0 0.0922 0.063 475Simvastatin 95,500 3.39 NA 14.2 0 0.0245 0.025 419Trazodone 202,000 312 NA 4.41 0 0.0871 0.07 372Triazolam 729 85.1 NA 0.261 0 0.101 0.161 343Zolpidem 16,300 140 NA 0.185 0 0.115 0.105 307Quinidine 615,000 4.00 0.348 0.00766 5.11�106 0.0491 0.299 325Verapamil 87,900 49.0 0.622 6.92 3.65�106 0.0516 0.104 455

NA, not applicable.

TABLE 4

Reported and prediction results of FG and observed and prediction results of F

Drug

FG CLint,HLMa FH,in vitro

b F

ml/min per pmolPredictedc Reportedd Predictede Observedf

Alfentanil 0.72 0.60a 0.890 0.613 0.440 0.43Alprazolam 0.99 0.99 NA NC NA NABuspirone 0.51 0.22 1.84 0.629 0.321 0.039Cisapride 0.56 0.55a 2.13 0.819 NA NACyclosporin 0.99 0.35 0.474 0.967 0.959 0.28Felodipine 0.02 0.53 15.5 0.155 0.003 0.16Lovastatin 0.04 0.09 35.4 0.154 0.006 0.05Midazolam 0.36 0.48 3.75 0.476 0.169 0.41Nifedipine 0.60 0.74 2.00 0.594 0.358 0.50Nisoldipine 0.03 0.11a 53.1 0.547 0.017 0.04Repaglinide 0.84 0.89a 0.737 0.913 0.771 0.60Rifabutin 0.85 0.21a 0.514 0.436 0.371 0.20Saquinavir 0.92 0.18a 51.3 0.090 0.083 0.07Sildenafil 0.77 0.54a 1.07 0.742 0.574 0.40Simvastatin 0.02 0.19 51.7 0.131 0.003 0.05Trazodone 0.90 0.83a 0.370 0.881 0.796 0.79Triazolam 0.98 0.45 NA NC NA NAZolpidem 0.99 0.81 0.0850 0.956 0.949 0.72

NA, not applicable.NC, not calculated.aGertz et al., 2011.bCalculated using eq. 35.cCalculated using eq. 34.dHisaka et al., 2014.eCalculated as the product of FAFG (eq. 33) and FH,in vitro.fAppendix of Parker et al., 2005.

Prediction of Gastrointestinal Absorption by Translocation Model 597

at ASPE

T Journals on M

ay 21, 2020dm

d.aspetjournals.orgD

ownloaded from

Calculation of Pharmacokinetics Parameters (FG and Bioavailability[F]). FAFG values were calculated by TLM as a ratio of the cumulative drugamount that flew into the portal vein to the administered dose (eq. 33).FG values were calculated by TLM as a ratio of FAFG to FA. The latter wascalculated by subtracting the ratio of the cumulative drug amount remaining inthe colon and rectum to the administered dose from 1

FAFG;predicted ¼ Aportal vein;cumulative

Doseð33Þ

FG;predicted ¼ FAFG;predicted

12Arectum;cumulative

Dose

: ð34Þ

Fig. 6. Effect of ME on lumen concentration (A), enterocyte concentration (B), lamina propria concentration (C), and FAFG (D) after 40 mg of oral verapamil. Dotted,dashed, and solid lines represent the simulation curve under 1, 5, or 20 of the ME condition.

Fig. 7. The prediction results of (A) the dose versus FAFG relationship of midazolam and (B) plasma time course of midazolam after oral administration of a therapeutic doseand microdose. (A) The dashed line represents the predicted dose FAFG, and closed circles represent the reported FAFG. (B) The closed and open circles represent theobserved plasma concentration after oral administration of the therapeutic dose and microdose, respectively. The solid line and dotted line represent the predicted plasmaconcentration after the therapeutic dose and microdose.

598 Ando et al.

at ASPE

T Journals on M

ay 21, 2020dm

d.aspetjournals.orgD

ownloaded from

By using in vitro hepatic intrinsic clearance (CLint,HLM) estimated using humanliver microsomes (Gertz et al., 2011), the blood-free fraction (fB), hepatic bloodflow rate (QH), and hepatic availability were estimated with in vitro informationusing the following equation (FH,in vitro):

FH;  in vitro ¼ QH

QH þ fBCLint; H;  in vitro: ð35Þ

By multiplying the predicted FAFG with FH,in vitro, F was predicted. The observedF values were collected from two reports (Varma et al., 2010; Appendix of Parkeret al., 2005).

Sensitivity Analysis. A sensitivity analysis of TLM was performed to estimatethe effect of the surface area of the apical membrane of enterocytes by usingverapamil as a model drug. Absorptions of 40 mg of verapamil was simulated byTLM, with ME values of 1, 5, and 20. Other parameters for TLMwere not changedin this sensitivity analysis. Then, changes in the intestinal lumen concentration,enterocyte concentration, lamina propria concentration in the absorption site, andFAFG were investigated by comparing the time course of these concentrations.

Evaluation of Precision and Accuracy of the Prediction. Precision andaccuracy of the prediction for FG and F were evaluated by using the within2-fold error, within 60.3 error, average fold error (AFE) (eq. 36), and rootmean square prediction error (RMSE) (eq. 37)

AFE ¼ 10+logðobserved=predictedÞ=N ð36Þ

RMSE ¼

ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi+h�predicted2 observed

�2iN

vuut ð37Þ

where N represents the number of drugs that were used for the calculation.

Prediction of Nonlinear Pharmacokinetics. A prediction of nonlinearpharmacokinetics was performed by simulating the FAFG-dose relationship formidazolam. Average FAFG values in the range of 0.01–10,000 mg of theadministered dose were plotted followed by confirmation of the similarity of thesimulated and observed values, which were calculated by using reported clinicaldata based on the assumption that the major site to induce the nonlinear systemicexposure after oral administration is the intestine (Bornemann et al., 1985; Misakaet al., 2010). For the calculation, body weight, RB, and hepatic blood flow ratewere assumed to be 70 kg, 1, and 25.5 ml/min per kg, respectively. Totalclearance and the fraction excreted in urine were cited from the following source(Appendix of Parker et al., 2005).

Systems and Application. Calculation of the ordinary differential equationsconstructed in this study was performed using Napp (http://square.umin.ac.jp/todaiyak/download.htm) with the Runge-Kutta–Fehlberg method (Hisaka andSugiyama, 1998).

Results

Construction of the TLM. The TLM was constructed based on theconcept that a drug is transferred from the lumen to the bloodstreamacross the enterocytes and lamina propria at an absorption site, which isrelocated, expanded, and decreased in size along the length of theGI tract in a time-dependent manner (Fig. 1). The locatable absorptionsite consisted of four compartments: solid formulation in the lumen,dissolved drug in the intestinal lumen, concentration in the enterocytes,and concentration in the lamina propria, including the capillaries (Fig. 2).The absorption site continuously traversed from the duodenum to theileum. In accordance with the physiologic structure, the stomach andcolon were defined as a separate compartment.Definition of Location and Distribution Functions. The location

and distribution functions that determine the movement of a drug ab-sorption site in the TLM were obtained using a fitting analysis of theresults of the gamma scintigraphy reported by Kimura and Higaki (2002).The movement of a nonabsorbable index drug in the gastrointestinal tractwas analyzed by using the location and distribution functions, and thenagreement was confirmed between the fitted and reported values (Fig.3A). By applying a simplified convolution with the mean residence time,these functions were extended to cover any drug input situation in theintestine, such as reduction of the gastric emptying rate, multiple inputs,and so forth (Fig. 10).Definition of Location-Dependent Parameters.All location-dependent

parameters defined for the absorption site of the TLM were describedby z, the location in the intestine. The equations that defined thelocation-dependent parameters are represented in Table 1. Changes ofthe parameters defined by z are shown in Fig. 4. The calculated totalsurface area and the total volume of the lumen, enterocyte, and laminapropria were 751,000 cm2 and 311, 67.6, and 378 ml, respectively.Approximation of psf and Vmax,Pgp. To determine the correspon-

dence of passive permeability between in vitro and in vivo systems,a fitting analysis was performed for the FA values of nine drugs(Supplemental Table 1). The value of psfpassive was determined as 2.23from the analysis. This result suggests that the passive membrane per-meability of enterocytes in vivo is apparently 2.23-fold higher than thatin vitro. By using this psfpassive, FA was estimated within 60.3 of theobserved value for all compounds (Supplemental Fig. 1).There were few reports on the in vivo Vmax value for P-gp. Therefore,

the product of psfPgp and Vmax,Pgp was obtained by a simultaneousfitting analysis to the human FAFG values for quinidine (0.1, 1, 10, and100 mg) and verapamil (0.1, 3, 16, and 80 mg) (Supplemental Fig. 2).The value was estimated as 1.66 g/h (CV = 11.1%), assuming that Vmax

is the same for quinidine and verapamil (Table 3).Sensitivity Analysis for ME. In the TLM, the model description of

the surface area of the enterocytes is different for the apical and basolateral

TABLE 5

Prediction results of FG and F and statistics for the prediction results

Parameters FG F

Within 2-fold (%) 50 53Within 60.3 (%) 72 93AFE 1.10 1.48RMSE 0.347 0.230

AFE, average fold error; RMSE, root mean square prediction error.

Fig. 8. Prediction result of bioavailability for 15 drugs by using TLM. Closedcircles represent the predicted F. The black dotted lines represent within 60.3 errorlines. The solid line represents the unity.

Prediction of Gastrointestinal Absorption by Translocation Model 599

at ASPE

T Journals on M

ay 21, 2020dm

d.aspetjournals.orgD

ownloaded from

membranes. The ratio of the surface area of the apical to basolateralmembrane was defined as ME, and ME was assumed to be 20 in thisstudy. This means the surface area of the apical membrane is 20-foldlarger than that of the basolateral membrane. The effect of ME on drugexposure in the lumen, enterocytes, lamina propria, and FAFG were con-firmed by sensitivity analysis using 40 mg of verapamil (Fig. 6). IncreasingME reduced the concentration in the intestinal lumen, whereas FAFG

was increased by an increase in ME.Simulation of Nonlinear Pharmacokinetics and Plasma Time

Profile of Midazolam. FAFG-dose relationships were simulated usingTLM for midazolam (Fig. 7A). The predicted dose-dependent increase ofFAFG agreed well with the observed FAFG. The plasma time profile waspredicted for oral midazolam after administration with a therapeutic doseor microdose by using pharmacokinetic parameters after intravenousadministration that were obtained from the same experiment with oraladministration (Fig. 7B). As a result, although plasma concentration afterthe microdose was slightly underestimated, overall, the time course ofdrug concentration after the therapeutic dose and microdose was wellpredicted.Prediction of Intestinal Availability and Bioavailability. To

validate the further utility of the TLM after confirming the validity of themodel’s construction, FG was calculated for 18 drugs and F wascalculated for 15 drugs so that the information about their affinity forCYP3A and P-gp are available (Table 3). The reported FG and F valuesand the predicted FG and F values are shown in Table 4 and Fig. 8,

respectively. The number of drugs within 2-fold errors for FG and F were50% and 53%, respectively, of drugs analyzed in this study. The ratiowithin 60.3 errors for FG and F were 72% and 93%, respectively. TheAFE and RMSE for FG prediction were 1.10 and 1.48, while the AFEand RMSE for F prediction were 0.347 and 0.230, respectively (Table 5).Simulation of Dose-Dependent Absorption. The time course of the

drug amount in the stomach, intestinal lumen, enterocyte, and laminapropria, cumulative drug amount in the portal vein and feces, andcumulative amount of the metabolized drug were simulated with inputconditions at 40 (clinically effective dose) and 0.1 mg (microdose) oforal verapamil using the TLM (Fig. 9). At 40 mg, most of the un-changed drug was rapidly absorbed and then 32.7% of the dosetraversed into the portal vein, while 57.9% was metabolized and 9.4%reached the colon. On the other hand, the extent of the drug absorbedinto the portal vein was lower at 0.1 mg. Only 16.6% of the dosetraversed into the portal vein, while 32.2% was metabolized and 51.2%reached the colon.Application to Multiple Administration. The time course of the

drug location (Fig. 10A) and drug concentration in the intestinallumen, enterocyte, and lamina propria (Fig. 10B) were simulated whenthe location and distribution functions were applied to multiple inputsof a low permeability drug. The simulation revealed that the drugabsorption site could move both back and forth as needed. A tendencyfor the drug concentration to decrease from the intestinal lumen to thelamina propria was observed appropriately.

Fig. 9. Simulation results of the amount of drug in the stomach (black solid line), lumen (black dashed spaced line), enterocytes (gray solid line), and lamina propria (blackdashed line), and cumulative amounts in the portal vein (gray dashed line), metabolite (gray dotted line), and rectum (black dotted line) after oral (A), (B) 40 mg, and (C and D)0.1 mg of verapamil. Normal plots are shown in (A and C), and logarithm scale plots are shown in (B and D), respectively.

600 Ando et al.

at ASPE

T Journals on M

ay 21, 2020dm

d.aspetjournals.orgD

ownloaded from

Discussion

The TLM was developed to construct a physiologically precise andrelatively simple absorption model, and its structure and parameters areeasily modifiable if needed. The model structure was simplified byintroducing a single absorption site, which is moved in accord with thelocation and variance functions. Previous models represent location-dependent changes of physiologic parameters with the fragmented in-testinal compartments (Yu and Amidon, 1999; Agoram et al., 2001;Jamei et al., 2009). Conversely, the TLM enabled them by describingeach parameter with a location-dependent function under various input/administration conditions.The location and distribution functions set in this study reproduced the

delay in gastric emptying under the fed state and the correspondingchanges in the distribution of a nonabsorbable index drug reported byKimura and Higaki (2002). Moreover, the profile of distribution inFig. 3B was similar to the results reported by Willmann et al. (2003).From these results, it was indicated that the location and variance func-tions used in this study could represent drug movement in the intestinallumen.Most of the reported models use effective permeability (ieff) as the

parameter for the passive permeability of the enterocytes (Yang et al.,2007; Badhan et al., 2009; Gertz et al., 2011). Peff is the parameter that iscalculated for human jejunum permeability and can be estimated by Pappin Caco-2 cells by using the result of the regression analysis betweenPapp and Peff (Sun et al., 2002). In the TLM, in vivo passive permeabilitywas integrated using the in vitro Papp of Caco-2 cells and multiplying theconstants of surface area expansion (PE for plicate, VE for villi, and MEfor microvilli). This direct integration of the passive permeability com-pared with the empirical approach of the extrapolation from Papp to Peffenables the user to take the intestinal physiologic factor and effect of thechanges in the intestinal pH into consideration. To confirm the utility ofdirect integration from single membrane permeability, Peff values fromthe TLM were simulated by modifying the TLM that mimicked the

reported experimental conditions of Loc-I-Gut (Petri et al., 2003). Thelocation and size of the absorption site (10 cm) was placed in thejejunum by TLM in exactly the same way. Finally, the reported andsimulated Peff values were compared (Supplemental Fig. 3). Our re-sults indicate that the passive permeability could be calculated adequatelyfrom the single membrane permeability.To confirm the effect of ME on drug absorption, a sensitivity analysis

was performed for verapamil after the oral administration of a ther-apeutic dose (40 mg). Results showed that the increase of ME reducedverapamil exposure to the lumen and increased the maximum con-centration in the enterocytes in the early time point, which might beimportant for first-pass metabolism due to the higher expression ofCYP3A in the proximal part of the small intestine. FAFG also increased,depending on the extent of the ME increase (Fig. 6). These resultsindicate that discrimination of the apical and basolateral membranes byusing an adequate ME would be advantageous to avoid potential un-derestimation of the systemic exposure.The obtained parameters were used for the prediction of the dose-

FAFG relationship of midazolam. As shown in Fig. 7A, the predictedand observed FAFG for midazolam were in close agreement, suggestingthe appropriateness of parameters obtained from the fitting analyses ofFA values for nine drugs. In addition to the nonlinearity in FAFG, theplasma time profile of midazolam was also predicted after oral admin-istration of the therapeutic dose and microdose (Fig. 7B). These resultssuggest that the time course of plasma concentration as well as non-linear absorption kinetics by TLM could be well predicted.In this research, the FG for 18 drugs and F for 15 drugs were sim-

ulated, as they are substrates of CYP3A and P-gp (Fig. 8). As a result, theFG and F of these drugs were reasonably predicted using the TLM fromin vitro parameters. As shown in Table 5, AFE and RMSE in all pre-dictions were acceptable, by which the potential of TLM for theprediction of clinical pharmacokinetics after oral administration ofvarious drug candidates from in vitro data from Caco-2 cell assays andmicrosomal metabolic kinetics studies during the drug discovery stagewas demonstrated.By using the TLM, the simulation of the local pharmacokinetics of

the intestine was performed at the clinical effective dose (40 mg) andmicrodose (0.1 mg) of verapamil. Lower availability was predictedwith a microdose than with 40 mg, which corresponds to the clinicalobservation (Maeda et al., 2011). The simulated maximum concen-tration in the enterocytes was approximately 6.7 mM, which wassmaller than the Km value of CYP3A4 (49.0 mM) and higher than theKm value for P-gp (0.622 mM). This would indicate a larger con-tribution of P-gp to the nonlinear drug absorption after 40 mg of oralverapamil.The recent commercially available applications for the prediction

of drug absorption are generally expensive and not easy to modifybecause of the complex and veiled structure. The TLM was developedby using Napp, a program for the analysis of pharmacokinetics, whichis available for free. This program would assist researchers to performdetailed analyses and predictions of drug absorption. Moreover, anadvantage of the TLM is that the model structure is completelytransparent to users.In conclusion, we successfully developed the TLM to describe the

physiology of drug absorption as precisely as possible, i.e., the move-ment of a drug in the intestine, permeation through the apical andbasolateral membranes, and the contributing blood flow. The structureof TLM is relatively simple compared with the previously availablemodels. It would be useful for the prediction of drug absorption duringnew drug development in the future. Moreover, various events duringthe absorption process would be analyzed more accurately with TLMby adjusting its structure and parameters, if necessary.

Fig. 10. Time course of (A) average location (solid line) and proximal and distal endsof the absorption site (dotted lines), and (B) concentration in the lumen (dotted line),enterocytes (dashed line), and lamina propria (solid line). Simulations were performedwith multiple input conditions for an absorbable drug with low permeability (apparentpermeability [Papp] = 0.0007 cm/h).

Prediction of Gastrointestinal Absorption by Translocation Model 601

at ASPE

T Journals on M

ay 21, 2020dm

d.aspetjournals.orgD

ownloaded from

Authorship ContributionsParticipated in research design: Hisaka, Ando, SuzukiConducted experiments: Ando, Hisaka.Contributed new reagents or analytic tools: Ando, Hisaka.Performed data analysis: Ando, Hisaka, Suzuki.Wrote or contributed to the writing of the manuscript: Ando, Hisaka,

Suzuki.

References

Agoram B, Woltosz WS, and Bolger MB (2001) Predicting the impact of physiological andbiochemical processes on oral drug bioavailability. Adv Drug Deliv Rev 50 (Suppl 1):S41–S67.

Badhan R, Penny J, Galetin A, and Houston JB (2009) Methodology for development ofa physiological model incorporating CYP3A and P-glycoprotein for the prediction of intestinaldrug absorption. J Pharm Sci 98:2180–2197.

Bergstrand M, Söderlind E, Eriksson UG, Weitschies W, and Karlsson MO (2012) A semi-mechanistic modeling strategy to link in vitro and in vivo drug release for modified releaseformulations. Pharm Res 29:695–706.

Bergstrand M, Söderlind E, Weitschies W, and Karlsson MO (2009) Mechanistic modeling ofa magnetic marker monitoring study linking gastrointestinal tablet transit, in vivo drug release,and pharmacokinetics. Clin Pharmacol Ther 86:77–83.

Bolger MB, Lukacova V, and Woltosz WS (2009) Simulations of the nonlinear dose dependencefor substrates of influx and efflux transporters in the human intestine. AAPS J 11:353–363.

Bornemann LD, Min BH, Crews T, Rees MM, Blumenthal HP, Colburn WA, and Patel IH (1985)Dose dependent pharmacokinetics of midazolam. Eur J Clin Pharmacol 29:91–95.

Bruyère A, Declèves X, Bouzom F, Ball K, Marques C, Treton X, Pocard M, Valleur P, BouhnikY, and Panis Y, et al. (2010) Effect of variations in the amounts of P-glycoprotein (ABCB1),BCRP (ABCG2) and CYP3A4 along the human small intestine on PBPK models for predictingintestinal first pass. Mol Pharm 7:1596–1607.

Chen J and Pang KS (1997) Effect of flow on first-pass metabolism of drugs: single pass studieson 4-methylumbelliferone conjugation in the serially perfused rat intestine and liver prepara-tions. J Pharmacol Exp Ther 280:24–31.

Cong D, Doherty M, and Pang KS (2000) A new physiologically based, segregated-flow model toexplain route-dependent intestinal metabolism. Drug Metab Dispos 28:224–235.

Cong D, Fong AK, Lee R, and Pang KS (2001) Absorption of benzoic acid in segmental regionsof the vascularly perfused rat small intestine preparation. Drug Metab Dispos 29:1539–1547.

DeSesso JM and Jacobson CF (2001) Anatomical and physiological parameters affecting gas-trointestinal absorption in humans and rats. Food Chem Toxicol 39:209–228.

Ekins S, Bravi G, Wikel JH, and Wrighton SA (1999) Three-dimensional-quantitative structureactivity relationship analysis of cytochrome P-450 3A4 substrates. J Pharmacol Exp Ther 291:424–433.

Englund G, Rorsman F, Rönnblom A, Karlbom U, Lazorova L, Gråsjö J, Kindmark A,and Artursson P (2006) Regional levels of drug transporters along the human intestinal tract:co-expression of ABC and SLC transporters and comparison with Caco-2 cells. Eur J PharmSci 29:269–277.

Gertz M, Harrison A, Houston JB, and Galetin A (2010) Prediction of human intestinal first-passmetabolism of 25 CYP3A substrates from in vitro clearance and permeability data. Drug MetabDispos 38:1147–1158.

Gertz M, Houston JB, and Galetin A (2011) Physiologically based pharmacokinetic modeling ofintestinal first-pass metabolism of CYP3A substrates with high intestinal extraction. DrugMetab Dispos 39:1633–1642.

Heikkinen AT, Baneyx G, Caruso A, and Parrott N (2012) Application of PBPK modeling topredict human intestinal metabolism of CYP3A substrates - an evaluation and case study usingGastroPlus. Eur J Pharm Sci 47:375–386.

Hénin E, Bergstrand M, Standing JF, and Karlsson MO (2012) A mechanism-based approach forabsorption modeling: the Gastro-Intestinal Transit Time (GITT) model. AAPS J 14:155–163.

Hisaka A, Nakamura M, Tsukihashi A, Koh S, and Suzuki H (2014) Assessment of intestinalavailability (FG) of substrate drugs of cytochrome p450s by analyzing changes in pharmaco-kinetic properties caused by drug-drug interactions. Drug Metab Dispos 42:1640–1645.

Hisaka A, Ohno Y, Yamamoto T, and Suzuki H (2010) Theoretical considerations on quantitativeprediction of drug-drug interactions. Drug Metab Pharmacokinet 25:48–61.

Hisaka A and Sugiyama Y (1998) Analysis of nonlinear and nonsteady state hepatic extractionwith the dispersion model using the finite difference method. J Pharmacokinet Biopharm 26:495–519.

Jamei M, Turner D, Yang J, Neuhoff S, Polak S, Rostami-Hodjegan A, and Tucker G (2009)Population-based mechanistic prediction of oral drug absorption. AAPS J 11:225–237.

Kajosaari LI, Laitila J, Neuvonen PJ, and Backman JT (2005) Metabolism of repaglinide byCYP2C8 and CYP3A4 in vitro: effect of fibrates and rifampicin. Basic Clin Pharmacol Toxicol97:249–256.

Kalampokis A, Argyrakis P, and Macheras P (1999) Heterogeneous tube model for the study ofsmall intestinal transit flow. Pharm Res 16:87–91.

Kimura T and Higaki K (2002) Gastrointestinal transit and drug absorption. Biol Pharm Bull 25:149–164.

Ku HY, Ahn HJ, Seo KA, Kim H, Oh M, Bae SK, Shin JG, Shon JH, and Liu KH (2008) Thecontributions of cytochromes P450 3A4 and 3A5 to the metabolism of the phosphodiesterasetype 5 inhibitors sildenafil, udenafil, and vardenafil. Drug Metab Dispos 36:986–990.

Lavrijsen KL, Van Houdt JM, Van Dyck DM, Hendrickx JJ, Woestenborghs RJ, Lauwers W,Meuldermans WE, and Heykants JJ (1988) Is the metabolism of alfentanil subject todebrisoquine polymorphism? A study using human liver microsomes. Anesthesiology 69:535–540.

Lin JH, Chiba M, and Baillie TA (1999) Is the role of the small intestine in first-pass metabolismoveremphasized? Pharmacol Rev 51:135–158.

Maeda K, Takano J, Ikeda Y, Fujita T, Oyama Y, Nozawa K, Kumagai Y, and Sugiyama Y(2011) Nonlinear pharmacokinetics of oral quinidine and verapamil in healthy subjects:a clinical microdosing study. Clin Pharmacol Ther 90:263–270.

Misaka S, Uchida S, Imai H, Inui N, Nishio S, Ohashi K, Watanabe H, and Yamada S (2010)Pharmacokinetics and pharmacodynamics of low doses of midazolam administered in-travenously and orally to healthy volunteers. Clin Exp Pharmacol Physiol 37:290–295.

Ohura K, Nozawa T, Murakami K, and Imai T (2011) Evaluation of transport mechanism ofprodrugs and parent drugs formed by intracellular metabolism in Caco-2 cells with modifiedcarboxylesterase activity: temocapril as a model case. J Pharm Sci 100:3985–3994.

Paine MF, Khalighi M, Fisher JM, Shen DD, Kunze KL, Marsh CL, Perkins JD, and ThummelKE (1997) Characterization of interintestinal and intraintestinal variations in human CYP3A-dependent metabolism. J Pharmacol Exp Ther 283:1552–1562.

Pang KS and Chow EC (2012) Commentary: theoretical predictions of flow effects on intestinaland systemic availability in physiologically based pharmacokinetic intestine models: the tra-ditional model, segregated flow model, and QGut model. Drug Metab Dispos 40:1869–1877.

Parker KL, Brunton LL, and Lazo JS, editors (2005) Goodman & Gilman’s The PharmacologicalBasis of Therapeutics, 11th ed, McGraw-Hill Companies, Inc., New York.

Petri N, Tannergren C, Holst B, Mellon FA, Bao Y, Plumb GW, Bacon J, O’Leary KA, KroonPA, and Knutson L, et al. (2003) Absorption/metabolism of sulforaphane and quercetin, andregulation of phase II enzymes, in human jejunum in vivo. Drug Metab Dispos 31:805–813.

Polasek TM, Sadagopal JS, Elliot DJ, and Miners JO (2010) In vitro-in vivo extrapolation ofzolpidem as a perpetrator of metabolic interactions involving CYP3A. Eur J Clin Pharmacol66:275–283.

Poulin P, Jones RD, Jones HM, Gibson CR, Rowland M, Chien JY, Ring BJ, Adkison KK,Ku MS, and He H, et al. (2011) PHRMA CPCDC initiative on predictive models of humanpharmacokinetics, part 5: prediction of plasma concentration-time profiles in human by using thephysiologically-based pharmacokinetic modeling approach. J Pharm Sci 100:4127–4157.

Rotzinger S, Fang J, and Baker GB (1998) Trazodone is metabolized to m-chlorophenylpiper-azine by CYP3A4 from human sources. Drug Metab Dispos 26:572–575.

Schiller C, Fröhlich CP, Giessmann T, Siegmund W, Mönnikes H, Hosten N, and Weitschies W(2005) Intestinal fluid volumes and transit of dosage forms as assessed by magnetic resonanceimaging. Aliment Pharmacol Ther 22:971–979.

Schulz R and Winne D (1987) Relationship between antipyrine absorption and blood flow rate inrat jejunum, ileum, and colon. Naunyn Schmiedebergs Arch Pharmacol 335:97–102.

Stephens RH, O’Neill CA, Warhurst A, Carlson GL, Rowland M, and Warhurst G (2001) Kineticprofiling of P-glycoprotein-mediated drug efflux in rat and human intestinal epithelia.J Pharmacol Exp Ther 296:584–591.

Sugano K (2009) Introduction to computational oral absorption simulation. Expert Opin DrugMetab Toxicol 5:259–293.

Sun D, Lennernas H, Welage LS, Barnett JL, Landowski CP, Foster D, Fleisher D, Lee KD,and Amidon GL (2002) Comparison of human duodenum and Caco-2 gene expression profilesfor 12,000 gene sequences tags and correlation with permeability of 26 drugs. Pharm Res 19:1400–1416.

Tachibana T, Kitamura S, Kato M, Mitsui T, Shirasaka Y, Yamashita S, and Sugiyama Y (2010)Model analysis of the concentration-dependent permeability of P-gp substrates. Pharm Res 27:442–446.

Tam D, Tirona RG, and Pang KS (2003) Segmental intestinal transporters and metabolic enzymeson intestinal drug absorption. Drug Metab Dispos 31:373–383.

Thomas S, Brightman F, Gill H, Lee S, and Pufong B (2008) Simulation modelling of humanintestinal absorption using Caco-2 permeability and kinetic solubility data for early drug dis-covery. J Pharm Sci 97:4557–4574.

Varma MV, Obach RS, Rotter C, Miller HR, Chang G, Steyn SJ, El-Kattan A, and Troutman MD(2010) Physicochemical space for optimum oral bioavailability: contribution of human in-testinal absorption and first-pass elimination. J Med Chem 53:1098–1108.

Willmann S, Schmitt W, Keldenich J, and Dressman JB (2003) A physiologic model for simu-lating gastrointestinal flow and drug absorption in rats. Pharm Res 20:1766–1771.

Willmann S, Schmitt W, Keldenich J, Lippert J, and Dressman JB (2004) A physiological modelfor the estimation of the fraction dose absorbed in humans. J Med Chem 47:4022–4031.

Winne D (1978) Blood flow in intestinal absorption models. J Pharmacokinet Biopharm 6:55–78.Yang J, Jamei M, Yeo KR, Tucker GT, and Rostami-Hodjegan A (2007) Prediction of intestinalfirst-pass drug metabolism. Curr Drug Metab 8:676–684.

Yu LX and Amidon GL (1999) A compartmental absorption and transit model for estimating oraldrug absorption. Int J Pharm 186:119–125.

Address correspondence to: Akihiro Hisaka, 1-8-1, Inohana, Chuo-ku, Chiba260-8675, Japan. E-mail: hisaka@chiba-u.jp

602 Ando et al.

at ASPE

T Journals on M

ay 21, 2020dm

d.aspetjournals.orgD

ownloaded from