Applied Concepts in PBPK Modeling

Citation: CPT Pharmacometrics Syst. Pharmacol. (2016) 5, 516–531; doi:10.1002/psp4.
12134
C 2016 ASCPT
V All rights reserved
TUTORIAL
Applied Concepts in PBPK Modeling: How to Build a

PBPK/PD Model
L Kuepfer1, C Niederalt1, T Wendl1, J-F Schlender1, S Willmann2, J Lippert2, M Block1, T Eissing1 and D Teutonico1,3*
The aim of this tutorial is to introduce the fundamental concepts of physiologically based pharmacokinetic/pharmacodynamic
(PBPK/PD) modeling with a special focus on their practical implementation in a typical PBPK model building workflow. To
illustrate basic steps in PBPK model building, a PBPK model for ciprofloxacin will be constructed and coupled to a
pharmacodynamic model to simulate the antibacterial activity of ciprofloxacin treatment.
CPT Pharmacometrics Syst. Pharmacol. (2016) 5, 516–531; doi:10.1002/psp4.12134; published online 21 September 2016.
GENERAL INTRODUCTION TO PBPK MODELING database of the PBPK software available. By contrast, the
Overview of PBPK drug parameters in the model are related to the compound
In recent decades, different types of computational models partition coefficient and permeability across biological mem-
have been applied in drug development programs. PBPK branes. In some cases such parameters can be measured
models combine information on the drug with independent in vivo or in vitro, but most often they are estimated from
prior knowledge on the physiology and biology at the organ- the physicochemical properties of the drug. In this context,
ism level to achieve a mechanistic representation of the physicochemical parameters act as surrogate parameters
drug in biological systems, allowing the a priori simulation for the calculation of the compound distribution. For this
of drug concentration–time profiles. Since PBPK models reason they can, in some cases, differ from the correspond-
explicitly consider different organs and tissues, it is possible ing measured value. This is particularly evident in the case
to obtain the quantitative characterization of concentration– of lipophilicity, since this parameter is usually measured
time profiles in the respective compartments. Such predic- under experimental in vitro conditions (e.g., octanol-water
tion is of high pharmacological relevance since it enables partitioning) which do not match in vivo conditions in the
the estimation of drug exposure not only in plasma but also biological environment. Usually the calculation of membrane
at the site of action, which may be difficult or impossible to permeabilities or partition coefficients is automatically per-
measure experimentally. formed by the PBPK software. The latter parameters are
A whole-body PBPK model (Figure 1a) contains an estimated from so-called distribution models quantifying
explicit representation of the organs that are most relevant equilibrium between plasma and the surrounding tissue. As
to the absorption, distribution, excretion, and metabolization an example, in Figure 1c the tissue and plasma concentra-
of the drug due to their physiological/pharmacological functions of theophylline estimated from its physicochemical
tion or their volume.1 These are typically heart, lung, brain, properties using different distribution models are illustrated.
stomach, spleen, pancreas, gut, liver, kidney, gonads, thy- As can be seen, the different models simulate different con-
mus, adipose tissue, muscle, bone, and skin. The tissues centrations in the different organs; this point will be dis-
are linked by the arterial and venous blood compartments, cussed in more detail in the section “Distribution models”
and each one of them is characterized by an associated (below).
blood-flow rate, volume, tissue-partition coefficient, and per- While passive processes can be calculated from such
meability. A major advantage of PBPK modeling is the model-based relationships, there is currently no similar meth-
availability of a comprehensive structural representation of od for a priori prediction of the kinetics of active processes
the physiology of an organism. The various parameters in (e.g., metabolism, target binding, or active transport of the
the model are either obtained from compilations of prior drug). One way to quantify the contribution of different organs
knowledge or may be calculated from specific and carefully to total clearance at the whole-body level is use of in vitro–in
validated formulas. From a general point of view, it is possi- vivo extrapolation.2 Furthermore, the relative tissue-specific
ble to distinguish between organism parameters on the one expression of genes or proteins might be considered.3 An
hand and drug parameters on the other. example of gene expression data is shown in Figure 1d,
Figure 1b represents the hierarchical relationship of the where the relative expression normalized to the tissue with
most important parameters in a PBPK model. The organ- the highest expression is represented for a group of enzymes,
ism parameters are usually used as direct input in the mod- in healthy individuals and cancer patients, respectively, as
el, representing the knowledge available a priori on the well as for receptors and transporters in healthy individuals.
anatomy and physiology. When dedicated software pack- The PBPK models are composed of different types of
ages are used, such information is usually contained in the information that are combined during model building and
1
Bayer Technology Services, Leverkusen, Germany; 2Bayer HealthCare, Wuppertal, Germany; 3Corresponding author current address: Clinical PK & Pharmacometrics,
Institut de Recherches Internationales Servier, 50, rue Carnot, 92284 Suresnes cedex, France. *Correspondence: D Teutonico ([email protected])
Received 13 April 2016; accepted 9 September 2016; published online on 21 September 2016. doi:10.1002/psp4.12134
Applied Concepts in PBPK Modeling
Kuepfer et al.
517
Figure 1 (a) Representation of the generic structure of a whole-body PBPK model. (b) Hierarchical representation of the main physio-
logical and drug parameters in a PBPK model. (c) Simulation of drug concentrations in different tissues using different distribution mod-
els for theophylline. This illustrates the impact of the choice of the distribution model on the simulated tissue concentration profiles
even though corresponding plasma concentrations are very similar. (d) Example of relative gene expression data for a group of
enzymes (in healthy and cancer patients), receptors, and transporters. Gene expression is represented as relative value obtained
through normalization to the tissue or organ with the highest expression.
that can be used to generate simulations of different treat- independent of organism physiology. Drug-biological proper-
ment scenarios. Such building blocks of information includ- ties (such as fraction of drug unbound, or tissue-plasma
ed in the model can be divided into organism properties, partition coefficient), on the other hand, are drug-specific
drug properties, and administration protocol and formulation but also defined by the interaction between the drug and
properties, respectively (Figure 2). the biological system itself, so they are dependent on both
Organism properties are, for instance, organ volumes, the drug and the organism properties.
organ composition, blood flows, surface areas, and expres- Finally, information on the administration protocol and for-
sion levels. Such properties are dependent on the species mulation properties is needed to define a PBPK simulation.
or population considered. For example, in the case of spe- Time-related special events such as gallbladder emptying
cial populations (e.g., diseased or pediatric populations) time or meal intake can also be included in the model and
these organism properties take into account physiological their impact on drug PK can be evaluated with the model.
differences between the special populations and the healthy The use of complex full-blown PBPK models was recently
adult reference population. By combining the drug proper- facilitated by the development of several commercial plat-
ties with the anatomical and physiological features of the forms that integrate physiological databases and implement
organism, it is possible to estimate the parameters for pas- PBPK modeling approaches, such as GastroPlus (Simula-
sive processes involved in distributing the drug in the body, tions Plus, Lancaster, PA), SimCyp (SimCyp, Sheffield,
such as, for example, permeation across membranes. UK), and PK-Sim and MoBi (Bayer Technology Services,
Drug properties include all parameters specific for the Leverkusen, Germany). These commercial PBPK modeling
compounds under study. Notably, physicochemical proper- platforms provide a generic model structure for the physiol-
ties such as compound lipophilicity, solubility, molecular ogy of predefined species and populations. They all include
weight (MW), and pKa values of a drug are fully physiological databases that are combined with compound-
www.wileyonlinelibrary/psp4
Kuepfer et al.
518
Figure 2 Representation of the general building blocks which can be part of a PBPK model. Some components may be optional
depending on the model considered.
specific information as well as biometric data, and are used integrative and iterative workflow such models can be used to
to parametrize a PBPK model on a whole-body level. generate working hypotheses from simulations. In turn, these
In previously published tutorials, Jones and Rowland-Yeo4 hypotheses can then be tested in specifically designed
introduced some of the basic concepts of PBPK as well as experiments that may lead to a better understanding of the
the use of PBPK in drug discovery and development, while underlying processes and a refinement of the model. In the
Maharaj and Edginton5 described the use of PBPK in pediat- case of pharmaceutical applications, the main goals of PBPK
ric clinical development. The aim of this tutorial is to cover modeling include (Figure 3): generating and testing a mecha-
the fundamental concepts of PBPK/PD models, focusing on nistic understanding of the physiological processes that gov-
their practical implementation in a typical PBPK model build- ern an observed drug behavior; translating the understanding
ing workflow. In the next sections an overview of the infor- to novel settings (e.g., to a different population); identifying an
mation contained in the different building blocks of the model ideal therapeutic regimen; and optimizing risk–benefit ratios.
(section “Building blocks in PBPK modeling”) will be pre- In fact, any deviation between the model simulation and the
sented and then the different parameters included in the data can provide insights into the mechanisms of the underly-
model structure will be discussed in more detail (section ing processes that may not yet be reflected in the existing
“Passive and active processes”). Finally, practical guidance model.
on the different steps of the model-building process will be Another important application of PBPK models are
provided (section “Best practices in PBPK model building”) extrapolations to novel clinical scenarios such as different
and an example of PBPK/PD model will be discussed (sec- treatment schedules or patient subgroups. Notably, transla-
tion “Case study: Building a PBPK model for ciprofloxacin”). tion to novel therapeutic designs can usually be achieved
by only changing single or limited sets of parameters.
Applications of PBPK modeling Extrapolations to different species or specific populations,
Computational models provide an ideal platform for knowl- in contrast, require knowledge of the sets of parameters
edge management since they offer the opportunity to collect, characterizing the physiological properties of the species or
integrate, analyze, and store information. Ideally, in an particular patient cohort of interest.6
CPT: Pharmacometrics & Systems Pharmacology

Kuepfer et al.
519
Figure 3 Schematic representation of the most common applications of PBPK modeling.
Typical applications of PBPK modeling include but are of such a model is discussed in the ciprofloxacin case
not limited to: study presented in section “Case study: Building a PBPK
model for ciprofloxacin” of this tutorial. Since different
• Pediatric extrapolations,5,7 where, with the support of the organs are explicitly represented in PBPK models, on- and
US Food and Drug Administration, a reference PBPK off-target tissue exposure can be directly quantified. This
model for an adult population is scaled to specific life enables a consideration of therapeutic or toxic effects by
stages in children. coupling of PD to the corresponding PBPK models. To this
• Extrapolations to diseased populations, e.g., hepatically end, tissue concentration profiles simulated with PBPK
impaired cirrhotic patients.8 models may be used as input for downstream PD models.
• Evaluating the effect of possible drug–drug interactions Notably, this coupling of PK and PD results in a multiscale
(DDI), e.g., by considering competitive inhibition of a PBPK/PD model that simultaneously describes drug ADME
metabolizing enzyme.9 at the whole-body level and the resulting drug effect at the
• Scaling between different treatments scenarios by simu- cellular or tissue scale.
lating specific dosing schemes.10 In the simplest case, PD models may represent simple pro-
• Assessing the impact of different drug formulations by liferation models quantifying, for example, bacterial growth in
modifying corresponding compound/formulation infection biology18 or tumor growth in oncology.19 Given the
parameters.11 ever-growing number of systems biology models at different
scales of biological organization, the PD disease or toxicity
Advanced applications include cross-species extrapola- models may be largely expanded. Examples of such detailed
tion,12,13 multiscale modeling,14–16 and statistical modeling PBPK-based models include glucose-insulin regulation in dia-
using methods such as Bayesian approaches.17 betes,10 management of endometriosis,20 acetaminophen
intoxication,15 drug-induced liver, injury21 or model-based
Pharmacodynamic modeling design and analysis of antithrombotic therapies.22 Ultimately,
PBPK models offer a mechanistic framework to quantify the such PBPK/PD models aim for a mechanistic and physiologi-
pharmacodynamic (PD) effect of a drug through coupling to cal representation of systems interactions within the body,23
simulated on- and off-target tissue concentrations. The final thereby providing an important toolbox for disease or toxicity
result in this case would be a PBPK/PD model. An example modeling in quantitative systems pharmacology.
Kuepfer et al.
520
BUILDING BLOCKS IN PBPK MODELING can enter the mucosa of the corresponding segment either
Anatomy and physiology by the transcellular route (passing via the enterocytes into
The physiology of the organism of interest is essential prior the intracellular space) or by paracellular diffusion into the
knowledge in PBPK modeling and relevant organs are interstitial space of the mucosa. The paracellular route has
explicitly included in the model. Each organ is represented been shown to contribute significantly to the overall intesti-
by its anatomical and physiological properties; for example, nal absorption of a few small molecules such as acyclovir,
volume, tissue composition, and perfusion blood flow rates. cimetidine, and ranitidine.29 However, the paracellular route
Furthermore, in some cases each compartment can be fur- is in general considered to contribute only marginally to the
ther divided into subcompartments, such as plasma, inter- intestinal absorption, mainly due to the small surface area
stitium, cellular space, and red blood cells. Structurally, fraction that allows paracellular permeation.35 During the
these subcompartments are the lowest level of structural transit through the lumen, the maximum concentration of
differentiation in PBPK models and, for example, enzymatic the drug might be limited by the local pH-dependent GI sol-
processes can be localized in the relevant intracellular com- ubility, which can be calculated using the Henderson–
partment to which the extracellular space can be connected Hasselbalch equation. The intestinal pH will also affect the
by passive diffusion and additionally by active drug trans- degree of dissociation of the compound, and consequently
port, where required. Without subcompartmentalization cer- the relative concentration of neutral vs. the charged form(s).
tain effects such as permeation limited metabolization In this case, the charged molecules have a lower perme-
within the liver may not be represented. Blood flow governs ability than the neutral form, and this should be also
mass transfer within the body, and its rate is specific to each accounted for in the model.36
organ. Where necessary, a more detailed mechanistic
Drug properties
description of a subset of organs can be included in the
The second major type of input into PBPK models are
model; for instance, to obtain a more accurate description of
compound-specific parameters. These include the physico-
the distribution in the brain,24 liver,25 kidney,26,27 or lung,28 or
chemical parameters of the compound (e.g., MW, lipophilic-
describe in detail the process of intestinal absorption.29–31
ity, solubility, and pKa values) that can usually be
Generally, two blood compartments, arterial and venous
determined in vitro. Lipophilicity is a key parameter that,
blood, connect the organ compartments in a PBPK model.
together with the MW, is used to calculate the membrane
Yet in clinical practice it is common to sample blood only
permeability of a drug. The MW is, in this context, a surro-
from the superficial veins of patients, e.g., the antecubital
gate measure for the size of the molecule. Since halogen
vein. Thus, the measurements collected in peripheral veins
atoms in particular contribute less to the molecular volume
present a concentration that may differ significantly from
than expected from their atomic weight, a correction term
both arterial and central vein compartments, making the
can be applied to obtain an effective MW, which is a better
comparison of model predictions with collected data diffi-
representation of the size.36 Particularly relevant for oral
cult. To solve this discrepancy, Levitt32 proposed a model in
administration, drug solubility determines the availability for
which an additional organ representing the tissues drained
absorption of a compound in the GI tract, while pKa values
by the antecubital vein was added such that the concentra-
are used to calculate pH-dependent changes in drug solubili-
tion in the peripheral venous blood can be described.
ty. Pharmacokinetic parameters, such as fraction unbound,
Based on this model, a peripheral venous blood compart-
are likewise compound-specific, yet they also depend on the
ment has also been included in some PBPK software
organism. These basic values are frequently used together
tools.12,33
with physiological parameters to calculate other parameters
Gastrointestinal transit and absorption models. For an orally that are much more difficult to measure but can be used
administered drug, a combination of physiological, com- immediately to quantify the drug mass balance in PBPK
pound, and formulation-related aspects comes into play. models. For example, tissue–plasma partition coefficients
These affect the relevant parameters such as solubility, dis- quantifying the distribution of a compound are calculated
solution rate, and intestinal permeability. from MW, lipophilicity, fraction unbound, and pKa values,
From a general perspective, the gastrointestinal (GI) tract depending on the distribution model used (see below). Frac-
can be divided into several segments based on their ana- tion unbound, the fraction of a drug not bound to plasma
tomical and functional role: stomach, small intestine (duo- proteins, strongly influences drug distribution and clearance
denum, upper and lower jejunum, upper and lower ileum) since only the available free fraction of a compound diffuses
and large intestine (cecum, colon ascendens, colon trans- into cells, where it might be metabolized. Also, only the free
versum, colon descendens, sigmoid, and rectum). Each of fraction of a compound is filtered for renal excretion. The
these segments can be further divided into zones repre- fraction unbound can either be measured experimentally,
senting the lumen and gut wall, which can in turn be divid- estimated from the drug concentration, or predicted from the
ed into mucosal and nonmucosal tissues. The separation of lipophilicity of the drug, respectively.37 Notably, all the afore-
the mucosal tissue from the remaining part of the gut wall mentioned compound-specific parameters contribute to the
allows the absorption processes to be described in a realis- quantification of passive processes such as membrane per-
tic manner, reflecting the localization of active processes meation and organ/plasma partitioning. On the other hand,
such as gut wall metabolism in the respective tissue.34 key parameters for quantifying active processes such as
After oral administration, the drug is transported along enzymatic activity (Vmax) must be obtained by a comple-
the lumen of the intestine according to a transit function. It mentary approach. For example, parameters for an active

Kuepfer et al.
521
process can be determined in vitro or fitted from in vivo data calculated from substance-specific surrogate markers, such
and then converted to the enzymatic rate or a different as lipophilicity, to obtain, for example, permeability or parti-
parameter. tion coefficients. Different concepts for quantifying passive
and active transport processes are described in the follow-
Formulation
ing sections.
The development of oral dosage forms for new chemical
entities is often the favored option in drug development. Estimation of passive processes from physicochemical
The design and development of such formulations can be properties for small molecules
supported by accurate mechanistic models that can simu- Distribution models. Small molecules can generally pene-
late in vivo bioavailability. In vitro dissolution profiles in bio- trate all kinds of tissues in the body. Experimentally, the
relevant media, such as the Fasted State Simulated quantification of this compound-specific distribution is very
Intestinal Fluid (FaSSIF) and the Fed State Simulated labor-intensive.41,42 A major advance in PBPK modeling
Intestinal Fluid (FeSSIF), have been established. When came with the use of calculation methods for organ/plasma
used in standardized in vitro dissolution test methods, such partition coefficients, which describe steady-state ratios of
media allow simulating the in vivo disintegration and disso- the concentrations in the blood or plasma and the sur-
lution behavior of orally administered dosage forms.38 This rounding tissue. Various concepts for mechanistic correla-
dissolution profiles together with drug PK profiles can be tions have been developed for the in silico estimation of
used to establish an in vitro–in vivo correlation (IVIVC). organ/plasma partition coefficients. Based on tissue compo-
Such correlation enables the estimation of in vivo concen- sition, these coefficients can account for the distribution
tration PK profiles from the in vitro dissolution profiles of dif- between drug-binding tissue constituents, such as proteins
ferent formulations, providing a tool for optimizing the or lipids, on the one hand, and water on the other. Although
administered dosage form with a minimal number of in vivo the principles are very similar in all cases, the calculation
animal experiments or clinical trials.39 methods deviate with respect to the kind of parameters
As an alternative to a direct correlation between in vitro used and resulting in different values of tissue concentra-
and in vivo concentrations, it is possible to include dissolution (Figure 1c). All known partition coefficient models
tion functions in the GI compartments of PBPK models to assume that tissue is composed of a limited number of
account for the disintegration and dissolution processes in components, and they all include partition coefficients for
the GI tract.40 The inclusion of such dissolution profiles in water/protein and lipid/water. These two partition coeffi-
GI physiological models and their combination with whole- cients are usually calculated from so-called surrogate in
body PBPK models has been described in the literature.30 vitro measurements. The total organ/plasma partition coeffi-
cient is then calculated as a weighted sum of the partition
Routes of administration and administration protocols
coefficients for all of the components; the weights are the
The administration protocol includes the (1) route of admin-
volume fractions of each component. Importantly, the distri-
istration, (2) amount of compound (dose), and (3) an
bution of a drug within aqueous and organic tissue compo-
administration scheme for multiple dosing. Usually, when
nents is always assumed to be homogenous and passive.
comparing model simulation to available data, the adminis-
The most widely used concepts for calculating organ/
tration protocol used in the model will match the one used
plasma partition are briefly introduced below:
in the study for which the data are under evaluation.
Various routes of administration can be considered in the • Poulin et al.43,44 calculate the lipo-hydrophilicity of tissue
model. This will define how the drug is absorbed and, part-
as a mixture of neutral lipids, phospholipids, and water. In
ly, how it will be distributed in the body. Common adminis-
addition to the volumetric tissue composition, fraction
tration routes are intravenous (i.v.) and oral (p.o.)
unbound (fu), lipophilicity (logP and logD), and pKa are
administration, but also alternative routes such as inhala-
used as compound-specific input parameters. Here as
tion, subcutaneous, intramuscular, topical, or ocular admin-
well as in all the following concepts for the calculation of
istration are used in PBPK models. Apart from i.v., when
organ/plasma partition coefficients, fu quantifies specific
considering other administration routes, also mechanistic
reversible binding to proteins in plasma and tissue,
consideration related to local drug release and absorption
whereas lipophilicity accounts for nonspecific binding to
from the administration site should be included in the PBPK
lipids.
model, for instance to describe the progressive drug
• Rodgers et al. extended the concepts of Poulin et al. to
release and diffusion from the injection site after intramus-
electrostatic interactions at physiological pH.45,46 These
cular administration.
include binding of ionized and unionized drugs to acidic
phospholipids and neutral lipids, respectively. Also elec-
PASSIVE AND ACTIVE PROCESSES trostatic drug interactions with extracellular proteins are
taken into account. Consequently, the partition coeffi-
In PBPK modeling, physiological parameters provide the cients are calculated taking into account the lipophilicity
structural scaffold of the model, but active and passive pro- and pKa value of the drug and the pH values of the
cesses describe the actual distribution, excretion, and tissues.
metabolization of the drug (mass balance). The most • Berezhkovskiy47 modified the calculation method by Pou-
important passive processes are tissue permeation and lin et al. by accounting for peripheral drug elimination that
organ/plasma partitioning. The underlying processes are results in a different volume of distribution.
Kuepfer et al.
522
• Willmann et al.48,49 extended the concept of Poulin et al. An important consideration in calculating the rate and
by additionally considering proteins as a tissue compo- extent of intestinal drug absorption is the intestinal perme-
nent. Moreover, Willmann et al. use membrane affinity ability. There are several experimental techniques that are
(logMA)50 to quantify partitioning between water and an able to provide a relatively reasonable assessment of the
artificial cellular membrane as a measure for lipophilicity intestinal permeability, e.g., artificial membranes,57 “intestine-
by using an empirical equation.51 This model is imple- like” cell lines (e.g., Caco-2 or MDK),58,59 Ussing chambers,60
mented in the PK-Sim software as the PK-Sim standard and in situ rat intestinal perfusion techniques (e.g., single-
distribution model. pass intestinal perfusion).61 In humans, effective permeability
• Schmitt52 developed a method for calculating the organ/ in different regions of the intestine can be measured using
plasma partition coefficient for organic compounds, by single-pass perfusion methods.35 An alternative approach for
compartmentalizing tissue as water, neutral lipids, neutral establishing the intestinal permeability is to use the physico-
phospholipids, acidic phospholipids, and proteins. This chemical properties as a basis for calculations.1 Several
concept accounts in particular for electrostatic interac- models are available for this approach; for instance, the mem-
tions between charged molecules at physiological pH and brane affinity of a compound (defined as its equilibrium parti-
acidic phospholipids. Also for this distribution method, the tion coefficient for water and immobilized lipid bilayers) and its
partition coefficients are calculated taking into account effective MW can be used for such calculations, using a semi-
the lipophilicity and pKa value of the drug and the pH val- empirical equation.51
ues of the tissues.
Estimation of active processes for small molecules
As shown in Figure 1c, different distribution models can Active processes allow the inclusion of additional mechanis-
be used to describe similar plasma concentration profiles of tic details in PBPK models, since they can frequently be
the drug but, because of the intrinsic nature of the models, assigned to biochemical processes at the molecular scale.
they will generate different tissue concentrations. If tissue Examples of such processes are metabolization reactions,
concentrations for the drug are not collected, when assess- in which a molecule is transformed and modified, as well as
ing target site drug exposure it should be kept in mind that transport processes, in which a compound is actively taken
the selection of the distribution model represents a funda- up or secreted into a specific tissue compartment. Quantifi-
mental assumption during model development. An inherent cation of active processes follows different concepts than
uncertainty should hence be taken into account in the pre- the ones of passive processes, relying on either careful
dictions. Ideally, different distribution models should be extrapolation of in vitro results to an in vivo situation or
available during model development to allow for full struc- knowledge-based adjustment of parameters that are difficult
tural flexibility in this regard. It should be kept in mind that to measure experimentally. In general, all clearance pro-
Figure 1c represents a specific example for the theophyl- cesses presented can be formulated as, for, instance first-
line molecule, and it should not be used to draw any con- order clearance kinetics or Michaelis–Menten kinetics. The
clusion concerning comparison of distribution models and Michaelis–Menten kinetics are often used to describe pro-
tissue predictions since the simulated concentrations are cesses that can be saturated by the substrate concentra-
largely dependent on the drug parameters used as input. tions; its equation is the following, with v representing the
Permeability models. Diffusion across the vascular wall reaction rate, Vmax the maximum reaction rate, Km being
from plasma to interstitial space of the organs or diffusion the Michaelis–Menten constant and the substrate concen-
across the cellular membranes of tissue cells or red blood tration [S]:
cells determines how fast drug distribution takes place and
can be rate limiting for distribution (permeation limited kinet- V max ½S
v5 (1)
ics). In general, the diffusion across the vascular wall is Km1½S
assumed to be fast for small molecules. However, this
assumption might not hold true for drugs with high protein If [S] is significantly lower than Km, elimination kinetics can
binding or which are highly hydrophilic.53 Also in the case be considered linear (first-order process) and the rate con-
of large molecules (e.g., protein therapeutics) this assump- stant for the clearance process can be approximated by the
tion does not hold true (cf. section “Passive and active pro- ratio Vmax/Km.
cess for large molecules”). Parameters included for active processes are typically
The cellular membrane is the diffusion barrier between rate constants for elimination or active transport processes
interstitial space and intracellular space or between blood (e.g., intrinsic hepatic clearance, Michaelis–Menten param-
plasma and blood cells. The rate constant describing the eters Km and Vmax, or parameters describing specific bind-
diffusion across the cellular membrane can be written as ing, for instance, to a receptor kon and koff). Examples of
the product of the drug permeability (P) and the effective such processes are metabolic elimination related to a par-
surface area (SA) of the cells. Methods to calculate the ticular enzyme, or transporter activity or drug distribution
drug-specific permeability depending on physicochemical due to binding partners. Parameters that define active pro-
properties like lipophilicity and MW can be found in the liter- cesses are usually extrapolated based on in vitro measure-
ature.54,55 In order to calculate the effective surface area ments,62,63 when the necessary extrapolation factors for
for different organ volumes, allometric scaling equations measurement settings are available, or fitted using in vivo
can be used.56 data.64 In vitro–in vivo extrapolation (IVIVE) can be applied

Kuepfer et al.
523
to the different stages of absorption, distribution, metabo- Vmaxj 5kcat E0;j 5kcat E0 erel;j 5kcat erel;j (2)
lism, and excretion (ADME) processes.
This reformulation allows describing the tissue-specific
abundance of an enzyme or transporter (E0,j) as the prod-
Intrinsic and total plasma clearance. Clearance is generally
uct of its overall concentration, Eo, and its relative expres-
used to quantify elimination rates in liver, kidney, or other
sion, erel,j, in a specific organ j. While Eo is obviously not
organs. At the body level, total plasma clearance (CLtot, i.e., accessible experimentally, relative expression profiles (erel,j)
volume of plasma cleared per time unit) describes the sum across different tissues may be obtained from public data-
of multiple clearance processes that occur simultaneously bases.3 In the above equation, kcat* is a new parameter
within multiple organs. Here, CLtot is the apparent rate at that combines the catalytic efficiency kcat and the overall
which a compound is removed from the systemic circulation. concentration Eo. This parameter needs to be estimated
In PBPK modeling, the relative contribution of each organ to through parameter optimization. Moreover, it should be not-
CLtot can be further differentiated by quantifying the specific ed that this reformulation reduces significantly the number
elimination rate in each organ. In this case, the intrinsic of independent model parameters, since only a single
clearance CLint is used to quantify the intracellular activity of parameter (kcat*) needs to be considered instead of
a metabolizing enzyme in different organs.65 assigning a catalytic efficiency Vmaxj separately for each
organ. Please note also that the relative tissue-specific
In vitro–in vivo extrapolation of clearance. For liver clear- abundance of an enzyme or a transporter may be approxi-
ance, numerous in vitro assays are available to estimate mated either through gene or protein expression, respec-
the enzymatic activity, and this clearance needs to be tively,3 even though correlations of both are a matter of
rescaled to the whole-organ in vivo situation. For example, debate. This is because any posttranscriptional effect can
activity can be derived from recombinant cytochromes, be assumed to be largely protein specific, as such only
impacting kcat*. The use of expression data to quantify
human hepatocytes, or human liver microsomes. For an
relative tissue-specific catalytic efficiency represents a com-
IVIVE, the experimentally measured specific activity needs
plementary approach to the IVIVE of clearance process-
to be multiplied by the amount of catalyzing agent (amount
es.68 Likewise, drug elimination in first-pass organs such as
of cytochrome, number of hepatocytes, or amount of micro-
liver, lung, or GI as well as other organs can be described.
somal protein) per gram of liver tissue to obtain the specific Figure 1d shows the relative gene expression for a
intrinsic clearance per gram of liver. The ratio of Vmax/Km group of enzymes, receptors, and transporters in healthy
describes the intrinsic metabolic clearance for unsaturated individuals or cancer patients, respectively. Note that the
conditions, which can be derived from in vitro studies in relative expression is dimensionless following normalization
human liver microsomes. In the past, IVIVE was routinely to the tissue with the highest expression. As expected,
conducted based on mean data disregarding interindividual cytochrome P450 expression is consistently largest in the
variability. As an alternative approach, it has been proposed liver and the intestines, while receptor expression is rather
to estimate enzyme abundance and its variability, measured ubiquitous. In turn, transporter availability is similar to cyto-
by immunoquantification, to allow a direct extrapolation to chrome P450 with an occasional small contribution in the
the population level.66,67 kidney (MRP2, BCRP). Interestingly, CYP2D6 expression is
The second main clearance route in the body is usually higher in the large intestine than in the liver in two indepen-
renal secretion. Total renal clearance can be derived from dent measurements (healthy subjects and cancer patients).
the fraction of a drug secreted in the urine. This value can To validate this observation protein expression data from
be further differentiated into glomerular filtration, which the literature could be alternatively used for model building,
describes the passive filtration rate of urine in the kidney, instead. Only a limited number of organs or tissues are rep-
and tubular reabsorption and tubular secretion, which are resented for illustration.
driven mainly by active transport. Receptor availability. The above-described concepts use
expression data to quantify catalytic efficiency but they are
Protein abundance in active processes. Although the liver applicable to both enzymatic processes and drug transport-
and kidney are the key secretory organs of the body, the ers as well. Notably, the approach can immediately also be
expression of many enzymes and drug transporters is not applied to estimate the abundance of receptors for different
limited to these two organs. One way to quantify the contri- drugs, including antibody-drug conjugates.69 Thus, the
bution of different organs to the total clearance at the availability of specific drug binding partners can be quanti-
whole-body level is to consider the relative tissue-specific fied simultaneously in multiple organs; this has enormous
expression of genes or proteins. Recently, a generic potential for mechanistic consideration of target-mediated
approach was proposed whereby gene expression is used drug disposition or clearance, as well as on/off-target
as a surrogate for tissue-specific protein abundance in effects.
PBPK modeling.3,12 The catalytic activity of each process is Passive and active processes for large molecules
generally quantified for each organ by Vmax (lmol/l/min). Therapeutic proteins are an increasingly important class of
Notably, Vmaxj is the product of the catalytic efficiency kcat drugs.70 PK/PD modeling of therapeutic proteins differs
(1/min) and the overall concentration of the catalyzing pro- from those of small-molecule drugs, mainly due to the dif-
tein E0,j (lmol/l) in the corresponding organ j: ferences in size.71–73 PBPK models as well must therefore
Kuepfer et al.
524
take into account the special mechanisms that govern the Besides these parameters, any further ADME information
pharmacokinetics of protein therapeutics, mechanisms that regarding, for instance, clearance processes, transporters,
can often be neglected for small molecules. Both the or specific binding partners are relevant in order to build
exchange of drug across the vascular endothelium and the physiologically plausible models. The appropriate level of
return of drug by the lymph flow from the interstitial space detail in representing the relevant processes and the crite-
of the organs to the systemic circulation are important phe- ria for model evaluation depend on the particular aims,
nomena for therapeutic proteins. These two processes objectives, and applications of the model as well as on
influence the volume of distribution for proteins, and are availability of data and information. Available experimental
generally considered in PBPK models of therapeutic PK data are used to identify unknown or uncertain parame-
proteins.74–84 ters as well as for model validation. Depending on their
Published PBPK models use various approaches to availability, data commonly used in PBPK model building
describe the extravasation of protein therapeutics. One includes time–concentration profiles of the drug in the plas-
commonly used is the two-pore formalism,85,86 which con- ma (for different preclinical species and/or humans, differ-
siders the barrier between the plasma and the interstitial ent application routes, different dosages, or applications
space as a membrane consisting of two types of pores: few schemes), time–concentration profiles of the drug in rele-
large ones and many small ones. Macromolecules are vant tissues, the percentages of drug excreted via the urine
assumed passing through these pores by convection as and feces, mass balance data, and metabolites data (time–
well as by diffusion; convection being the predominant concentration profiles, excretion data). A scheme of the
extravasation mechanism for large protein therapeutics PBPK model-building workflow for small molecule drugs is
such as antibodies or albumin. provided in Figure 4.
Another relevant process for antibodies or other proteins Establishment of the PBPK model for i.v.
(e.g., albumin fusion proteins) is the catabolism within administration
endosomal space and the protection from catabolism by As a first step, a method for calculation of organ/plasma
neonatal Fc receptor (FcRn), which is also often taken into partition coefficients, frequently referred to as distribution
account by more recently published PBPK models.76–84 model, needs to be chosen to describe the drug distribution
In describing the lysosomal degradation of drugs and behavior. The distribution model selected for a certain com-
their recycling by FcRn, the organ representation used for pound should be able to consistently describe the com-
small molecules is usually extended by adding a compart- pound distribution independently of the considered species
ment for the endosomal space. This space represents the or administration protocol. By comparing simulations with
region within the endothelial capillary walls where catabo- PK measurements in vivo after i.v. administration, different
lism and high-affinity binding to the FcRn occurs (acidic distribution models can be compared and clearance param-
environment). The fraction of drug that is bound to FcRn is eters can be estimated. The clearance parameters can be
recycled to the plasma and interstitial space, whereas the estimated from experimental in vivo data, i.e., plasma
unbound fraction is subject to endosomal clearance. Addi- concentration–time profiles. Additionally, mass balance data
tional processes can also be important and implemented if might be used if different elimination pathways should be
needed. Examples include target-mediated disposition and informed. Rates for metabolism processes might also be
clearance84 and immunogenicity.87,88 estimated using IVIVE approaches (see section “In vitro–in
vivo extrapolation of clearance”). Surrogate compound
BEST PRACTICES FOR PBPK MODEL BUILDING parameters, like lipophilicity, are usually slightly adapted at
this stage to obtain a good agreement with experimental
Due to increasing application of PBPK modeling in drug i.v. data. There are no established rules regarding the
development and regulatory submissions, there have been extent of change that is reasonable, since this also
recent efforts by regulatory agencies, industry, and aca- depends on which lipophilicity measure is used as starting
demia to discuss and develop best practices for establish- value. Even though a PBPK model may comprise several
ing as well as reporting of PBPK modeling.89–94 In this hundreds of ordinary differential equations, the number of
section, general recommendations for PBPK model building independent model parameters for a new compound is usu-
are provided (for information specific to applications for ally small (in most cases, fewer than five per compound),
pediatric populations, see Refs. 5,95). due to the large amount of prior, independent physiological
information that is incorporated (Figure 1b). As in the case
Compilation of available data and information of the distribution models, also these compound-specific
As a first step in model development, all available informa- parameters are usually kept unchanged across different
tion on the drug regarding its ADME properties are gath- species or administration protocols. If the selection of distri-
ered. This includes the drug-specific parameters, which are bution models and adjustment of parameters are not suffi-
used as input parameters, and the characterization of the cient to describe the i.v. data, this can be an indication that
organism or population. Default physiological parameters relevant processes are not yet taken into account in the
should only be changed if there is a mechanistic rationale, model. In such cases, for instance, active transport or bind-
for instance, in the case of special populations. The physi- ing partners can be incorporated based on current knowl-
cochemical drug properties (see section “Drug properties”) edge or as testable hypothesis. Such additional processes
are necessary parameters to inform the drug distribution. might also be needed to explain features such as dose

Kuepfer et al.
525
Figure 4 Flowchart illustrating the steps usually used in PBPK model building.
nonlinearity, the observed extent of clearance, or the drug distribution or metabolism/excretion should be further modi-
distribution. When introducing additional processes, it is fied and only those parameters that influence the oral
best to include as much experimental data as possible absorption should be varied. Typical parameters to be esti-
(e.g., Km and Kd, values, abundance of enzyme, binding mated during development of a p.o. model are intestinal
partner). It should be noted that data for different doses permeability and parameters related to meal events. In
have to be available (showing dose-nonlinearity) in order to addition, formulation-related parameters describing drug
be able to identify Michaelis-Menten parameters, e.g., for release as well as solubility might be adjusted, especially if
transporters or clearances. aqueous solubility rather than FaSSIF and FeSSIF solubility
were used as starting values. After that, additional
Establishment of the PBPK model for oral absorption-related processes, such as enterohepatic recy-
administration cling (EHC), might be included if relevant for the drug under
Once an i.v. model is established, a model for p.o. adminis- consideration. The stepwise approach of i.v. and p.o. (or
tration (or for other extravascular routes) can be estab- any extravascular administration) model building is described
lished. In this step none of the parameters that influence to stress that during the establishment of a model including
Kuepfer et al.
526
drug absorption, the information regarding distribution and impaired patients, are available, the model can also be
clearance processes from the i.v. data should be used in evaluated using these data after changing the physiological
order to be able to identify the parameters relevant for parameters accordingly. The model might additionally be
absorption. Note that in some cases the proposed strictly con- evaluated for consistency across compounds for which the
secutive building of i.v. and p.o. PBPK might not be the best same processes are relevant for the ADME properties. For
approach; for instance, if the i.v. model incorporates process- example, if a saturable receptor binding was implemented
es that are also relevant for oral absorption, e.g., uptake as a relevant process into a PBPK model, it could be
transporters. In such cases, it might be preferable to perform checked if the receptor concentration is the same in a
parameter identification using the i.v. and p.o. data PBPK model of a reference compound binding to the same
simultaneously. receptor.
As mentioned, i.v. data are highly valuable, since they
allow informing distribution and clearance processes inde- Sensitivity analysis or best- and worst-case scenarios. In
pendent from the drug absorption. Depending on the order to assess uncertainties in model results, it is recom-
modeling purpose and the needed model precision, a mended to perform a sensitivity analysis for relevant param-
PBPK model might also be developed using p.o. data only. eters. Such sensitivity analysis can be performed, for
Taking into account data on drug absorption, metabolism, instance, on uncertain parameters for active processes
and mass balance data, it is necessary to rely on the included in the PBPK model or to all parameters in the mod-
PBPK or IVIVE methods regarding prediction of absorption el, in order to identify the most sensitive parameters for a
or clearance. If human i.v. data are missing, a human p.o. specified model output (e.g., plasma concentration or PK
model might be scaled from an animal model that was parameters). Additionally best- and worst-case scenarios
established using i.v. as well as p.o. data. Without using i.v. can be simulated in order to evaluate the effect of changing
data, the model uncertainty can be considerably higher uncertain parameters to extreme values within the experi-
depending on the BCS classification of the drug and if mental and physiological uncertainties. Uncertainties regard-
transporters or gut-wall metabolism are involved. ing underlying mechanisms can be assessed by simulating
the respective model alternatives. The results of the sensitiv-
Overall model evaluation
After model development, the model quality should be eval- ity analysis or best- and worst-case scenarios can be used
uated. This step should address whether the model fits its to assess if the conclusions of modeling work are robust.
purpose. As such, the outcome of the evaluation depends
on the goal of the modeling project. The following criteria or
CASE STUDY: DEVELOPING A PBPK/PD MODEL FOR
tests can be used in evaluating the model.
CIPROFLOXACIN
Agreement of modeling outcome with experimental data.
Usually, a model is evaluated by visually comparing simu- In order to illustrate the various steps of PBPK model
lated vs. experimental concentration–time profiles (for plas- development in a practical example, the construction of a
ma and, if available, for other tissues), focusing on the PBPK model for the antibiotic ciprofloxacin (CIP) will be
absolute concentrations and the dynamic shape of the PK described in this section. Following the best-practices sec-
profile or by means of error functions such as root-mean- tion above, the experimental data needed for parametriza-
square deviation (RMSD), the area under the curve (AUC) tion of the basic model structure will be discussed and it
error,12 or the concordance correlation coefficient.16 Addi- will be shown how both i.v. and p.o. administration of the
tionally, typical PK parameters like Cmax, tmax, AUC, and t1/2 drug may be systematically considered during model estab-
can be compared between data and simulation during mod- lishment. Finally, an empirical PD model of CIP treatment
el evaluation. While comparing model simulations and of E. coli infections18 will be coupled to the PBPK model.
experimental data, it should be considered that both are This resulting PBPK/PD model will be then used to simulate
subject to uncertainty, and also experimental data might be the therapeutic effect of different dosing schemes. The indi-
critically reevaluated. In addition to data used for model vidual steps in model development are provided as a
building, further data can also be considered as external hands-on tutorial in the Supplementary Materials. The
validation to assess consistency of model prediction. final modeling example shows prototypical demands and
Consistency of PBPK models for different doses, across applications of PBPK modeling in a pharmaceutical devel-
species, special populations, and compounds. An important opment program.
cross-check for model validation is the consideration of dif- The first step in PBPK model building is gathering and
ferent doses. Deviations in estimations for dose levels that assembling existing information. In this regard, the basic
had not been considered during model development point physicochemistry of the drug is of particular importance
to structural shortcomings of the model. The drug- (Table 1A). The physicochemical information in Table 1A
dependent parameters as well as the calculation methods can be plugged directly into a PBPK software tool, where it
for distribution and cellular permeability should be the same is used to calculate all indirect PBPK model parameters
across all species for a certain drug. If this is not possible, from the models, as explained above. Notably, the informa-
a plausible physiological explanation, as, for instance, tion provided in Table 1A, together with the exhaustive col-
species-specific processes, should be discussed. If PK lections of physiological parameters integrated in PBPK
data for special populations, such as hepatically or renally software tools, is sufficient to obtain a preliminary, yet fully

Kuepfer et al.
527
Table 1A Physicochemical parameters PBPK model, the distribution, metabolization, and excretion of
Parameter Value Parameter Value CIP is structurally represented. The standard distribution of
logP 0.95 pKa (acid) 6.1 PK-Sim was identified as the most accurate distribution mod-
MW 331 g/mol pKa (base) 8.6 el, and the associated model parameters are quantified. Since
MW (effective)* 314 g/mol fu in human 0.67 the main focus is on intravenous administration during initial
Solubility 6.18 mg/ml establishment of the model, as mentioned, it is reasonable to
These parameters represent the a priori input parameters for the drug.
ignore oral absorption at this time. Note that a mean patient is
*CIP contains a fluorine atom that leads to a reduction of the effective MW. usually represented by the collection of physiological PBPK
model parameters used as such quantifying the basic model
parametrized initial model regarding absorption and distri- structure. The only information missing at this point is the CIP
bution processes. dose administered intravenously. In accordance with the
In the case of CIP, various clinical studies have been per- experimental PK data, an i.v. dose of 200 mg CIP is consid-
formed to identify physiological processes governing drug ered here. Figure 5a shows the result of a 24-hour simulation
ADME. The following assumptions were made in the model: of the i.v. CIP PBPK model, as well as the corresponding
(1) CIP is metabolized via CYP1A2,96 (2) CIP is secreted observed PK plasma data.100 In an actual pharmaceutical
into bile ducts in the liver,97 and (3) CIP is subject to signifi- development program, this single step can potentially involve
cant renal elimination.98 In this regard, it is interesting to several iterations of parameter optimization, and bring about
note that the rate at which CIP is excreted in the kidney can- the identification of structural changes in the model.101
not be explained by passive glomerular filtration alone; active In the next step, the oral administration of CIP was con-
tubular secretion must also contribute to the excretion pro- sidered. The dataset from the study by Davis et al.100 is
cess. In the case of a novel compound, the parametrization particularly well suited for this purpose because it contains
of such active physiological processes requires either IVIVE a crossover study in which CIP was administered to the
of physiological parameters in combination with dedicated same patients both i.v. and p.o. The missing model param-
scaling factors68 or identification of the model parameter eters for oral administration are relevant only to absorption
using targeted experimental mass balance and excretion in the GI tract. First, the specific drug formulation needs to
data. For the CIP example discussed here, it was assumed be considered. A Weibull function was assumed to quantify
that the contribution of each of the three above-mentioned dissolution of the tablet (Table 1C). Together with formulation
physiological processes has been determined before release, the physicochemical solubility of CIP (Table 1A)
(Table 1B). Because CIP demonstrates dose linear PK,99 all and the estimation of the intestinal permeability (Table 1C)
three processes are quantified by first-order kinetics. are sufficient to quantify absorption of a CIP tablet in the GI
Assuming a mean standard individual (i.e., an individual tract. As previously discussed, the physiological processes
with average demographic covariates) and a given administra- implemented above for the i.v. administration remain
tion scheme, the resulting PBPK model can be used for a first unchanged, such that CIP ADME can now be fully repre-
simulation of drug plasma concentrations. As outlined in the sented with the PBPK model. Figure 5b shows the result of
best-practice section above, the establishment of a PBPK a 24-hour simulation of the p.o. CIP PBPK model, as well as
model for an i.v. administration is usually a reasonable first the corresponding PK plasma data for an oral CIP dose of
step, since it makes it possible to ignore all effects related to 750 mg.100 The resulting PBPK model can now be used to
absorption in the GI tract. Here, clinical data for i.v. administra- assess different administration schemes, to simulate virtual
tion of 200 mg CIP are considered.100 For initial PBPK model populations, or to extrapolate the data to special popula-
building, the physicochemical properties of CIP (Table 1A) tions.5 As a validation of model predictions, the model was
are first integrated into the PBPK software to calculate the used to simulate the administration of 500 mg b.i.d. and
basic distribution model. In a second step, parameters for 1,000 mg q.d., and the model predictions were compared
physiological metabolization and excretion processes were with the data available in the publication by Schuck et al.18
estimated from experimental PK data (Table 1B). Whereas As can be seen from Figure 5c,d, the model is able to
renal excretion and biliary secretion are solely linked to single describe both doses as well.
organs, i.e., kidney and liver, respectively, enzyme-mediated Having established a PBPK model for i.v. and p.o. admin-
metabolization may simultaneously occur in tissues through- istration of CIP, as next step such model was applied to fur-
out the body. For the CIP model, relative CYP1A2 abundance ther PD analyses. As mentioned above, CIP is routinely
was quantified, using tissue-specific gene expression as a sur- used in clinical practice for the treatment of E. coli
rogate for protein availability3 (Figure 1d). In the resulting
Table 1C Oral absorption.
Table 1B Metabolization and excretion Process Parameter/Function
Process Parameter Drug dissolution in the GI tract (Weibull) time (50%):4 min
CYP1A2 metabolization 17 ml/min (intrinsic clearance) lag time: 0 min
Biliary secretion 1.03 ml/min/kg Shape: 0.8
Renal excretion GFR specific: 0.266 ml/min/g of organ Transcellular intestinal permeability 1E-06 cm/min
TBS: 0.57 l/min Small intestine transit time 4h
These parameters have been estimated from the plasma drug concentration. These parameters have been estimated from the plasma drug
GFR, glomerular filtration rate; TBS, tubular secretion concentration.
Kuepfer et al.
528
Figure 5 (a) PBPK simulations for 200 mg CIP (i.v.).100 (b) PBPK simulation for 750 mg CIP (p.o.).100 (c) PBPK simulation for 500 mg b.i.d.
(p.o.).18 (d) PBPK simulation for 1,000 mg q.d. (p.o.).18 (e) PD simulations with an adaptive Emax model that describes time-kill profiles of E.
coli (11775) in the context of various in vitro doses of CIP.18 (f) PBPK/PD simulations. q.d., once-a-day dosing; b.i.d., twice-a-day dosing.
infections. An adaptive in vitro Emax model for CIP treat- rate constant, C is the lung interstitial concentration, i.e.,
ment of infections with the microbial strain E. coli (11775) the site of infection, and z accounts for the initial lag phase
was previously established.18 The model allows the (Table 1D). Cr is the drug concentration inducing adaptive
description of microbial growth as well as CIP-mediated resistance (Table 1E):
inhibition such that time-kill profiles can be mechanistically

analyzed. The model hence quantifies growth in the Cr 5C0 e 2ke ðt2tlag Þ 2e 2kecr ðt2tlag Þ (4)
absence as well as in the presence of CIP treatment, and
shows an initial phase of rapid killing followed by the devel-
In this equation, C0 is the initial CIP concentration in the
opment of an adaptive resistance that slows the killing rate
in the presence of CIP. experiment, ke is the simulated elimination rate constant, kecr
is the decline of the adaptive resistance (Table 1E), tlag is
0 1 the lag time until development of adaptive resistance sets in,
dN @ k1 ð12 IC50C1C
r
1k2
5 k2
r
C A N 12e zt (3) EC50 is the concentration of CIP that gives half-maximal
dt EC50 1C response, and IC50 is the concentration at 50% of maximum
resistance. The adaptive Emax model has been validated
Here, N is the bacterial count (CFU), k is the growth rate, using in vitro data on the growth of E. coli in the presence of
k1 is the initial kill rate constant, k2 is the permanent kill different in vitro CIP concentrations18 (Figure 5e).

Kuepfer et al.
529
Table 1D Bacterial growth model 1. Peters, S.A. Physiologically-Based Pharmacokinetic (PBPK) Modeling and Simula-
k (h21) k1 (h21) k2 (h21) EC50 (mg/l) IC50 (mg/l) z (h21) tions: Principles, Methods, and Applications in the Pharmaceutical Industry (John
Wiley & Sons, Hoboken, NJ, 2012).
7.76 10.8 8.4 0.0035 0.00047 0.09 2. Proctor, N.J., Tucker, G.T. & Rostami-Hodjegan, A. Predicting drug clearance from
These parameters are provided in the literature (95). recombinantly expressed CYPs: intersystem extrapolation factors. Xenobiotica 34,
151–178 (2004).
3. Meyer, M., Schneckener, S., Ludewig, B., Kuepfer, L. & Lippert, J. Using expression
Table 1E Adaptive resistance data for quantification of active processes in physiologically based pharmacokinetic
tlag (h) kecr (h21) ke (h21) modeling. Drug Metab. Dispos. 40, 892–901 (2012).
4. Jones, H. & Rowland-Yeo, K. Basic concepts in physiologically based pharmacoki-
4 10 0.014 netic modeling in drug discovery and development. CPT Pharmacometrics Syst.
These parameters are provided in the literature (95). Pharmacol. 2, e63 (2013).
5. Maharaj, A.R. & Edginton, A.N. Physiologically based pharmacokinetic modeling and
simulation in pediatric drug development. CPT Pharmacometrics Syst. Pharmacol. 3,
The CIP PBPK model and the PD model for CIP treatment e150 (2014).
of respiratory tract infections of E. coli infections can now be 6. Jadhav, P.R. et al. A proposal for scientific framework enabling specific population
combined to generate a full-blown PBPK/PD model. In a previ- drug dosing recommendations. J. Clin. Pharmacol. 55, 1073–1078 (2015).
7. Jiang, X.L., Zhao, P., Barrett, J.S., Lesko, L.J. & Schmidt, S. Application of physio-
ous PK/PD study, an initial CIP concentration was calculated logically based pharmacokinetic modeling to predict acetaminophen metabolism and
from a simulated 24-hour area AUC vs. time curve (AUC24)/ pharmacokinetics in children. CPT Pharmacometrics Syst. Pharmacol. 2, (e80)
MIC ratio (C0’ 5 2.43 mg/l).18 Using the lung interstitial con- (2013).
8. Edginton, A.N. & Willmann, S. Physiology-based simulations of a pathological condi-
centration in the PBPK/PD as input in the adaptive Emax mod- tion: prediction of pharmacokinetics in patients with liver cirrhosis. Clin. Pharmacoki-
el, the therapeutic effect of 1,000 mg once-daily (q.d.) and net. 47, 743–752 (2008).
500 mg twice-daily (b.i.d.) of CIP was simulated for this bacte- 9. Sinha, V.K. et al. From preclinical to human -prediction of oral absorption and drug-
rial strain (Figure 5f). As reported before, the time kill curve drug interaction potential using physiologically based pharmacokinetic (PBPK)
modeling approach in an industrial setting: a workflow by using case example. Bio-
for both dosing regimens show an equally effective decrease pharm. Drug Dispos. 33, 111–121 (2012).
in CFU/ml >5 log units over the first 24 hours.18 As also 10. Schaller, S. et al. A generic integrated physiologically based whole-body model of
reported in the literature, this is related to the use of an E. coli the glucose-insulin-glucagon regulatory system. CPT Pharmacometrics Syst. Phar-
macol. 2, e65 (2013).
strain extremely sensitive to CIP, for which the 500 mg b.i.d. 11. Willmann, S., Thelen, K., Becker, C., Dressman, J.B. & Lippert, J. Mechanism-based
dosing regimen was already high enough to promote maxi- prediction of particle size-dependent dissolution and absorption: cilostazol pharmaco-
mum bacterial killing. Notably, the presented PBPK/PD model kinetics in dogs. Eur. J. Pharm. Biopharm. 76, 83–94 (2010).
could now be the starting point for further investigations involv- 12. Thiel, C. et al. A systematic evaluation of the use of physiologically based pharma-
cokinetic modeling for cross-species extrapolation. J. Pharm. Sci. 104, 191–206
ing, for example, dedicated disease models of bacterial infec- (2015).
tions in specific tissues or usage of interstitial or intracellular 13. Jones HM, Parrott N, Jorga K, Lave T. A novel strategy for physiologically based
target tissue concentrations as effective input concentrations. predictions of human pharmacokinetics. Clin. Pharmacokinet. 45, 511–542 (2006).
14. Eissing, T. et al. A computational systems biology software platform for multiscale
modeling and simulation: integrating whole-body physiology, disease biology, and
molecular reaction networks. Front. Physiol. 2, 4 (2011).
SUMMARY 15. Krauss, M. et al. Integrating cellular metabolism into a multiscale whole-body model.
PLoS Comput. Biol. 8, e1002750 (2012).
In conclusion, PBPK models can be a valuable support in 16. Schwen, L.O. et al. Spatio-temporal simulation of first pass drug perfusion in the liv-
drug development, in particular considering the possibility to er. PLoS Comput. Biol. 10, e1003499 (2014).
17. Krauss, M. et al. Using Bayesian-PBPK modeling for assessment of inter-individual
generate a priori simulations. As a consequence, they are variability and subgroup stratification. In Silico Pharmacol. 1, 6 (2013).
increasingly used in pharmaceutical companies and regulato- 18. Schuck, E.L., Dalhoff, A., Stass, H. & Derendorf, H. Pharmacokinetic/pharmacodynamic
ry agencies like the US Food and Drug Administration. Such (PK/PD) evaluation of a once-daily treatment using ciprofloxacin in an extended-release
dosage form. Infection. 33(suppl. 2), 22–28 (2005).
model development is possible, on the one hand, thanks to 19. Simeoni, M. et al. Predictive pharmacokinetic-pharmacodynamic modeling of tumor
the collection of physiological and anatomical information as growth kinetics in xenograft models after administration of anticancer agents. Cancer
well as models that relate the physicochemical properties of a Res. 64, 1094–1101 (2004).
20. Riggs, M.M., Bennetts, M., van, der Graaf P.H. & Martin, S,W. Integrated pharmaco-
molecule to its permeability and partition properties. On the
metrics and systems pharmacology model-based analyses to guide GnRH receptor
other hand, during the model building process model assump- modulator development for management of endometriosis. CPT Pharmacometrics
tions should be carefully considered and components included Syst. Pharmacol. 1, e11 (2012).
in the model should be in agreement with physiological and 21. Thiel, C. et al. Model-based contextualization of in vitro toxicity data quantitatively
predicts in vivo drug response in patients. Arch. Toxicol. (2016) [Epub ahead of
pharmacological information available for the considered com- print].
pound. Such assumptions should also be clearly kept in mind 22. Burghaus, R. et al. Computational investigation of potential dosing schedules for a
when using the model for simulation and extrapolation. switch of medication from warfarin to rivaroxaban-an oral, direct Factor Xa inhibitor.
Together with assumptions regarding the processes included Front. Physiol. 5, 417 (2014).
23. Danhof M. Systems pharmacology — Towards the modeling of network interactions.
in the model, also assumptions relative to the physiological Eur. J. Pharm. Sci.; e-pub ahead of print 2016.
and anatomical information incorporated from the database 24. Liu, X. et al. Use of a physiologically based pharmacokinetic model to study the
should be carefully evaluated, in particular when considering time to reach brain equilibrium: an experimental analysis of the role of blood-brain
barrier permeability, plasma protein binding, and brain tissue binding. J. Pharmacol.
diseased populations for which in many instances such infor- Exp. Ther. 313, 1254–1262 (2005).
mation is not available or extremely variable. 25. Hoehme, S. et al. Prediction and validation of cell alignment along microvessels as
order principle to restore tissue architecture in liver regeneration. Proc. Natl. Acad.
Sci. U. S. A. 107, 10371–10376 (2010).
Conflict of Interest. L.K., C.N., T.W., J-F.S., M.B., T.E., and D.T.
26. Niederalt, C. et al. Development of a physiologically based computational kidney
are employees of Bayer Technology Services, S.W. and J.L. are employ- model to describe the renal excretion of hydrophilic agents in rats. Front. Physiol. 3,
ees of Bayer HealthCare. 494 (2012).
Kuepfer et al.
530
27. Thomas, S.R. Kidney modeling and systems physiology. Wiley Interdiscipl. Rev. 56. Kawai, R. et al. Physiologically based pharmacokinetic study on a cyclosporin deriv-
Syst. Biol. Med. 1, 172–190 (2009). ative, SDZ IMM 125. J. Pharmacokinet. Biopharm. 22, 327–365 (1994).
28. Tawhai, M.H. & Bates, J.H. Multi-scale lung modeling. J. Appl. Physiol. 110, 57. Sugano, K., Nabuchi, Y., Machida, M. & Aso, Y. Prediction of human intestinal
1466–1472 (2011). permeability using artificial membrane permeability. Int. J. Pharm. 257, 245–251
29. Thelen, K. et al. Evolution of a detailed physiological model to simulate the gastroin- (2003).
testinal transit and absorption process in humans, part 1: oral solutions. J. Pharm. 58. Artursson, P. & Karlsson, J. Correlation between oral drug absorption in humans
Sci. 100, 5324–5345 (2011). and apparent drug permeability coefficients in human intestinal epithelial (Caco-2)
30. Thelen, K., Coboeken, K., Willmann, S., Dressman, J.B. & Lippert, J. Evolution of a cells. Biochem. Biophys. Res. Commun. 175, 880–885 (1991).
detailed physiological model to simulate the gastrointestinal transit and absorption 59. Irvine, J.D. et al. MDCK (Madin-Darby canine kidney) cells: A tool for membrane
process in humans, part II: extension to describe performance of solid dosage permeability screening. J. Pharm. Sci. 88, 28–33 (1999).
forms. J. Pharm. Sci. 101, 1267–1280 (2012). 60. Grass, G.M. & Sweetana, S.A. In vitro measurement of gastrointestinal tissue per-
31. Agoram, B., Woltosz, W.S. & Bolger, M.B. Predicting the impact of physiological meability using a new diffusion cell. Pharm. Res. 5, 372–376 (1988).
and biochemical processes on oral drug bioavailability. Adv. Drug Del. Rev. 61. Salphati, L., Childers, K., Pan, L., Tsutsui, K. & Takahashi, L. Evaluation of a
50(Suppl 1), S41–67 (2001). single-pass intestinal-perfusion method in rat for the prediction of absorption in man.
32. Levitt, D.G. Physiologically based pharmacokinetic modeling of arterial — antecubital J. Pharm. Pharmacol. 53, 1007–1013 (2001).
vein concentration difference. BMC Clin. Pharmacol. 4, 2 (2004). 62. Kumar, S. et al. Extrapolation of diclofenac clearance from in vitro microsomal
33. Musther, H. et al. Are physiologically based pharmacokinetic models reporting the metabolism data: role of acyl glucuronidation and sequential oxidative metabolism of
right Cmax? Central venous versus peripheral sampling site. AAPS J. 17, the acyl glucuronide. J. Pharmacol. Exp. Ther. 303, 969–978 (2002).
1268–1279 (2015). 63. Kusuhara, H. & Sugiyama, Y. In vitro-in vivo extrapolation of transporter-mediated
34. Thelen, K. & Dressman, J.B. Cytochrome P450-mediated metabolism in the human clearance in the liver and kidney. Drug Metab. Pharmacokinet. 24, 37–52 (2009).
gut wall. J. Pharm. Pharmacol. 61, 541–558 (2009). 64. Lippert, J. et al. A mechanistic, model-based approach to safety assessment in clini-
35. Lennernas, H. Intestinal permeability and its relevance for absorption and elimina- cal development. CPT Pharmacometrics Syst. Pharmacol. 1, e13 (2012).
tion. Xenobiotica 37, 1015–1051 (2007). 65. Wilkinson, G.R. Clearance approaches in pharmacology. Pharmacol. Rev. 39, 1–47
36. Willmann, S., Schmitt, W., Keldenich, J., Lippert, J. & Dressman, J.B. A physiologi- (1987).
cal model for the estimation of the fraction dose absorbed in humans. J. Med. 66. Rostami-Hodjegan, A. & Tucker, G.T. Simulation and prediction of in vivo drug metabo-
Chem. 47, 4022–4031 (2004). lism in human populations from in vitro data. Nat. Rev. Drug Disc. 6, 140–148 (2007).
37. Kilford, P.J., Gertz, M., Houston, J.B. & Galetin, A. Hepatocellular binding of drugs: 67. Jamei, M., Dickinson, G.L. & Rostami-Hodjegan, A. A framework for assessing inter-
correction for unbound fraction in hepatocyte incubations using microsomal binding individual variability in pharmacokinetics using virtual human populations and inte-
or drug lipophilicity data. Drug Metab. Dispos. 36, 1194–1197 (2008). grating general knowledge of physical chemistry, biology, anatomy, physiology and
38. Dressman, J.B., Amidon, G.L., Reppas, C. & Shah, V.P. Dissolution testing as a genetics: A tale of ’bottom-up’ vs ’top-down’ recognition of covariates. Drug Metab.
prognostic tool for oral drug absorption: immediate release dosage forms. Pharm. Pharmacokinet. 24, 53–75 (2009).
Res. 15, 11–22 (1998). 68. Harwood, M.D. et al. Absolute abundance and function of intestinal drug transport-
39. Cardot, J., Beyssac, E. & Alric, M. In vitro-in vivo correlation: importance of dissolu- ers: a prerequisite for fully mechanistic in vitro-in vivo extrapolation of oral drug
tion in IVIVC. Dissolut. Technol. 14, 15 (2007).
absorption. Biopharm. Drug Dispos. 34, 2–28 (2013).
40. Costa, P. & Sousa Lobo, J.M. Modeling and comparison of dissolution profiles. Eur.
69. Block, M. Physiologically based pharmacokinetic and pharmacodynamic modeling in
J. Pharm. Sci. 13, 123–133 (2001).
cancer drug development: status, potential and gaps. Expert Opin. Drug Metab. Tox-
41. Blakey, G.E., Nestorov, I.A., Arundel, P.A., Aarons, L.J. & Rowland, M. Quantitative
icol. 11. 743–756 (2015).
structure-pharmacokinetics relationships: I. Development of a whole-body physiologi-
70. Wang, W., Wang, E.Q. & Balthasar, J.P. Monoclonal antibody pharmacokinetics and
cally based model to characterize changes in pharmacokinetics across a homolo-
pharmacodynamics. Clin. Pharmacol. Ther. 84, 548–558 (2008).
gous series of barbiturates in the rat. J. Pharmacokinet. Biopharm. 25, 277–312
71. Agoram, B.M., Martin, S.W. & van der Graaf, P.H. The role of mechanism-based
(1997).
pharmacokinetic-pharmacodynamic (PK-PD) modelling in translational research of
42. Jones, H.M., Gardner, I.B. & Watson, K.J. Modelling and PBPK simulation in drug
biologics. Drug Discov. Today. 12, 1018–1024 (2007).
discovery. AAPS J. 11, 155–166 (2009).
72. Baumann, A. Early development of therapeutic biologics—pharmacokinetics. Curr.
43. Poulin, P., Schoenlein, K. & Theil, F.P. Prediction of adipose tissue: plasma partition
Drug Metab. 7, 15–21 (2006).
coefficients for structurally unrelated drugs. J. Pharm. Sci. 90, 436–447 (2001).
73. Lobo, E.D., Hansen, R.J. & Balthasar, J.P. Antibody pharmacokinetics and pharma-
44. Poulin, P. & Theil, F.P. A priori prediction of tissue:plasma partition coefficients of
codynamics. J. Pharm. Sci. 93, 2645–2668 (2004).
drugs to facilitate the use of physiologically-based pharmacokinetic models in drug
74. Baxter, L., Zhu, H., Mackensen, D., Butler, W. & Jain, R. Biodistribution of monoclo-
discovery. J. Pharm. Sci. 89, 16–35 (2000).
nal antibodies: scale-up from mouse to human using a physiologically based phar-
45. Rodgers, T., Leahy, D. & Rowland, M. Physiologically based pharmacokinetic
modeling 1: predicting the tissue distribution of moderate-to-strong bases. J. Pharm. macokinetic model. Cancer Res. 55, 4611–4622 (1995).
75. Baxter, L., Zhu, H., Mackensen, D. & Jain, R. Physiologically based pharmacokinetic
Sci. 94, 1259–1276 (2005).
46. Rodgers, T. & Rowland, M. Physiologically based pharmacokinetic modelling 2: pre- model for specific and nonspecific monoclonal antibodies and fragments in normal tis-
dicting the tissue distribution of acids, very weak bases, neutrals and zwitterions. J. sues and human tumor xenografts in nude mice. Cancer Res. 54, 1517–1528 (1994).
Pharm. Sci. 95, 1238–1257 (2006). 76. Chabot, J.R., Dettling, D.E., Jasper, P.J. & Gomes, BC. Comprehensive
47. Berezhkovskiy, L.M. Volume of distribution at steady state for a linear pharmacoki- mechanism-based antibody pharmacokinetic modeling. Conference Proceedings:
netic system with peripheral elimination. J. Pharm. Sci. 93, 1628–1640 (2004). Annual International Conference of the IEEE Engineering in Medicine and Biology
48. Willmann, S., Lippert, J. & Schmitt, W. From physicochemistry to absorption and dis- Society IEEE Engineering in Medicine and Biology Society Annual Conference,
tribution: predictive mechanistic modelling and computational tools. Expert Opin. 4318–4323 (2011).
Drug Metabol. Toxicol. 1, 159–168 (2005). 77. Chen, Y. & Balthasar, J.P. Evaluation of a catenary PBPK model for predicting the
49. Willmann, S. et al. PK-SimV R : a physiologically based pharmacokinetic ‘‘whole-body’’ in vivo disposition of mAbs engineered for high-affinity binding to FcRn. AAPS J.
model. Biosilico. 1, 121–124 (2003). 14, 850–859 (2012).
50. Loidl-Stahlhofen, A. et al. Multilamellar liposomes and solid-supported lipid mem- 78. Davda, J., Jain, M., Batra, S., Gwilt, P. & Robinson, D. A physiologically based
branes (TRANSIL): screening of lipid-water partitioning toward a high-throughput pharmacokinetic (PBPK) model to characterize and predict the disposition of mono-
scale. Pharm. Res. 18, 1782–1788 (2001). clonal antibody CC49 and its single chain Fv constructs. Int. Immunopharmacol. 8,
51. Leahy, D., Lynch, J. & Taylor, D. Mechanisms of Absorption of Small Molecules. 401–413 (2008).
Novel Drug Delivery and Its Therapeutic Application (Wiley, New York, 1989). 79. Ferl, G., Kenanova, V., Wu, A. & DiStefano, J. III. A two-tiered physiologically based
52. Schmitt, W. General approach for the calculation of tissue to plasma partition coeffi- model for dually labeled single-chain Fv-Fc antibody fragments. Mol. Cancer Ther.
cients. Toxicol. In Vitro 22, 457–467 (2008). 5, 1550 (2006).
53. Levitt, D.G. PKQuest: capillary permeability limitation and plasma protein binding — 80. Ferl, G., Wu, A. & DiStefano, J. A predictive model of therapeutic monoclonal anti-
application to human inulin, dicloxacillin and ceftriaxone pharmacokinetics. BMC body dynamics and regulation by the neonatal Fc receptor (FcRn). Ann. Biomed.
Clin. Pharmacol. 2, 7 (2002). Eng. 33, 1640–1652 (2005).
54. Levin, V.A., Dolginow, D., Landahl, H.D., Yorke, C. & Csejtey, J. Relationship of 81. Fronton, L., Pilari, S. & Huisinga, W. Monoclonal antibody disposition: a simplified
octanol/water partition coefficient and molecular weight to cellular permeability and PBPK model and its implications for the derivation and interpretation of classical
partitioning in s49 lymphoma cells. Pharm. Res. 1, 259–266 (1984). compartment models. J. Pharmacokinet. Pharmacodyn. 41, 87–107 (2014).
55. Lieb, W.R. & Stein, W.D. Biological membranes behave as non-porous polymeric 82. Garg, A. & Balthasar, J. Physiologically-based pharmacokinetic (PBPK) model to
sheets with respect to the diffusion of non-electrolytes. Nature. 224, 240–243 predict IgG tissue kinetics in wild-type and FcRn-knockout mice. J. Pharmacokinet.
(1969). Pharmacodyn. 34, 687–709 (2007).

Kuepfer et al.
531
83. Sepp, A., Berges, A., Sanderson, A. & Meno-Tetang, G. Development of a physio- 94. Zhao, P., Rowland, M. & Huang, SM. Best practice in the use of physiologically
logically based pharmacokinetic model for a domain antibody in mice using the two- based pharmacokinetic modeling and simulation to address clinical pharmacology
pore theory. J. Pharmacokinet. Pharmacodyn. 42, 97–109 (2015). regulatory questions. Clin. Pharmacol. Ther. 92, 17–20 (2012).
84. Shah, D.K. & Betts, A.M. Towards a platform PBPK model to characterize the plas- 95. Leong, R. et al. Regulatory experience with physiologically based pharmacoki-
ma and tissue disposition of monoclonal antibodies in preclinical species and netic modeling for pediatric drug trials. Clin. Pharmacol. Ther. 91, 926–931
human. J. Pharmacokinet. Pharmacodyn. 39, 67–86 (2012). (2012).
85. Rippe, B. & Haraldsson, B. Fluid and protein fluxes across small and large pores in 96. McLellan, R.A., Drobitch, R.K., Monshouwer, M. & Renton, KW. Fluoroquinolone
the microvasculature. Application of two-pore equations. Acta Physiol. Scand. 131, antibiotics inhibit cytochrome P450-mediated microsomal drug metabolism in rat and
411–428 (1987). human. Drug Metab. Dispos. 24, 1134–1138 (1996).
86. Rippe, B. & Haraldsson, B. Transport of macromolecules across microvascular walls: 97. Ball, C.S., Manson, J.M., Reid, F. & Tweedle, D.E. The pharmacokinetics of
the two-pore theory. Physiol. Rev. 74, 163–219 (1994). the biliary excretion of ciprofloxacin. HPB Surg. 1, 319–326 discussion 26-27
87. Chen, X., Hickling, T.P. & Vicini, P. A mechanistic, multiscale mathematical model (1989).
of immunogenicity for therapeutic proteins: part 1-theoretical model. CPT Pharmaco- 98. Drusano, G.L. An overview of the pharmacology of intravenously administered cipro-
metrics Syst. Pharmacol. 3, e133 (2014). floxacin. Am. J. Med. 82, 339–345 (1987).
88. Chen, X., Hickling, T.P. & Vicini, P. A mechanistic, multiscale mathematical model 99. Bergan, T., Thorsteinsson, S.B., Kolstad, I.M. & Johnsen, S. Pharmacokinetics of ciprofloxa-
of immunogenicity for therapeutic proteins. Part 2. model applications. CPT Pharma- cin after intravenous and increasing oral doses. Eur. J. Clin. Microbiol. 5, 187–192 (1986).
cometrics Syst. Pharmacol. 3, e134 (2014). 100. Davis, R.L. et al. Pharmacokinetics of ciprofloxacin in cystic fibrosis. Antimicrob.
89. European Medicines Agency Committee for Medicinal Products for Human Use: Agents Chemother. 31, 915–919 (1987).
Concept Paper on Qualification and reporting of physiologically-based pharmacoki- 101. Kuepfer, L., Peter, M., Sauer, U. & Stelling, J. Ensemble modeling for analysis of
netic (PBPK) modeling and analyses. EMA/CHMP/211243/2014 (2014). cell signaling dynamics. Nat. Biotechnol. 25, 1001–1006 (2007).
90. Jones, H.M. et al. Physiologically based pharmacokinetic modeling in drug discovery
and development: a pharmaceutical industry perspective. Clin. Pharmacol. Ther. 97,
C 2016 The Authors CPT: Pharmacometrics & Systems
V
247–262 (2015).
91. Sager, J.E., Yu, J., Raguenau-Majlessi, I. & Isoherranen, N. Physiologically based Pharmacology published by Wiley Periodicals, Inc. on
pharmacokinetic (PBPK) modeling and simulation approaches: a systematic review behalf of American Society for Clinical Pharmacology and
of published models, applications and model verification. Drug Metab. Dispos.; e-
pub ahead of print 2015.
Therapeutics. This is an open access article under the
92. Shepard, T., Scott, G., Cole, S., Nordmark, A., & Bouzom, F. Physiologically based mod- terms of the Creative Commons Attribution-NonCommer-
els in regulatory submissions: output from the ABPI/MHRA forum on physiologically based cial-NoDerivs License, which permits use and distribution
modeling and simulation. CPT Pharmacometrics Syst. Pharmacol. 4, 221–225 (2015). in any medium, provided the original work is properly cit-
93. Wagner, C. et al. Application of physiologically based pharmacokinetic (PBPK)
modeling to support dose selection: report of an FDA Public Workshop on PBPK. ed, the use is non-commercial and no modifications or
CPT Pharmacometrics Syst. Pharmacol. 4, 226–230 (2015). adaptations are made.
Supplementary information accompanies this paper on the CPT: Pharmacometrics & Systems Pharmacology website
(http://www.wileyonlinelibrary.com/psp4)

Applied Concepts in PBPK Modeling

Uploaded by

Copyright:

Available Formats

Applied Concepts in PBPK Modeling

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Applied Concepts in PBPK Modeling

Uploaded by

Copyright:

Available Formats

Citation: CPT Pharmacometrics Syst. Pharmacol. (2016) 5, 516–531; doi:10.1002/psp4.

Applied Concepts in PBPK Modeling: How to Build a

CPT: Pharmacometrics & Systems Pharmacology

Figure 3 Schematic representation of the most common applications of PBPK modeling.

CPT: Pharmacometrics & Systems Pharmacology

CPT: Pharmacometrics & Systems Pharmacology

CPT: Pharmacometrics & Systems Pharmacology

CPT: Pharmacometrics & Systems Pharmacology

Biliary secretion 1.03 ml/min/kg Shape: 0.8

TBS: 0.57 l/min Small intestine transit time 4h

CPT: Pharmacometrics & Systems Pharmacology

CPT: Pharmacometrics & Systems Pharmacology

You might also like