28 BioPharm International eBook March 2021 www.biopharminternational.com
"well-stirred" model (6), with clear-
ance attributed to the liver and
kidney. In advanced PBPK models,
gut absorption is sub-compartmen-
talized into different regions of the
gastrointestinal tract. Patients tak-
ing drugs with food can impact
drug absorption and thus PK prop-
erties of the drugs, which can also
be simulated (7). PBPK modeling
may require a variable degree of
pa ra meter izat ion using in-vit ro
and in-vivo data depending on the
question the model is intended
to address. Essential drug-specific
pa ra meters required for model
execution will vary with the drug
absorption, distribution, metab-
ol i sm, a nd e xc r e t ion (A DM E)
properties and the formulation in
which it is administered. Typically,
sensitivity analysis is employed to
assess the model uncertainty and
identify and investigate parame-
ters with the greatest impact on
the simulation out put (3), and
clinical data are used to refine
the mechanistic PBPK models fur-
ther. In practice, a combination
of approaches (ref inement and
simulation) occurs (8) resulting in
a top-down, bottom-up, or mid-
dle-out approach to model build-
ing. Once these models are built
a nd cont i nuously ref i ned a nd
verified, they can have a signifi-
cant impact in addressing various
challenges in pharmaceutical drug
development as outlined below.
IMPACT OF PBPK MODELING
ON ASSESSING
DRUG-DRUG INTERACTIONS
As part of the clinical development
process, the potential for pharma-
cokinetic drug-drug interactions
(DDIs) is studied and used to deter-
mine optimal doses and prescribing
information to achieve the safest
and most effective therapy possible.
Pharmacokinetic DDIs are those
that impact the processes of absorp-
tion, distribution, or elimination of
other medications. It is also import-
ant to consider that a DDI arises
when one drug, called the 'perpe-
trator', modulates the underlying
physiological or biochemical mech-
anism of either the PK or pharma-
codynamics (PD) of a 'victim' drug.
One method to study DDIs is apply-
ing mechanistic PBPK approaches
where 'virtual humans' are dosed
to accurately predict the clinical
outcome of potential pharmaco-
kinetic DDIs with co-medications.
Such PBPK models mechanistically
integrate physiology, pharmacol-
ogy, and biochemistry to predict
how the administration of one drug
may affect the absorption, distribu-
tion, and elimination processes of
another drug.
PBPK modeling and simulation
has also been recommended by
regulatory authorities to inform
clinical DDI study design and esti-
mate the magnitude of DDIs. In a
review by IQ Consor tium (IQ) member
companies (8), it was emphasized
that inclusion of PBPK modeling
and simulation has increased in
both regulatory submissions and
approved dr ug labels w ith the
majority involving DDI predictions.
In addition, a number of approved
drug labels include guidance to
physicians on specific DDIs based
on predictive modeling and simu-
lation using verified and validated
PBPK models (4). Notably, the sub-
stantial impact of IQ is evident by
peer-reviewed publications empha-
sizing the utilization of PBPK mod-
eling across the industry (8) and
those focusing on DDI modeling
(9–10) being cited in recent regulatory
DDI guidance (11).
P B P K mo d e l i n g i s a u s e f u l
tool to quantitatively rationalize
obser ved dr ug interactions and
understand the underlying mech-
anisms of DDI. It is highly valued
for the elucidation of complex DDI
mechanisms such as the contri-
bution of perpetrator metabolites
a nd t ra nspor ter- en z y me i nter-
play (12). In this IQ-sponsored manu-
script, three case examples of DDIs
involving inhibitory metabolites
using PBPK modeling were investi-
gated, and guidance was provided
for its strategic application. PBPK
modeling was used to mechanisti-
cally understand the quantitative
contribution of the metabolites to
the DDIs. The examples presented
emphasized how perpetrator prop-
erties of a parent drug alone can
sometimes fail to predict the actual
clinical DDI without incorpora-
tion of metabolite formation and
distribution k inetics as well as
their perpetrator properties. What
is important is the ability to cap-
ture the time- and tissue-depen-
dent formation of these inhibitory
metabolites to more accurately pre-
dict clinical DDIs in PBPK models.
For i nsta nce, i n one of t he
case examples, the prediction of
amiodarone clinical DDIs w ith
enzyme substrates could not be
explained by parent drug in-vitro
DDI parameters alone. Inclusion of
formation of the abundantly circu-
lating metabolite (mono-N-deseth-
ylamiodarone) with its perpetrator
properties was necessary to simulate
Regulatory Sourcebook Quality Collaboration
PBPK modeling
is a useful tool
to quantitatively
rationalize observed
drug interactions
and understand
the underlying
mechanisms of DDIs.