Physiologically-based pharmacokinetic (PBPK) modelling is an established in silico
platform for simulating pharmacokinetics in humans and pre-clinical species. These
models are widely used within drug metabolism and pharmacokinetic (DMPK)
research to facilitate the design of drug molecules with human pharmacokinetic (PK)
profiles that lead to effective and safe therapeutic treatments. This research focuses
on the mechanistic approaches within PBPK models to parameterise drug
distribution, and their influence on plasma concentration-time (Cp) profiles. As such,
these in silico methodologies utilise properties of the drug and physiologies of the
organism to predict tissue-to-plasma partition coefficients (Kp); which, when
combined with regional blood flows allows the movement of drug within the body to
be modelled. However, Kp values can also be generated in vivo, where it is generally
accepted that these measured values (Kp(meas)) lead to improved pharmacokinetic
simulations with respect to both the accuracy of the Cp profile and mechanistic
estimates of the volume of distribution. Twenty-two novel compounds were
identified for this research covering a diverse range of GSK chemistries; including,
acidic, basic, neutral and zwitterionic compounds. Overall, the standard GSK PBPK
modelling approach, incorporating in silico Kp values derived from the preferred
Lukacova method, demonstrated moderate to high levels of PK simulation accuracy
in the rat for the majority of compounds across these classes (> 90 % for basic, neutral
and zwitterionic molecules); however, refinements to these methodologies were
proposed leading to greater model performance. Conversely, the Lukacova PBPK
methodology showed varied levels of simulation accuracy (poor to high) in the rat for
the acidic compound class which was attributed to the influence of transporter
mediated drug disposition (TMDD), and the apparent lack of sensitivity to reduced
acidic strength. Incorporating rat Kp(meas) values showed varying levels of
improvements to model performance, which was dependent upon compound class,
and the accuracy of the initial Kp predictions. These observations provided the
foundation for an optimised modelling strategy that could be applied to specific
compounds classes, affording GSK and the wider modelling community with a robust
set of go-to PBPK methodologies that can be applied early in drug discovery to predict
systemic pharmacokinetics in human and support clinical dose projections.
| Date of Award | 7 Mar 2024 |
|---|
| Original language | English |
|---|
| Awarding Institution | - University Of Strathclyde
|
|---|
| Sponsors | University of Strathclyde |
|---|
| Supervisor | Geoffrey McKay (Supervisor), Marie Boyd (Supervisor) & Nigel Mottram (Supervisor) |
|---|