Catherine Ollagnier 1, Thomas Lewin 1, Achi Haider 1, Caterina Bissantz 1, Davide Bassani 1, Serge Summermatter 2, Nicolas Frances 1
1 Roche Pharma Research and Early Development, Pharmaceutical science, Roche Innovation Center Basel (, Switzerland), 2 Roche Pharma Research and Early Development, CMI2O DTA, Roche Innovation Center Basel (, Switzerland)
Introduction/Objectives
The therapeutic landscape for metabolic disorders has been transformed by the increasing
number of incretins, which demonstrate varying degrees of efficacy regarding body weight (BW)
loss. While extensive in vitro potency data and numerous preclinical studies exist, a definitive
quantitative link between in vitro receptor activation and in vivo weight loss across different
chemical modalities (small molecules vs. peptides) and species remains elusive. This lack of a
standardized translational framework complicates the early-stage selection of candidates with
the highest clinical potential. The primary objective of this study was to develop a robust
translational modeling platform capable of connecting in vitro potency metrics to longitudinal
body weight loss in obese animal models. By integrating pharmacokinetics (PK) and
pharmacodynamics (PD), we aimed to characterize the relationship between receptor activation
and phenotypic outcomes for GLP-1 receptor agonists.
Methods
As of now, two GLP-1 receptor agonists were selected for model development: exenatide and
semaglutide. In vitro potency was assessed using in-house data and literature-derived values
[1]. Preclinical PK and BW loss data from obese animals (Diet-Induced Obese [DIO] mice) were
integrated from internal studies and published literature [2].
Pharmacokinetic models were tailored to each compound: a 2-compartment model with first-
order absorption/elimination for semaglutide, using literature population PK parameters as
posteriori estimates [3] to fit sparse in-house data and a 1-compartment model for exenatide.
The PD was characterized using a turnover inhibition model, where the drug inhibits the rate of
weight gain k in . Three dose levels were utilized to evaluate the in vivo IC 50 ,of exenatide, while for
semaglutide only one dose was available. The model predicted percentage weight loss over
time, which allows us to compare results across species. In vitro EC 50 values were determined in
cell lines expressing the GLP-1 receptor, measuring cAMP stimulation at low receptor densities.
Finally, we evaluated the relationship between in vivo and in vitro potency (EC 50 in nM) while
accounting for the free fraction (F up ) of each compound.
Results
The developed PKPD models successfully described the longitudinal weight loss profiles for
both compounds. The estimated in vivo IC 50 values were 32.3 nM for semaglutide and 0.4 nM
for exenatide. These were compared against literature-reported in vitro EC 50 values in GLP-1R
cAMP assays (semaglutide: 0.813 nM; exenatide: 0.464 nM).
The observed variations between in vivo and in vitro potency were partially reconciled by
incorporating the plasma protein binding differences. Indeed, semaglutide demonstrated low
free fractions (F up = 0.28%), in contrast to exenatide, which was entirely unbound (F up = 100%).
When corrected for F up , the translational potency gap narrowed significantly, with the ratio R
in_vivo_IC 50 / in_vitro_EC 50 remaining between 0.11 and 0.86.
Conclusions
This study establishes a quantitative correlation between in vitro GLP-1 receptor activation and
in vivo body weight reduction in preclinical species. By accounting for compound-specific PK
properties and protein binding, the platform allows for the prospective evaluation of new GLP-1
agonists using early in vitro data. The consistency of the model across different compounds
highlights its utility in drug discovery. While these results are encouraging for preclinical
development, further research is required to validate the predictive validity of this model for
clinical responses in human subjects and to account for the potential differences in receptor-
effector coupling between species.
References:
[1] Sloop et al., Sci Transl Med (2024), DOI: 10.1126/scitranslmed.adp5765.
[2] Iwasaki et al., J Pharmacol Exp Ther (2017), DOI: 10.1124/jpet.117.242651.
[3] Kim et al., J Pharm Investig (2024), DOI: 10.1007/s40005-024-00717-3.
Reference: PAGE 34 (2026) Abstr 11862 [www.page-meeting.org/?abstract=11862]
Poster: Drug/Disease Modelling - Other Topics