FT Musuamba (1), VL di Iorio (1), M Danhof (1), O Della Pasqua (1,2)
(1) Leiden Academic Center for Drug Research, Division of Pharmacology, Leiden, The Netherlands; (2) Clinical Pharmacology Modelling & Simulation, GlaxoSmithKline, United Kingdom
Background: Neuropathic pain arises as direct consequence of a lesion or disease affecting the somatosensory system. Despite the wide range of experimental models available for drug screening, neuropathic pain remains an unmet medical need, with available treatments showing limited or no efficacy in patients. Whilst many questions regarding the construct validity of animal models are still unanswered, the use of pharmacokinetic-pharmacodynamic modelling may facilitate the translation and extrapolation of preclinical findings.
Objectives: The aim of this study was to develop a semi-mechanistic model to characterise the time course of allodynia and hyperalgesia in two models neuropathic pain in rats. A secondary objective was to identify model parameters that best describe drug effects, providing a more robust basis for the ranking of candidate molecules during the screening of novel compounds.
Methods: Allodynia and hyperalgesia data from experiments based on chronic construction injury and spinal nerve ligation were used for the purpose of our investigation. Model building was performed taking into account the use of drug- and system-specific parameters. The effect of different compounds (duloxetine, gabapentin, pregabalin, and venlafaxine) was then evaluated in terms of potency and maximal pain inhibition. Parameter estimates were subsequently ranked and compared to the exposure levels observed at currently approved clinical doses. Analysis was performed in NONMEM v7.2. R was used for data manipulation, statistical and graphical summaries.
Results: A semi-mechanistic model enabled characterisation of the different phases of allodynia. This model was sensitive to drug effects on different disease parameters. Goodness of fit and validation procedures show appropriate model performance, despite high variability in the data. In addition, clear differences in potency were observed for the different compounds, which seem to reflect pharmacological activity. The drug ranking based on the relative potencies was comparable for both models, but discrepancies exist between clinical exposure and drug levels observed in these models.
Conclusion: Current protocol sampling and dosing schemes represent a major limitation for accurate characterisation of drug effects in preclinical experiments. A meta-analytical approach is essential to ensure that system-specific parameters are estimated independently from treatment effect. Differences in construct validity of these models are compounded by poor precision and inaccuracy in parameter estimation.
Reference: PAGE 22 (2013) Abstr 2918 [www.page-meeting.org/?abstract=2918]
Poster: Other Drug/Disease Modelling