Authors :
Lyndsey Anderson, PhD – Praxis Precision Medicines, Boston, MA, USA
Presenting Author: Kris Kahlig, PhD – Praxis Precision Medicines, Boston, MA, USA
Marcio Souza, PharmD, MBA – Praxis Precision Medicines, Boston, MA, USA
Steven Petrou, PhD – Praxis Precision Medicines, Boston, MA, USA
Rationale:
Approximately 3.5 million people in the United States are diagnosed with epilepsy, almost a third of whom are refractory to conventional anti-seizure medications (ASMs). Central to the development of novel treatments is testing of anticonvulsant activity in preclinical seizure models. While various models exist, the predictive validity of each across the spectrum of epilepsy indications is less clear. The Praxis Analysis of Concordance (PAC) framework was recently developed to assess the translational concordance of common preclinical seizure models, demonstrating three acute seizure models with highest predictive validity for focal onset seizures: audiogenic, maximal electroshock (MES) and 6-Hz (32mA). Here, we sought to establish concordance between commonly used preclinical models and generalized epilepsies, thus extending our work to capture the spectrum of human epilepsies.
Methods:
Performance of 32 approved ASMs across established preclinical seizure models was evaluated based on reported TD50 and ED50 values. Protective index values were calculated and a weighted scale representing relative antiseizure efficacy was used to grade preclinical ASM response for each ASM in each seizure model. ASM clinical use and perceived efficacy were similarly evaluated, and a weighted scale used to grade clinical ASM response for each ASM in each seizure indication. Predictive validity of preclinical models was assessed using the Praxis Analysis of Concordance (PAC) scoring matrix. Concordance scores were assigned spanning the spectrum from complete discordance (-1) to complete concordance (1) between preclinical and clinical ASM responses. Scores were then summed and normalized to generate a global translational concordance score.
Results:
Genetic rat models of absence seizures (GAERS and WAG/Rij) exhibited high concordance with absence and myoclonic generalized seizure types. MES, audiogenic, and amygdala kindling models showed high concordance with primarily generalized tonic-clonic seizures, with findings based on robust data depth (assessed based on the number of ASMs tested in each model). Models with the highest concordance for atonic or tonic seizures were intra-amygdala kainate and hippocampal kindling; albeit based on limited data depth. MES, audiogenic, mouse 6 Hz (32 and 44mA) and kindling models demonstrated high concordance with status epilepticus.
Conclusions: Using the newly developed PAC framework, findings from this study extend our insights into the predictive validity of commonly used seizure models across the spectrum of human epilepsies. Notably, we demonstrate differential translational concordance between preclinical seizure models and generalized epilepsy types, as well as variability in data depth across utilized models. We anticipate these findings to have important implications for supporting ongoing research efforts as well as promoting efficient resourcing for novel ASM development for generalized epilepsies.
Funding: Praxis Precision Medicines.