Abstracts

Rationale: P-Glycoprotein (PGP) is probably an important component of the multidrug resistance mechanism in epilepsy. It is a very large membrane protein making definitive structure determination difficult. Some biochemical data are available on this protein. Our objective in this study was two fold : 1) To computationally model the structure of PGP using available data and use molecular dynamics simulations to sample at least some of the possible conformations and 2) To clinically assess the utility of verapamil induced inhibition of PGP. Methods: For the first part of the project, using the 2006 Vandeuver et. al. model as a starting structure and all the available biochemical data including the chemical crosslinking as well as the thiol protection data, we modified the structure using the moleculr modeling software O. We subjected this to molecular dynamics simulations using CHARMM to obtain a family of structures which were analyzed with respect to agreement to available biochemical data and other available structures. For the second part of the project, verapamil was added to the existing anti-epileptic medication regimen of patients with medically refractory epilepsy (more than 3 seizures a month despite adequate doses of at least two anti-epileptic medications) for 3 months. This was done with the University of Massachusetts IRB approval. The efficacy of this on seizure control was then analyzed. Results: The family of structures obtained by our molecular dynamics simulations were very similar. The energy variations did not show a significant spread and the binding pocket structure overlapped to a large extent. The models show a bundle of parallel helices with a pseudo-two fold symmetry parallel to the helices. The NBDs are in close interaction which is in agreement with the EM structure. The lateral openings between helices 5 and 8 and between 2 and 11 noted in the initial structure can also be identified. We are in the process of analyzing the structure in more detail and studying the binding of ligands. For the second part of the project, we added verapamil 40mg three times a day to the AED regimen of 4 patients with refractory epilepsy. All four patients needed to come off the verapamil within one month due to increased fatigue. There was no decrease in the seizure frequency. We then closed the study to further recruitment. Conclusions: From the molecular dynamics simulations, we have an idea of the allowed conformations for PGP as well as the orientation of the transmembrane and nucleotide binding domains. We are analyzing the structure as well as the binding of ligands. From the verapamil study, despite the small sample size, it seems like verapamil induced inhibition of PGP is not effective in seizure control. It is unclear if the increased fatigue is a result of the verapamil induced inhibition of AED-PGP interaction. We are planning to study lower doses of verapamil. More specific inhibition of PGP is an option that probably should be explored.