Epilepsy Benchmark IC3

2005 Report submitted by Benchmark Steward(s):
Edward Bertram, M.D. (University of Virginia School of Medicine)

Background of the benchmark goal: 
The development of new therapies is greatly facilitated when molecular targets or processes are understood.  Current anti-epileptic drugs (AEDs) target a variety of ion channels and receptors, with varying success in the clinic.  Industry and academia generate tens of thousands of new compounds a year, and it is likely that among those many compounds there are several that have the potential for a significant impact on epilepsy.  However, it is unlikely that many of these compounds will ever be fully evaluated for a potential for epilepsy treatment, in part because there is no easy way to screen thousands of compounds for epilepsy therapy.  Current acute seizure models or proposed chronic seizure models that have high predictive value are too labor intensive and slow to serve as initial screens for all new compounds.   A true high throughput biochemical or biological identification process that can screen thousands of compounds in a matter of days is needed to identify candidate compounds that should progress to preclinical testing.  Reliable, predictive biomarkers must be found in order to develop a high throughput screening process for epilepsy therapeutics.

Epileptogenesis and epilepsy are actually distinct parts of a dynamic process, which may vary significantly depending on the type of epilepsy.  Changes in the expression of ion channels and receptors, synaptic rearrangements, inflammation and changes in glia are some of the hypothesized contributors to epilepsy and epileptogenesis. Some changes are regionally quite discretely localized.   However, to date few if any of these changes have been demonstrated as critical to the process or condition.   Targets for the prevention of epileptogenesis likely change in the period that precedes the onset of the first seizure as the process proceeds, so that appropriate targets at one point may not be relevant at another.  The development of appropriate markers is also complicated by the multiple types of epilepsy and an at best rudimentary  knowledge of the processes that lead to chronic epilepsy and the initiation of spontaneous seizures.  This deficit should not be viewed as a roadblock to the identification of a appropriate markers, but must be taken into consideration. 

The need for models with good parallels to the human condition has been underlined by the growing number of observations of epilepsy specific changes and physiology that may require therapy that is developed specifically to meet the needs of epilepsy.  Although some potential markers may be identified in human based studies, the importance of the marker can only be confirmed with studies in animal models.  It is likely that  some markers may serve only to denote risk for future epilepsy.  Although such markers could not be used in high throughput screens as a target for therapy identification, they could be useful for the timely identification of key points in epileptogenesis that would be amenable to intervention.

Identification of critical and realistic time points for intervention is essential in the prevention of epilepsy.  When are risk factors for the eventual development of epilepsy sufficiently clear and of high enough predictive value to warrant treatment with potentially deleterious drugs?  When are the risk factors likely to be detected clinically?  At the time of routine genetic screening?  Following a defined initial precipitating injury?  From defined patterns on a neuroimaging study?  It will not help ultimately to identify mechanisms and treatments at times that would or could never be used clinically.

The development of this benchmark is clearly tied to the development of other benchmarks, especially those in the 1C group, but it will also draw on genetics, imaging and epidemiology. 

Current Status of Field:
Preclinical screening for potential antiepileptic compounds remains based on primarily acute in vivo models.  Compounds screened may have a known mechanism of action, but often do not.  There is now a program in place to develop chronic models but the results are not in, and the role for these models in a large scale screening program, should any of the models prove viable, is unknown.  Although the models may turn out to be highly predictive, they are too labor intensive to be of value in an initial screen.

There are a number of high throughput screens which are capable of determining whether compounds have a specific binding or functional profile.  These profiles are usually obtained for generic binding affinities, agonist or antagonist functions.  The subtype specific effects that will likely be necessary can be used, but are usually performed on a smaller scale, and often not in commercial laboratories at the moment.  

The development of models for epilepsy and epileptogenesis from which biomarkers can be derived is well established for the post status epilepticus/limbic epilepsy model and some genetic forms of generalized epilepsy.  There are many forms of common epilepsy that have no good model.  In addition there is no established method for objectively comparing the claims for new models with the actual human condition.  Lack of this process may lead to misleading claims about the potential of a model to contribute to our knowledge about a specific syndrome. Realistic models are essential to our developing appropriate in vitro biomarkers.

Activities update: 
There continues to be steady progress made in identifying a number of changes in receptors and channels associated with epilepsy, especially limbic epilepsy.  Epilepsy specific regional expression of particular channel or receptor subtypes is regularly described, as is the unique pharmacology and physiology of these proteins.  The findings have emphasized the importance of using epilepsy specific changes in the discovery of new therapies.  To date these observations have not been translated into clinical screens, but once the role/importance of them is established, they may become part of such a screening process. 

The recent models workshops have established a process for selecting promising models for drug resistance.  Although the emphasis has been on in vivo models, once the models are established and accepted, they will be able to serve as a source for biomarkers for therapy resistance.  With regard to markers for epileptogenesis, the best markers to date have been in the post status epilepticus models, in which the early appearance of neuronal loss has a high predictive value for later epilepsy. There is significant activity using gene array technology examining chronic and acute changes, and it is expected that in the next several years this pathway will provide critical new insights about potential biomarkers that could be developed into screens.

Top priorities for next 5-10 years:

• Identify appropriate realistic animal models of epileptogenesis and epilepsy that are clinically relevant.
• Determine key changes in these models that likely underlie epileptogenesis or the generation of spontaneous seizures. 
• Using a mechanism based approach, use compounds of known and appropriate mechanisms of action that have access to the central nervous system to identify mechanisms that may be effective. Guide the selection of compounds by the evolving knowledge of the contributory changes and processes associated with epilepsy and epileptogenesis.
• Take successful mechanisms from the in vivo models and develop a profile for compounds that are likely to be clinically successful.  Use the in vitro tests to identify and refine appropriate compounds for eventual clinical trials.

Roadblocks to progress:

• There are a limited number of animal models with strong parallels to the various human epilepsies.  It is unclear whether the current limited number of models will be sufficient to provide an appropriate number of markers for screening for the spectrum of human epilepsy.
• We are missing a complete catalog of the changes associated with any one disorder, let alone all of the disorders.  Further, the true role of the known changes is frequently not understood.   For choosing among the known changes as potential targets for therapy, we must have a better sense of which are most likely to provide the most therapeutically useful targets.
• Once key or potentially important targets are identified, they have to be translated into an in vitro screen.  Although that process is becoming more routine, it may be difficult to encourage those commercial enterprises that are involved in large scales screening to create an epilepsy specific profile.
• Will industry use the results of new epilepsy specific screens to develop compounds for clinical trial?  There will likely need to be an initial successful proof of concept for this approach to therapy discovery to be more widely accepted.
• The relative importance of the known and to be discovered epilepsy associated changes is unknown.  A process of ranking the likely contributions of the changes to the process of epileptogenesis or epilepsy must be established to be able to differentiate etiological from secondary (but non-contributing) alterations.
• The translational program may be an appropriate route for the development of screening tools/assays.