Anticonvulsant Screening Program Report - May 29, 2015

A report from the NINDS Anticonvulsant Screening Program Working Group, of the National Advisory Neurological Disorders and Stroke (NANDS) Council

May 29, 2015



Findings of the 2015 Working Group


Working Group Charge and Roster


Program History and Description

The NINDS Anticonvulsant Screening Program (ASP) was established in 1975, as part of a larger Antiepileptic Drug Development (ADD) program that included basic and clinical research, and extra- and intramural components. The ASP arm was added to address a critical need for new anti-seizure medications and to promote industry interest in their development, and it is the only remaining component of the original ADD program today. Since its inception, the program has been supported through a contract to the University of Utah, which NINDS initially awarded after open competitions in 1974, 1980, and 1984, and renewed thereafter on a sole-source basis (principal investigator: H. Steve White, PhD; average annual funding for the contract, 2011-2014: $3.2M). An internal NINDS ASP program office with five staff members, led since November 2012 by John Kehne, PhD, oversees the contract and the overall program (annual cost approximately $1M).

The anticonvulsant screening process involves researchers from academia and industry in the U.S. and abroad (“ASP participants”) submitting compounds to the ASP for tests of anticonvulsant efficacy in a series of animal (rodent) seizure models. These tests are performed on a blinded and confidential basis through a written agreement between the NINDS and the ASP participant at no cost to the ASP participant. The NINDS ASP program office reports test results to participants and provides advice on future development steps for promising compounds, while protecting confidentiality and intellectual property. Approximately 30,000 compounds have been screened to date, and the ASP has contributed to bringing 10 antiseizure drugs to market since 1990, with a major role in some and a minor role in others. NINDS also maintains a database of chemical structures and test results that is designed to provide insights on structure-activity relationships. A growing subset of these data is made publicly available through an online database called PANAChE, for Public Access to Neuroactive & Anticonvulsant Chemical Evaluations.

ASP Working Group and Review

The ASP was reviewed by different external groups in 1982, 1987, 1992, 1996, and most recently in 2011, by a prior working group of the National Advisory Neurological Disorders and Stroke (NANDS) Council, which delivered a final report in February 2012 (see summary below).

The current NANDS Council Working Group (charge and roster attached) was formed in July 2014, and was composed of a subset of the 2011 working group along with four new members. This group was asked to assess the program’s responses to the 2012 report and other updates to the program’s operations, outputs, and impacts; and to develop a renewed set of findings and recommendations on the effectiveness of the ASP and the value of the program within the current landscape of epilepsy research and drug development.

The group reviewed data and background information on the ASP compiled by NINDS and University of Utah contract staff and convened for a one-day meeting on February 17, 2015, in Rockville, MD.  During the meeting, they heard presentations from and held discussions with the NINDS Acting Director and Deputy Director, the NINDS Associate Director for Translational Research, NINDS ASP and extramural program staff, and ASP contract site staff. They also began to formulate their recommendations, which they developed further via conference call and email correspondence in April 2015. This report summarizes the working group’s findings and final recommendations, as presented to the NANDS Council on May 28, 2015. Comments from Council included a statement of the need for better animal models for pediatric epilepsy and for better understanding the impact of epilepsy and ongoing seizures on the developing brain, which has implications for ultimately treating or preventing comorbid conditions. Further discussion helped clarify that the working group’s general position on screening for compounds’ effects on comorbidities reflects the current lack of strong animal models for these conditions. Council members also discussed different possibilities for a new program structure, consistent with the working group’s recommendation to consider alternatives. Following discussion, the NANDS Council voted to accept the report’s recommendations.

Summary of recommendations in 2012 Working Group Report

The 2012 Working Group Report acknowledged the success of the ASP in facilitating the development of new antiseizure drugs with fewer drug-drug interactions, less detrimental hypersensitivity, and improved seizure control in some patients. However, the group noted that despite newer drugs coming to the market, the fraction of people with epilepsy who are not successfully treated with available medications (approximately one third) remained unchanged, and that treatments to prevent or modify the course of epilepsy remained lacking. The working group recommended that support for the ASP continue at its current level, but that the ASP shift its focus to meet the most urgent current needs, including epileptogenesis and disease modification, treatments for pharmacoresistant epilepsies, true comorbidities of epilepsy, and targeted and optimized treatments for specific epilepsy subtypes and affected populations. At the time of the 2011 review, NINDS was beginning the process to recruit a new director for the ASP. The 2012 Working Group report recommended this position be filled with a strong leader with experience in drug discovery who would understand the current landscape of epilepsy drug development, which had changed considerably since the ASP first began. In addition, the 2012 report recommended changes in program operations to improve the quality of compounds screened and data obtained; increased integration with other NINDS programs; more effective communication with contract site staff to better leverage their expertise; and an external advisory group for ongoing oversight.



Findings of the 2015 Working Group

Responses to 2012 recommendations and other program improvements

The 2015 Working Group finds that NINDS has made significant progress in implementing recommendations from the 2012 report, in particular at the operational level. First, NINDS recruited a capable new leader for the ASP in Dr. John Kehne, who has overseen a complete transition in NINDS ASP staff, including new staff with pharmaceutical industry experience.  Under Dr. Kehne’s leadership, and that of Dr. Rajesh Ranganathan, NINDS Associate Director for Translational Research, the ASP has become more integrated within the NINDS Office of Translational Research (OTR) and with other NINDS epilepsy research programs. Enhanced transparency and communication between ASP staff and NINDS extramural program directors for epilepsy has allowed for more effective use of scientific and programmatic expertise and efforts to make ASP participants aware of other NINDS resources and their eligibility requirements. Similarly, communication and transparency have also improved markedly between NINDS and the University of Utah contract site staff, allowing the Utah leadership team to provide scientific input and improving operational flow and information exchange. The launch of PANAChE is also a significant accomplishment and an example of increased transparency that will broaden the impact of the ASP. It will be important for the program to continue to build this resource by adding data from more compounds. 

In terms of shifting the focus of the ASP toward priority areas recommended by the 2012 Working Group report, the program has made efforts to implement or explore new models for drug refractory epilepsy, epileptogenesis/disease modification and epilepsy comorbidities. Recently adopted models include a model of mesial temporal lobe epilepsy (mTLE) supported through the program’s first subcontract to an outside vendor (SynapCell, Grenoble, France) and models of cognitive function that are currently used for assessing cognitive side effect liabilities of screened compounds. Additional models under evaluation include a model of viral-induced epilepsy (Theiler’s murine encephalitis virus, TMEV) and another model of chronic seizures that will allow video-EEG monitoring and non-invasive chronic drug dosing.

The ASP made substantial changes to improve quality control and decision-making for how compounds are selected for screening in the program. For example, participants are asked to provide more data about their compounds (e.g., biological rationale, novelty, pharmacokinetics data, etc.) and about their research goals. Screening is not always conducted in the absence of scientific rationale or evidence of compound purity and identity, and in cases where participants have many compounds in a chemical series, participants are asked to prioritize among them rather than submit all for screening. Analyses presented by ASP staff suggest the increased stringency in quality control and compound selection are making an impact: since 2012, the program is screening fewer compounds overall, but a significantly greater fraction of screened compounds are showing positive results in the standard screening flow for anticonvulsant efficacy. In addition, compounds screened from new participants are more likely to be structurally unrelated to existing antiseizure drugs. The ASP has revised and defined its testing flow to include decision points but to allow flexibility to accommodate novel approaches as well. More systematic use of this screening flow has helped the program to identify additional future modifications, such as eliminating redundant tests when data show highly correlated results. Finally, changes were also made to screening procedures to improve experimental rigor and screening data quality.

Overall Conclusions

The 2015 Working Group commends the ASP for making significant improvements to program leadership, operating procedures and quality control, and transparency and communication within NINDS and between NINDS and ASP contract staff. Moving forward, the ASP director should continue to leverage the scientific expertise of the contract staff but to also aggressively manage the contract in order to effectively steer the program in new strategic directions.

Consistent with the prior review in 2011-2012, this Working Group is unambiguously in favor of prioritizing the identification of treatments for drug refractory epilepsy, epileptogenesis, and disease modification over the identification of additional undifferentiated acute seizure drugs. The Working Group is encouraged to see that the program has adopted new models for drug refractory epilepsy and has begun to explore other models relevant to drug refractory epilepsy, epileptogenesis, and disease modification. However, the shift toward new directions has not yet gone far enough, and the program also should rapidly revisit its strategies for testing related to different types of comorbid conditions associated with epilepsy and its treatment. Because clinically validated models and a demonstrated path forward to drug development do not yet exist in these areas, the program may need to look beyond its traditional mode of operation to successfully accelerate advances. For example, in addition to profiling potential drug compounds, the program should consider the value of target validation; and, in addition to accepting externally initiated compound submissions, the program should consider a means for prioritizing and actively soliciting agents for screening. Finally, this Working Group reiterates that the program would benefit from an Advisory Board of expert consultants who could provide guidance and support to the program on an ongoing basis as it implements new strategic directions. The following recommendations present the Working Group’s views in more detail.



  1. Establish an Advisory Board or similar group for the program, whose members can provide advice on an ongoing basis. The program would benefit from regular input and oversight on strategy and implementation from external experts, including input related to selecting new models, developing and monitoring screening flows, prioritizing targets and mechanisms for interrogation, and establishing milestones for program progress. As a link to the epilepsy research community, such a group would help the program to be more proactive in making adjustments and maintaining relevance as a unique community resource. The membership of this group should allow for some consistency over time as well as for ad hoc participation of individuals with expertise appropriate for specific programmatic matters. At the discretion of the NINDS Director, program oversight also should continue to include periodic overall reviews by a working group of the NANDS Council.
  2. Accelerate the shift in emphasis toward treatments for drug refractory epilepsy, antiepileptogenesis, and disease modification.  An early step will be to develop a new name for the program that reflects these directions and sends a clear message to the community about the program’s priorities.
  3. Align screening with specific goals for different types of treatments (drug refractory epilepsy, epileptogenesis, and disease modification). For each, the program should incorporate etiologic models relevant to human pathophysiology. The Working Group expects the program to actively monitor the research literature to identify promising new models and to implement a clear process for selecting models to incorporate, with advice from external experts and input from the epilepsy research community. This selection process should take into account gaps or shortcomings within existing screening flows and reflect the program’s goal to provide unique value, rather than duplicate investments in other research programs supported by NIH or others. The Working Group did not endorse particular models for consideration, but members discussed models of specific refractory pediatric epilepsy syndromes as an example. In addition, with rapid advances in identifying genetic causes of epilepsy, models of genetic epilepsies are also among possibilities for incorporation into the program. Consideration of genetic models should take into account an understanding that any one genetic cause represents a small fraction of all epilepsies, and that, depending on the mechanisms involved, screening results in one model may not be generalizable across other causes of epilepsy. As noted below (Recommendation VII), incorporating new models may necessitate a change to the current structure of the program since the best expertise and resources to conduct testing in newer models may reside beyond the present contract facility.

Screening for different types of treatments will likely require separate screening flows and decision trees, but the program should allow for flexibility so that opportunities are not missed by assessing compounds too narrowly based on certain purported mechanisms of action. For each screening flow, the program should define the profile of results that would support further development toward a new drug. This will enable the program to better interpret results and communicate clear guidance to participants regarding potential next steps. Below, the working group provides suggestions for revising existing screening flows and building new ones. These suggestions are not meant to be fully prescriptive, but to serve as illustrative examples. Program staff should, on an ongoing basis, build comparative pharmacology data on models, monitor the effectiveness of screening flows and decision trees, and make modifications as needed, based on data and expertise obtained over time.

  1. For drug refractory epilepsy, further streamline the current standard screening flow into only two levels (“Identification” and “Differentiation”) with two associated go/no go decisions by removing redundant assays and reordering others so that less promising compounds for drug refractory epilepsy can be ruled out earlier in the process.
    Specifically, the Working Group recommends:
    1. The program currently employs the 6Hz psychomotor seizure test as part of a first “Identification” step in its screening flow. The Working Group recommends eliminating the 32mA version of this test, and only using the 44mA version in the Identification step, together with the Maximal Electroshock (MES) and Subcutaneous (sc) Metrazol Seizure Threshold tests, one of the two kindling models (hippocampal and/or corneal kindling), and rotarod impairment in mice and minimal motor impairment in rats.
    2. Unless a strong rationale can be provided, the Working Group did not see value added for a second “Differentiation” step including the Bicuculline/Picrotoxin-induced Seizures test, Intravenous Metrazol Seizure Threshold test, Frings Audiogenic Seizures model, Long Term Potentiation (LTP) assay, Morris Water Maze (MWM) test in rats, and Novel Object Recognition (NOR) test in mice. These tests are therefore recommended to be eliminated from the drug refractory epilepsy screening flow. Instead, a single “Differentiation” step should include the Lamotrogine Resistant Amygdala Kindled Rat model, Pilocarpine-Induced Refractory Status Epilepticus model, SynapCell mTLE model, and the kainic acid-induced chronic seizure model currently under development. This latter model, which allows for chronic video/EEG monitoring and chronic drug dosing, should be formally adopted into the program as soon as possible.
  2. For epileptogenesis and disease modification, clinically validated models are also lacking, and while some models are already being pursued by the program, further effort should be made to deliberately consider others, as described above, with input from external experts and the epilepsy research community.
  1. Revise the program’s approach to assessing the impact of compounds on comorbid conditions beyond seizures, including cognitive impairment and psychiatric conditions such as anxiety and depression. The Working Group recognizes that comorbid conditions reduce quality of life for many people with epilepsy and that these conditions remain an important area of need for the development of effective epilepsy treatments. However, assays currently used in the program to assess cognitive liabilities of compounds have so far demonstrated limited predictive value. These assays, as well as assays for neuroprotection, are proposed to be eliminated from the screening flow for drug refractory epilepsy (see above), where the primary goal of screening should be to identify superior efficacy over currently available antiseizure drugs. Similarly, these assays should not be part of core screening for disease modification and antiepileptogenesis.

    Instead, testing for compounds’ effects on comorbid conditions would be more appropriate as part of late-stage characterization of drug candidates for drug refractory epilepsy or disease modification, and in particular when a drug’s mechanism suggests the potential for inducing, worsening, or ameliorating a comorbid condition. For example, it may be helpful to know whether a compound with promising screening results for disease modification might also limit cognitive impairment associated with chronic epilepsy, or whether a promising candidate for drug refractory epilepsy may be anxiolytic or likely to have adverse cognitive effects. The Working Group anticipates that companies and other participants in the program could conduct their own late stage studies to characterize effects on comorbidities, when likely to influence the ability to move a drug to the clinic. However, the program should also be able to conduct these studies when a compound’s mechanism strongly suggests the potential for an effect and when doing so would not exceed available bandwidth in terms of resources and personnel. Working Group members acknowledged limitations of existing models for predicting cognitive or psychiatric effects in humans, but based on available evidence, they favored in vivo behavioral tests over in vitro assays, such as LTP and in vitro neuroprotection assays, which have not yet been shown to be predictive markers. In addition, the Working Group noted the complex causal relationships comorbid conditions have with the underlying causes of epilepsy, long-term consequences of chronic seizures, and effects of seizure treatments. Moving forward, models and assays in the program should reflect current knowledge about these relationships so that effects of compounds on different types of comorbidities can be appropriately characterized.
  1. Look beyond the traditional role of the program to facilitate and drive advances. To date, the program has acted primarily as a service provider to screen compounds submitted by companies and researchers, and success in the program has been evidence of anticonvulsant activity for an individual compound. However, another way to de-risk development efforts and attract industry interest would be to provide initial proof-of-concept evidence for new targets or mechanisms, particularly in areas where clinically validated models are not available. In addition, while the ability to react to externally initiated submissions should be maintained, the program also should consider playing a greater role in driving what compounds and targets are pursued.
    1. Prioritize and self-nominate compounds and targets for testing, with input from the epilepsy research community and external program consultants. Recent programmatic changes toward enhanced selectivity in accepting compounds for screening have led to a decrease in the total number of compounds screened, and the Working Group recommends using some program resources resulting from this decrease to actively solicit compounds to interrogate prioritized targets and mechanisms. For example, the program could develop a matrix of targets, models, and test compounds based on promising hypotheses that have emerged in the research literature, and then prioritize among them and solicit compounds from academic and industry researchers.
    2. Profile tool compounds to develop a “reference library” that relates activity against a particular target to activity in the program’s core models of refractory epilepsy, and then ultimately in models identified as important for a given disease modification objective. This pharmacology assessment would be of immense value to the field, as it would facilitate interest in developing druggable compounds against promising targets and limit the direction of resources toward targets that do not seem worthwhile. Moreover, while positive screening results for an individual compound might de-risk further development for a single company, validation of novel targets or mechanisms could promote the interest of multiple companies. To develop this mechanistic reference library, the program should leverage sources of tool compounds from pharmaceutical companies and academia, such as very well-annotated chemogenomic compound sets, as well as existing drugs available for repurposing. In addition, where appropriate, program staff could mine existing screening data. The Working Group recognizes that proof-of-concept data are also produced though research supported by grants from NINDS and others, representing an area of existing investment for NINDS. However, the group felt there could be value in providing these data in a standardized way, by a group independent of the initial discovery, and across a broader range of potential targets than may be considered by individual investigator-initiated research projects.
  2. Take further steps to ensure efficacy data are not confounded by ADME-tox issues. A compound’s potential to become an effective drug may be limited by liabilities unrelated to the compound’s mechanism of action, such as likelihood for drug-drug interactions, brain access, p450 inhibition or induction, and P-glycoprotein (P-gp) substrate characteristics; or by liabilities that are related to the mechanism of action, including both acute and chronic toxicity. Limited information about these liabilities could lead to false negative or false positive results in efficacy screening for different types of epilepsy treatments. The Working Group recommends a stronger requirement for ADME-tox data, as well as plasma half-life and brain to plasma ratio (to inform dosing), before moving compounds from early screening into testing in more resource-intensive assays. In general, providing these data should be the responsibility of program participants, but NINDS may play a facilitating role as needed and appropriate.   
  3. Hold an open competition for the next period of support for the program. The screening contract of the ASP has been awarded to the University of Utah since the program’s inception and on a sole-source basis since 1989. This site has developed a unique capacity for successfully identifying anticonvulsant compounds and has laid important groundwork toward exploring screening for drug refractory epilepsy and models relevant to epileptogenesis, disease modification, and comorbidities. However, a broader source of expertise should be considered for continuing to implement the paradigm shift in recommended directions. In addition, NINDS should consider different contracting models for supporting the program, including for example, a hub and spoke model that would allow multiple sub-contracts under a master agreement.



Charge and Roster, July 2014


In 1975, the National Institute of Neurological Disorders and Stroke (NINDS) established the Anticonvulsant Screening Program (ASP) to promote the development and evaluation of new antiseizure drugs. The ASP provides a resource for academic and industry investigators from the U.S. and abroad (“ASP participants”) for screening potential drug compounds in several standardized animal seizure models. Screening is performed via contract at the University of Utah, and NINDS staff provides test results and advice to ASP participants regarding further development of tested compounds, while protecting confidentiality and intellectual property rights. NINDS also maintains a database of chemical structures and test results that is designed to provide insights on structure-activity relationships. A growing subset of these data is made publicly available through an online database called PANAChE, for Public Access to Neuroactive & Anticonvulsant Chemical Evaluations.

Purpose of the Working Group:

The ASP was reviewed by different external groups in 1982, 1987, 1992, 1996, and most recently in 2011 by a working group of the National Advisory Neurological Disorders and Stroke (NANDS) Council. The current working group of the NANDS Council, composed of a subset of the 2011 working group along with four new members, will review the program’s responses to the 2011 review and other updates to the program’s operations, outputs, and impacts; and will develop a renewed set of findings and recommendations on the effectiveness of the ASP and the value of the program within the current landscape of epilepsy research and drug development.  

With important contributions from the ASP itself, more antiseizure drugs are available today as compared to when the ASP began. There have also been significant advances in relevant scientific areas, including the identification of new genetic etiologies for several types of epilepsy, as well as changes in research and drug development technology and in the nature of drug development programs supported by the private sector and by NIH. The working group should review the ASP in this context and with respect to evolving goals within the epilepsy research community, which place increased emphasis on pharmacoresistant epilepsy, preventing the development and progression of epilepsy, and comorbid conditions associated with epilepsy. 

In particular, the working group should:

  • Consider the value of the ASP, as currently structured, within the epilepsy drug development landscape, and advise on the scope, configuration, and scientific strategies of future NINDS commitments in preclinical epilepsy drug development, to include any future role of the ASP.   
  • Assess the response of the ASP (including both contract site and NINDS components) to recommendations provided in 2012, and advise on recommendations and findings that may warrant further action or revision.

Working Group Composition:

The working group will include eight members, all from outside the NIH, including one member of the NANDS Council. Their collective experience brings expertise in adult and pediatric epilepsy, in drug development generally, and in the interests of the epilepsy patient advocacy community. A roster of members and their relevant affiliations is attached at the end of this document.

Working Group Activities and Deliverables:

  • The working group will work closely with NINDS staff to identify data or other background requirements that will help the group formulate its recommendations. NINDS staff will serve as the group’s liaison to the Institute and to the ASP contract facility, providing data and other support to the group.
  • The group will hold one in-person meeting, on or near the NIH Campus in Bethesda, MD, on February 17, 2015. NINDS will facilitate a conference call for the group prior to the in-person meeting, to review this charge and initial data requests. NINDS will facilitate further calls prior to or after the in-person meeting as requested by the group.
  • The group may discuss and deliberate among themselves via conference calls, email, or other means as they deem appropriate, with or without NINDS staff participation. If the group determines that an additional in-person meeting is necessary, NINDS will organize such a meeting. The group may also conduct additional virtual or in-person interviews with NINDS or ASP site staff as the group determines appropriate.
  • The group shall submit minutes or summaries of all meetings to be filed with NANDS Council records.
  • The final product of the group will be a written summary report and a set of recommendations to be presented to the NANDS Council for deliberation in open session in during the May 28-29, 2015, Council meeting.


Chair: Robert Pacifici, PhD, CHDI Management /CHDI Foundation

Amy Brooks-Kayal, MD, University of Colorado Denver, The Children's Hospital

Ray Dingledine, PhD, Emory University

Henrik Klitgaard, PhD, UCB Pharma

Elizabeth Kovar, PhD, University of Chicago

Dan Lowenstein, MD, University of California, San Francisco

Roy Twyman, MD, Johnson and Johnson

Michelle Welborn, PharmD, ICE Epilepsy Alliance