Therapy Discovery for Pharmacoresistant Epilepsy and for Disease‐modifying Therapeutics: Summary of the NIH/NINDS/AES Models II Workshop
2003; Wiley; Volume: 44; Issue: 12 Linguagem: Inglês
10.1111/j.0013-9580.2003.32803.x
ISSN1528-1167
AutoresJames P. Stables, Ed Bertram, F. Edward Dudek, Greg Holmes, Gary W. Mathern, Asla Pitkänen, H. Steve White,
Tópico(s)Computational Drug Discovery Methods
ResumoThe National Institutes of Health (NIH), through the Antiepileptic Screening Project (formerly known as the Antiepileptic Drug Development Program), has been a key element in the discovery and introduction of new pharmacologic therapies since the institution of the program more than 25 years ago. The program was established with the goals of facilitating the identification of new drugs that were safer and more efficacious epilepsy treatments. Through these efforts and leadership, the program has been highly successful, with the introduction of many new antiepileptic compounds, providing new and welcomed treatment options to epilepsy patients worldwide. In spite of these successes, these new drugs have yet to make a significant impact in the most difficult to treat conditions, or for those patients who are incompletely responsive to available pharmacologic treatments. It is estimated that more than one third of patients with epilepsy (∼750,000 Americans) belong in this pharmacoresistant category. Because of this large number of pharmacoresistant patients with epilepsy, a developing consensus mandates that the current process of therapy discovery and development must be reevaluated and enhanced in light of the evolution of available models and our understanding of epilepsy. Several large meetings (notably the White House–initiated conference, “Curing Epilepsy: Focus on the Future,” which was held in March 2000, and the NIH Workshop for Models for Epilepsy and Epileptogenesis, held in March 2001) started the process. One of the primary recommendations of these conferences was that the existing process of therapy discovery and development should be enhanced with a focus on developing preclinical models that would be more predictive of clinical success in stopping the seizures of patients currently incompletely controlled with existing therapies. In addition, strong recommendations were made to develop mechanisms to identify treatments that will prevent the development of epilepsy (epileptogenesis). A key element in this latter recommendation is the identification of appropriate models for epileptogenesis and the creation of a process that will use these models for antiepileptogenic therapy discovery. In response to these recommendations, a workshop was held in September 2002 to identify useful models for therapy discovery for treatment of resistant epilepsy and the actual prevention of the disease. The first goal of this conference focused on recommendations of models to identify potential new therapies and to predict the clinical efficacy of these therapies. The second was the development of a process that can be used to evaluate and add new and promising models as they come along. The following paragraphs are a summary of the meeting with an emphasis on the recommendations that were made regarding therapy resistance and epileptogenesis. Two working groups were created for the meeting: one to concentrate on issues of therapy-resistant epilepsy, and the other to focus on disease-modifying therapeutics (DzM). The two focus groups were asked to consider the potential of candidate models in predicting clinical success of any particular treatment. Other major considerations required in the evaluations were the relative ease of use and the potential parallels among the different human epilepsies. An important point of emphasis was that the evaluation and eventual recommendations must be focused on the potential for the models to be predictive for clinical success, not for the relative value of the models for basic epilepsy research, while acknowledging the critical need for the latter. This process of evaluation should at no time be considered a general ranking system of epilepsy models. The many models, in vitro and in vivo, possess important scientific value for our understanding of the mechanisms of epilepsy. Therefore the overall goal of this workshop was to apply these criteria toward model evaluation to identify the most scientifically feasible screening tools that might be currently applied for discovery of new therapies. The issues concerning animal models of and protocols for therapy resistance and disease modification are generally similar. Therefore the discussions during the breakout sessions focused on the most appropriate experimental approaches or protocols, in addition to the particular animal models themselves. Current therapy screening has been developed on the assumption that epilepsy treatments could be identified by using short-term induced seizures in normal brains. This approach has many advantages including the ready availability of the models and the ability to use these models in a serial screening process for multiple compounds. We wish to augment this approach to therapy screening by using animals that have abnormal brains and seizures. The animals or tissues used would have an increased propensity for seizures or recurrent spontaneous seizures. Although we believe this approach will provide new targets for therapy development, neither the models nor the approach has been validated for the ability to identify effective treatments. The community has historically approached antiepileptic drug (AED) discovery with the general assumption that the underlying mechanisms causing different types of epilepsy and their associated seizures are similar enough to each other that testing protocols used in several types of seizures induced in normal animals would generically reveal new and more effective AEDs. This approach assumes that seizures in all types of epilepsy have relatively few mechanisms that underlie the development of epilepsy or the initiation of seizures. Clearly, however, the many epilepsy syndromes have multiple different mechanisms and different natural histories, with some syndromes likely having multiple contributing mechanisms. If we expect therapeutics to be maximally effective for individuals, then it may be necessary to develop animal models for the different forms of pharmacoresistant epilepsy and to target treatments toward the specific mechanisms that lead to seizures in the different epilepsy syndromes. This approach does not assume that there is a single treatment for each epilepsy syndrome, and that polytherapy addressing multiple epileptogenic mechanisms may be necessary in some syndromes. A logical assumption exists that therapies discovered by screening with models that more closely approximate chronic epilepsy would complement current screening methods. This assumption is based on the premise that the models that more closely represent the human disorder will identify therapies with additional mechanisms of action compared with the current models. In addition, these models could uncover potential toxicities that may not be apparent in normal animals or demonstrate reduced toxicities in epileptic animals compared with controls. These considerations would decrease costs and development time. However, these hypotheses remain to be tested, and these significant issues should be considered in our community's attempt to augment therapy discovery directed at the prevention of epilepsy and at pharmacoresistant patients. Pharmacoresistance in humans is generally defined as failure to achieve complete control of seizures. Although no absolute determination regards how many drugs should be tried, the general consensus is that the failure to respond to trials with two or three drugs with complete seizure control represents pharmacoresistance, even if a reduction in seizure frequency occurs as a result of the treatments. Pharmacoresistance in animal models, based on human experience, can be minimally defined as persistent seizure activity that does not respond to monotherapy at tolerable doses with at least two current AEDs. An important need for the therapy-screening process is to find animal models that do not respond, or respond poorly, to two currently available AEDs. Pharmacoresponsiveness can be considered a major subset of pharmacoresistance. Pharmacoresponsiveness can be defined as a partial suppression of seizures. Although this can be considered a statistical success, it often does not result in significant benefit to the patients' overall condition or quality of life. It is important to determine pharmacoresponsiveness in the selected animal models. An animal model for pharmacoresistant epilepsy should have good parallels to human forms of pharmacoresistant epilepsy. Such models should demonstrate a response to a new therapy that will later be found to be effective in humans resistant to existing AEDs. The new treatment should result in a significant decrease or cessation of spontaneous seizures. A goal should be to develop new animal models for therapy that reflect the most common forms of pharmacoresistance in humans. This will require better characterization of the human syndromes/epilepsies that are pharmacoresistant by using appropriate research tools, including epidemiology. The response to a new therapy should be significantly greater than the response to existing AEDs in the pharmacoresistant models. Another goal will be to create in vivo models that are good representations of the many types of human epilepsy that will facilitate therapy development. In vitro systems related to in vivo models of resistant epilepsy also merit exploration. It is likely, however, that the in vivo systems that are highly predictive of clinical success will have to be developed first, and in vitro systems that can predict success in the in vivo systems will follow. Many models, in vivo and in vitro, were considered. The primary weight given to the final recommendations were parallels that the models may have to human pharmacoresistant epilepsy. Based on current knowledge and the models available, models of injury-induced epilepsy based on electrically or chemically induced status epilepticus are ready to be tested for pharmacosensitivity and resistance. This group of models was considered to have the greatest parallels with a common form of human drug-resistant epilepsy, the mesial temporal lobe or limbic epilepsy syndrome. Additional recommendations were made that resources be directed specifically for the development of models of pharmacoresistance in adult and developing brains that represent other pharmacoresistant syndromes (see section on model characteristics). Suggested strategies for pharmacoresistant model development included the use of models that were based on genetically prone animal strains. These animals are more “seizure susceptible” and might be more prone to the development of epilepsy than are normal strains (i.e., Genetically Epilepsy Prone Rats, the E1 mouse strain.). Animal models with a high frequency of seizures may be useful for initial serial screening for multiple compound screening in a potentially pharmacoresistant model. The question of what defines an adequate response to a potential AED, such as seizure reduction or a change in the character of the seizures, should be left to a future meeting after more data are available. With present technologies, these experiments will be difficult and resource intensive, but this goal was considered extremely important. Practical technologic advances would substantially improve the use of models that are pharmacoresistant to current AEDs and enhance the discovery of new antiepileptic therapies. Technology needs include prolonged drug delivery and monitoring and miniaturized systems for long-term recording from animals with spontaneous seizures. Although it is recognized that the goal in human treatment is elimination of seizures, the most practical approach now is to search for a significant decrease in seizure frequency and/or severity, particularly in a model that shows pharmacoresistance to two appropriate presently used AEDs. The degree of acceptable seizure reduction will likely vary by model, but considerable improvement over the currently available treatments should ensue. In addition to the effect on seizures, behavioral side effects should be considered in the therapy discovery process at therapeutic concentrations. Finally, the models should consider the difference between the developing and mature brain. The translation of a pharmacoresistant therapy identified in an animal model to human therapy is critical. History has shown that successful compounds identified for epilepsy have broad applications for other disorders. In the event that a pharmacoresistant therapy is found that industry does not support because of limited application, funding mechanisms (private and/or public) will be necessary to develop the therapy for release to enhance the translational process. An Advisory Committee of experts in the field should be formed to monitor NIH-sponsored therapy discovery for pharmacoresistant models of epilepsy. The results of these studies should be made available to all researchers within a reasonable period, and the process of model development, utilization, and therapy discovery should be accessible to and monitored by the epilepsy community. Epileptogenesis has acquired many definitions over time. For the purposes of this discussion, it means the process of developing chronic epilepsy. This definition implies that an interval exists between the appearance of a potential causative factors (genetic, malformation, injury) and the appearance of the chronic epileptic condition. The latent period is a key concept in epileptogenesis. It is the interval that exists before the onset of seizures that implies that an active process leads to changes that eventually cause epilepsy. Because epilepsy is frequently associated with other problems (e.g., behavioral and cognitive), assessment of antiepileptogenic treatments will require careful attention to a variety of outcome markers that can be considered changes that occur in the latent period that may herald the eventual development of epilepsy and its associated problems. These outcome markers could include, but will not be limited to, behavioral seizures, electroencephalographic (EEG) seizures, interictal epileptiform discharges, cognitive and behavioral outcomes, and pathology. The identification of these and other surrogate markers in models may facilitate the development of assays with increased throughput capacities that in turn will simplify the identification of potential antiepileptogenic treatments. Disease modification under antiepileptogenic treatment was considered an important issue as well. This issue deals with the many problems associated with epilepsy other than actual seizures themselves: mood disorders, psychiatric disorders, cognitive problems, seizure severity, and postictal state severity. This concept was introduced because improving these issues, separate from preventing epilepsy itself, can have a significant impact on the patients' quality of life. As used in this document, antiepileptogenic treatments refer to pharmacologic compounds and other interventional devices, such as electrical stimulation. Two groups of models were recommended as potentially useful tools for antiepileptogenic treatment screening: status epilepticus–induced recurrent seizures and kindling. A number of approaches can be used within each group. For the post–status epilepticus models, a variety of different chemoconvulsants and intracerebral electrical stimulation patterns have been used to induce status epilepticus, which is followed, after a latent period, by spontaneous recurrent seizures. Of the various systemic chemoconvulsants, kainic acid and pilocarpine have been the best characterized in regard to seizure phenomenonology, EEG features, cognitive outcome, and neuropathology. With titration of the dosage of the chemoconvulsant and use of anticonvulsants, status severity and duration can be controlled. This manipulation may allow investigators to vary the percentage of animals that develop spontaneous seizures systematically and to control the size and extent of neuropathologic lesions involving both limbic and nonlimbic structures after the initial insults associated with status epilepticus. A variety of direct brain electrical-stimulation patterns also have been used to induce status epilepticus and subsequent spontaneous seizures. Whereas this variation of the post–status epilepticus epilepsy model requires the surgical placement of an intracerebral electrode, no toxins are necessary. All of the status epilepticus models have the advantage of a latent period during which spontaneous seizures do not occur. After the latent period, spontaneous recurrent seizures typically escalate in frequency over time. The latent periods offer an opportunity to introduce therapy and measure its effect on prevention. These models are frequently associated with cognitive impairment. Another appealing feature of these models is their similarity to human temporal lobe epilepsy with partial seizures with or without secondary generalization. The kindling model also was recommended for investigation as a potential antiepileptogenic screening. This model, which was developed in 1969 by Graham Goddard, is a model in which repeated excitatory stimuli initially induce partial seizures. With continued stimulation, a progressive increase occurs in the severity of the seizures. With a sufficient number of stimulations, delivered by particular protocols, spontaneous seizures may occur, although this finding is not seen universally. The model has been extensively evaluated by investigators worldwide and provides the opportunity for investigators to study the step-wise progression of various neurobiologic alterations that underlie the epileptogenic process. Disease-susceptibility genes play an important role in epileptogenesis. Whether the brain develops seizures after an insult may be partially genetically determined. It is recommended that investigators take advantage of the advance in rodent genomics and existing genetically prone strains such as the genetically epilepsy prone rat in assessing antiepileptogenic treatments. Models of epileptogenesis should not be confined to one strain of mouse or rat. To achieve the goals addressed with the animal models selected, methods of drug delivery and seizure monitoring have to be developed to make this very labor-intensive process a little easier. In addition, the timing of therapeutic intervention, the duration of treatment, and length of follow-up require careful attention. Because a considerable number of animal models exist in which epileptogenesis can be studied, it is the consensus of the subcommittee that this is an ideal time for the investment of resources into screening potential antiepileptogenic compounds in established rodent models. Two approaches to screening compounds should be undertaken with the status epilepticus models. Initially investigators should use models that have a high incidence of developing spontaneous seizures so that it is possible to develop uniform outcome measures. This approach will allow promising treatments to be studied in a standardized and reproducible manner. Agents reducing or preventing seizures with these stringent requirements may have a high potential for antiepileptogenic treatments in humans. However, it also is likely that treatments powerful enough to prevent or reduce seizures in this situation will be toxic with regard to neurologic recovery after injury or to neurologic development. It is recommended that the second step in screening antiepileptogenic treatments use status epilepticus models in which the initial insult is less severe and the percentage of rats with spontaneous seizures lower. It is possible that this variability will allow treatments with less potency but less toxicity to be identified. At the moment, however, such models are not well established, so that efforts at characterization and standardization are needed before this latter approach could be implemented. We need to develop new technologies that will allow investigators to screen more compounds in a timely and scientifically sound manner. The goal of antiepileptogenesis is to prevent epilepsy. An important outcome measure in the recommended rodent models will most likely be spontaneous seizures, because such paradigms are more representative of human disease. EEG and video monitoring of rodents is technically difficult, time consuming, and expensive. Development of innovative technologies that improve accuracy of automatic seizure detection while reducing time commitment of personnel is highly desirable. Identification of surrogate markers will be aided by development of technologies that can detect early, and predictable, indicators of epileptogenesis. Continued efforts in new model development are needed. A variety of epileptogenesis models allows specific mechanistic hypothesis testing and provides insights into potentially new therapeutic treatments. In addition, a paucity of models exists that mimic clinical situations such as the catastrophic epilepsies (e.g., infantile spasms, Lennox–Gastaut syndrome, or severe myoclonic epilepsy of childhood). The highest incidence of epilepsy occurs at the extremes of life, during infancy and in the elderly, yet little attention has been given to developing models that are comparable to these age-related syndromes. Likewise, epileptogenesis could be influenced by hormones. For this reason, investigators should strive to evaluate potential antiepileptogenic treatments in both male and female models. The ultimate goal of this process is to provide a framework for therapy evaluation that will allow the targeted development of treatments that control or prevent the seizures of epilepsy in all patients. It is recommended that the screening programs be conducted through a consortium of investigators with expertise in the proposed models or through a contract with a commercial enterprise. Techniques should be standardized in the laboratories, with quality control monitored by the NIH. Because the mechanisms underlying epileptogenesis are complex and multifactorial, a process for selecting types of therapeutic intervention in these models should be established. The models that have been recommended require significant amounts of work, and the process of evaluating any compound will be relatively long, until such time as appropriate high-throughput/in vitro screening models can be developed. For this reason, a regular review by external advisors must occur with regard to the particular models used and the compounds tested. As an example for the implementation of the screening process, the models could be placed in the following steps of evaluation. Demonstrate resistance to currently available medications. Modify the screening process used in step 1 to create maximal efficiency without sacrificing the ultimate predictiveness of the model. Once the models are established as broadly pharmacoresistant, baseline response rates to existing medications are established, and the methods for using the models for drug testing are established and optimized, select models to evaluate new compounds prospectively. Use the models to evaluate compounds on a mechanistic basis to establish the potential of a particular mechanism of action (e.g., calcium channel blocker, neuropeptide Y agonist). Establish a publicly accessible database for the results of the mechanistic screens to assist in the development of effective compounds and to help with the understanding of the mechanisms underlying seizure generation. An advisory board should be established to review the ongoing evaluations at regular intervals and to make recommendations, based on basic science information and the availability of compounds, for the inclusion of a particular mechanism in the screening project. Confidentiality rules will be established regarding the screening of proprietary drugs from the private sector. In the recommended models, verify the lack of efficacy reported in human trials of epilepsy prevention. Determine intervention points in the models that are clinically relevant and realistic (e.g., after an injury that has a high risk of resulting in epilepsy and after the first seizure, when the chances for development of chronic epilepsy may be greatest). Define outcome measures in the models. These targets could include the prevention of epilepsy, improved behavioral or cognitive outcomes, the prevention of pharmacoresistance, or neuronal loss. Establish appropriate tests for demonstrating the effect of intervention. An advisory board should be established to review the ongoing evaluations at regular intervals and to make recommendations, based on basic science information and the availability of compounds, for the inclusion of a particular mechanism in the screening project. Rules will be established regarding the screening of proprietary drugs from the private sector. As noted in both the pharmacoresistance and epileptogenesis sections, a great need exists for new models that have strong parallels to various types of human epilepsy. Because many claims are made for models, the committee wished to outline features of some of the important epilepsy syndromes that are characterized by pharmacoresistance, a period of epileptogenesis, or both. The creation of appropriate models not only would facilitate the development of syndrome-specific treatments but also would greatly enhance our understanding of the underlying mechanisms. In this section, important features of several of the syndromes that should be modeled are defined so the model can be validated as relevant to a particular epilepsy syndrome. Certainly there could be crossover among some syndromes (e.g., infantile spasms and neonatal hypoxia ischemia). Infantile spasms: age limited, early onset; abnormal EEG interictally; cognitive abnormalities; electrographic seizures not necessary; spasms; pharmacoprofile similarities; no specific pathology required; etiology variable; syndrome evolves. Hypoxic/ischemic: age limited (neonatal); typical hypoxic/ischemic pathology; chronicity of seizures; abnormal interictal EEG. Cortical dysplasia: pathological parallels; abnormal interictal EEG; chronicity of seizures; pharmacoresistance desirable. Lennox–Gastaut: age limited with some evidence for evolution of syndrome, no pharmacologic profile required; pharmacoresistance; abnormal interictal EEG; chronicity; multiple seizure types desirable; pathology variable. Adult brain injury: acute inciting event; pathological parallels; chronic seizures; latent period; behavioral and cognitive changes; pharmacoresistance; etiologic parallels (e.g., stroke, trauma; status epilepticus). Tumor associated: pathology parallels; seizures. Parasite associated: pathology and etiology parallels; seizures. Genetic: human genetic parallels, similar associated pathology, mutation, EEG, seizure type, and clinical course. Mesial temporal lobe: pathological and physiological parallels; spontaneous seizures; EEG similarities; pharmacologic response parallels. In summary, this workshop recommended the development of the post–status epilepticus models of chronic limbic epilepsy into screening tools that will be used to identify new therapies for pharmacoresistant epilepsy and for the prevention of chronic epilepsy. The kindling model also should be examined for its potential to identify antiepileptogenic treatments. Recommendations were made for the process of identifying new compounds to be used in these models, including the creation of an advisory board and the development of a public database for the posting of the results in a reasonable period. Finally, a strong recommendation was made for the development of models from important therapy-resistant and devastating epilepsy syndromes that could be used in expanded therapy-discovery programs. Both groups recognized the difficulties in using the recommended models, especially with regard to maintaining effective drug levels throughout a prolonged testing period. However, the groups agreed that, in addition to bringing better treatments to the clinic, the potential to learn about the mechanisms of epilepsy and epileptogenesis in a program that was carefully designed to evaluate compounds of defined mechanisms could have broader benefits to epilepsy research. Edward Bertram, M.D. University of Virginia Health Science Center Charlottesville, VA F. Edward Dudek, Ph.D. Colorado State University Fort Collins, CO Gregory Holmes, M.D. Harvard Medical School Boston, MA Margaret Jacobs National Institutes of Health Bethesda, MD Gary Mathern, M.D. University of California, Los Angeles (UCLA) Los Angeles, CA Asla Pitkanen, M.D., Ph.D. University of Kuopio Kuopio, Finland James Stables National Institute of Neurological Disorders and Stroke National Institutes of Health Bethesda, MD H. Steve White, Ph.D. University of Utah Anticonvulsant Screening Project Salt Lake City, UT Amy Brooks-Kayal, M.D. Children's Hospital of Philadelphia Philadelphia, PA Douglas Coulter, Ph.D. Children's Hospital of Philadelphia Philadelphia, PA Marc Dichter, M.D., Ph.D. University of Pennsylvania Philadelphia, PA David Hosford, M.D., Ph.D. Target Disease Associations GlaxoSmithKline R&D Research Triangle Park, NC Diane Howden National Institute of Neurological Disorders and Stroke National Institutes of Health Bethesda, MD Frances Jensen, M.D. Children's Hospital and Harvard Medical School Boston, MA Phillip Jobe, Ph.D. University of Illinois College of Medicine Peoria, IL Russell Katz, M.D. Food and Drug Administration Rockville, MD Harvey Kupferberg, Ph.D., Pharm.D. Kupferberg Consultants, LLC Potomac, MD Fred Lado, M.D., Ph.D. Albert Einstein College of Medicine Bronx, NY Brian Meldrum, Ph.D., M.B. Kings College London London, United Kingdom Luiz Mello, M.D., Ph.D. Universidade Federal de São Paulo São Paulo, Brazil Solomon Moshe, M.D. Albert Einstein College of Medicine Bronx, NY Astrid Nehlig, Ph.D. INSERM Strasbourg Cedex, France Audrey Penn, M.D. National Institutes of Health Bethesda, MD Roger Porter, M.D. Wyeth-Ayerst Research Philadelphia, PA Michael Rogawski, M.D., Ph.D. National Institutes of Health Bethesda, MD Chris Rundfeldt, Ph.D., DVM. Elbion AG Radebeul, Germany Philip Schwartzkroin, Ph.D. University of California, Davis School of Medicine Davis, CA Thomas Sutula, M.D., Ph.D. University of Wisconsin Madison, WI John Swann, Ph.D. Baylor College of Medicine Houston, TX Roy Twyman, M.D. Johnson & Johnson Pharmaceutical Research/Development Raritan, NJ Annamaria Vezzani, Ph.D. Mario Negri Institute for Pharmacology Milano, Italy Claude Wasterlain, M.D., L.Sc. VA Medical Center West Los Angeles, CA Karen Wilcox, Ph.D. University of Utah Salt Lake City, UT 84112 Alan Willard, Ph.D. National Institutes of Health Bethesda, MD Susan Axelrod Citizens United for Research in Epilepsy (CURE) Chicago, IL, Suzanne Berry, C.A.E. American Epilepsy Society West Hartford, CT Dianne Flescher The Epilepsy Foundation Landover, MD Madhumita Banerjee, M.D. National Institutes of Health Bethesda, MD Christina King National Institutes of Health Bethesda, MD Yuan Liu, Ph.D. National Institutes of Health Bethesda, MD Paul Scott, Ph.D. National Institutes of Health Bethesda, MD
Referência(s)