Steve H. White MD*
José H. Woodhead MD**
Karen S. Wilcox PhD***
James P. Stables BS(pharm)****
Harvey J. Kupferberg MD*****
Harold H. Wolf MD******
* Professor, Department of Pharmacology and Toxicology; and Director, Anticonvulsant Screening Project, University of Utah, Salt Lake City, Utah
** Department of Pharmacology and Toxicology, University of Utah, Salt Lake City, Utah
*** Research Assistant Professor, Department of Pharmacology and Toxicology, University of Utah, Salt Lake City, Utah
**** Program Director, Anticonvulsant Screening Project, Technology Development Cluster, NINDS, National Institute of Health, Rockville, Maryland
***** Epilepsy Branch, National Institute of Neurological Disorders and Stroke, National Institutes of Health, Bethesda, Maryland
****** Professor, Department of Pharmacology and Toxicology, University of Utah, Salt Lake City, Utah
In 1974, the National Institute of Neurological Disorders and Stroke established the Anticonvulsant Drug Development (ADD) Program. Since then, this program has been instrumental in stimulating the discovery and development of new chemical entities for the symptomatic treatment of human epilepsy. The ADD Program serves as an excellent example of a successful collaboration between government, the pharmaceutical industry, and academia. Since its inception, this National Institutes of Health (NIH)-sponsored program has accessioned over 24,000 investigational antiepileptic drugs (AEDs) from academic and pharmaceutical chemists worldwide. The initial identification and characterization of their anticonvulsant activity has been established through a contract with the University of Utah Anticonvulsant Screening Project (ASP). The success of this collaboration is exemplified by the worldwide approval of several new AEDs since 1993 and the continued clinical evaluation of numerous other promising candidates.
The long-standing mission of the ASP has been to identify and characterize the anticonvulsant activity of those chemical entities that display a reasonable degree of separation between their anticonvulsant and behaviorally toxic doses. The characterization of a drug's anticonvulsant and behavioral toxicity profile is established using a battery of well defined animal models. Herein lies one of the most frequently discussed issues in the current AED discovery process: What is the most appropriate animal model to use when attempting to screen for efficacy against human epilepsy? The two primary screens of the ASP continue to be the maximal electroshock seizure (MES) test and the subcutaneous pentylenetetrazol (Metrazol; s.c. MET) seizure test. With one exception (i.e., levetiracetam), all of the AEDs approved since 1993 have been found to be active in one or both of these tests. Levetiracetam appears to represent a truly unique compound that is inactive in the traditional MES and s.c. MET tests, yet active in partial and primarily generalized seizure models (1, 2, 3, 4, 5, 6). Activity in these models provided the rationale supporting the early clinical evaluation of levetiracetam in patients with epilepsy. In three pivotal trials, levetiracetam was found to be effective as add-on therapy for the management of partial seizures (7, 8, 9, 10, 11, 12). In this regard, the identification and development of levetiracetam as an efficacious AED for the treatment of partial seizures demonstrates the need for flexibility when screening for efficacy and the need to incorporate levetiracetam-sensitive models into the early evaluation of an investigational AED.
OVERVIEW OF THE ANTICONVULSANT SCREENING PROJECT TESTING PROTOCOL
The ASP uses a combination of nonmechanistic, mechanistic, and syndrome-specific animal models to identify and characterize the anticonvulsant profile of an investigational AED. The nonmechanistic approach is very well suited for
P.37
the early evaluation of anticonvulsant activity because it assumes that the pharmacodynamic activity of a drug is independent of its mechanism of action. For the most part, the model systems used by the ASP display clear and definable seizure end points and require minimal technical expertise. Furthermore, this approach is ideally suited to the large number of chemically diverse entities that are evaluated annually by the ASP.
|
FIGURE 3.1. Anticonvulsant screening project testing paradigm. MES, maximal electroshock seizure; s.c. MET, subcutaneous Metrazol. |
The current in vivo testing protocol of the ASP has evolved over the last 27 years to include a variety of animal seizure models that have proven to be valuable in identifying clinically effective AEDs (13). On receipt at the laboratories of the ASP, the test substance is subjected to a large number of testing procedures according to the paradigm summarized in Figure 3.1. As shown in Figure 3.1, there are four major phases of the ASP testing protocol: identification, quantification, differentiation, and advanced testing. The specifics of these major phases are outlined in Table 3.1, and the details of many of the anticonvulsant procedures outlined in the table are discussed in the following sections.
MATERIALS AND METHODS
Experimental Animals
Adult male CF No. 1 albino mice (18 to 25 g) and adult male Sprague-Dawley albino rats (100 to 150 g) are used as experimental animals. These particular strains are preferred for anticonvulsant studies because they are docile and easy to handle. Moreover, CF No. 1 mice rarely succumb to induced seizures (14). Animals of the same sex, age, and weight are used to minimize biologic variability (15). The animals are maintained on a 12-hour light/dark cycle and allowed free access to food and water, except during the short time they are removed from their cages for testing. Animals newly received in the laboratory are allowed 24 hours to compensate for the food and water restriction incurred during transit. This is necessary because such restriction increases the severity of MES (16). All animals are maintained and handled in a manner consistent with the recommendations in the U.S. Department of Health, Education and Welfare publication (NIH) No. 8623, Guide for the Care and Use of Laboratory Animals. Animals usually are used only once and then disposed of in a humane manner. In those instances where they are used a second time, at least a 1-week interval is allowed for the animal to eliminate the test drug.
Convulsant Chemicals
For tests based on chemically induced convulsions, the convulsant chemical is prepared in a concentration that induces convulsions in more than 97% of animals when injected in
P.38
mice in a volume of 0.01 mL/g body weight or in rats in a volume of 0.02 mL/10 g body weight. For mice, MET and picrotoxin (PIC) are dissolved in 0.9% saline sufficient to make a 0.85% and 0.032% solution, respectively. For rats, MET is given in a concentration of 2.82%. Bicuculline (BIC) is dissolved in 1.0 mL of warmed 0.1 N hydrochloric acid with the aid of a micromixer and sufficient 0.9% saline added to make a 0.027% solution. The solution is used within 30 minutes. All chemical convulsants are administered s.c. into a loose fold of skin in the midline of the neck. No other drugs or chemicals are injected in the same s.c. site. The judicious selection of injection sites avoids false-positive results induced by vasoconstrictor substances retarding the absorption of the convulsant agents. Because the doses used in the aforementioned tests induce convulsions in over 97% of animals, it is unnecessary to run control groups simultaneously with the test groups.
|
TABLE 3.1. OVERVIEW OF TESTING PROCEDURESa |
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Preparation of Test Drugs
Regardless of their water solubility, all drugs are either dissolved or suspended in 0.5% methylcellulose. The test substance is given in a concentration that permits optimal accuracy of dosage without the volume contributing excessively to total body fluid. Thus, the volume used in mice is 0.01 mL/g body weight, and in rats, 0.04 mL/10 g body weight. Test drugs are routinely administered intraperitoneally (i.p.) or orally (p.o.), as indicated in Table 3.1.
Determination of Acute Toxicity
Abnormal neurologic status disclosed by the rotorod test (17) commonly is taken as the end point for minimal behavioral impairment in mice. Abnormal neurologic status disclosed by the positional sense test, muscle tone test, or the gait and stance test is taken as the end point for minimal behavioral impairment in rats. Inability of a rat to perform normally in at least two of these tests indicates that the animal has some neurologic deficit. The names assigned to these tests are those used in the authors' laboratories, and do not necessarily refer to the specific neurologic reflexes involved (18).
Rotorod Test.
The rotorod test is used exclusively in mice to assess minimal motor impairment. When a normal mouse is placed on a rod 1 inch in diameter that rotates at a speed of 6 rpm, the mouse can maintain its equilibrium for long periods. Inability of the mouse to maintain its equilibrium in three trials during 1 minute on this rotating rod is used as an indication of such impairment.
Positional Sense Test.
If the hind leg of a normal mouse or rat is gently lowered over the edge of a table, the animal will quickly lift its leg back to a normal position. Neurologic deficit is indicated by inability of the animal to correct rapidly such an abnormal position of the limb.
Gait and Stance Test.
Neurologic deficit is indicated by a circular or zigzag gait, ataxia, abnormal spread of the legs, abnormal body posture, tremor, hyperactivity, lack of exploratory behavior, somnolence, stupor, catalepsy, and the like.
P.39
Muscle Tone Test.
Normal animals have a certain amount of skeletal muscle tone that on handling is apparent to the experienced technician. Neurologic deficit is indicated by a loss of skeletal muscle tone characterized by hypotonia or flaccidity.
Anticonvulsant Identification
At present, no single laboratory test, in itself, establishes the presence or absence of anticonvulsant activity or fully predicts the clinical potential of a test substance. In the ASP, three tests are used for the routine identification of anticonvulsant activity: the MES test, the s.c. MET seizure threshold test, and the 6-Hz psychomotor seizure test.
Maximal Electroshock Seizure Test and Subcutaneous Metrazol Seizure Threshold Test
In the MES test, 60-Hz alternating current (mice, 50 mA; rats, 150 mA) is delivered for 0.2 second through corneal electrodes by means of an apparatus similar to that originally designed by Woodbury (19). At the time of administration of the test substance, a drop of 0.5% tetracaine in saline is applied to the eyes of all animals assigned to any electroshock test. Immediately before the placement of corneal electrodes, a drop of electrolyte (saline) is placed on each eye. The animals are restrained by hand and released immediately after stimulation to permit observation of the seizure throughout its entire course. Abolition of the hind limb tonic extensor component is taken as the end point for this test. Tonic extension is considered abolished if the hind limbs are not fully extended at 180° with the plane of the body. Absence of this component suggests that the test substance has the ability to prevent the spread of seizure discharge through neural tissue.
In the s.c. MET test, a convulsive dose (CD97) of MET (85 mg/kg in mice, 56.4 mg/kg in rats) is injected s.c. The animals are placed in isolation cages and observed for the next 30 minutes for the presence or absence of an episode of clonic spasms persisting for at least 5 seconds. Absence of a clonic seizure suggests that the test substance has the ability to raise the seizure threshold.
In identification studies involving mice, the ASP routinely uses the MES and s.c. MET tests. Sixteen mice are randomly divided into three groups of four, eight, and four mice each; each group is then given 30, 100, or 300 mg/kg, respectively, of the test substance i.p. Thirty minutes after administration of the test substance, all animals are subjected to the rotorod test; one animal in the 30 and 300 mg/kg group and three animals in the 100 mg/kg group are then subjected to the MES test, and one animal in each group to the s.c. MET test. Four hours after drug administration, all remaining animals in each group are subjected to the rotorod test; these animals are then subjected to the MES and s.c. MET test as indicated previously. Thus, it requires only 16 mice to cover the dose range of 30, 100, and 300 mg/kg and the periods of 30 minutes and 4 hours.
|
TABLE 3.2. ANTICONVULSANT IDENTIFICATION IN MICE AFTER INTRAPERITONEAL ADMINISTRATION |
||||||||||||||||||||||||||||||||||||||||||||||||||||
|
||||||||||||||||||||||||||||||||||||||||||||||||||||
An example of results obtained with the MES and s.c. MET screening procedures is shown in Table 3.2. As can be seen, these preliminary results suggest that the test substance is effective in toxic doses in the MES and s.c. MET test. Furthermore, they demonstrate that the minimal behaviorally toxic dose is >30 mg/kg but <100 mg/kg. The test substance also appears to have a relatively rapid onset and short duration of action because both the anticonvulsant and neurotoxic effects are greater at 30 minutes than at 4 hours.
In identification studies using rats, 30 mg/kg (MES) or 50 mg/kg (s.c. MET and toxicity) of the test substance is administered orally to 10 groups (5 groups for MES and 5 for s.c. MET) of 4 rats per group. At various times after administration (0.25, 0.5, 1, 2, and 4 hours), animals are evaluated for neurologic deficit and then subjected to the MES or s.c. MET tests. The ratios of animals protected or toxic to animals tested are determined. The initial identification studies in rats provide information as to whether the test substance is active or toxic in a dose of 30 or 50 mg/kg after p.o. administration. It also discloses the time of onset, the approximate time of peak effect (TPE), and the duration of anticonvulsant activity or neurotoxicity. Anticonvulsant identification results obtained in rats with the same test substance described in Table 3.2 are shown in Table 3.3. As can be seen from these data, the test substance is active in rats by the MES test within 15 to 30 minutes after p.o.
P.40
administration. Similarly, the total duration of action of this compound appears to be quite short; however, additional studies would be necessary to define the duration of action profile. The results summarized in Table 3.3 also suggest that the test compound is active in rats against clonic seizures induced by s.c. MET. In fact, the activity observed after p.o. administration to rats is more impressive than that observed in the mouse identification studies. At the doses tested (30 and 50 mg/kg), no evidence of neurologic deficit (toxicity) is observed. For the test substance profiled in Tables 3.2 and 3.3, the results suggest that further experimental work is justified because the favorable anticonvulsant profile supports possible clinical usefulness in generalized tonic-clonic seizures, complex partial seizures, and perhaps generalized myoclonic seizures (see later for discussion). This particular example also demonstrates the importance of not basing a “go/no-go” decision on the results obtained from one species after one route of administration. In this case, the p.o. rat data clearly are more favorable than those obtained from the mouse after i.p. administration, and provide the framework for additional testing.
|
TABLE 3.3. ANTICONVULSANT IDENTIFICATION IN RATS AFTER ORAL ADMINISTRATION |
||||||||||||||||||||||||||||||||||||||||||
|
||||||||||||||||||||||||||||||||||||||||||
The 6-Hz Psychomotor Seizure Test
The MES and s.c. MET tests have become the two most widely used seizure models for the early identification and high-throughput screening of investigational AEDs. These tests, albeit extremely effective in identifying new AEDs that may be useful for the treatment of human generalized tonic-clonic seizures and generalized myoclonic seizures, respectively (20), may miss novel AEDs that may be useful for the treatment of therapy-resistant partial seizures (e.g., levetiracetam). Although ineffective in the traditional MES and s.c. MET tests, levetiracetam has been found to be highly effective [median effective dose (ED50), 19 mg/kg, i.p.] in the 6-Hz seizure model that was originally described in the early 1950s (21,22). In light of the marked sensitivity of levetiracetam to the 6-Hz test, the ASP now routinely screens investigational AEDs found to be inactive in either the MES or s.c. MET tests for their ability to block seizures induced by a low-frequency (6-Hz), long-duration (3-second) stimulus delivered through corneal electrodes.
For this test, 20 mice are pretreated i.p. with 100 mg/kg of the test substance. At varying times (0.25, 0.5, 1, 2, and 4 hours) after treatment, individual mice (4 at each time point) are challenged with sufficient current (32 mA at 6 Hz for 3 seconds) delivered through corneal electrodes to elicit a psychomotor seizure. Typically, the seizure is characterized by a minimal clonic phase that is followed by stereotyped, automatistic behaviors that are not unlike automatistic behaviors of human patients with complex partial seizures. Animals not displaying this behavior (e.g., jaw chomping, whisker movement) are considered protected.
Results are expressed as the number of animals protected out of the number of animals tested over time. Drugs that are active (i.e., at least two of four animals protected at two or more time points) in the 6-Hz test are evaluated quantitatively, and active compounds may then become candidates for the kindled rat test. Drugs that are found to be active only in the 6-Hz identification test and subsequently found to be active in the kindled rat test represent potentially novel anticonvulsant substances for the treatment of therapy-resistant seizures. Such compounds are worthy of further investigation.
The results from these identification tests (MES and s.c. MET in mice and rats, 6-Hz in mice, and neurologic deficit in mice and rats) provide important preliminary information pertaining to oral bioavailability, species variation, duration of action, toxicity, efficacy, and overall potential of novel anticonvulsant substances.
QUANTIFICATION OF EXPERIMENTAL RESULTS
Anticonvulsant quantification details the ED50 by the MES, s.c. MET, or 6-Hz tests; the median toxic dose (TD50) by the rotorod test; the 95% confidence intervals; and protective indices (PI; TD50/ED50). The TPE data provide further insight into the time of onset and the duration of anticonvulsant activity and behavioral impairment. In addition to i.p. studies in mice, anticonvulsant quantification also is conducted in rats after p.o. administration to delineate anticonvulsant activity and behavioral impairment by a different route of administration in another rodent species, and to develop dose information prerequisite to subsequent chronic toxicity studies.
Time of Peak Effect
All quantitative studies are performed at the TPE. To determine the TPE for anticonvulsant activity, five groups of four animals each are administered an appropriate dose of test drug and subjected to the MES, s.c. MET, or 6-Hz test at 0.25, 0.5, 1, 2, or 4 hours. For toxicity determinations, a single group of eight animals is injected and tested for minimal motor impairment at the same time intervals. The time interval showing the greatest animal response is taken as the TPE.
Median Effective or Median Toxic Doses
Eight animals are injected with the dose used in the determination of the TPE and subjected to the respective anticonvulsant or behavioral impairment test. The number of animals responding is recorded and another dose level, usually one-half or double the initial dose, is selected. This procedure is repeated until a minimum of four dose levels has been established with at least two points between the dose level that induces 0% animal response and the dose level that induces 100% animal response.
The various ED50s and TD50s are calculated by a FOR-TRAN probit analysis program. This program also provides
P.41
the 95% confidence intervals, the slope of the regression lines, and standard error of the slopes. Reasonable estimates of these values may be determined by the log-probit method of Litchfield and Wilcoxon (23). This statistical treatment provides the kind of data essential for a critical evaluation of anticonvulsant activity and toxicity in structure-activity relationship studies.
ANTICONVULSANT DIFFERENTIATION
After the efficacy of a test substance against MES-, s.c. MET-, or 6-Hz-induced seizures has been quantitated, a battery of tests is used to characterize further the anticonvulsant potential of the substance. Other chemoconvulsants used include the type A γ-aminobutyric acid receptor (GABAA) antagonist BIC and the chloride channel blocker PIC. Ability of a candidate substance to alter seizure threshold also is assessed by the intravenous MET seizure threshold (i.v. MET) test (24). Candidate substances also may be tested for their ability to block the expression of stage 5 seizures in fully kindled animals, to block sound-induced seizures in the genetically susceptible Frings mouse model of reflex epilepsy (25), and to influence spike-wave electrographic seizure profile of the γ-hydroxybutyrate (GHB) model of absence (26).
In contrast to other seizure models, the audiogenic seizure-susceptible mouse model (not detailed later) is of no particular predictive value because it is nondiscriminatory with respect to clinical categories of anticonvulsant drugs (27). Nonetheless, it provides useful information regarding efficacy in a genetically susceptible model. The GHB model (also not described later), like the i.v. MET test, is used in the screening project to ascertain the proconvulsive potential of test substances that exhibit a phenytoin-like anticonvulsant profile in the other models examined, and to differentiate further those compounds that may be potentially useful for the treatment of spike-wave seizures.
Subcutaneous Bicuculline and Picrotoxin Tests
The CD97 of BIC (2.70 mg/kg) and PIC (3.15 mg/kg) is injected s.c. at the previously determined TPE for the test substance. The mice are placed in isolation cages and observed for the presence or absence of a clonic seizure. BIC-treated animals are observed for 30 minutes. PIC-treated animals are observed for 45 minutes because of the slower absorption of this convulsant. Absence of a clonic seizure indicates that the substance has the ability to elevate the seizure threshold to chemoconvulsants that act by antagonizing GABAAreceptors and blocking chloride channels, respectively. The activity of substances that exhibit anticonvulsant efficacy in these tests is quantitated by determining ED50s, as described previously.
Timed Intravenous Infusion of Metrazol
This test measures the minimal seizure threshold of each animal that has received the ED50 of the test substance (24). At the TPE, the convulsant solution (0.5% MET in 0.9% saline containing 10 USP units/mL of heparin sodium) is infused into the tail vein at a constant rate of 0.34 mL/min. The time in seconds from the start of the infusion to the appearance of the first twitch and the onset of clonus is recorded for each experimental and control animal. The times to each end point are converted to milligrams per kilogram of MET for each mouse in the vehicle control and the test drug group (10 mice/group), and the mean doses and standard errors are calculated.
Kindled Rats
None of the tests described thus far is very useful for identifying new drugs that are likely to be useful for the treatment of difficult-to-control seizure types and epilepsy syndromes such as adult complex partial seizures. In recent years, the kindling model has been a useful adjunct to the more traditional anticonvulsant tests for identifying a substance's potential utility for treating complex partial seizures. Kindled seizures provide not only an experimental model of focal seizures, but a means of studying complex brain networks that may contribute to seizure spread and generalization from a focus (28).
Of the various kindling paradigms described in the literature, the rapid hippocampal kindling model of Lothman et al. (28) appears to offer some distinct advantages for the routine screening and evaluation of new anticonvulsant substances. One potentially important advantage of the rapidly recurring hippocampal seizure model is its ability to provide a framework for assessing, in a temporal fashion, drug efficacy in a focal seizure model. Thus, this model has been incorporated into the ASP protocol.
For these studies, the candidate substance is evaluated for its ability to block the kindled motor seizure (seizure scores of 4 and 5) and limbic behavioral seizure (seizure score between 1 and 3) and to effect changes in the electrical afterdischarge duration. The procedures for surgical implantation of a bipolar electrode have been described previously (20). After a 1-week recovery period, animals are kindled to a stage 5 behavioral seizure (20) using a stimulus consisting of a 50-Hz, 10-second train of 1-millisecond, biphasic 200-µA pulses delivered every 30 minutes for 6 hours (12 stimulations per day) on alternating days for a total of 60 stimulations (5 stimulus days). Drug testing usually is initiated after a 1-week stimulus-free period. On each day of a drug trial, animals receive two to three suprathreshold stimulations delivered every 30 minutes before drug treatment. During these control blocks, the stability of the behavioral seizure stage and afterdischarge duration is assessed. Fifteen minutes after the last control block, a single dose of the test substance is administered i.p. After 15 minutes, each rat is
P.42
then stimulated every 30 minutes for 3 to 4 hours. After each stimulation, individual seizure scores and afterdischarge durations are recorded. The group mean and standard error of the mean are calculated for each parameter. After the last stimulation, animals are allowed 4 to 5 drug- and stimulus-free days between subsequent drug tests.
Using this approach, a drug that reduces the seizure score from 5 to 3 without any effect on afterdischarge duration would be presumed useful against secondarily generalized seizures. In contrast, those drugs that reduce the seizure score from 5 to less than 1, as well as reduce the electrographic afterdischarge, would be anticipated to be effective against focal seizures.
The kindled animal also is an important tool that is used in a limited capacity to identify drugs that prevent or attenuate the development of a seizure focus (i.e., antiepileptic versus anticonvulsant drugs). In an acquisition paradigm, animals begin receiving the test substance before initiation of the kindling process. A parallel control group receives vehicle, and at the TPE individual rats are challenged with the kindling stimulation protocol. On each of the kindling days, dosing always precedes the kindling stimulus. This paradigm is continued until the animals in the vehicle control group become fully kindled. After a 7- to 10-day stimulus-and drug-free period, animals in both groups are challenged with the kindling stimulus. If the seizure score and afterdischarge duration of the drug-treated animals remain significantly lower than in the control group, the treatment would be considered to have delayed or prevented the development of kindling. In addition to expanding the anticonvulsant profile of a candidate substance to its utility in a model of focal seizures, the kindling model is particularly useful for characterizing the anticonvulsant and antiepileptic potential of those “few” drugs that display activity only in the 6-Hz identification test.
MECHANISM OF ACTION STUDIES
One of the goals of the advanced testing conducted by the ASP is to conduct pilot electrophysiologic studies that may identify the molecular mechanism of action of the candidate substance. Although numerous molecular targets exist wherein anticonvulsants may exert an effect, the final common pathway appears to be through modulation of voltage-gated or neurotransmitter-gated ion channels (29, 30, 31, 32). Most of the prototype anticonvulsants are thought to exert their primary action by (a) reducing sustained, high-frequency, repetitive firing of action potentials by modulating voltage-dependent sodium (Na+) channels; (b) enhancing GABA-mediated inhibitory neurotransmission through a receptor-gated chloride channel; or (c) modulating neurotransmitter release and neuronal bursting through an effect on voltage-gated and receptor-gated calcium (Ca2+) channels. In addition, one of the newer anticonvulsant substances (i.e., retigabine) under clinical development has been found to activate a potassium channel comprising KCNQ2/Q3 potassium channel subunits (33, 34, 35). Loss-of-function mutations in this particular channel are thought to provide the basis for human benign familial convulsions (36, 37, 38) The common link among the various proposed mechanisms involves the ability of an anticonvulsant to modulate ion channel function. In light of this, the ASP uses the whole-cell patch-clamp technique (39) to assess the effect of promising candidate substances on current flow through voltage-gated and receptor-gated ion channels.
Mouse neuroblastoma cells (N1E-115) are recorded in conditions designed to assess selectively the effect of novel compounds on voltage-gated Na+ channels. Sodium currents are elicited by brief steps to 0 mV from a variety of holding potentials. The same voltage step paradigms are then performed in the presence of a range of concentrations of the candidate substance. Results from these studies provide information concerning the voltage sensitivity of drugs found to inhibit Na+ currents.
To assess the effects of the candidate substance on current flow through inhibitory and excitatory receptor-gated ion channels, the whole-cell patch-clamp technique is used to record currents evoked by exogenous application of subsaturating concentrations of GABA, kainate, or N-methyl-D-aspartate (NMDA). The respective ligand-gated currents are induced in cultured murine cortical neurons and agonists are applied either alone or in combination with the candidate substance, thus providing insight into how a compound exerts its anticonvulsant properties.
For some of the more recently developed anticonvulsants (e.g., levetiracetam), it has not been possible to identify the molecular activity contributing to its anticonvulsant action. For others, (e.g., felbamate, topiramate, and zonisamide), their anticonvulsant effect appears to be mediated by more than one molecular action. In this case, it is possible that one or more of a drug's mechanism of action interacts to dampen excitability in the neural networks that underlie seizure generation and propagation in ways that are not immediately obvious by examining individual receptor-gated and voltage-gated ion channels. Thus, to understand better the role that novel anticonvulsants play in modifying circuit behavior, the ASP has begun to examine the effects of anticonvulsants on synaptic transmission between pairs of monosynaptically connected neurons maintained in cell culture (40; Otto et al., personal communication). Inhibitory synapses formed by pairs of neurons in this culture system consist of postsynaptic GABAA receptors, whereas excitatory synapses have both non-NMDA and NMDA receptors that are colocalized to the synapse. Patch clamping both the presynaptic and postsynaptic neuron of a monosynaptically connected pair of neurons allows for the simultaneous examination of the effect of a compound on a number of presynaptic and postsynaptic neuronal parameters, including input resistance, resting membrane potential, action potential generation, neurotransmitter
P.43
release, uptake mechanisms, postsynaptic receptor-gated ion channels, presynaptic receptors, and short-term plasticity. Thus, pair recordings represent a powerful paradigm in which to assess the sum total of the actions of an anticonvulsant in a simple neuronal circuit.
EVALUATION OF ANTIEPILEPTIC POTENTIAL
The anticonvulsant potential of a candidate AED usually is assessed by comparing the results obtained with the test substance in well standardized test procedures with those obtained with the established AEDs. For the purpose of discussion, data for six established and seven second-generation AEDs subjected to several of the previously above described animal models after i.p. and p.o. administration to mice and rats are summarized in Tables 3.4 and 3.5. The TD50s and ED50s provide important information, but they reveal little when viewed alone. For the 13 drugs shown in Table 3.4, the TD50s range from 0.3 to >500 mg/kg. For the active compounds, the ED50s by the MES test range from 5.6 to 263 mg/kg, and those by the s.c. MET test range from 0.02 to 220 mg/kg. Considerably more can be learned from a comparison of PIs, or the TD50/ED50 ratio. The PIs, however, are based on the assumption that the slope of the
P.44
two regression lines (toxicity and anticonvulsant potency) are parallel. If the regression lines are parallel, the calculated PI is the same at any particular point on the regression lines. If the regression lines are not parallel, the PI is valid only at the median effective and toxic level. Above or below this median level, the PI may be either higher or lower. Therefore, in terms of drug tolerability, the calculated PI may be misleading. Ideally, an anticonvulsant drug should be capable of suppressing experimental seizures in all animals at dose levels devoid of even minimal toxic effects. Thus, it is more informative to calculate a “safety ratio” (TD3/ED97) for the candidate substance and to compare this with similar ratios for prototype drugs.
|
TABLE 3.4. ANTICONVULSANT PROFILE OF ESTABLISHED AND NEWER AEDs FOLLOWING I.P. ADMINISTRATION TO MICE |
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
TABLE 3.5. NEUROTOXICITY AND PROFILE OF ANTICONVULSANT ACTIVITY OF ORALLY ADMINISTERED PROTOTYPE ANTIEPILEPTIC DRUGS IN RATS |
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Albeit important, a comparative evaluation of a compound based on its PI and safety ratio provides little information about the overall anticonvulsant profile of a candidate drug compared with that of other AEDs tested in the same models. For example, in mice, the anticonvulsant profile of phenytoin, lamotrigine, topiramate, and zonisamide is relatively narrow compared with that of valproic acid, phenobarbital, tiagabine, and felbamate. Interestingly, in mice the anticonvulsant profile of carbamazepine includes activity against MES and PIC, whereas the anticonvulsant profile of gabapentin includes activity in the MES and s.c. MET tests. In addition, all of the sodium channel blockers listed (carbamazepine, phenytoin, and lamotrigine) display less than maximal efficacy in the 6-Hz seizure test.
In contrast, clonazepam and ethosuximide are both active in mice against clonic seizures induced by the chemoconvulsants MET, BIC, and PIC, but are virtually inactive (at nontoxic doses) against tonic extension seizures induced by MES. In contrast to the sodium channel blockers, those AEDs that act by elevating seizure threshold (e.g., clonazepam, ethosuximide, felbamate, phenobarbital, tiagabine, and valproic acid) are fully efficacious in the 6-Hz seizure test.
The anticonvulsant profile of levetiracetam clearly is unique among all of the AEDs listed in Tables 3.4 and 3.5. As discussed earlier, levetiracetam is inactive in the two traditional screening tests (i.e., MES and s.c. MET); however, levetiracetam displays potent activity in both the s.c. BIC and 6-Hz seizure tests (Table 3.5). Given the clinical utility of levetiracetam in partial seizures, these findings strongly support the inclusion of the s.c. BIC or the 6-Hz seizure test into the initial identification protocol of laboratories evaluating the anticonvulsant activity of investigational AEDs. In the case of levetiracetam, these data provide the rationale for pursuing more advanced testing in other “syndrome-specific” models such as kindling.
An evaluation of an investigational AED in the mouse tests summarized in Table 3.4 usually provides sufficient information to conclude whether a drug possesses anticonvulsant activity after i.p. administration and whether it is likely to have a narrow or broad spectrum of activity. However, it is not clear from these studies if an AED will have activity after p.o. administration. For this purpose, the ASP also evaluates each promising candidate substance in the rat MES and s.c. MET tests after p.o. administration (Table 3.5). Often, drugs are found to be more potent after p.o. administration in the rat MES test (e.g., topiramate) and less potent, or inactive, in the rat s.c. MET test. For example, both gabapentin and felbamate display activity in the mouse s.c. MET test (Table 3.4), yet both are inactive in this test after p.o. administration to rats (Table 3.5). Loss of activity in the rat s.c. MET test is not observed for all drugs active in the mouse s.c. MET test (e.g., clonazepam, ethosuximide, phenobarbital, tiagabine, and valproic acid). From a comparison of the results summarized in Tables 3.4 and 3.5, it appears that activity in the rat s.c. MET test is preserved for those AEDs that also are active in at least one of the other chemoconvulsant mouse models (i.e., s.c. PIC and s.c. BIC). The significance of this observation is not known.
Activity of a test substance in one or more of the electrical and chemical tests described previously provides some insight into the overall anticonvulsant potential of the test substance.
P.45
However, a concern voiced recently is that the procedures currently used in the search for novel AEDs are likely to identify “me too” drugs and are unlikely to discover those drugs with different mechanisms of action. In an attempt to address this concern, the ASP evaluates all compounds found to be active in one or more of the initial identification screens for their ability to limit focal and secondarily generalized seizures in the rapid hippocampal kindling model (28). For these studies, the two primary end points (i.e., seizure score and afterdischarge duration) are plotted as a function of time. Because of the relatively short refractory period observed in the hippocampal kindled rat, a time-effect curve can be obtained for each animal. Thus, in a limited number of kindled rats, the duration of action and degree of efficacy against focal or secondarily generalized seizures can be determined quite quickly. Drugs that reduce the seizure score to <1 and significantly attenuate the afterdischarge duration over a prolonged time are considered ideal candidates for further evaluation in this and other models of focal epilepsy.
The correlation between animal models of epilepsy, mechanisms of action, and clinical effectiveness of currently marketed AEDs has been reviewed by Macdonald and Kelly (30) and White (32). Most of the clinically effective AEDs decrease membrane excitability by interacting with ion channels, neurotransmitter receptors, metabolism, or uptake. The only real exception to this generalization is levetiracetam. Very little is known about the precise mechanism through which this AED exerts its effect. The ASP uses standard patch-clamp electrophysiologic techniques to evaluate the ability of an investigational AED to modulate voltage- and receptor-gated ion channels. In addition, the recently implemented pair recording experiments will allow us directly to evaluate the action of AEDs on a number of parameters that are involved in synaptic transmission and plasticity at both excitatory and inhibitory synapses in a simple neural circuit. These studies are conducted in an effort to differentiate the mechanistic profile of the investigational AED from that of the established AEDs. For example, when a drug is demonstrated to possess an anticonvulsant profile that is similar to that of phenytoin and lamotrigine, it would not be surprising to find that it also inhibits voltage-sensitive Na+ currents. The subsequent demonstration that it lacks activity at the voltage-sensitive Na+ channel or other molecular sites through which the traditional AEDs are thought to work would be indicative of a novel action. This information, albeit negative, becomes important when attempting to identify the “truly novel” AED. Several of the second-generation AEDs possess rather broad and unique mechanistic profiles compared with the older, established AEDs. For example, felbamate and topiramate are unique among all of the AEDs in that they exert a negative modulatory effect on glutamate currents mediated by the NMDA and α-amino-3-hydroxy-5-methyl-4-isoxazole propionate (AMPA) receptors, respectively. In addition to these activities, both substances also negatively modulate voltage-sensitive Na+ and Ca2+ currents and positively modulate GABAA receptor function. Topiramate also possesses the ability to inhibit type II and IV carbonic anhydrase. It is likely that the broad mechanistic profile of some of the newer AEDs contributes to their corresponding broad clinical profiles.
|
TABLE 3.6. CORRELATION BETWEEN CLINICAL UTILITY AND EFFICACY IN EXPERIMENTAL ANIMAL MODELS OF THE ESTABLISHED AND SECOND-GENERATION ANTIEPILEPTIC DRUGS |
|||||||||||||||||||||||||||||||||||||
|
|||||||||||||||||||||||||||||||||||||
The results obtained from the detailed evaluation of a given drug in the battery of seizure and epilepsy models described previously provide the preclinical basis for the clinical evaluation of a candidate antiepileptic substance. Furthermore, they provide some insight into the overall potential clinical utility of a candidate compound. As summarized in Table 3.6, drugs found to be active in the MES
P.46
test are likely to be useful for the management of generalized tonic-clonic seizures; whereas drugs active against fully expressed kindled seizures have all demonstrated activity in clinical trials against partial seizures. In the past, positive results obtained in the s.c. MET seizure test were considered suggestive of potential clinical utility against generalized absence seizures. This interpretation was based largely on the finding that drugs active in the clinic against partial seizures (e.g., ethosuximide, trimethadione, valproic acid, the benzodiazepines) were able to block clonic seizures induced by MET, whereas drugs such as phenytoin and carbamazepine that were ineffective against absence seizures also were inactive in the s.c. MET seizure test. Based on this argument, phenobarbital, gabapentin, and tiagabine should all be effective against spike-wave seizures, and lamotrigine should be inactive against spike-wave seizures. However, clinical experience has demonstrated that this is an invalid prediction. Thus, the barbiturates, gabapentin, and tiagabine all aggravate spike-wave seizure discharge, whereas lamotrigine has been found to be effective against absence seizures. As such, the overall utility of the s.c. MET test in predicting activity against human spike-wave seizures is limited. Before any firm conclusion concerning potential utility against spike-wave seizures is warranted, positive results in the s.c. MET test should be corroborated by positive findings in other models of absence, such as the GHB seizure test, the genetic absence epilepsy in rats from Strasbourg (GAERS) model, and thelhlh mouse (Table 3.6). With this exception, the data obtained in laboratory models correlate reasonably well with the clinical use of these agents and provide a basis for appropriately designed clinical trials that will ultimately determine the overall clinical potential of any given antiepileptic substance.
A further review of the data summarized in Table 3.6 demonstrates the importance of using multiple models in any screening protocol when attempting to identify and characterize the overall potential of a candidate antiepileptic substance. For example, levetiracetam is inactive in the traditional MES and s.c. MET tests, yet it demonstrates excellent efficacy in the GAERS model of primary generalized seizures and in the kindled rat model (1). Likewise, the efficacy of tiagabine and vigabatrin against human partial seizures was not predicted by the MES test, but by the kindled rat model (31,41). Furthermore, as mentioned previously, exacerbation of spike-wave seizures would not have been predicted by the s.c. MET test but by the other models (i.e., GHB, GAERS, and the lhlh mouse), wherein both drugs have been shown to increase spike-wave discharges (42). These examples serve to illustrate the limitations of some of the animal models while emphasizing their overall utility in predicting both clinical efficacy and potential seizure exacerbation. What is clear from this discussion is that there is a need to evaluate each investigational AED in a variety of seizure and epilepsy models. Only then will it be possible to gain a full appreciation of the overall spectrum of activity for a given investigational drug.
THE FUTURE OF ANTIEPILEPTIC DRUG DISCOVERY AND DEVELOPMENT
Since its inception in 1975, the ASP has screened over 24,000 investigational AEDs. In addition to the compounds that have been successfully developed, a number of additional compounds are in various stages of clinical development. Each of these drugs has brought about substantial benefit to the patient population in the form of increased seizure control, increased tolerability, and better safety and pharmacokinetic profiles. Unfortunately for 25% to 40% of patients with epilepsy, there remains a need to identify therapies that will more effectively treat their therapy-resistant seizures. As such, there is a continued need to identify and incorporate more appropriate models of refractory epilepsy into the AED screening process. There are several model systems that could be suggested, including the phenytoin-resistant kindled rat (43); the carbamazepine-resistant kindled rat (44); the 6-Hz psychomotor seizure model (21; Barton et al., personal communication); and the in vitro low-magnesium hippocampal slice preparation (45). Unfortunately, it will take the successful clinical development of a drug that is effective for the management of refractory epilepsy before any one of these (or other) model systems is clinically validated. Nonetheless, this should not prevent the community at large from continuing the search for a more effective therapy using the available models. In fact, until there is a validated model, it becomes even more important to characterize and incorporate several of the available models into the drug discovery process, while continuing to identify new models of refractory epilepsy.
There are no known therapies that are capable of modifying the course of acquired epilepsy. Attempts to prevent the development of epilepsy after febrile seizures, traumatic brain injury, and craniotomy with the older, established drugs have been disappointing (46). At the same time, discoveries at the molecular level have provided greater insight into the pathophysiologic process of certain seizure disorders. As such, it may be possible in the not-so-distant future to identify a treatment strategy that slows or halts the progression of epilepsy and prevents the development of epilepsy in susceptible individuals. However, any successful human therapy will necessarily be identified and characterized in a model system that closely approximates human epileptogenesis. There are several potential animal models wherein spontaneous seizures develop secondary to a particular insult or genetic manipulation. If we are to be successful in identifying a novel, disease-modifying therapy in the near future, we must become intentional in our efforts to characterize and incorporate such models of epileptogenesis into our screening protocols.
P.47
SUMMARY
In this chapter, the technical procedures used by the NIH-sponsored ASP are described. Special attention has been directed to the order in which these procedures are used and the use of these tests in the detection and quantification of anticonvulsant activity and minimal behavioral toxicity. Data obtained by subjecting six of the prototype AEDs (phenytoin, carbamazepine, ethosuximide, valproate, clonazepam, and phenobarbital) and seven of the second-generation AEDs (felbamate, gabapentin, lamotrigine, levetiracetam, tiagabine, topiramate, and zonisamide) to the anticonvulsant identification and quantification procedures are presented. Attention also is given to the role of anticonvulsant efficacy, acute behavioral toxicity, and protective indices in the evaluation of anticonvulsant potential. Last, the limitations associated with the current approaches are discussed. The rationale and need to broaden the scope of AED screening protocols to include models of therapy resistance and epileptogenesis also is discussed in context with the continuing need to identify more efficacious drugs for the 25% to 40% of patients who remain refractory to the currently available AEDs.
ACKNOWLEDGMENT
This work was supported by contracts (N01-NS-5-23-2, N01-NS-1-2347, N01-NS-4-2361, NO1-NS-9-2328, NO1-NS-4-2311, and NO1-NS-9-2313) from the National Institute of Neurological Disorders and Stroke, National Institutes of Health.
REFERENCES
P.48