Progress report on new antiepileptic drugs: A summary of the Ninth Eilat Conference (EILAT IX)

Introduction and rationale for development 

Brivaracetam (ucb 34714) is a novel chiral (with two asymmetric centers denoted above with an asterisk *) high-affinity synaptic vesicle protein 2A (SV2A) ligand which is currently in Phase III development for epilepsy. It has a 10-fold higher affinity for SV2A than levetiracetam (Keppra®) and, in contrast to levetiracetam, also displays inhibitory activity at neuronal voltage-dependent sodium channels (Table 3). The function of SV2A is not well established, but relates to synaptic vesicle exocytosis and neurotransmitter release (Xu and Bajjalieh, 2001). A strong functional correlation exists between SV2A binding affinity and anticonvulsant potency in animal models of both partial and generalized epilepsy (Kaminski et al., 2008).

Pharmacology 

Preclinical studies have shown that brivaracetam is more potent and efficacious than levetiracetam in animal models of seizures and epilepsy (Matagne et al., 2008). The activity of brivaracetam in a wide range of animal models suggests potential efficacy in a broad spectrum of seizure types (Table 1).

In vitro, brivaracetam (1–10μM) significantly suppresses evoked epileptiform responses recorded in the CA3 area of rat hippocampal slices following perfusion with a high potassium–low calcium containing fluid at concentrations 10-fold below those of levetiracetam (3.2μM vs. 32μM) (Matagne et al., 2008). Importantly, brivaracetam (3.2μM) also reduces the occurrence of spontaneous bursts, while levetiracetam (32μM) is inactive against this drug-refractory epileptiform marker.

The anticonvulsant profile of brivaracetam in both mice and rats is summarized in Table 1, Table 2. Brivaracetam is more potent than levetiracetam in protecting against secondarily generalized motor seizures in corneally kindled mice with ED50 values of 1.2mg/kg vs. 7.3mg/kg intraperitoneally (i.p.) and against clonic convulsions in audiogenic seizure-susceptible mice (ED50 values of 2.4mg/kg vs. 30mg/kg, i.p.) (Matagne et al., 2008). Brivaracetam differs from levetiracetam by its ability to protect against seizures induced by maximal electroshock (MES) or a maximal dose of pentylenetetrazol (PTZ). In contrast to levetiracetam, brivaracetam-induced nearly complete suppression of both motor seizure severity and after-discharge duration in amygdala-kindled rats and of spike-and-wave discharges in Genetic Absence Epilepsy Rats from Strasbourg (GAERS) (Matagne et al., 2008). In the mouse corneal-kindling model, chronic pretreatment twice daily with levetiracetam (1.7–54mg/kg, i.p.) and with 10-fold lower doses of brivaracetam (0.21–6.8mg/kg, i.p.) prior to corneal stimulation resulted in a similar suppression of kindling acquisition (Matagne et al., 2008). When compared to levetiracetam, continued corneal stimulations after termination of treatment revealed that brivaracetam possesses a persistent ability to counteract the kindling process.

Brivaracetam demonstrated potent and nearly complete seizure suppression in a rat model of acute, partially drug-resistant, self-sustaining status epilepticus induced by perforant path stimulation (Wasterlain et al., 2005). Brivaracetam shortened the cumulative duration of active seizures in a dose-dependent manner. For example, 20mg/kg brivaracetam, 200mg/kg levetiracetam, and 10mg/kg diazepam decreased the total duration of convulsive seizures to 11%, 35%, and 15% of control, respectively. Diazepam (1mg/kg) or brivaracetam (1mg/kg) alone had little effect, but when combined, reduced the duration of active seizures to 3% of controls. This study demonstrates the potent anticonvulsant effect of brivaracetam and suggests a marked synergy with diazepam in an animal model of status epilepticus (Wasterlain et al., 2005).

Toxicology 

Brivaracetam demonstrates low acute oral toxicity in mice, rats and dogs. Transient dose-related CNS effects were seen in mice, rats and dogs, generally at oral doses of 100mg/kg and above. No significant effects on the cardiovascular, respiratory or gastrointestinal systems were noted in these studies (UCB, data on file). Chronic oral toxicity studies showed that the target organ of brivaracetam toxicity is the liver and biliary tract.

Clinical pharmacokinetics 

Brivaracetam is rapidly and almost completely absorbed following oral administration with a median time to peak (tmax) of approximately 1–2h (Sargentini-Maier et al., 2007, Rolan et al., 2008). Pharmacokinetics is linear over a large dose-range (10–600mg after a single dose and 400mg/day after multiple doses). The compound is weakly bound to plasma proteins (≤20%) (Sargentini-Maier et al., 2008) and its volume of distribution is close to that of total body water (0.6L/kg). The terminal half-life of brivaracetam is approximately 8h and does not vary with dose.

Brivaracetam is eliminated primarily by metabolism, with the major metabolic pathways including hydrolysis of the acetamide group, cytochrome P450 (CYP)-mediated hydroxylation and a combination of these (Sargentini-Maier et al., 2008). The resulting metabolites are not pharmacologically active. The main circulating species is unchanged brivaracetam, amounting to 90% of the total plasma radioactivity following oral intake of 14C-brivaractam to healthy humans. Although the renal clearance of the parent compound is low (0.06mL/(minkg)), its metabolites have a high renal clearance and >95% of a radioactive dose is recovered in urine within 72h.

Population pharmacokinetic analysis from the two Phase II studies showed that most of the interindividual variability in brivaracetam pharmacokinetics, in an ethnically diverse population of patients, was accounted for by differences in body weight and concomitant use of enzyme inducing AEDs. Enzyme inducing AEDs resulted in a 30% reduction of brivaracetam AUC. Since the identified covariates had a modest influence on pharmacokinetic parameters, brivaracetam is deemed to have a predictable exposure in individual subjects. These results suggest that no dose adjustments due to drug interactions are required (Lacroix et al., 2007).

Pharmacokinetic studies in elderly and renally impaired subjects showed a pharmacokinetic profile of brivaracetam similar to that in healthy subjects (UCB, data on file). A study in subjects with hepatic impairment showed an increase in exposure to brivaracetam by up to 50–60% in severely impaired subjects (Child-Pugh Class C). The tolerability profile of brivaracetam in these special population studies was similar to that observed in healthy adults.

A randomized, double-blind, placebo- and positive-controlled thorough QT study of multiple doses of brivaracetam in 184 healthy subjects demonstrated the absence of any effects of brivaracetam on cardiac repolarization at both high therapeutic (150mg/day) and supra-therapeutic (800mg/day) doses (Rosillon et al., 2008).

Drug interactions 

Plasma concentrations of concomitant AEDs were monitored in the two Phase II studies (Otoul et al., 2007). Baseline and evaluation period measurements were obtained for 245 patients who received brivaracetam concurrently with stable doses of other AEDs, most commonly carbamazepine (41%), lamotrigine (27%), valproic acid (19%), levetiracetam (18%) or oxcarbazepine (15%). Brivaracetam did not appear to modify the steady-state plasma concentrations of carbamazepine, lamotrigine, levetiracetam, oxcarbazepine, topiramate or valproic acid in this population. Carbamazepine-epoxide was increased in a dose-dependent manner during evaluation vs. baseline (placebo 5.8%; brivaracetam 5mg 8.5%, 20mg 11.9%, 50mg 38.3%, 150mg 69.6%). The potential effect of brivaracetam on phenytoin was equivocal. These data suggest that no dose adjustment is required when brivaracetam is added to any of these AEDs, although caution might be considered when high doses of brivaracetam are added to patients who are on high doses of carbamazepine.

Efficacy data 

A phase IIa, subject-blind, placebo-controlled, single-dose study has shown that brivaracetam (10, 20, 40, or 80mg) suppresses generalized photoparoxysmal electroencephalographic (EEG) responses in patients with photosensitive epilepsy (Kasteleijn-Nolst Trenité et al., 2007). Eight of 18 patients (44%) had a response to placebo, whereas 17 of 18 patients (94%) had a response to brivaracetam. Complete abolishment of standard photosensitivity ranges (SPRs) was achieved in 14 of 18 (78%) brivaracetam-treated patients, and in none of the placebo-treated patients. All subjects receiving the brivaracetam 80mg single-dose exhibited complete suppression, lasting an average of 60h. The area under the effect curve was significantly correlated with the area under the plasma concentration curve of brivaracetam.

Two phase IIb, double-blind, randomized, placebo-controlled, parallel-group dose-ranging studies were conducted to evaluate the efficacy and tolerability of adjunctive brivaracetam therapy in patients aged 16–65 years with partial-onset seizures, with or without secondary generalization (French et al., 2007, van Paesschen and Brodsky, 2007). According to eligibility criteria, patients had to be inadequately controlled despite treatment with one or two AEDs, and to experience at least four partial-onset seizures during a 4-week prospective baseline. In study N01193, patients who met the eligibility criteria were randomized to placebo (n=54) or brivaracetam 5 (n=50), 20 (n=52) or 50 (n=52) mg/day administered twice daily without titration for 7 weeks (French et al., 2007). In study N01114, patients were randomized to placebo (n=52) or brivaracetam 50 (n=53) or 150 (n=52) mg/day administered twice daily over 3-week titration and 7-week stable-dose periods (van Paesschen and Brodsky, 2007). The retention rate in the brivaracetam groups was high in both studies, with between 92% and 98% of patients completing the study, compared with 91% and 92% in the placebo groups (Brodsky et al., 2007).

In study N01193 (French et al., 2007; UCB, data on file), a strong dose–response relationship was shown in (i) percent reduction over placebo in the partial-onset seizure frequency/week (primary efficacy analysis; 22.1%, p=0.004 at 50mg/day; 14.9%, p=0.062 at 20mg/day; 9.8%, p=0.240 at 5mg/day); (ii) reduction in partial-onset seizure frequency/week from baseline (53.1%, p<0.001 at 50mg/day; 42.6%, p=0.014 at 20mg/day; 29.9%, p=0.086 at 5mg/day vs. 21.7% on placebo); and (iii) 50% responder rate (55.8%, p<0.001; 44.2%, p=0.002; 32.0%, p=0.047 for brivaracetam 50, 20 and 5mg/day vs. 16.7% for placebo); and (iii) median percent reduction in partial-onset seizure frequency/week from baseline (53.1%, p<0.001 at 50mg/day; 42.6%, p=0.014 at 20mg/day; 29.9%, p=0.086 at 5mg/day vs. 21.7% on placebo). Seizure freedom over the full treatment period was achieved in 7.7%, 7.7%, 8.0% and 1.9% of patients at 50, 20, 5mg/day brivaracetam and placebo, respectively.

In study N01114 (van Paesschen and Brodsky, 2007; UCB, data on file), statistical significance vs. placebo was not achieved in the primary parametric efficacy analysis: the reduction in partial-onset seizure frequency/week over placebo during the 7-week maintenance period was 14.7% (p=0.093) on brivaracetam 50mg/day and 13.6% (p=0.124) on brivaracetam 150mg/day. The 50% responder rates were 39.6% (p=0.077) and 33.3% (p=0.261) on brivaracetam 50 and 150mg/day, respectively, vs. 23.1% on placebo). However, the 50mg/day dose reached significance for median percent reduction from baseline in seizure frequency/week (38.2% vs. 18.9% on placebo, p=0.017), which is indicative of a treatment effect. The higher brivaracetam dose (150mg/day) was not associated with greater efficacy (median percent reduction in partial-onset seizure frequency/week from baseline was 30.0%, p=0.113 vs. placebo). Seizure freedom over the 10-weektreatment period, including the up-titration phase, was achieved in 9.4% and 5.8% of patients on 50 and 150mg/day, respectively, compared with 1.9% on placebo.

Dose– and exposure–response population modelling based on pooled data from the two phase IIb studies demonstrated a dose–response relationship for adjunctive brivaracetam in 73% of patients with uncontrolled partial-onset seizures, with an ED50 of 21mg/day (Laveille et al., 2007). The maximum effective dose was predicted to be approximately 100mg/day, whereas a low dose of 5mg was predicted to result in small improvement over placebo. Age, body weight, gender, number of concomitant AEDs or the use of enzyme-inducing AEDs had no significant effect on the percentage of responders or on the magnitude of the response.

These studies suggest 20–50mg/day of brivaracetam as the appropriate therapeutic dose range, with possible additional benefit at doses up to 100mg/day.

Tolerability and adverse effect profile 

Adjunctive brivaracetam (5–150mg/day) was very well tolerated in adults with uncontrolled partial-onset seizures, despite the lack of an up-titration period in study N01193 (Brodsky et al., 2007). In the pooled analysis of the two phase IIb studies, the proportion of patients with treatment-emergent adverse events (TEAEs) was similar in the brivaracetam (all doses; 154/259, 59.5%) and placebo (64/106, 60.4%) groups. The most frequent TEAEs with an incidence of >5% in any group were nausea, vomiting, fatigue, nasopharyngitis, anorexia, convulsion, dizziness headache, somnolence, and insomnia. No difference in incidence greater than 3% was observed between the pooled brivaracetam and placebo groups for any of these adverse events. Eight brivaracetam patients (3.1%) and three placebo patients (2.8%) reported TEAEs that led to discontinuation of study medication. No major differences between the brivaracetam and placebo groups were found for any laboratory parameters or vital signs.

Two studies of brivaracetam as adjunctive treatment of myoclonus in patients (50 and 56, respectively) with Unverricht-Lundborg disease have recently been completed. Both trials failed to meet the primary endpoint of symptom relief of action myoclonus, but showed beneficial effects in a subset of patients.

Ongoing clinical studies 

Three phase III, randomized, double-blind, placebo-controlled, multicentre, multinational studies are ongoing in adults with epilepsy. Two of these are pivotal studies, designed to evaluate the efficacy and safety of brivaracetam (5, 20 and 50mg/day, and 20, 50 and 100mg/day without a titration period) over 12 weeks in patients with partial-onset seizures inadequately controlled by up to two concomitant AEDs. The aim of the third study is to investigate flexible doses of adjunctive brivaracetam (from 20 to 150mg/day) over 16 weeks in patients with inadequately controlled partial-onset or primary generalized seizures despite treatment with up to three concomitant AEDs.

Introduction 

Carisbamate (RWJ-333369; S-2-O-carbamoyl-1-o-chlorophenyl-ethanol) is a novel anticonvulsant (with one chiral center denoted above with asterisk *), initially developed by SK Biopharmaceuticals, under development for the treatment of epilepsy. It shows a broad spectrum of activity in preclinical models of epilepsy and has demonstrated a favorable efficacy and tolerability profile in a Phase II clinical trial. Phase III clinical trials are in progress.

Pharmacology 

Carisbamate exhibited potent and broad activity in acute rodent models of seizures and epilepsy including audiogenic seizures and seizures induced by MES, PTZ, bicuculline, and picrotoxin (Table 1; White et al., 2006). Carisbamate was also effective in reducing seizure severity in corneal-kindled rats and reducing seizure severity and after-discharge duration in the hippocampal-kindling model of partial epilepsy at doses well below those producing motor impairment (Table 2). In mice, carisbamate significantly elevated the threshold for i.v. PTZ-induced seizures even when tested at a dose high enough to produce motor impairment (White et al., 2006). The results from the i.v. PTZ and the MES tests suggest that carisbamate has the ability to control seizures by both elevating seizure threshold and preventing seizure spread. In the GAERS model, carisbamate (30 and 60mg/kg i.p.) significantly suppressed the duration of spike-and-wave discharges (Francois et al., 2008). Carisbamate (10 and 30mg/kg i.p.) significantly reduced the frequency of spontaneous recurrent seizures in the kainate post-status epilepticus model of temporal lobe epilepsy; in this model, carisbamate was more effective than topiramate, completely suppressing spontaneous recurrent seizures in a larger proportion of the rats studied (7/8 rats at 30mg/kg i.p. for carisbamate vs. 1/8 rats at 100mg/kg i.p. for topiramate) (Grabenstatter and Dudek, 2004, Grabenstatter and Dudek, 2008). In contrast to phenytoin, lamotrigine, and topiramate, carisbamate showed activity in two additional models of pharmacoresistant epilepsy, the 6-Hz seizure model and the lamotrigine-resistant amygdale-kindled rat, at doses that were devoid of associated motor impairment. In the 6-Hz seizure model, carisbamate retained its anticonvulsant activity as the stimulus intensity was increased from 22 to 44mA With ED50 values between 20.7 and 27.6mg/kg (Table 1; White et al., 2006). In the lamotrigine-resistant amygdala-kindled rat, carisbamate caused a dose-dependent reduction in behavioral seizure score and after-discharge duration (Table 2). When administered during the kindling acquisition phase, carisbamate has also been demonstrated to delay the development of amygdala kindling. In addition, unlike lamotrigine, carisbamate is still effective in treating fully kindled seizures even when the rat has been kindled in the presence of carisbamate (Klein et al., 2007). In the lithium-pilocarpine model of post-status epilepticus-induced epileptogenesis, carisbamate, when dosed 1 and 8h after the onset of status epilepticus, then twice daily for 6 days, reduced neuron loss in cortex, hippocampus, amygdala, and thalamus, and dose-dependently delayed or prevented the development of spontaneous recurrent seizures (Francois et al., 2005). These results suggest that carisbamate may have disease-modifying effects in epilepsy.

Overall, carisbamate possesses a broad-spectrum anticonvulsant profile in rodent models of generalized and partial epilepsy at non-toxic doses. The molecular actions of carisbamate that contribute to its broad-spectrum anticonvulsant activity have not been fully elucidated and remain under investigation.

Toxicology 

Carisbamate demonstrated anticonvulsant efficacy at doses well below those that produce CNS toxicity. In rotorod and inclined-screen tests in mice and observed minimal motor impairment in rats, the single-dose oral median toxic dose (TD50) for motor impairment was ≥137mg/kg, resulting in a protective index (TD50/ED50) of ≥17 (ED50=7.7mg/kg for mice and 4.4mg/kg for rats in the MES test) (White et al., 2006). The maximum tolerated single-dose ranged from 360 to 600mg/kg i.v. in mice and rats; these doses were associated with decreased activity, ataxia, sedation, and prostration.

Repeated-dose oral toxicity studies were conducted in adult rats up to 6 months and in adult dogs up to 12 months. No significant organ toxicity was observed in the repeated-dose studies at maximum plasma concentration (Cmax) and area under the curve (AUC) exposures 2- to 3-fold greater, respectively, than anticipated human therapeutic exposures. Toxicity was also evaluated in juvenile rats orally administered carisbamate from postnatal days 12 to 54. CNS adverse events were elicited only at doses of 80 and 160mg/kg in a pattern similar to adults. Moreover, carisbamate treatment of juvenile rats had no effect on learning, memory, or reproductive function.

Carisbamate tested negative for genetic toxicity in the bacterial reverse mutation test, the in vitro human lymphocyte chromosomal aberration assay, and the in vivo oral mouse bone marrow micronucleus assay.

Clinical pharmacokinetics 

Maximum plasma concentrations of carisbamate are achieved within 1–3h after single oral dosing. Bioavailability is approximately 94% and is not impacted by concomitant food intake (Yao et al., 2006). Carisbamate does not distribute extensively to peripheral tissues, as indicated by an apparent volume of distribution of about 50L. Plasma protein binding is approximately 44% and is concentration-independent. Total recovery in urine is approximately 94% of the administered dose, with approximately 2% as unchanged carisbamate.

The primary routes of metabolism of carisbamate include O-glucuronidation and carbamate ester hydrolysis with subsequent oxidation of the aliphatic side chain (Mannens et al., 2007). Minor routes include chiral inversion to the (R)-enantiomer followed by O-glucuronidation, and hydroxylation of the aromatic ring followed by sulfation. With the use of very sensitive liquid chromatography–mass spectrometry/mass spectrometry techniques, only traces of aromatic (pre)mercapturic acid conjugates were detected in urine (each <0.3% of the dose), suggesting a low potential for reactive metabolite formation. Atropaldehydes, reactive intermediates believed to be associated with the toxicity of the dicarbamate felbamate, were not found with carisbamate.

Carisbamate has a low apparent oral clearance (CL/F) of 3.4–4.2L/h, which equals <5% of liver blood flow. Thus, orally administered carisbamate to healthy subjects exhibits minimal hepatic first-pass metabolism (Yao et al., 2006). The Cmax and AUC of carisbamate are dose proportional, when given as single doses up to 1500mg and a regimen of up to 750mg twice daily (b.i.d). Steady-state conditions are achieved within 3 days, which is in agreement with a half-life of about 12h. The 12-h half-life will enable twice daily (b.i.d.) dosing with an immediate-release formulation (Yao et al., 2006). Plasma concentrations increased by a factor of 2 (compared with single dose) when carisbamate reached steady-state on a b.i.d. regimen. Consistent pharmacokinetics following single and repeat administration suggests that carisbamate does not induce or inhibit its own metabolism.

Drug interactions 

Administration of carisbamate 250 or 500mg b.i.d. had minimal impact on the activity of the CYP2D6 and CYP3A4 P450 isozymes, but minimally inhibited CYP2C9. Similar regimens produced no clinically significant changes in the pharmacokinetics of lamotrigine, carbamazepine, valproic acid, ethinylestradiol, or norethindrone. The plasma concentrations of carisbamate were reduced by 36% following co-administration with carbamazepine (Chien et al., 2006) and by approximately 10–20% with an oral contraceptive, presumably due to induction of glucuronidation. Lamotrigine and valproic acid had no clinically significant impact on the pharmacokinetics of carisbamate (Chien et al., 2007). When carisbamate was co-administered with phenytoin, mean AUC values at steady-state of carisbamate were 47% lower relative to the values in subjects who received carisbamate alone. This interaction was unidirectional, because carisbamate did not alter the pharmacokinetics of phenytoin (data on file). Administration of carisbamate 200mg b.i.d. had no impact on the pharmacokinetics or pharmacodynamics of warfarin or ethanol (data on file).

Only clinically insignificant differences were observed in mean plasma Cmax and AUC0–48 values of carisbamate in subjects with moderate and severe renal impairment compared with subjects with normal renal function. Two glucuronide metabolites of carisbamate were noted to accumulate in subjects with moderate and severe renal impairment, but hemodialysis reduced the mean plasma Cmax and AUC values of each metabolite by approximately 50% (data on file).

Following a single 200mg dose of carisbamate, mean AUC was approximately 16% greater in subjects with mild hepatic impairment and 107% greater in subjects with moderate hepatic impairment compared with subjects with normal hepatic function. The pharmacokinetic profile of carisbamate in subjects with moderately impaired liver function is also characterized by a prolongation in the mean terminal half-life (20.7h) compared with healthy subjects (10.5h) (Moore et al., 2008).

Efficacy data 

A randomized, double-blind, placebo-controlled, dose-ranging Phase IIb study in refractory epilepsy patients with partial-onset seizures treated adjunctively (n=537 randomized) explored daily dosages of 100, 300, 800, and 1600mg. Carisbamate was efficacious at doses of 300, 800, and 1600mg (Fig. 1). Initial seizure frequency reduction occurred during titration, within the first month of treatment, and efficacy remained apparent through the 16-week study. Two subsequent identically designed, randomized, double-blind, placebo-controlled Phase III studies examined daily dosages of 200 and 400mg. The 400mg dosage was efficacious in one of the two studies, but the 200mg dosage did not separate from placebo in either study. An additional analysis of the efficacy of carisbamate in the presence and in the absence of UGT-inducing AEDs revealed that, for both the 200 and 400mg daily dosages in both of these studies, noninduced subjects had a greater reduction in seizure rate and a greater responder rate than induced subjects.

View Large Image Download to PowerPoint

Figure 1. Percentage reduction from baseline in partial-onset seizure frequency. Median values for percent reduction are noted above the bars for each group. For carisbamate (CRS) vs. placebo: p=0.07 (100mg/day), p<0.001 (300mg/day), p=0.006 (800mg/day), and p<0.001 (1600mg/day). *Statistically significant vs. placebo.

Tolerability and adverse effect profile 

The tolerability of carisbamate at efficacious doses was excellent in the dose-ranging trial. Discontinuation due to TEAEs was less than placebo for doses of 100 and 300mg/day (5% and 6%, respectively, for carisbamate vs. 8% for placebo). The most common drug-related TEAEs were dose-dependent and similar to previous clinical experience, and included headache (24%, placebo; 27%, 100mg/day; 23%, 300mg/day; 28%, 800mg/day; 42%, 1600mg/day), dizziness (5%, placebo; 7%, 100mg/day; 10%, 300mg/day; 17%, 800mg/day; 28%, 1600mg/day), somnolence (7%, placebo; 12%, 100mg/day; 13%, 300mg/day; 12%, 800mg/day; 17%, 1600mg/day), and nausea (6%, placebo; 7%, 100mg/day; 10%, 300mg/day; 13%, 800mg/day; 14%, 1600mg/day). These were all generally mild and transient, or resolved with dose reduction or discontinuation.

Elevations of serum alanine aminotransferase (ALT) 3 times above normal values occurred in 4-carisbamate-treated subjects in the dose-finding trial; 1 was receiving 800mg/day and 3 were receiving 1600mg/day. Similar elevations have been seen in other subjects receiving carisbamate at dosages ranging from 400 to 2000mg/day. Two of the 752 carisbamate-treated patients (0.3%) in the 2 Phase 3 studies had clinically significant elevations of ALT, including 1 patient with acute hepatitis B infection. These subjects have been asymptomatic, and all ALT elevations have resolved, in some cases with continued treatment with carisbamate at the same or lower dosage. No other patterns of clinically significant abnormalities in laboratory evaluations were detected in this study.

Ongoing clinical studies 

More than 80% of subjects who completed the double-blind phase of the randomized phase II trial described above entered an open-label extension. The longest exposure in subjects with epilepsy has been more than 4 years, demonstrating the favorable retention of carisbamate in Phase II studies. Based upon these results, Phase III trials of carisbamate for adjunctive use in partial-onset seizures are in progress to further explore the therapeutic potential of carisbamate in the treatment of epilepsy.

Introduction and rationale for development 

2-Deoxy-d-glucose is a chemical analogue of glucose currently in preclinical development as a novel anticonvulsant and antiepileptic compound. 2-Deoxy-d-glucose differs from normal glucose only by removal of a single oxygen atom at the 2 position, and has been used for decades as a tracer in auto radiography and positron emission tomography (PET) imaging. As a close chemical analogue of glucose, 2-deoxy-d-glucose is taken up into cells by normal glucose transport mechanisms, but unlike glucose cannot undergo metabolism and thereby acts as an inhibitor of glycolysis.

The anticonvulsant and antiepileptic properties of 2-deoxy-d-glucose were discovered during experiments investigating the mechanisms of action of the ketogenic diet, a high-fat, adequate-protein, carbohydrate-restricted diet which is effective in reducing seizures in up to half of drug-refractory patients. A remarkable feature of the diet is that ingestion of even a small amount of carbohydrate by patients who achieved seizure control can rapidly reduce the diet's effectiveness and result in recurrence of seizures. With this background clinical observation, preliminary experiments in hippocampal slices revealed that removal of glucose and substitution of alternative energy substrates such as lactate or pyruvate reversibly suppressed epileptic discharges in the CA3 region of the hippocampus, and administration of 2-deoxy-d-glucose, a known inhibitor of glycolysis, was subsequently demonstrated to have anticonvulsant effects in vitro and in vivo against a variety of acute and chronic models of seizures (Stafstrom et al., submitted for publication).

Pharmacology 

Pharmacokinetics and cerebral uptake mechanisms 

2-Deoxy-d-glucose is rapidly absorbed and distributed after oral or parenteral administration, and the kinetics of tissue distribution and clearance have been extensively studied for examination of glucose utilization and use of 2-deoxy-d-glucose as an imaging agent (Sokoloff et al., 1977, Newman et al., 1990). In pioneering studies of 2-deoxy-d-glucose for measurement of tissue glucose utilization in rats, 2-deoxy-d-glucose was rapidly distributed into tissues and brain after i.v. administration with a plasma half-life of about 5min and peak brain distribution obtained at about 10min (Sokoloff et al., 1977). Studies of human administration by the oral route for adjuvant therapy of cancer have reported rapid absorption with a peak plasma concentration at 0.5–1h (Raez et al., 2007) and a half-life of 48min after i.v. administration of 50mg/kg in children (Hansen et al., 1984).

2-Deoxy-d-glucose undergoes activity-dependent acute loading into neurons as a function of neural activity, which potentially is highly advantageous for anticonvulsant therapeutic effects in regions of neural circuitry activated by seizures. As an analogue of glucose, 2-deoxy-d-glucose freely passes the blood–brain barrier and is delivered preferentially to brain regions in response to energy demand, specifically regions of circuitry generating seizures. Glucose and 2-deoxy-d-glucose are delivered to brain regions in response to energy demand through an exquisitely regulated system of neurovascular coupling that increases regional blood flow in response to neural activity (Berwick et al., 2008). The activity-dependent physiological process of neurovascular coupling depends on a system of regulation involving vascular cells, perivascular neurons, and astrocytes. This system operates with an impressive dynamic range such that local vascular perfusion and glucose supply are increased in response to neural activity within seconds and within a few hundred microns of activity in neural circuits (Berwick et al., 2008). As a consequence of local neurovascular coupling, delivery of glucose and 2-deoxy-d-glucose is increased in regions of high neural activity, as occurs in circuitry undergoing synchronization during seizures.

As an analogue of glucose, 2-deoxy-d-glucose uptake into cells in response to neural activity occurs through glucose transporters and is preferentially increased in cells with increased energy consumption and metabolic demands. Like glucose, 2-deoxy-d-glucose is a substrate for hexokinase, and undergoes phosphorylation at the 6 position to 2-deoxy-d-glucose-6-phosphate. Unlike glucose-6-phopsphate, 2-deoxy-d-glucose-6-phosphate cannot undergo isomerization and is not further metabolized by glucose-6-phosphate isomerase to fructose-6-phosphate. 2-Deoxy-d-glucose-6-phosphate becomes “trapped” in cells due to relatively slow dephosphorylation by phosphatases, thereby inhibiting subsequent steps of glycolysis (Sokoloff et al., 1977, Newman et al., 1990). As 2-deoxy-d-glucose-6-phosphate is “trapped” in cells with high rates of glucose utilization, radiolabeled substitutes such 3H and 14C substituted 2-deoxy-d-glucose have been used for autoradiographic and quantitative measurement of glucose utilization (Sokoloff et al., 1977, Newman et al., 1990).

Toxicology 

There is an extensive experience with administration of radiolabeled positron emitting 18F-2-deoxy-d-glucose in humans during decades of use as a PET imaging tracer. The doses required for anticonvulsant and antiepileptic effects in preclinical animal models (30–200mg/kg) are greater by orders of magnitude, and administration of 2-deoxy-d-glucose in this dose range has received approval for Phase 1 clinical trials as an adjuvant therapy for cancer (Raez et al., 2007). In addition, preliminary preclinical chronic toxicity including studies of spatial memory in the Morris water maze have revealed lack of effects on spatial memory at doses 1g/kg b.i.d. which are nearly 30 times greater than the minimum dose which reduces kindling progression. Doses of 500mg/kg administered daily for 6 months in rats had no overt adverse systemic effects (Stafstrom and Sutula, personal communication, manuscript in preparation).

Ongoing and planned studies 

Continuing efficacy and preclinical toxicology studies are planned in anticipation of filing of an IND with the FDA. With the unusual property of focal activity-dependent loading of 2-deoxy-d-glucose into regions of circuitry during seizures or in response to focal stimulation with indwelling electrodes, and rapid absorption and uptake in solid dosage forms and parenteral formulations, it is anticipated that clinical trials of 2-deoxy-d-glucose could include protocols of administration immediately after seizures, during status epilepticus, and in conjunction with device therapies, in addition to conventional trial designs with scheduled oral administration.

Introduction 

Eslicarbazepine acetate (ESL) is a third-generation, single-enantiomer (with one chiral center denoted above with asterisk *) member of the long-established family of first-line dibenz/b,f/azepine AEDs represented by carbamazepine (first-generation) and oxcarbazepine (second-generation) (Bialer, 2006). ESL shares with carbamazepine and oxcarbazepine the dibenzazepine nucleus bearing the 5-carboxamide substitute but is structurally different at the 10,11-position. This molecular variation results in differences in metabolism and is expected to result in improved tolerability and ease of administration (once daily dosing). Unlike carbamazepine, ESL is not metabolized to carbamazepine-10,11-epoxide and is not susceptible to induction of its own metabolism. Unlike oxcarbazepine, which is metabolized to both eslicarbazepine (also called S-licarbazepine, S-MHD or BIA 2-194) and (R)-licarbazepine (also called R-MHD or BIA 2-195), ESL is a prodrug of eslicarbazepine, which is the drug entity responsible for ESL pharmacological effects (Almeida and Soares-da-Silva, 2007).

Pharmacology 

As summarized in Table 1, Table 2, ESL has been evaluated in several animal models predictive of anticonvulsant efficacy. It blocks tonic extension seizures in the MES test in rats and mice and limbic seizures in the corneal kindled mouse and amygdala-kindled rat (Almeida et al., 2008, Almeida et al., in press). ESL displays weak effects against clonic seizures induced by PTZ, bicuculline, picrotoxin, and 4-aminopyridine. In mice, ESL shows analgesic activity in the formalin paw test and in the chronic constriction nerve injury pain model of neuropathic pain.

The precise mechanism of action of ESL is not known. In vitro electrophysiological studies indicate that both ESL and eslicarbazepine competitively interact with site 2 of the inactivated state of a voltage-gated sodium channel, preventing its return to the active state and repetitive neuronal firing (Almeida and Soares-da-Silva, 2007). Much like carbamazepine and oxcarbazepine, ESL inhibits release of the neurotransmitters or neuromodulators glutamate, GABA, aspartate and dopamine in rat striatal slices (Parada and Soares-da-Silva, 2002, Almeida and Soares-da-Silva, 2007). Effects at the voltage-gated sodium channels are likely to contribute to ESL ability to limit sustained repetitive firing, ictogenesis and seizure spread (Table 3).

Toxicology 

There were no findings of special concern for human use in conventional preclinical studies of safety pharmacology, toxicology, genotoxicity, reproductive toxicity and carcinogenicity. In acute toxicology studies, the estimated lethal dose of ESL, in the mouse and in the rat, was 500mg/kg when administered by oral gavage, and 100 and <50mg/kg when administered by i.v. bolus, respectively. An increase in activated partial thromboplastin time (aPTT) was reported in chronic toxicity studies in the Beagle dog but no similar changes were reported in other species, including man. An increase in plasma cholesterol was reported in rats and dogs, but this was not reported in other species, including man (Almeida et al., 2008, Almeida et al., in press).

Clinical pharmacokinetics 

Eslicarbazepine is the main active metabolite of ESL and represents about 95% of the total systemic drug exposure following oral administration of ESL. After oral administration, the plasma levels of ESL usually remain below the limit of quantification. Thus, the pharmacological activity of ESL is primarily exerted through eslicarbazepine (Almeida et al., 2008, Almeida et al., in press).

Peak eslicarbazepine concentrations are attained at 2–3h post-dose. Bioavailability is considered high because the amount of metabolites recovered in urine corresponds to more than 90% of an ESL dose. Food has no effect on ESL pharmacokinetics (Almeida et al., 2008, Almeida et al., in press). The binding of eslicarbazepine to plasma proteins is relatively low (<40%) and independent of concentration (Almeida and Soares-da-Silva, 2007).

ESL is rapidly and extensively biotransformed to eslicarbazepine by hydrolytic first-pass metabolism. In studies in healthy subjects and adults with epilepsy, the terminal half-life of eslicarbazepine was 10–20 and 13–20h, respectively. Peak plasma concentrations of eslicarbazepine are attained at 2–3h post-dose and steady-state concentrations are attained after 4–5 days of once daily dosing, consistent with an effective half-life in the order of 20–24h. Minor metabolites in plasma are (R)-licarbazepine and oxcarbazepine, which were shown to be active, and the glucuronide conjugates of ESL, eslicarbazepine, (R)-licarbazepine and oxcarbazepine. The pharmacokinetics of eslicarbazepine is linear and dose-proportional, both in healthy subjects and in patients with epilepsy (Almeida et al., 2008, Almeida et al., in press).

The metabolites of ESL are eliminated from the systemic circulation primarily by renal excretion, in unchanged form and as glucuronide conjugates. In total, eslicarbazepine and its glucuronide conjugates account for more than 90% of the metabolites excreted in urine. Approximately two thirds of urinary eslicarbazepine is in free (unconjugated) form, and one third is conjugated with glucuronic acid (Almeida et al., 2008, Almeida et al., in press).

A study in patients with mild to severe renal impairment showed that eslicarbazepine clearance is dependent on renal function (Maia et al., 2008). A dosage adjustment is necessary in patients with creatinine clearance below 60mL/min. Haemodialysis was shown to be effective in removing ESL metabolites from plasma. Moderate liver impairment has no relevant effects on ESL pharmacokinetics.

Drug interactions 

In vitro studies have shown that the plasma protein binding of eslicarbazepine was not affected to a major extent by the presence of warfarin, diazepam, digoxin, phenytoin and tolbutamide. The binding of warfarin, diazepam, digoxin, phenytoin and tolbutamide was not significantly affected by the presence of eslicarbazepine (Almeida et al., 2008, Almeida et al., in press).

In in vitro studies in human liver microsomes, eslicarbazepine had no relevant inhibitory effect on the activity of CYP1A2, CYP2A6, CYP2B6, CYP2D6, CYP2E1, CYP3A4 and CYP2C9, and only a moderate inhibitory effect on CYP2C19. The 50% inhibitory concentration (IC50) value for inhibition of CYP2C19 activity by eslicarbazepine was 232mg/L. Studies with eslicarbazepine in fresh human hepatocytes showed no significant induction of CYP1A2, CYP3A and phase II enzymes involved in the glucuronidation and sulfatation of 7-hydroxy-coumarin (Almeida et al., 2008, Almeida et al., in press).

Incubation of 14C-ESL in the presence of acetazolamide, clobazam, clonazepam, gabapentin, lamotrigine, phenobarbital, phenytoin, primidone and valproic acid showed no relevant inhibition of ESL metabolism by these AEDs. Co-administration of ESL 1200mg once daily with warfarin showed a mild, although significant, decrease in exposure to (S)-warfarin, with no significant effect on the (R)-warfarin pharmacokinetics. Since (S)-warfarin clearance is mediated almost entirely by CYP2C9, whereas (R)-warfarin clearance is dependent on multiple CYP pathways (CYP2C19, CYP3A4 and CYP1A2), these results suggest that ESL may have a mild inducing effect on CYP2C9 (Almeida et al., 2008, Almeida et al., in press).

Despite the fact that glucuronidation is the major metabolic pathway for both eslicarbazepine and lamotrigine, studies in healthy subjects and population pharmacokinetic data in patients showed no relevant effect of ESL on lamotrigine pharmacokinetics. Population pharmacokinetics analysis of data from Phase III studies in adults with epilepsy also showed no relevant effect of ESL on the clearance of carbamazepine, phenytoin, topiramate, clobazam, gabapentin, phenobarbital, levetiracetam and valproic acid. Overall, no need for dose adjustment for any of the above AEDs is anticipated when ESL is added on to the therapeutic regimens of patients receiving these medications (Almeida et al., 2008, Almeida et al., in press).

Efficacy data 

Following an early phase II placebo-controlled, adjunctive-therapy study in which ESL was found to be efficacious and well tolerated in adults with refractory partial-onset seizures (Elger et al., 2007), three Phase III trials have been completed recently in a total of 1050 patients enrolled at 125 sites in 23 countries. All three studies used a multicentre, randomised, double-blind, placebo-controlled parallel-group design and included patients with at least four simple or complex partial-onset seizures per 4 weeks despite treatment with one to three AEDs. Each study consisted of an 8-week baseline period, followed by double-blind 2-week titration and a double-blind 12-week maintenance period. There were three ESL dose groups (400, 800 or 1200mg once daily) in two studies, but only two ESL dose groups (800 or 1200mg once daily) in one study. Patients completing the double-blind treatment could enter an optional 1-year open-label ESL extension phase (Almeida et al., 2008, Almeida et al., in press).

Between 64% and 75% of patients in each of the Phase III studies were using two concomitant AEDs, the most common of which was carbamazepine (approximately 60% of subjects, overall). After carbamazepine, the next most common concomitant AEDs were valproic acid, lamotrigine, levetiracetam, topiramate, phenytoin and phenobarbital and clobazam, which were used in up to 30% of subjects in any treatment group (Almeida et al., 2008, Almeida et al., in press).

The main efficacy endpoints were seizure frequency standardized per 4 weeks (primary endpoint in all studies), median relative reduction in standardized seizure frequency, and 50% responder rate (proportion of patients with ≥50% decrease in seizure frequency) in the intent-to-treat population. The adjustment is as follows: number of seizures/number of days of treatment×28 days. At the dosages of 800 and 1200mg once daily, ESL was associated with a statistically significant decrease in standardised seizure frequency (p<0.0001 for both dosages vs. placebo) and median relative reduction in seizure frequency (placebo: 8.5%; 800mg: 29.4%, p<0.0001; 1200mg: 30.6%, p<0.0001). Statistically significant differences in responder rates in comparison to placebo were found in each of the three studies for the 800 and 1200mg treatment arms; no difference between the 400mg and placebo arms was found in any study (Almeida et al., 2008, Almeida et al., in press).

Pharmacokinetic/pharmacodynamic analysis in the pooled population included in Phase III epilepsy studies showed continuous relationships, with moderate inter-subject variability, between antiepileptic efficacy and eslicarbazepine plasma concentrations. These relationships were not affected by concomitant AEDs. In a completed 1-year open-label extension, significant improvements in quality of life (QOLIE-31) and depressive symptoms (MADRS) were found in comparison to baseline (Almeida et al., 2008, Almeida et al., in press).

Tolerability and adverse effect profile 

In the pooled population of adults with epilepsy included in placebo-controlled studies, 388 (45.3%) treated with ESL and 82 (24.4%) treated with placebo reported possibly related TEAEs. Possibly related TEAEs with an incidence >2% were (ESL vs. placebo) dizziness (18.8% vs. 5.7%), somnolence (11.2% vs. 7.4%), nausea (6.5% vs. 2.4%), diplopia (6.3% vs. 1.2%), headache (5.5% vs. 2.1%), vomiting (4.8% vs. 1.2%), coordination abnormal (4.4% vs. 1.8%), vision blurred (3.5% vs. 0.9%), vertigo (2.1% vs. 0%) and fatigue (2.1% vs. 1.8%). The incidence of possibly related TEAEs was dose-dependent.

In the Phase III studies, the overall incidence of TEAEs leading to discontinuation was low (4.5% with placebo, 8.7% with ESL 400mg, 11.6% with 800mg and 19.3% with 1200mg). The incidence of rash was low: 1 (0.3%) patient on placebo, 1 (0.5%) on ESL 400mg, 3 (1.1%) on 800mg, and 9 (3.2%) on 1200mg. Hyponatraemia <125mmol/L was only reported in 4 patients: 1 (0.5%) on ESL 400mg, 2 (0.7%) on 800mg and 1 (0.4%) on 1200mg. TEAEs were usually mild to moderate in intensity and occurred predominantly during the first weeks of treatment. After 6 weeks, no relevant differences on the incidence of TEAEs were apparent between patients treated with ESL and patients treated with placebo. The incidence of behavioral or psychiatric adverse events was low (<1% of patients for any possibly related TEAE reported in either the ESL or placebo groups) (Almeida et al., 2008, Almeida et al., in press).

The favorable efficacy and safety profiles of ESL 800 and 1200mg, including sustained efficacy for at least 1 year and lack of concerns related to rash, hyponatraemia or body weight increase, indicate that ESL is a valuable addition to the current adjunctive-therapy armamentarium for partial seizures. Although the incidence of TEAEs was approximately similar in the ESL 800 and 1200mg groups, TEAEs in the ESL 800mg group were less likely to be possibly related to the study medication and to lead to discontinuation of study medication. Thus, the ESL 800mg dose appears to offer the best benefit to risk ratio.

Ongoing and planned studies 

Currently there are ongoing studies with ESL in children with epilepsy and in patients with neuropathic pain.

Introduction 

Ganaxolone (3α-hydroxy-3β-methyl-5α-pregnan-20-one) is the 3-β-methyl analog of the neurosteroid allopregnanolone, a metabolite of progesterone. Ganaxolone does not have detectable classical nuclear hormone activity and, unlike allopregnanolone, cannot be biotransformed to metabolites with such activity. Four Phase II clinical trials are ongoing, two controlled trials and two open-label trials. The first controlled trial is being conducted in adults with partial-onset seizures and the second is in pediatric patients with infantile spasms (Nohria and Giller, 2007).

Pharmacology 

Ganaxolone is a positive allosteric modulator of GABAA receptors with potency and efficacy comparable to those of its endogenous neurosteroid analog allopregnanolone (Carter et al., 1997). Neurosteroids have two separate actions on GABAA receptors: they potentiate the action of GABA and directly activate the receptor at two distinct sites that are different from the benzodiazepine or barbiturate sites (Hosie et al., 2006). Although neurosteroids modulate both synaptic and extrasynaptic GABAA receptors, their modulatory action is enhanced for extrasynaptic GABAA receptor isoforms that contain a δ subunit (Herd et al., 2007).

Ganaxolone has protective activity in diverse rodent seizure models, including clonic seizures induced by PTZ and bicuculline, limbic seizures in the 6Hz model, and amygdala and cocaine-kindled seizures (Carter et al., 1997, Leśkiewicz et al., 2003, Kaminski et al., 2003, Kaminski et al., 2004, Rogawski and Reddy, 2004). In chronically treated rats, tolerance does not occur to the anticonvulsant activity of ganaxolone (Reddy and Rogawski, 2000).

In addition to the anticonvulsant activity, there is evidence that neurosteroids can retard the development of spontaneous recurrent seizures in some animal models of epileptogenesis, and therefore they have antiepileptogenic actions in such models (Biagini et al., 2006, Biagini et al., in press). In a rapid kindling epileptogenesis model in adolescent rats (P35), pre-treatment with ganaxolone (20 and 40mg/kg), delayed the occurrence, and decreased the number of full limbic seizures (stage 4–5) in a dose-dependent manner. Continuous treatment with ganaxolone in the drinking water beginning on day 4 after pilocarpine-induced status epilepticus in adult rats did not alter the onset of spontaneous recurrent generalized tonic–clonic seizures but transiently reduced their duration. Moreover, ganaxolone-treated animals had markedly reduced damage in the CA3 hippocampal region (G. Biagini, unpublished).

Toxicology 

Acute ganaxolone treatment is associated with reversible, dose-related sedation. In studies performed to assess CNS side effects in mice, ganaxolone showed less of an interaction with ethanol than did valproic acid on the hanging wire-mesh test of ataxia and ganaxolone affected cognitive function in an animal passive-avoidance paradigm only at doses causing ataxia (Marinus Pharmaceuticals, data on file). Chronic dog and rat toxicity studies as well as embryo-fetal and peri-postnatal development studies have not revealed any significant toxicity associated with ganaxolone. The maximum tolerated doses in these studies have been limited by sedation. Chronic toxicology studies continue to support the long-term administration of ganaxolone (Marinus Pharmaceuticals, data on file).

Clinical pharmacokinetics 

Pharmacokinetic studies in healthy subjects have been carried out with ganaxolone formulated as a β-cyclodextrin complex suspension in a variety of paradigms, including single doses up to 1600mg and rising-dose multiple dosing studies using the β-cyclodextrin complex suspension (once daily doses up to 500mg for 17 days, and 300mg b.i.d. doses for 10 days). Ganaxolone showed a linear (dose proportional) pharmacokinetics after single doses in the range of 50–600mg. Plasma levels over 300ng/mL were usually associated with significant somnolence in healthy volunteers, although some subjects with levels as high as 612ng/mL did not report any treatment-related adverse events.

After oral ingestion, ganaxolone is rapidly distributed into tissues and exhibits an effective half-life of approximately 10h and a terminal elimination half-life of 35–40h. In humans, ganaxolone is primarily metabolized by CYP3A4/3A5. The occurrence of CNS adverse effects appears to correlate with tmax as well as with the absolute Cmax value. Ganaxolone-cyclodextrin suspensions have exhibited pronounced food effects, with plasma exposure (AUC) values after intake with food being 5–15-fold higher than exposure values recorded after intake in the fasting state.

Marinus Pharmaceuticals has created several unique formulations of ganaxolone to address the variable absorption under different dietary conditions. These include two solid capsule forms, one providing immediate release and the other providing pH-sensitive delayed release. In addition, a new suspension form (50mg/mL) has been developed for pediatric studies. None of these formulations contain cyclodextrins.

In a Phase I study in healthy subjects with the new suspension there was greater bioavailability in the fasted state than with cyclodextrin complexed forms. Exposures after intake with food were comparable to those seen with the historical cyclodextrin complex and AUC values after intake with food were no more than 2–3-fold greater than those observed after intake in the fasted state. In adults with partial-onset seizures, doses up to 500mg t.i.d., the highest dose regimen currently being tested, were well tolerated. Similarly, doses of up to 54mg/(kgday) (18mg/kg t.i.d.) in 4–24-month-old infants with infantile spasms were well tolerated.

Preliminary pharmacokinetic data from a study in healthy volunteers with the immediate-release capsule dosage form (400mg, oral) demonstrated comparable performance in the fed and fasted state to that of the new suspension, with a slightly prolonged tmax value to the new suspension formulation. A favorable tolerability profile was obtained in healthy volunteers with a 400mg immediate-release solid dose form Preliminary modeling data of 400mg b.i.d. doses of immediate-release ganaxolone capsules estimates a Cmax of approximately 150ng/mL and Cmin of 50ng/mL at steady-state. Cmin values in responders during the historical pre-surgical trial and preliminary analysis of responders (at least a 50% reduction in seizures over a 10 weeks period compared to baseline) in the ongoing adult refractory partial-onset trial gave mean Cmin values of 30–50ng/mL. This analysis indicates that b.i.d. dosing is possible with the new immediate-release ganaxolone capsules. Completion of the ongoing partial-onset trial and an ongoing steady-state pharmacokinetic trial in healthy volunteers are needed to support this initial analysis.

Drug interactions 

In vitro drug–drug interaction studies have revealed no significant interactions with other AEDs. Ganaxolone has a high plasma protein binding (>99%) but it did not demonstrate any protein binding interactions with valproic acid in vitro. In rats, enzyme induction with phenobarbital attenuates the exposure of a ganaxolone dose by approximately 5-fold.

Formal human in vivo drug–drug interaction studies have not yet been carried out. The ongoing Phase II trials will evaluate drug–drug interactions between ganaxolone and other AEDs. Analyses of available data from historical pediatric add-on studies provided no evidence for significant drug–drug interactions between ganaxolone and concomitant AEDs in that subjects receiving known CYP inducers such as phenobarbital, carbamazepine and phenytoin had response rates similar to those recorded in subjects not receiving these drugs.

Clinical efficacy 

In an 8-day treatment-refractory pre-surgical trial, ganaxolone monotherapy (1875mg/day) or placebo was administered to 52 patients, of whom 24 received ganaxolone and 28 placebo (Laxer et al., 2000, Monaghan et al., 1997). Efficacy of ganaxolone compared to placebo was supported by an intent-to-treat survival analyses showing a trend toward significance (p=0.0795; log-rank test). Fifty percent (12/24) of those treated with ganaxolone completed the full study period compared to 25% (7/28) of those on placebo. Lack of statistical significance was attributed to insufficient power and a higher than usual placebo response rate. However, a covariate analyses revealed a significant treatment effect on the survival time in males (p=0.0306). A post hoc chi square test showed that significantly more ganaxolone-treated patients completed the full 8- or 10-day (the first 2 subjects were enrolled before a protocol amendment and received 10 days of dosing) study period as per protocol than did placebo patients (p=0.04).

In three Phase II open-label, adjunctive-therapy studies, 79 pediatric subjects (ages 6 months to 15 years) received ganaxolone. All children had refractory seizures despite up to 3 concurrent AEDs and the majority of them had a history of infantile spasms. On average, each subject had been previously treated with 7 AEDs, including ACTH, vigabatrin, valproic acid and benzodiazepines.

In one of the pediatric studies, 20 patients aged 0.6–7 years with refractory infantile spasms or with continuing seizures after a history of infantile spasms received ganaxolone for up to 12 weeks at a dose titrated up to 36mg/(kgday) (Kerrigan et al., 2000). Sixteen subjects completed the study. The 4 early discontinuations were due to noncompliance (1), withdrawal of consent (1), flu (1), and mild leucopenia (1). The child who withdrew due to mild leukopenia reported no tonic seizures in the 7 weeks with ganaxolone treatment compared to 4 tonic seizures during the 3-week baseline. The responses of the 15 patients with infantile spasms or a history of infantile spasms (1 of the 16 completers did not have a history of infantile spasms) who completed the study were considered substantial (>50% reduction in spasms) in 5 (one of whom became spasm-free), and moderate (25–50% reduction) in 5. Five patients were considered nonresponders (<25% reduction).

In a second pediatric study, 15 patients aged 5–15 years with partial or generalized seizures refractory to conventional AEDs received ganaxolone for approximately 10 weeks (2 weeks titration, 8 weeks maintenance) with a starting dose of 1mg/kg b.i.d. to a maximum dose of 12mg/kg t.i.d. (Pieribone et al., 2007). Thirteen of the 15 patients were evaluated at maintenance week 4. Two subjects withdrew from the study, one due to non-compliance and one due to an adverse event (hyperexcitation, auto-hetero-aggression, and hallucination). Four subjects were considered substantial (>50% reduction in spasms) responders (1 seizure-free), 3 moderate (25–50% reduction) responders, and 6 nonresponders (<25% reduction). Of the 8 subjects evaluated at maintenance week 8, 4 were substantial responders (1 of whom was a nonresponder at maintenance week 4), 2 moderate responders and 2 nonresponders. Twelve subjects continued treatment beyond a year, and one subject continued into compassionate use for 3.8 years.

In the third pediatric study, 45 subjects aged 2–15 years with partial or generalized seizures refractory to conventional AEDs received ganaxolone for approximately 17 weeks (9 weeks titration, 8 weeks maintenance) with a starting dose of 1mg/kg b.i.d. to a maximum dose of 12mg/kg t.i.d. Twenty-seven (60%) of the 45 subjects completed the study to the end of maintenance week 8 and provided baseline and maintenance seizure data. Twelve (27%) subjects were considered responders (≥50% reduction in seizure frequency, which included all seizure types). Thirty-five of the 45 subjects enrolled experienced epileptic spasms. Of these, one showed complete remission at maintenance week 8, and 10 additional subjects had a reduction in spasm frequency of >50%. Thus, approximately 31% (11/35) showed a response to treatment.

In total, 29 pediatric subjects continued ganaxolone treatment in compassionate use protocols. Ten of the 29 subjects reported behavioral improvement as one of the reasons for continuing ganaxolone treatment. Eighteen subjects were treated for over 1 year: 4 for 4–5 years, 2 for 3–4 years, 1 for 2–3 years, and 11 for 1–2 years. The improvement in seizure frequency, along with observations of behavioral improvement, support the conduct of a controlled trial in infantile spasms.

Tolerability and adverse effect profile 

The majority of adverse events in Phase I studies were judged as mild (74%) or moderate (18%) in intensity, and all resolved spontaneously. Eighty-one percent (198/243) of subjects reported at least 1 adverse event (180/216 ganaxolone subjects; 18/27 placebo subjects). The principal dose-limiting adverse event observed in all clinical studies has been somnolence. No pattern of serious adverse events has been seen to date. Ganaxolone appears to be safe and well tolerated at the doses being tested currently (Marinus Pharmaceuticals, data on file). The ongoing trials will further define the adverse event profile.

Ongoing and planned studies 

Approximately 200 subjects are expected to be enrolled in an ongoing double-blind, randomized, placebo-controlled, add-on study in adults with uncontrolled partial-onset seizures. Subjects are randomized to ganaxolone or placebo in a 2:1 ratio after an 8-week baseline and undergo a 2-week dose escalation up to a dose of 1500mg/day (500mg t.i.d.), followed by an 8-week maintenance period. The primary efficacy assessment is based on a comparison between mean weekly seizure frequencies (ganaxolone vs. placebo) during the maintenance period. An exploratory analysis will be performed to assess effects on catamenial seizures. This study is expected to end in October 2008. Initial results suggest that ganaxolone is well tolerated based on the reported adverse events and low dropout rate (10%) after randomization at this point in the trial.

In the second ongoing trial, infants and young children (ages 4–24 months) diagnosed with infantile spasms are being enrolled in a double-blind, placebo-controlled, dose-escalation study of the new ganaxolone suspension to a target dose of 54mg/(kgday) (18mg/kg t.i.d.). Subjects are randomized to either ganaxolone or placebo in a 2:1 ratio for the first 8–10 days of treatment (period 1), after which all subjects will receive ganaxolone (incomplete crossover) for an additional 8–10 days (period 2). Assessment of the effects on spasm frequency will be based on 24-h video-EEG monitoring at baseline and after each treatment period. The primary efficacy measure is a comparison between spasm frequency (ganaxolone group vs. placebo group) after period 1. Fifty-six subjects enrolled into this study and all 56 have completed. Initial results indicate that subjects tolerate the target dose based on the safety profile observed in the crossover portion of the trial.

Further Phase I studies are planned to support ongoing formulation work, and additional clinical trials will be planned based on the findings of the ongoing studies.

Introduction 

Huperzine A is a sesquiterpene Lycopodium alkaloid isolated from Chinese club moss (Huperzia serrata), also known as the Chinese folk medicine Qian Ceng Ta, traditionally used in China for swelling, fever and inflammation, blood disorders and schizophrenia (Zangara, 2003, Ward and Caprio, 2006). Huperzine A is approved and used in China for the treatment of Alzheimer's disease. It is classified as a dietary supplement by the FDA and is widely available in synthesized form in health food stores or via the Internet, labeled as a memory aid.

Pharmacology 

Huperzine A was active against subcutaneous (s.c.) PTZ- but not MES-induced seizures following p.o. administration to Swiss-Webster mice, with peak anticonvulsant activity at 1h (White et al., 2005). At doses of 1, 2, and 4mg/kg, a maximum of 62.5% protection was observed. Impairment on the rotarod test was observed in 75% and 100% of mice tested at doses of 2 and 4mg/kg, respectively, with a TD50 value of 0.83mg/kg.

In the 6-Hz model, ED50 values for i.p. Huperzine A were 0.28, 0.34 and 0.78mg/kg for 22, 32, and 44mA, respectively (Schachter et al., 2006). Repeat ED50 for 44mA was 0.58 (0.40–0.82)mg/kg, i.p.

Planned study 

A pilot dose-ranging trial to evaluate the tolerability and efficacy of adjunctive Huperzine A in patients with medically refractory epilepsy is planned.

Acknowledgment 

The investigators are grateful for the support of the NINDS Anticonvulsant Screening Program and a grant from The Epilepsy Research Foundation (S.S.).

Introduction 

JZP-4 (3-(2,3,5-trichloro-phenyl)-pyrazine-2,6-diamine) is a novel, potent sodium and calcium channel blocker, structurally related to lamotrigine and endowed with broad-spectrum anticonvulsant activity. Its pharmacokinetic properties may allow it to be titrated rapidly to the appropraite therapeutic dose range with a low risk of serious cutaneous reactions.

Pharmacology 

Toxicology 

JZP-4 was negative in genotoxicity, mutagenicity, fertility and teratogenicity tests. In completed repeated dosing studies ranging from 26 weeks in rats and 39 weeks in beagle dogs, JZP-4 had no overt toxicity beyond the ataxia and sedation produced by high, toxicological doses. Although high concentrations of JZP-4 had effects in the hERG and Purkinje fiber studies, the safety margin compared with expected therapeutic serum concentrations was wider than that of lamotrigine. In further support of these conclusions, ECG studies in rats and beagles showed no significant change in any ECG parameter.

Clinical pharmacokinetics 

A mass balance study to assess the metabolism and excretion of JZP-4 in healthy non-elderly subjects showed that the oral suspension form of JZP-4 was rapidly absorbed and readily eliminated from plasma. Total recovery was approximately 90%, with the primary route of excretion in urine. The majority of circulating total radioactivity was potentially associated with metabolites. JZP-4 recovery in urine was low, with less than 0.1% of the administered dose recovered as unchanged JZP-4. JZP-4 and its metabolites are not highly associated with red blood cells. Oral clearance, terminal phase half-life and apparent terminal-phase volume of distribution in healthy non-elderly adults are about 45L/h, 10h and 640L, respectively. Disposition is biexponential, with the terminal phase appearing at about 12h after a single dose.

Drug interactions 

In vitro studies showed that JZP-4 had no inhibitory effect on CYP2E1 and weak inhibitory effects (IC50>100μM) on CYP2A6, CYP2C9, CYP2C19, CYP2D6, and CYP3A4. JZP-4 exhibited IC50 values that were greater than 50μM for CYP1A2 and greater than 10μM for CYP2C8 and CYP2B6. JZP-4 had no significant inhibitory effect on UGT1A1, UGT1A4 or UGT1A9 at the maximum tested concentration of 100μM, and exhibited IC50 values >100μM for UGT1A6, UGT2B7 and UGT2B15. Since clinically efficacious plasma concentrations of JZP-4 are expected to be substantially less than 10μM, JZP-4 is believed to have a very low potential to affect the pharmacokinetics of co-administered drugs that are substrates of these enzymes. JZP-4 was also evaluated for its enzyme induction potential in isolated human hepatocytes. The results suggested that JZP-4 is not an enzyme inducer.

A crossover clinical pharmacokinetic study in healthy volunteers was conducted in which a single oral 50mg dose of JZP-4 was administered alone and on day 8 of a 10 days course of valproic acid (500mg b.i.d). Results showed that JZP-4 Cmax and AUC increased by about 15% during co-administration with valproic acid, however half-life remained unchanged. The single doses of JZP-4 did not affect the pharmacokinetics of valproic acid. These results are unlikely to be clinically significant.

Planned and ongoing studies 

JZP-4's antiepileptic properties are currently being evaluated in two proof-of-concept studies in humans. Single oral doses of JZP-4 ranging from 50 to 200mg are compared to a single oral dose of Lamotrigine (325mg) in their ability to decrease cortical excitability following transcranial magnetic stimulations in healthy male volunteers (Funk et al., 2008). In a second study involving photosensitive epilepsy patients on background anti-epileptic drugs, single oral doses of JZP-4 ranging from 25 to 200mg are compared to a single oral dose of levetiracetam (1000mg) or valproic acid (1000mg) in their ability to decrease photo-paroxysmal EEG responses (Kasteleijn-Nolst Trenité et al., 2008).

Additional clinical drug–drug interaction studies and in vitro interaction studies with p-glycoprotein are also in progress.

Introduction 

Lacosamide (SPM 927, formerly known as harkoseride), or (R)-2-acetamido-N-benzyl-3-methoxypropionamide, is a synthetic chiral derivative (with one asymmetric carbon denoted above with asterisk *) of the amino acid D-serine. In initial studies, lacosamide demonstrated potent antiepileptic activity in various animal models used at NINDS (Stoehr et al., 2007).

Lacosamide has been developed for the adjunctive treatment of partial-onset seizures in adults (age ≥16) and is being further developed for additional indications in epilepsy. It is also undergoing clinical evaluation for the monotherapy treatment of diabetic neuropathic pain, fibromyalgia and migraine prophylaxis.

Lacosamide is being developed in multiple bioequivalent formulations. An oral tablet is complemented by an isotonic i.v. solution (10mg/mL) which may be administered without dilution or the need for dose-adjustment and is intended as temporary replacement therapy for those unable to tolerate oral medicine. Additionally, lacosamide is available as oral syrup (15mg/mL).

Pharmacology 

Lacosamide protects against a variety of seizure types in animal models, including primary generalized seizures (PTZ test), reflex epilepsy (Frings audiogenic seizure-susceptible mouse), refractory partial-onset seizures (6-Hz psychomotor model, hippocampal kindling), secondarily generalized seizures (MES model) and status epilepticus (pilocarpine-, cobalt-homocysteine, and electrical-stimulation models) (Stoehr et al., 2007; Table 1, Table 2). Furthermore, lacosamide demonstrates potential neuroprotective properties in status epilepticus, as well as potential antiepileptogenic properties (amygdala kindling model, status epilepticus-induced chronic epilepsy model) (Brandt et al., 2006).

As summarized in Table 3, lacosamide demonstrates a novel mechanism of action profile by selectively enhancing the slow inactivation of voltage-gated sodium channels, as well as modulating the activity of collapsin response mediator protein-2 (CRMP-2) (Errington et al., 2006). This latter mechanism has been suggested to mediate a putative neuroprotective and neuroregenerative effect (Czech et al., 2004), although its exact role in lacosamide's anticonvulsant mechanism of action remains to be determined.

Toxicology 

Lacosamide at high doses was associated with exaggerated CNS effects that were dose-limiting in all species (mice, rats, rabbits and dogs). However, preclinical data revealed no special hazards for humans based on conventional studies of safety pharmacology, repeated dose toxicity, genotoxicity and carcinogenic potential when used at plasma concentrations similar to those observed in clinical trials. At maternally toxic doses, embryotoxicity was noted in rats and rabbits, but there was no evidence of teratogenicity or reproductive defects. There was no evidence for age-specific toxicity in a study with juvenile rats.

Clinical pharmacokinetics 

Drug interactions 

Lacosamide does not significantly bind to plasma proteins (≤15%). Although there is limited metabolism by CYP2C19, studies in healthy subjects involving CYP2C19-inhibiting drugs and poor CYP2C19-metabolizers demonstrated no clinically relevant effects on lacosamide plasma concentrations. Specific drug-interaction studies involving carbamazepine, valproic acid, omeprazole, metformin, digoxin and an oral contraceptive (ethinylestradiol and levonorgestrel) also demonstrated no relevant interactions influence on the pharmacokinetics of these drugs or lacosamide. Hence, lacosamide is considered to have a low potential for drug–drug interactions.

Efficacy data 

The efficacy of lacosamide 200–600mg/day as adjunctive therapy for refractory partial-onset seizures in adults (ages 16–70) has been demonstrated in three large phase IIb/III randomized, placebo-controlled trials (Ben-Menachem et al., 2007, Halasz et al., 2006, Chung et al., 2007). All three of these trials were similar in design, thus allowing for pooled efficacy analyses. Each trial was powered to evaluate efficacy based on dual primary endpoints comparing the baseline to the maintenance phase (i.e. 50% responder rate and reduction in seizure frequency), and thus suitable for evaluation by regulatory bodies in both the E.U. and U.S.

In the initial pivotal dose-ranging phase IIb trial (SP667, E.U./U.S.; Ben-Menachem et al., 2007), 421 patients taking 1–2 concomitant AEDs were randomized (1:1:1:1) to receive either placebo or lacosamide 200, 400, or 600mg/day. Following an 8-week baseline phase, subjects were titrated over a period of 6-week (i.e. 100mg/week), proceeding to a 12-week maintenance phase, which was then followed by a 2-week transition or 3-week taper period, with a subsequent option for continued open-label treatment. The 50% responder rate in the SP667 intent-to-treat analysis population was higher in the lacosamide 200mg/day group (32.7%, p=0.090) than in the placebo group (21.9%), and the difference in responder rate vs. placebo reached statistical significance for both the lacosamide 400mg/day (41.1%, p=0.004) and lacosamide 600mg/day (38.1, p=0.014) groups. Percent reduction in seizure frequency per 28 days over placebo was 14.6% in the lacosamide 200mg/day group (p=0.101) and reached statistical significance for both the lacosamide 400mg/day (28.4%, p=0.002) and the lacosamide 600mg/day (21.3%, p=0.008) groups.

In the most recently completed large phase III trial (SP754, U.S.; Chung et al., 2007), 405 patients taking 1–3 concomitant AEDs were randomized (1:2:1) to receive placebo, lacosamide 400mg/day or lacosamide 600mg/day. Following an 8-week baseline phase, subjects were titrated over a period of 6-week (i.e. 100mg/week) proceeding to a 12-week maintenance phase which was then followed by a 2-week transition/taper period, with a subsequent option for continued open-label treatment. The 50% responder rate in the SP754 intent-to-treat analysis population demonstrated statistically significant results for both the lacosamide 400mg/day (38.3%, p<0.004) and lacosamide 600mg/day (41.2, p<0.001) groups compared to placebo (18.3%). Statistically significant percent reductions in seizure frequency per 28 days over placebo were seen for both lacosamide 400mg/day (21.6%, p=0.008) and lacosamide 600mg/day (24.6%, p=0.006).

Another large phase III trial (SP755, E.U./Australia; Halasz et al., 2006) was conducted in 485 patients taking 1–3 concomitant AEDs who were randomized (1:1:1) to receive placebo, lacosamide 200mg/day or lacosamide 400mg/day. Following an 8-week baseline phase, subjects were titrated over a period of 4-week (i.e. 100mg/week), proceeding to a 12-week maintenance phase which was then followed by a 2-week transition/taper period, with a subsequent option for continued open-label treatment. The 50% responder rate in the SP755 intent-to-treat analysis population was higher in the lacosamide 200mg/day group (35.0%, p=0.070) than in the placebo group (25.8%), and the difference in responder rate vs. placebo reached statistical significance in the 400mg/day group (40.5%, p=0.006). Statistically significant percent reductions in seizure frequency per 28 days over placebo were seen for both lacosamide 400mg/day (15.0%, p=0.033) and lacosamide 200mg/day (14.4%, p=0.022).

Subsequent analysis of pooled efficacy data from these randomized placebo-controlled trials further supports the overall efficacy of lacosamide at doses of 200–600mg/day. Pooled analysis demonstrate that complete seizure freedom during the maintenance period was achieved in 2.7%, 3.3% and 4.8% of patients randomized to lacosamide 200, 400, and 600mg/day, respectively, compared with 0.9% in the placebo group. These results were obtained in a refractory population, with 84% of subjects taking two or three concomitant AEDs (including substantial numbers on newer AEDs such as topiramate and levetiracetam) and 17% being additionally treated with the vagal nerve stimulator. Overall, subjects reported an average duration of epilepsy of 24 years and median baseline seizure frequencies of 10–17/28 days. Nearly half (45%) reported a lifetime use of ≥7 other AEDs.

Data from open-label extension trials suggests that the efficacy of lacosamide is maintained over time and that tolerance to its seizure suppressing effects does not occur. In long-term assessment studies, patients completing 1, 2 and 3 years of long-term lacosamide therapy demonstrated improvement in their mean number of seizure-free days (48, 59 and 68 days/year, respectively) when compared to their pre-lacosamide baseline (on average 219 seizure-free days).

Tolerability and adverse effect profile 

The lacosamide epilepsy development program includes 9 phase II/III trials involving ≥1300 adults (ages 16–70) with refractory partial-onset seizures who received oral lacosamide doses up to 800mg/day, representing >2200 patient-years of exposure. From this population, 199 subjects also received lacosamide i.v. solution at doses up to 800mg/day over infusion durations of 10–60min.

The use of lacosamide has been associated with reports of CNS and gastrointestinal TEAEs which are typically mild-moderate in nature. Overall, those reported most frequently (i.e. ≥5%) as compared to placebo in forced-titration trials, including subjects on up to 3 concomitant AEDs, can be related to dose and include dizziness (22.4%), diplopia (8.6%), nausea (7.0%), vomiting (6.8%), abnormal coordination (6.5%) and blurred vision (5.9%).

In randomized, controlled trials overall discontinuation rates for the treatment period were 8.1%, 17.2% and 28.6% for lacosamide doses of 200, 400 and 600mg/day as compared to 5.2% for placebo and were generally associated with the CNS and gastrointestinal events described above.

Lacosamide is typically not associated with clinically meaningful changes in laboratory parameters, vital signs or body weight. The use of lacosamide in controlled trials in patients with epilepsy was associated with a small (1.4–6.6ms), dose-related increase in the PR-interval of the ECG, which was not associated with an increase in cardiac-related adverse events and occurred in the absence of other meaningful ECG changes. A definitive, randomized controlled QTc trial in healthy volunteers also showed this small dose-related PR-interval change, as well as no effect on the QT interval and no other notable ECG findings.

In general, TEAEs reported during long-term open-label trials did not differ from those reported in shorter-term randomized trials, suggesting that long-term lacosamide therapy has not yet been associated with emergence of unexpected adverse events, and that the incidence of overall adverse events does not increase with continuing exposure to lacosamide.

The tolerability and side effect profile of lacosamide i.v. solution is similar to that of the oral product. Occasionally, injection-site reactions were reported, which were mild and self-limited in nature. Treatment with i.v. lacosamide can replace oral lacosamide therapy without the need for dilution or dose adjustments.

Planned studies 

Lacosamide is being thoroughly developed for use in patients with epilepsy, including ongoing or planned trials in pediatric populations and in patients with primary generalized seizures, and monotherapy trials.

In addition, lacosamide is being studied in areas outside of epilepsy, including diabetic neuropathic pain, fibromyalgia and migraine prophylaxis.

Introduction and rationale for development 

Galanin is an endogenous neuropeptide in the CNS which has been recognized as a potential anticonvulsant agent since the pioneer work of Mazarati and coworkers (Mazarati et al., 1992, Mazarati et al., 1998, Mazarati et al., 2001, Mazarati et al., 2006, Mazarati and Lu, 2005).

When injected directly into the lateral brain ventricle or hippocampus, galanin decreased the severity of picrotoxin-induced kindled convulsions in rats (Mazarati et al., 1998). In one animal model of status epilepticus, perihilar injection of galanin, before or after perforant path stimulation, shortened the duration of seizures (Mazarati and Lu, 2005). These effects were reversed by co-application of galanin antagonists. However, galanin's marginal metabolic stability and inability to penetrate the blood–brain barrier when delivered systemically has precluded its development as an AED. Recently, this compound has been modified to yield a series of galanin agonists that demonstrate anticonvulsant activity following systemic administration (Bulaj et al., 2008). NAX 5055 is one of the prototype compounds that has been evaluated extensively following systemic administration. Proof-of-concept studies with NAX 5055 have validated the technology platform by demonstrating efficacy in rodent seizure, epilepsy and pain models following oral, i.v., i.p. and s.c. administration. NAX 5055 has exhibited potent, long-lasting and dose-dependent activity in these animal models (Table 1, Table 2).

Peptide- and protein-based drugs have long been recognized as having distinct advantages over their small molecule chemical counterparts Peptides, as natural products endogenous in the human body, are often more potent and specific to their in vivo receptor subtype targets, potentially resulting in fewer adverse effects. NAX 5055, for instance, maintains high affinity for galanin receptor subtypes GalR1 and GalR2 after systemic delivery through the blood–brain barrier.

Pharmacology 

Structure-activity relationship studies demonstrated that the N-terminal fragment GAL(1-16) retains potent agonist activity at hippocampal galanin receptors. A series of truncated galanin analogs in which some non-essential amino acid residues at the C-terminal part of the Gal(1-16) fragment were replaced by cationic and lipoamino acid residues were rationally synthesized. Based on available structure-activity relationship data, the following modifications were introduced to the GAL(1-16) analogs: (i) the Gly1 residue was replaced by Sarcosine. N-methylation of Gly1 does not affect galanin receptor affinity, but the capping of the N-terminal amino group is likely to decrease the rate of metabolic degradation by aminopeptidases; (2) residues following critical Tyr9 were replaced by a combination of Lys residues and lipoamino acids, such as lysine-conjugated to long-chain carboxylic acids.

A number of analogs, including NAX 5055, were chemically synthesized on a solid support using standard Fmoc protocols and automated peptide synthesizer. The peptides were purified by reversed-phased HPLC separations and their identities were confirmed by mass spectrometry, prior to pharmacological testing. NAX 5055 is one of the prototype galanin analogs that has undergone extensive pharmacologic evaluation.

Planned studies 

Ongoing evaluations are attempting to further define the therapeutic potential of NAX 5055 and/or one of its analogs for the treatment of pharmacoresistant partial seizures. Once an IND candidate has been identified, preclinical toxicology studies will be initiated in anticipation of an IND filing with the FDA.

Introduction and rationale for development 

Valproic acid is considered the first choice drug for many patients with generalized onset seizures and, because of its broad-spectrum efficacy, it is also useful in difficult to classify epilepsies. In the recent SANAD trial, valproic acid was indeed better tolerated than topiramate and more efficacious than lamotrigine for initial treatment of patients with generalized and unclassifiable epilepsies (Marson et al., 2007). The clinical usefulness of valproic acid, however, is limited in women of child-bearing age because of teratogenicity, and in infants and young children because of potential liver toxicity (Bialer and Yagen, 2007). Propylisopropyl acetamide (PID) was selected from a drug discovery programme aimed at identifying new potentially non-teratogenic and non-hepatotoxic CNS agents with equal or greater potency than valproic acid.

PID is a chiral (with one chiral center denoted above with asterisk *) CNS-active constitutional isomer of valpromide, the CNS-active corresponding amide of valproic acid (Bialer, 1991). PID is currently being developed by Jazz Pharmaceuticals, CA, USA. Unlike valpromide that undergoes extensive biotransformation to valproic acid in mice, rats or dogs, PID is not metabolized in animals to its corresponding acid (propylisopropylacetic acid) and therefore can be regarded as a metabolically stable constitutional valpromide isomer (Spiegelstein et al., 1999, Isoherranen et al., 2003, Bialer and Yagen, 2007). In contrast to valpromide, which serves as a prodrug to valproic acid in humans, PID is likely to act as a drug on its own.

Pharmacology 

Teratogenicity 

Unlike valproic acid and similar to (R)-PID and (S)-PID (Spiegelstein et al., 1999), the corresponding acids of the PID enantiomers ((R)- and (S)-propylisopropylacetic acid) failed to exert a teratogenic effect in SWV-Fnn mice sensitive to valproic acid-induced teratogenicity following a single dose (600mg/kg, ip) at day 8.5 of gestation.

Pharmacokinetics in animals and stability studies in vitro 

Following i.p. administration of individual enantiomers in rats, (S)-PID had lower clearance and volume of distribution and a shorter half-life than (R)-PID. However, following administration of (R,S)-PID, both enantiomers had similar clearance and volume of distribution, but (R)-PID retained a longer half-life. The pharmacokinetics of PID were enantioselective following ip administration of individual enantiomers to mice, with (R)-PID having a lower clearance (0.8L/(hkg) vs. 1.4L/(hkg)) and longer half-life than (S)-PID (Isoherranen et al., 2003). Following i.v. administration of the individual PID enantiomers to dogs, (R)-PID had a significantly lower mean clearance (0.3L/(hkg) vs. 0.6L/(hkg)) and longer half-life. In rats (i.v.), mice (i.p.) and dogs (i.v.), no enantioselectivity in the pharmacokinetics of PID was observed following administration of (R,S)-PID, which may be due to enantiomer-enantiomer interaction (Spiegelstein et al., 1999).

The metabolic stability relative to valproic acid of (R)- and (S)-PID enantiomers was determined by incubation with various human liver enzyme sources (hepatocytes, S9 fraction, microsomes, and mitochondria) at substrate concentrations of 1–50μM. Hepatocyte and S9 studies used positive controls of testosterone, desipramine and tolbutamide at the same concentration as valproic acid and PID to represent high, moderate, and low clearance compounds, respectively. All samples were analyzed using liquid chromatography–mass spectrometry/mass spectrometry techniques. After 120min incubation with hepatocytes (0.5million viable hepatocytes/mL) at concentrations of 1–10μM, no degradation/metabolism was detected for (R)-PID. At the end of 120min incubation, 76.9% of (S)-PID remained non-metabolized.

A second hepatocyte study using 50μM substrate and 1.5millioncells/mL was conducted. Apparently only valproic acid was metabolized by the hepatocytes (only 2% remaining after 2h). No/minimal degradation or metabolism could be observed in the incubations with (R)- or (S)-PID. A glucuronide metabolite could only be detected for valproic acid, consistent with literature reports that glucuronidation is the major pathway of valproic acid metabolism. The valproic acid-glucuronide was rapidly formed within 30min and then further metabolized to an undetectable and unidentified compound.

Studies with S9 fraction, microsomes (supplemented with NADPH regenerating system), mitochondria (supplemented with NAD+), and a combination of the latter two (supplemented with both) produced similar results. No metabolism was observed in S9 with (R)-PID, and very little (about 6.5%) consumption was observed with (S)-PID. No metabolism could be observed with valproic acid. No metabolism of (R)- or (S)-PID enantiomers or valproic acid could be measured in microsomes or mitochondria.

These results indicate that PID enantiomers have good metabolic stability relative to valproic acid and are consistent with in vivo studies reported in the literature (Isoherranen et al., 2003) indicating that in mice and dogs (S)-PID has a shorter half-life than (R)-PID.

Planned studies 

Additional studies to explore the potential and to support the development of PID enantiomers for epilepsy are being planned.

Introduction 

Retigabine is a novel compound currently being developed as adjunctive treatment for partial epilepsy by Valeant Pharmaceuticals. A unique feature of retigabine among potential AEDs in clinical development is its innovative mode of action, which consists primarily in activation of neuronal M-current mediated by KCNQ (Kv7) voltage-gated potassium channels.

Pharmacology 

Toxicology 

In general, across-species acute toxicity was limited to dose-related CNS effects, such as hyperkinesia, hypokinesia, disturbed coordination, stilted gait, tremor and convulsions. In repeated-dose studies in rats and to a lesser extent in dogs, retigabine administration was associated with bladder and minor renal changes. These may have reflected inhibition of bladder contractility and urinary retention secondary to retigabine's effects on KCNQ channels in the detrusor muscle of the bladder. Retigabine did not prolong QTc interval in isolated guinea pig hearts and had no effect on ECG parameters monitored continuously by telemetry in dogs given daily oral doses up to 38mg/kg for 7 days.

No retigabine-related reproductive effects were observed in male or female rats. No teratogenic effects of retigabine were detected in rats or rabbits. Perinatal/postnatal administration of retigabine to mated female rats was not associated with developmental toxicity in offspring except for delayed growth at the highest dose exposure. Evaluation of retigabine safety in juvenile animals is ongoing. Retigabine is not mutagenic or carcinogenic.

Pharmacokinetics 

The absolute bioavailability of orally administered retigabine capsules is 60%, primarily due to first-pass metabolism. Absorption is rapid with peak plasma concentrations achieved within 1–2h. Plasma concentrations and AUC values AUC increase linearly with daily doses up to at least 1200mg retigabine in epilepsy patients. Peak plasma concentrations are modestly increased when retigabine is taken with a high-fat meal, though AUC remains unchanged. The volume of distribution at steady-state is about 2–3L/kg. Binding to human serum albumin and plasma proteins does not exceed ∼80%, and is not concentration-dependent over the ranges evaluated (0.1–2.0μg/mL).

Retigabine is extensively metabolized by hydrolysis/acetylation and glucuronidation without involving oxidative CYP-mediated metabolic pathways. A circulating metabolite of retigabine in humans is an N-acetyl derivative that displays weak pharmacologic effects and inconsistent antiseizure activity in preclinical models. Both retigabine and the N-acetyl metabolite of retigabine undergo glucuronidation.

The plasma half-life of retigabine and of the N-acetyl metabolite of retigabine is 8–10h. Total body clearance (CL/F) following i.v. dosing is approximately 0.57L/(hkg). Retigabine does not induce or inhibit its own metabolism. Elimination of retigabine and its metabolites is primarily renal. Clearance is reduced approximately 30% in elderly subjects when compared with younger subjects, possibly reflecting age-related changes in renal function. In patients with moderate or severe hepatic impairment, retigabine clearance is 30% and 50% lower, respectively, and exposure is increased 50% and 100% when compared with healthy individuals and those with mild hepatic impairment. Relative to healthy subjects, retigabine clearance was reduced approximately 25% in individuals with mild renal dysfunction and approximately 50% in those with moderate or severe renal disease or those who required dialysis.

Drug interactions 

Studies in human liver microsome preparations showed that retigabine co-administration is not expected to result in clinically significant inhibition of the metabolism of drugs metabolized by CYP enzymes. At expected peak concentrations of approximately 6μM occurring clinically at therapeutic doses, retigabine has only a modest potential to inhibit CYP2A6, and low or no potential to inhibit other tested CYP isoforms (CYP1A2, CYP2C8, CYP2C9, CYP2C19, CYP2D6, CYP2E1, CYP3A4/5, CYP4A9/11). No clinically relevant interactions due to inhibition or competition for glucuronidation are expected based on in vitro studies in human liver microsomes with imipramine, lamotrigine, or valproate, all of which are extensively glucuronidated.

In healthy subjects, no interaction was found between retigabine and phenobarbital, a known inducer of glucuronidation, when phenobarbital 90mg/day was administered for 28 days before retigabine administration. In a study designed to evaluate pharmacokinetics of retigabine as adjunctive therapy and as monotherapy in patients with epilepsy, retigabine clearance was increased 36% and 27%, respectively, by co-therapy with the enzyme-inducing AEDs phenytoin (n=9) and carbamazepine (n=8); valproic acid (n=4) or topiramate (n=5) co-administration did not significantly alter retigabine pharmacokinetics in this study. Lamotrigine did not significantly affect retigabine clearance in a study in healthy volunteers (n=15). In contrast, population pharmacokinetic analyses of data from all clinical studies involving more than 800 patients did not identify any effect of enzyme-inducing AEDs (i.e. carbamazepine, phenytoin, or phenobarbital) and no clinically meaningful effects of non-enzyme-inducing AEDs on retigabine pharmacokinetics. Retigabine co-therapy in patients with epilepsy did not alter the pharmacokinetics of phenytoin, carbamazepine, valproic acid, or topiramate. In Phase III studies, retigabine co-administration was not associated with clinically meaningful changes in the trough concentrations of concomitant AEDs although trough lamotrigine concentrations were reduced 20%. Covariates having meaningful effects on retigabine pharmacokinetics in the population pharmacokinetic analysis included body surface area, age, and creatinine clearance.

Retigabine is unlikely to alter the pharmacokinetics of other drugs (non-AEDs) metabolized by CYP isozymes. Therapeutic dosages of retigabine (a 28-day regimen including maintenance at 750mg/day for 19 days) did not reduce contraceptive hormone exposure in women taking a combination oral contraceptive (0.035mg ethinyl estradiol+1mg norethindrone), suggesting no potential for reduced contraceptive efficacy due to retigabine-induced metabolism of contraceptive hormones. Although the study was not designed to assess the effect of oral contraceptives on retigabine pharmacokinetics, retigabine AUC values following two oral contraceptive cycles were within the range of AUC values observed in other studies, suggesting that combination oral contraceptives do not markedly alter the pharmacokinetics of retigabine.

Efficacy 

A Phase II, multicenter, randomized, double-blind, placebo-controlled dose-ranging trial (Study 205) evaluated retigabine 600, 900, or 1200mg/day as adjunctive therapy in adults with partial-onset seizures (Porter et al., 2007). Patients (n=396) were 16–70 years of age with inadequately controlled partial-onset seizures despite stable therapy with 1–2 AEDs. Minimum monthly seizure frequency was 4 partial seizures with or without secondary generalization, with no seizure-free period >30 days. Pre-retigabine seizure frequency was established during the 8-week prospective baseline phase, after which patients were randomized to placebo or retigabine 600, 900, 1200mg/day. Retigabine was titrated to the assigned dosage over 2–6 weeks with limited dose adjustments (100-mg dose decrements at weeks 7 and 8). The 300mg/day starting dose (administered t.i.d. as immediate-release tablets) was increased weekly in 150mg/day increments. Minimum fixed dosages during the 8-week maintenance phase were retigabine 400, 700, 1000mg/day.

Retigabine significantly (p<0.001) reduced partial seizure frequency vs. placebo across all treatment arms in a dose-dependent manner. Responder rates (patients with ≥50% seizure frequency reduction) in the double-blind phase were 23% with 600mg/day, 32% with 900mg/day (p=0.021) and 33% with 1200mg/day (p=0.016) vs. 16% in the placebo arm. Of patients completing the double-blind study, 80% (222/279) entered the open-label extension in which retigabine and background AED dosages were adjusted according to clinical response. In patients initially assigned to placebo or 600-mg retigabine arms, retigabine was up-titrated during the blinded conversion phase to 900mg/day as a target dose. Retigabine was down-titrated to 900mg/day in patients initially assigned to 1200mg retigabine and maintained at the double-blind maintenance dose for those initially assigned to the 900mg/day arm. Maximum retigabine dose was 1200mg/day. In patients from the 900 and 1200mg arms, seizure control achieved during the double-blind trial was maintained during long-term, open-label treatment. Seizure control improved in patients from the placebo and retigabine 600-mg arms when transitioned to 900mg/day as an initial target dose. At the end of open-label treatment, 78% of patients were receiving ≤900mg/day retigabine. In 46% of patients seizure frequency was reduced ≥50% from the baseline of the preceding double-blind study. Nine percent (9%) had seizure-free intervals ≥6 months during retigabine treatment. No additional AEDs were added during the seizure-free intervals.

Two recently completed double-blind, placebo-controlled Phase III studies have confirmed the dose-dependent efficacy of 600–1200mg/day retigabine and demonstrated that 600–900mg/day is an appropriate initial target dose range for retigabine as adjunctive therapy in adults with partial-onset seizures. These studies were identical in design, except for target dosages and titration phase duration. One study (RESTORE 2/Study 302) compared placebo, 600 and 900mg/day retigabine (immediate-release tablets administered t.i.d.) titrated over 2 or 4 weeks, with no downward dosage adjustments allowed. The second study (RESTORE 1/Study 301) compared placebo and 1200mg/day retigabine (administered t.i.d.) titrated over 6 weeks but allowed a dosage adjustment at the start of the maintenance phase (minimum maintenance dosage, 1050mg/day retigabine). In both studies, the starting dosage of 300mg/day retigabine was increased weekly in 150mg/day increments; the maintenance period was 12 weeks. Patients were 18–75 years of age with refractory partial-onset seizures. Refractoriness was defined as ≥2 years since epilepsy diagnosis and past failure of ≥2 approved AEDs, alone or together, in adequate doses for sufficient period of time to evaluate clinical response. Patients had to have ≥4seizures/month despite stable dosages of 1–3 AEDs, with or without vagal nerve stimulation during an 8-week prospective baseline phase.

Patient characteristics at baseline were very similar across the two studies. Patients had been diagnosed with epilepsy 22–24 years before study entry; had failed 3–4 (median) AEDs in the 3 years prior to study entry; and were having 11–12seizures/month (median). More than 75% of patients were receiving 2 or 3 AEDs. Mean age was 37–38 years and 52–54% were female.

The intent-to-treat populations comprised 538 patients in RESTORE 2/Study 302 (placebo, n=179; 600mg retigabine, n=181; 900mg retigabine, n=178) and 303 patients in RESTORE 1/Study 301 (placebo, n=150; 1200mg retigabine, n=151). During double-blind treatment, retigabine at all doses significantly reduced median seizure frequency vs. placebo (Table 4). The dose-related pattern of efficacy was also observed in the responder rate. Retigabine was associated with an early onset of therapeutic effect as demonstrated by the significant improvement in seizure control during titration.

Table 4. Efficacy results from Phase III double-blind placebo-controlled trials of 600, 900, and 1200mg/day retigabine in refractory partial-onset seizures.

		RESTORE 2 (Study 302)			RESTORE 1 (Study 301)
		Placebo	Retigabine (mg/day)		Placebo	Retigabine (mg/day) 1200
			600	900
	Median % seizure reduction
	Overall	15.9%	27.9%	39.9%	17.5%	44.3%
	p-Value		<0.01	<0.001		<0.001
	Titration	11.0%	26.1%	32.1%	10.4%	30.4%
	p-Value		<0.01	<0.001		<0.001
	Maintenance	17.4%	35.3%	44.3%	19.0%	54.5%
	p-Value		<0.01	<0.001		<0.001
	Responder rate (≥50% reduction in baseline seizure frequency)
	Overall	17.3%	31.5%	39.3%	18.0%	45.0%
	p-Value		<0.01	<0.001		<0.001
	Titration	17.3%	28.7%	33.1%	17.3%	37.7%
	p-Value		<0.01	<0.001		<0.001
	Maintenance	18.9%	38.6%	47.0%	23.2%	55.5%
	p-Value		<0.01	<0.001		<0.001

Tolerability and adverse effect profile 

Consistent with AEDs as a class, the most common adverse events associated with retigabine are generally non-specific CNS effects such as dizziness, somnolence, and fatigue (Table 5) that are mild-to-moderate in nature. The temporal pattern of these events suggests that they may often be peak-dose effects of the immediate-release retigabine tablets. Among non-CNS events, an increased incidence of bladder-related adverse events (e.g. urinary hesitancy) relative to placebo was observed with retigabine, primarily with 1200mg. These events were rarely serious or cause for discontinuation. Bladder ultrasound revealed a modest increase in mean post-void residual volume at the 1200-mg dose but not at lower doses, a finding that is of uncertain clinical significance. Other clinical and laboratory monitoring, including urinalysis, ECG and vital signs showed no clinically significant changes related to retigabine administration during double-blind, placebo-controlled trials or long-term, open-label treatment.

Table 5. Most common (≥10%) adverse events during double-blind, placebo-controlled Phase III trials of 600, 900, and 1200mg/day retigabine in refractory partial-onset seizures.

		Placebo (n=331)	600 (n=181)	Retigabine (mg/day)
				900 (n=178)	1200 (n=153)
	Dizziness	9%	17%	26%	40%
	Somnolence	13%	14%	26%	31%
	Fatigue	5%	17%	15%	16%
	Confusion	1%	2%	5%	14%
	Dysarthria	1%	5%	2%	12%
	Headache	16%	11%	17%	12%
	Ataxia/gait disturbance	4%	3%	5%	12%
	Urinary tract infection	5%	1%	2%	12%
	Tremor	3%	2%	8%	11%
	Vision blurred	2%	1%	5%	11%
	Nausea	5%	6%	7%	11%

Based on double-blind, placebo-controlled trials, which featured forced-titration and very little dosing flexibility to reduce CNS TEAEs, the tolerability of retigabine is dose-related and influenced by titration/dose-management. In Phase III trials, discontinuations due to TEAEs increased according to assigned retigabine dosage: 600mg/day, 14%; 900mg/day, 26%; and 1200mg/day, 27% compared with 8% in those assigned to placebo; more than two-thirds of discontinuations occurred during forced-titration. TEAEs leading to retigabine discontinuation in >3% of patients were dizziness (4%), fatigue (4%), somnolence (4%), and confusion (3%). Double-blind, placebo-controlled studies as well as open-label, long-term studies suggest that the optimal dosage for most patients who benefit from retigabine will be 600–900mg/day.

Ongoing/planned studies 

A clinical development program to evaluate efficacy and safety studies in pediatric patients with partial-onset seizures and patients with Lennox-Gastaut syndrome is planned. Future studies also include development of a modified-release formulation to improve tolerability and dosing convenience.

Introduction 

T2000 (1,3-dimetoxymethyl-5,5-diphenylbarbituric acid), a barbiturate derivative under development for the treatment of epilepsy and essential tremor by Taro Pharmaceuticals and discussed more extensively in the subsequent section, is considered to act mainly as a prodrug to 5,5-diphenylbarbituric acid. The latter compound has been relatively extensively investigated on its own merits for a number of years, and a brief overview of its properties is summarized below.

5,5-Diphenylbarbituric acid was originally prepared by McElvain in 1935 and evaluated as a potential hypnotic in rats (McElvain, 1935). Due to the lack of hypnotic efficacy, the interest in further development of 5,5-diphenylbarbituric acid was low until the discovery of the anticonvulsant activity of phenytoin by Merritt and Putnam in 1938. It was then suggested that phenyl groups were important in conferring anticonvulsant effects. Subsequently, 5,5-diphenylbarbituric acid was evaluated for its effects on seizure threshold induced by electrical stimulation in cats, rabbits, and rats. Those studies revealed weak to moderate anticonvulsant activity (Alles et al., 1947, Knoefel and Lehmann, 1942, Merritt and Putnam, 1945). However, 5,5-diphenylbarbituric acid was not evaluated in a systematic manner and the validity of the experimental methods used in those early studies could be questioned.

Given its structural similarity to phenytoin and phenobarbital, the interest in 5,5-diphenylbarbituric acid was renewed by A. Raines and his colleagues some 30 years later. Subsequent studies provided strong evidence that 5,5-diphenylbarbituric acid shares pharmacological actions and properties with phenytoin and barbiturates. Such combination of pharmacological effects is considered to offer potential therapeutic advantages.

Pharmacology 

The pharmacological effects of 5,5-diphenylbarbituric acid in vitro and in vivo resemble those of both phenytoin and barbiturates. In electrophysiological studies, 5,5-diphenylbarbituric acid suppressed high frequency repetitive neuronal discharges in cat motor nerve terminals in a manner similar to that of phenytoin, whereas it depressed synaptic discharges in the cat spinal cord in a manner similar to that of phenobarbital (Zavadil et al., 1985). Furthermore, 5,5-diphenylbarbituric acid enhanced slow outward current and suppressed repetitive firing in Aplysia giant neurons with similar efficacy to and higher potency than phenobarbital (Huguenard and Wilson, 1985). In one study diphenylbarbituric acid lacked the ability to enhance the activity of GABA in rat hippocampal slices (Huguenard and Wilson, 1985). Thus, how exactly 5,5-diphenylbarbituric acid interacts with different binding sites at the GABAA receptor complex or if it has other pharmacological actions remain to be evaluated.

In in vivo studies, 5,5-diphenylbarbituric acid was found to be effective in the PTZ and MES tests in mice (Raines et al., 1973). Specifically, it protects against PTZ-induced clonic seizures with ED50 values of 57 and 26mg/kg after p.o. and i.p. administration, respectively. In a parallel study, phenobarbital was effective (ED50=12.5mg/kg p.o.) whereas phenytoin was ineffective against PTZ-induced seizures. 5,5-Diphenylbarbituric acid is also effective against MES-induced seizures (ED50=320mg/kg p.o. and 63mg/kg i.p.). In the rotarod test, no signs of neurotoxicity were apparent with doses of 5,5-diphenylbarbituric acid up to 19.2g/kg p.o. The dose that produced neurotoxicity in 50% of mice after i.p. administration was 250mg/kg. Thus, 5,5-diphenylbarbituric acid demonstrated favorable therapeutic indices after administration both p.o. (>337 and >60 in the PTZ and MES tests, respectively) and i.p. (9.6 and 4.0 in the PTZ and MES tests, respectively).

5,5-Diphenylbarbituric acid was also efficacious against seizures in the PTZ and MES test in rats (ED50 p.o.=130 and 11.5mg/kg, respectively) (Raines et al., 1975). Maximal efficacy was observed between 3 and 6h after p.o. administration. The pharmacokinetic profile also appeared to be favorable. Exposure to the drug was dose-dependent and there was a good correlation between brain and plasma concentrations with a brain-to-plasma concentration ratio ranging from 0.1 to 0.5 after doses of 12.5, 25, and 50mg/kg p.o. As with mice, no signs of neurotoxicity were produced by 5,5-diphenylbarbituric acid doses p.o. up to 2.8g/kg in the rotarod test in rats. Thus, protective index values in rats were greater then 21.5 and 243 in the PTZ and MES tests, respectively. However, 5,5-diphenylbarbituric acid (50–100mg/kg, i.p.) was ineffective in reducing the lethality induced by the organophosphate agent diisopropylfluorophosphate (Dretchen et al., 1986). In the same study, phenytoin was effective, whereas carbamazepine (12.5–25mg/kg, i.p.) and phenobarbital (30mg/kg, i.p.) were not.

Czuczwar and colleagues also reported that 5,5-diphenylbarbituric acid significantly potentiated the effects of valproic acid against MES-induced seizures in mice (unpublished data), which would suggest potential beneficial effects of this barbiturate in add-on therapy. There is also some evidence that 5,5-diphenylbarbituric acid may affect epileptogenic processes accompanying repeated exposures to convulsive stimuli (Gasior, unpublished data). In particular, 5,5-diphenylbarbituric acid (100mg/kg, i.p.) was found to decrease the development of PTZ-induced kindled seizures in mice, and to exhibit antiepileptogenic properties in kindled seizures induced by repeated administrations of low doses of cocaine in mice. Interestingly, 5,5-diphenylbarbituric acid was devoid of any acute efficacy against seizures induced by high doses of cocaine and yet retarded the development of cocaine kindled seizures. This property against cocaine-induced kindled seizures implies its potential utility in some neuropsychiatric syndromes associated with a dysfunction of the dopaminergic neurotransmitter system. There are no reports on the propensity of diphenylbarbituric acid to produce tolerance to its pharmacological effects that is often seen with GABAA agonists. (Gasior, personal communication).

In summary, although how diphenylbarbituric acid interacts with sodium channels, GABAA receptors, and/or other molecular targets remains to be elucidated, it clearly shows several favorable properties of potential value for the treatment of epilepsy and perhaps other CNS disorders. A prodrug of 5,5-diphenylbarbituric acid, T2000, is discussed in some detail in the section below.

Introduction 

T2000 is a member of the barbiturate class of drugs, a class which has been in wide clinical use for over 100 years (Moros and Rutman, 2007). T2000 is a prodrug and is rapidly metabolized to monomethoxymethyl-5,5-diphenylbarbituric acid (MMMDPB) and 5,5-diphenylbarbituric acid (McKay et al., 2006a). T2000 and its metabolites exhibit useful pharmacologic properties at dosages not limited by sedation. T2000 is being investigated for the treatment of essential tremor, myoclonus dystonia and epilepsy.

Pharmacology 

Earlier studies performed on isolated neural systems in aplysia and the hippocampus of the rat have shown that 5,5-diphenylbarbituric acid, the major metabolite of T2000, suppresses neural repetitive firing in both systems at concentrations lacking significant effects on GABA neurotransmission. These effects may be attributable to enhancement of outgoing membrane potassium current. Studies of 5,5-diphenylbarbituric acid on cat motor nerve terminal function show a suppression of repetitive discharges similar to the effects of phenytoin, a compound which acts on sodium channels in the nerve membrane.

Receptor binding studies indicate that binding of 3H-flunitrazepam to the GABAAα1 receptor is inhibited by 67% by MMMDPB (25μg/mL) and by 10% by 5,5-diphenylbarbituric acid (75μg/mL). Binding of 3H-WIN,35,428 (2β-carbomethoxy-3β-(4-fluorophenyl)tropane) to the dopamine transporter is inhibited by 58–71% by 5,5-diphenylbarbituric acid (75μg/mL) and by 20% by MMMDPB (25μg/mL).

Toxicology 

In rats, single oral doses of T2000 up to 750mg/kg had no effect on alertness, neurological function, muscle tone and general behavior. T2000 did not exhibit the discriminative stimulus effects of pentobarbital, and produced no physiological or behavioral signs of physical dependence in rats.

The maximum tolerated acute oral dose of T2000 in rats and dogs was 2000mg/kg. Chronic 26-week oral dose toxicity studies were conducted in rats with doses up to 750mg/(kgday). After 750mg/(kgday) for 26 weeks in rats, the total barbiturate peak plasma concentration was 69μg/mL in males and 114μg/mL in females. Chronic 52-week oral dose toxicity studies were conducted in dogs with doses up to 450mg/(kgday). Because of toxicity (severe anorexia, weight loss), dose levels of 450 and 225mg/(kgday) were reduced to 150mg/(kgday). Total barbiturate peak plasma concentration for these dose groups (males/females) were 41/43 and 60/33μg/mL, respectively. The principal effect was a dose-dependent increase in liver weight, hepatocellular hypertrophy, and induction of hepatic CYPs, reversible after 4 weeks recovery. The No-Observed-Adverse-Effect-Level (NOAEL) was <50mg/(kgday) in the rat and 25mg/(kgday) in the dog.

The cardiovascular safety of T2000, 5,5-diphenylbarbituric acid (as sodium salt) and MMMDPB were evaluated for potential QT interval prolongation. T2000 had no effect in vitro on hERG current up to 30μg/mL. 5,5-Diphenylbarbituric acid and MMMDPB inhibited hERG current at 3000 and 150μg/mL, respectively.

T2000 has shown no genotoxic effects in vitro and in vivo, based on bacterial reverse mutation, mammalian cell gene mutation, mammalian chromosome aberration tests, and mammalian erythrocyte micronucleus tests in mice and rats.

In the fertility and general reproductive toxicity study in the rat (Segment I), the paternal No-Observed-Effect-Level (NOAEL) for general and reproductive toxicity was 300mg/(kgday), and maternal NOAEL for general toxicity was 60mg/(kgday). In the development toxicity studies in the rat and rabbit (Segment II), the NOEL was <20mg/(kgday) in rats, and 15mg/(kgday) in rabbits. T2000 was identified as a teratogen in rats, but not in rabbits. In the developmental and perinatal/postnatal reproduction toxicity study (Segment III), the reproductive NOEL was 200mg/(kgday) and maternal NOEL was 20mg/(kgday).

Clinical pharmacokinetics 

Six phase I clinical studies have been conducted in a total of 191 healthy male and female subjects. The pharmacokinetics of T2000 and its primary metabolites 5,5-diphenylbarbituric acid and MMMDPB show near linearity up to 1200mg daily (McKay et al., 2006a). Plasma protein binding values for T2000, 5,5-diphenylbarbituric acid and MMMDPB are 67%, 35% and 51%, respectively (Yacobi et al., 2006). T2000 is given with meals since food enhances its absorption (Raines et al., 2007). After single and multiple oral doses of T2000 (25–1200mg) for up to 2 weeks in healthy subjects, mean terminal half-lives of T2000, 5,5-diphenylbarbituric acid and MMMDPB range between 9–29, 27–65 and 8–27h, respectively. Steady-state is reached after 2 weeks of dosing in young healthy subjects. After 1200mg daily (600mg b.i.d.) oral doses of T2000 for 2 weeks, mean peak plasma concentrations of T2000, 5,5-diphenylbarbituric acid and MMMDPB were 0.79, 68 and 47μg/mL, respectively (Ganes et al., 2006, Tremblay et al., 2006, Foreman et al., 2008).

After single 400mg oral doses of T2000 in young and elderly (65 years and greater) healthy subjects, peak plasma concentrations and terminal half-lives (young/elderly) of T2000, 5,5-diphenylbarbituric acid and MMMDPB were 1.03/1.29, 3.39/3.51 and 3.26/3.63μg/mL, respectively, and 23/27, 25/32 and 12/19h, respectively.

T2000 and its primary metabolites 5,5-diphenylbarbituric acid and MMMDPB are further metabolized via hydroxylation and form glucuronide conjugates that are eliminated in urine (McKay et al., 2006a, McKay et al., 2006b). In vitro studies using recombinant human CYP isoforms indicate that diphenylbarbituric acid, MMMDPB, and hydroxy-MMMDPB are formed by CYP2C19, while hydroxy-diphenylbarbituric acid is formed primarily by CYP2C9.

Drug interactions 

In vitro studies using human liver microsomes and probe substrates indicate that MMMDPB is a competitive inhibitor of CYP2C19 and CYP2C9 and that 5,5-diphenylbarbituric acid is a competitive inhibitor of CYP2C9 and CYP3A4 at therapeutic concentrations (Nguyen et al., 2006). 5,5-Diphenylbarbituric acid and MMMDPB are potent inducers of CYP3A4. T2000, 5,5-diphenylbarbituric acid and MMMDPB are not substrates of P-glycoprotein in Caco-2 cells (Nguyen et al., 2007).

Efficacy 

Four phase II studies have been completed in the treatment of essential tremor (Melmed et al., 2007, Moros and Rutman, 2007). In two brief randomized, placebo-controlled, parallel-group, double-blind, single center trials, the effect of T2000 on essential tremor, when given at doses of 400 and 300mg b.i.d, was assessed in 22 and 12 patients, respectively. Using a two-factor mixed ANOVA model to evaluate within group and between group changes, the effect of T2000 was significantly different from that of the placebo group (p=0.03) in the group receiving 400mg b.i.d., but not for the group receiving the lower dosage. Combining the placebo groups provided adequate statistical power to evaluate the treatment effect of each dose by adjusting for the difference in baseline tremor score. When the group receiving 800mg/day was compared to this larger placebo group using the two-factor mixed ANOVA model followed by a split trend analysis of the treated and placebo regression line slopes, the treated group differed significantly from placebo (F=12.35, p=0.0025). When the two treated groups were compared, the 800mg/day group differed significantly from the 600mg/day group (F1,21=6.32, p=0.0202). Some treated patients in each study, but no placebo patients, experienced marked tremor improvement. Mean steady-state trough concentrations of T2000, 5,5-diphenylbarbituric acid and MMMDPB (600–800mg daily) were 0.46–0.51, 31–36 and 18–20μg/mL, respectively.

The third study was an open-label study in 5 patients who participated in the earlier phase II studies who either responded to T2000 or did not respond to placebo. Patients were given escalating 200–800mg daily oral doses up to 12 weeks. There was a statistically significant decrease in total tremor score at the end of treatment compared to baseline. Mean steady-state trough concentrations of T2000, 5,5-diphenylbarbituric acid and MMMDPB were 0.14, 60 and 16μg/mL, respectively. In the fourth randomized, double-blind, placebo-controlled, sequential dose escalation phase II study, 10 patients were given escalating 600–1000mg daily oral doses over 5 months. Trough concentrations of T2000, 5,5-diphenylbarbituric acid and MMMDPB at the end of treatment ranged between 0.03–0.24, 30–77 and 10–33μg/mL, respectively. In that study, rapid initiation of high dose T2000, 600mg as a single daily dose, in an elderly population resulted in higher total barbiturate levels (34–110μg/mL) and adverse sedative effects. Six patients showed >25% improvement in tremor score at 1 or more visits compared to baseline.

Two additional phase II studies are ongoing, an open-label 1000mg daily oral dose continuation study for up to 18 months in essential tremor (1 patient) and an open-label escalating 200–1000mg daily oral dose study up to 12 weeks in myoclonus dystonia (n=up to 6 patients).

Tolerability and adverse effect profile 

Of 198 subjects (normal healthy volunteers or tremor patients) who received T2000 in completed single or multiple dose studies, a total of 9 subjects developed a skin rash or pruritus requiring discontinuation of medication, antihistamine or antiinflammatory treatment. Rashes, including rashes requiring treatment, were also noted in the placebo groups in controlled studies (4 out 41 patients), including one patient who required treatment. Skin rashes are a recognized side effect of barbiturates.

Other adverse effects reported more often in patients taking T2000 than in those taking placebo include nausea (29%), vomiting (9%), abdominal pain (8%), inflammation of eye (18%), inflammation of throat (12%), and difficulty sleeping (6%). Some male patients taking T2000 reported erectile dysfunction (5%) that resolved either during continued treatment with T2000 or after treatment was discontinued.

Planned studies 

Next studies planned include: Phase IIb safety and efficacy study in patients with essential tremor (n=60) patients; carcinogenicity in rats and transgenic mise; mass balance/metbolic profiling of 14C-T2000 in rat, dog and man; drug interaction and special population studies.

Introduction and rationale for development 

Tonabersat [(3S-cis)-N-(6-Acetyl-3,4-dihydro-3-hydroxy-2,2-dimethyl-2H-1-benzopyran-4-yl)-3-chloro-4-fluorobenzamide], formerly know as SB-220453, is a chiral (with two asymmetric centers denoted above with asterisk *) novel benzoylaminobenzopyran compound with potent anticonvulsant activity. In its chemical structure, tonabersat is very similar to carabersat (SB-20469), previously developed as an AED, and contains one additional chlorine on the aromatic ring Currently under development at Minster Pharmaceuticals, tonabersat is under investigation as a prophylactic anti-migraine therapy; however, tonabersat's unique anticonvulsant properties offer a real potential for an effective treatment of epilepsy.

Tonabersat selectively and specifically binds a unique stereoselective site in the CNS, thought to be at the neuronal gap junction. As such, tonabersat represents a ‘first-in-class’ neurotherapeutic that does not act via any established anticonvulsant mechanisms (Table 3).

Pharmacology 

Toxicology 

Long term toxicity studies in the rat (6 months treatment at doses up to 2000mg/kg) and dog (1 year at doses up to 200mg/kg) identified no effects expected to limit the proposed dosing options in humans and toxicokinetic data provide evidence of an adequate safety margin with respect to the systemic exposure likely to be achieved with a dose of up to 80mg/day. There was no evidence of treatment-related carcinogenicity in studies in the rat and mouse and no evidence of genotoxicity in a standard battery of tests. Similarly there was no evidence of embryotoxicity. Pharmacological safety investigations showed that tonabersat was without effect on cardiovascular, renal or respiratory function in rats receiving 300mg/kg p.o. or in dogs receiving 60mg/kg p.o. Effects on hERG channels were noted only at high concentrations, with an IC50 of around 18μM. Tonabersat had no effect on isolated human blood vessels at concentrations up to 100μM. Safety studies conducted to date have identified no significant effects on major organ function in vivo or in vitro; the compound was well tolerated and there was no indication from the side effect profile of any specific risks. Tonabersat lacks the cardiovascular actions that have limited the clinical utility of other anti-migraine compounds and this unique profile is attributable to a selective interaction at a unique stereospecific CNS binding site.

Clinical pharmacokinetics 

The pharmacokinetics of tonabersat has been evaluated in six studies in healthy subjects and one in patients with migraine. Following oral administration of a single dose (2–80mg), tonabersat was fairly rapidly absorbed with a median tmax of between 0.5 and 3h post-dose. As expected, absorption was delayed by a high-fat meal (approximately 3h increase in tmax) but food had no effect on the overall systemic exposure. Systemic exposure (AUC) and peak plasma concentrations were generally dose proportional between 2 and 40mg. Following single and repeated dosing, tonabersat was generally quantifiable in plasma up to at least 96h post-dose.

Tonabersat is highly bound to human plasma proteins (98.2%) in a concentration-independent manner. In vitro and animal studies identified CYP3A induction in rat and mouse models; however, no evidence of CYP3A induction has been demonstrated in man. The primary metabolic clearance route of tonabersat occurs via the hepatic system and its urinary concentrations are negligible. In human tissues, the primary metabolite of tonabersat is a reduced ketone generated by cytosolic carbonyl reductase enzymes. This metabolite was detectable in plasma throughout the sampling period when tonabersat was administered at doses of ≥10mg. Acetyl hydroxylation via the CYP3A family of enzymes accounted for a minor hydroxyl metabolite, which a radiolabelled metabolism study identified at a level of approximately 5.8% of the primary metabolite. Tonabersat plasma concentrations decline in a biexponential manner with an average terminal half-life of 24–40h. This relatively long half-life supports a once daily dosing regimen.

Drug interactions 

Two drug interaction studies have been performed to date to investigate tonabersat's potential as a long-term prophylactic agent. The first study addressed concomitant use of an oral contraceptive. When Microgynon® (active ingredients ethinylestradiol and levonorgestrel) was co-administered with 80mg tonabersat for 7 days (days 4–10 of the menstrual cycle), there was an increase in ethinylestradiol peak plasma concentration (mean 14%, 90% CI: 1.0–28.0%) and AUC (mean 21%, 90% CI: 1.3–30.0%). Tonabersat had no effect on the metabolism of levonorgestrel. These data show that tonabersat does not affect the metabolism of the oral contraceptive pill to a major extent and therefore no change in contraceptive effectiveness is anticipated. Sumatriptan is a widely used acute treatment for migraine and might be taken concomitantly with tonabersat. A study in healthy subjects showed no significant effects of either drug on the pharmacokinetic profile of the other.

Efficacy data 

The efficacy and tolerability of tonabersat as a prophylaxis agent for the prevention of migraine headache has recently been evaluated in a placebo-controlled phase II proof-of-concept study in the prophylaxis of migraine headache. The study included a 4-week baseline assessment and a 12-week treatment period at 20mg o.d. for 2 weeks increasing to 40mg o.d. for 10 weeks with 1 week follow-up. Baseline migraine attack frequency was 4.6 in the tonabersat group (n=59) and 4.2 in the placebo group (n=65). The proportion of patients defined as responders (≥50% reduction in migraine attacks) during month 3 was significantly greater in the tonabersat group (62% vs. 45%; p<0.05) and there was a 90% greater reduction in the number of days on which rescue medication was taken. The results of this study support further evaluation of the prophylactic efficacy of tonabersat.

The efficacy of tonabersat in treating acute migraine attacks has been investigated in several studies. In one study, single oral doses of tonabersat 15 and 40mg were associated with a significant increase in the number of patients reporting headache relief after 2h (37% and 41%, respectively, compared with 21% in the placebo group). This superiority over placebo was not seen consistently across studies however, and this was likely due to the confounding effects of the prior use of sumatriptan in these trials.

Overall, these studies confirm that tonabersat possesses pharmacological activity when administered orally in man. Its relatively slow absorption and 30h plasma half-life make it potentially well suited for use as a once daily prophylactic treatment.

Tolerability and adverse effect profile 

Tonabersat has been administered to more than 1000 subjects in several clinical trials. Both healthy subjects and migraine patients received either single oral doses (up to 80mg) or repeated doses (up to 80mg o.d. for up to 7 days). Tonabersat was generally well tolerated in all studies with scope for increasing the dose in the future. Headache, nausea, dizziness and somnolence were the most commonly reported TEAEs. The majority of events were described as mild or moderate and rapidly resolved. No significant changes in vital signs, ECGs or laboratory assessment were reported after tonabersat administration.

Planned studies 

Further phase IIb studies are planned to investigate dose–response, efficacy and tolerability initially for prophylactic treatment og migraine.

Introduction 

Valrocemide (N-valproyl glycinamide) was selected from a series of N-valproyl derivatives of GABA, glycine and taurine because of its favorable pharmacokinetic and anticonvulsant activity profile in preclinical screening models (Isoherranen et al., 2001). Valrocemide pharmacokinetic-based design was aimed at enhancing brain penetration compared to valproic acid by converting valproic acid carboxylic acid to a carboxamide moiety, and utilizing the glycinamide-glycine biotransformation as major metabolic pathway.

Pharmacology 

Valrocemide has a wide spectrum of anticonvulsant activity in various animal models, including the 6Hz psychomotor seizure mice model and the hippocampal-kindled rat model (Table 1, Table 2). Valrocemide is more potent than valproic acid in following models: (a) the MES test (mice and rats); (b) Frings mice and, (c) bicuculline- and picrotoxin-induced-seizures in mice (Bialer et al., 2007, Isoherranen et al., 2001). Valrocemide is also more active than valproic acid in the spinal ligation model for neuropathic pain and in the amphetamine-induced hyperactivity rat model for mania (Bialer et al., 2007). In all these animal models valrocemide acts as a drug on its own and not as a prodrug to valproic acid, since there is no biotransformation of valrocemide to valproic acid in rodents and valrocemide is mainly metabolized to an inactive metabolite valproyl glycine in these species.

Clinical pharmacokinetics 

In man, valrocemide shows linear pharmacokinetics over a dose range from 250 to 4000mg (single oral doses between 250 and 4000mg and multiple dosing ranging between 250 and 1000mg, t.i.d.) (Bialer et al., 2002, Bialer et al., 2004, Bialer et al., 2007).

Following oral administration to humans, 57–75% of a valrocemide dose is excreted in urine as valproyl glycine and 10–20% of the dose is excreted unchanged in urine. Valproic acid is a minor metabolite of valrocemide in humans and the fraction of valrocemide biotransformed to valproic acid has been estimated to be 4–6% (Bialer et al., 2007). Following repetitive dosing of valrocemide (2000mg b.i.d.) to patients with epilepsy, mean plasma valproic acid levels were 22mg/L, i.e. significantly lower than valproic acid's therapeutic range (50–100mg/L). Valrocemide thus might be developed into a product with efficacy comparable to valproic acid, potentially without the dose-related side effects of valproic acid.

A controlled-release formulation of valrocemide has been developed, and its AUC has been found to be bioequivalent to that of a valrocemide immediate-release formulation.

Drug interactions 

Preliminary studies in patients with epilepsy indicated that the biotransformation of valrocemide to valproyl glycine may be moderately induced by enzyme inducing AEDs.

Following repetitive dosing of valrocemide to healthy subjects, valrocemide did not affect the pharmacokinetics of CYP1A2, CYP2C9, CYP2C19 and CYP2D6 probe drugs. Valrocemide reduced the AUC of midazolam, a CYP3A4 probe, by 33%, which is a moderate effect compared with the 94% reduction of midazolam AUC caused by carbamazepine and phenytoin (Bialer et al., 2007).

Efficacy and adverse effect profile 

There have been no further studies in addition to those described in the progress reports of the last Eilat Conferences (Bialer et al., 2007).

Ongoing and planned studies 

In July 2005, Teva Pharmaceuticals returned valrocemide to Yissum, the Technology Transfer Company of The Hebrew University of Jerusalem. In June 2006, Yissum and Shire LLC entered into a worldwide license whereby Shire LLC will continue the development of valrocemide. In December 2007 Shire LLC terminated its agreement with Yissum because of portfolio decision. Subsequently, in February 2008, valrocemide was in-licensed from Yissum by Desitin (Hamburg, Germany), which acquired the rights for commercialization in Europe. Currently, Desitin and Yissum are looking for a partner for co-development which will acquire the rights in North America. Because of these events, there has been only modest progress in valrocemide development since the compound was reviewed at the previous Eilat Conference (Bialer et al., 2007).

Introduction 

YKP3089 is a novel compound (with one chiral center demotes above with asterisk*) with broad-spectrum anticonvulsant activity under clinical development at SK Life Science. In addition to being a candidate for use in epilepsy therapy, YKP3089 has the potential to be a versatile CNS drug with multiple therapeutic uses.

Pharmacology 

Toxicology 

In repeated-dose toxicity studies in various rodents and non-rodents, YKP3089 at very high doses displayed predominantly adverse effects related to exaggerated pharmacology in all species tested.

YKP3089 was negative in all genotoxicity assays tested (Ames bacterial reverse mutation, mouse lymphoma and in vivo rat bone marrow micronucleus studies). In reproductive studies in rats and rabbits there was no evidence of teratogenicity. YKP3089 did not affect cardiovacular functions in a telemetry cardiovascular study in monkeys.

Clinical pharmacokinetics 

In phase I studies completed in the U.S., exposure of YKP3089 following single oral administration was dose proportional with an observed half-life of 30–75h, which would support once daily dosing. In single-dose studies, YKP3089 pharmacokinetics was linear over a large dose range (5–750mg), with a median tmax of 1.5–3.5h, a mean volume of distribution (Vd/F) of 37–55L, and a mean apparent oral clearance of 0.63L/h. In a multiple-dose 14-day study with once daily dosing, both peak plasma concentrations and AUC correlated linearly with the dose (50–150mg/day).

In a food effect study, there appeared to be no significant differences in pharmacokinetic profile after intake with a high-fat diet compared with the fasted condition.

Tolerability and adverse effect profile 

YKP3089 was well-tolerated in phase I studies after single doses ranging from 5 to 750mg and multiple doses ranging from 50 to 150mg/day (daily dosing of 14 days), with a low incidence of adverse effects. Most common TEAEs were CNS related, were of mild to moderate intensity and resolved rapidly. No clinically significant changes in ECGs or laboratory parameters were observed.

Ongoing and planned studies 

In view of its good tolerability, possibility of once daily dosing and broad-spectrum activity in animal models of epilepsy and other disorders, phase II studies aimed at various indications including epilepsy are currently ongoing.