Introduction

A pharmacodynamic interaction occurs when the pharmacology (effect) of one drug alters the pharmacology of another drug without altering plasma concentrations. Theoretically, the interaction can occur at the receptor or sites of action or indirectly by affecting other physiological mechanisms. This is in contract to pharmacokinetic interactions, where one drug alters the plasma concentrations of another, resulting in altered effects. Pharmacodynamic interactions can result in a change in the desired outcome (i.e., efficacy) or in undesirable outcomes (i.e., toxicity).

Ultimately, the pharmacological effect of a drug is a consequence of interaction(s) with specific molecular target(s), which induces downstream changes in ionic membrane gradients, intracellular signaling pathways, and even transcriptional regulation of specific genes. Thus, to fully understand pharmacodynamic drug interactions, one must first appreciate how each drug might independently affect neuronal function, and then examine the effects of coadministration on the usual monotherapy responses expected of each agent. Hence, pharmacodynamics really begins with an understanding of fundamental mechanisms of action.

Mechanisms of Actions of AEDs

Despite hundreds of basic studies addressing the mechanisms of action of antiepileptic drugs (AEDs), the relevance of these observations to the clinical effectiveness of these agents remains unclear. This problem is in large part due to the fact that epileptic seizures are manifestations of an abnormal network of neural and glial elements across multiple brain regions, and a mechanistic effect of an AED in one area may or may not be related to blockade of seizure activity.¹ Additionally, controlled mechanistic studies can only be performed in vitro, with tissue removed from the milieu in which in vivo effects are observed. Thus, it is not always clear whether a laboratory finding truly translates to the intact epileptic brain. Notwithstanding these limitations, by correlating efficacy studies in humans and in animal models with primarily cellular electrophysiological effects in isolated neurons and in brain slices, a widely accepted conceptual framework has emerged regarding putative mechanisms of AED action.^2–5

It should be understood at the outset that no single mechanistic finding is sufficient to explain all the clinical effects of a particular AED. Moreover, it is becoming clearer that every AED possesses multiple potential mechanisms of action, and that such diverse effects are dependent on numerous variables, including brain region, cell type, molecular composition of receptor targets, and drug concentration.^2,3,5 And there is an emerging consensus that the multiplicity of molecular action for any given AED is perhaps predictive and correlative with a spectrum of clinical activity across multiple seizure types. For example, topiramate’s actions on voltage-gated sodium and calcium channels, combined with the effects on voltage-gated sodium channels, γ-aminobutyric acid_A (GABA_A), and glutamate receptors, are consistent with efficacy against both partial and several forms of generalized seizures.⁵

Dozens of AEDs have been approved for clinical use since the introduction of the barbiturate phenobarbital in 1912. These drugs have been products of extensive testing in a variety of animal seizure models and in human clinical trials.⁶ However, the paradigms for drug discovery—until very recently—have been sharply biased toward the identification of candidate drugs similar to traditional agents such as phenytoin and phenobarbital. This is in large measure due to testing of compounds in a normal brain against acutely provoked seizures, rather than more appropriate screening in epileptic models that more closely mirror the human condition.⁶ Moreover, despite their clinical efficacy, current AEDs fail to “cure,” prevent, or modify the disease process. Rather, they eliminate the major symptom (i.e., seizures) by dampening neuronal excitation, synchronization, and spread of seizure activity. No clinical data exist to support the notion that these drugs are truly “antiepileptogenic” or “antiepileptic” (i.e., prevent the development or maintenance of the epileptic state).⁷

Two traditional models employed in routine screening and identification of new anticonvulsants are the maximal electroshock (MES) and subcutaneous pentylenetetrazol (PTZ or Metrazol) tests, conducted in rodents.⁶ The former tests the ability of a drug to block tonic extension evoked by an electrical stimulus, whereas the latter tests an agent’s ability to inhibit a generalized clonic seizure induced by subcutaneous administration of PTZ, a GABA_A receptor antagonist. MES seizures can be blocked by AEDs such as phenytoin and carbamazepine, which are effective against partial-onset seizures. In contrast, AEDs that are efficacious in the treatment of generalized absence seizures (e.g., ethosuximide) can inhibit PTZ-induced seizures. Valproate, a broad-spectrum agent, is active in both MES and PTZ tests and is clinically effective against most seizure types.⁸ However, these traditional screening tests may fail to identify drugs that may act through novel mechanisms. An example is that of levetiracetam, a newer AED that is clearly effective in the treatment of partial-onset seizures, but was originally found to be inactive in both MES and PTZ models, and thus initially discarded.⁶ The later observation that levetiracetam could retard kindling in rodents resurrected this unusual compound as an AED candidate.⁹

In general, three major classes of molecular targets are believed to be relevant for limiting seizure activity.² These include: (1) voltage-gated sodium and calcium channels, (2) GABA_A receptors, and (3) ionotropic glutamate receptors (i.e., NMDA or N-methyl-D-aspartate, AMPA, or α-amino-3-hydroxy-5-methyl-4-isoxazoleproprionic acid) and kainate receptors. Clinically useful AEDs exert their effects principally on one or more of these targets. Although a host of other targets could affect neuronal excitability (see following discussion), the validity of these voltage-dependent and ligand-gated ion channels toward AED action has stood the test of time.

For decades, the primary AEDs of choice for the treatment of partial-onset epilepsy have been phenytoin and carbamazepine. Both phenytoin and carbamazepine cause voltage-, frequency-, and use-dependent block of sodium channels in a wide variety of neuronal preparations.¹⁰ Sustained, high-frequency, repetitive firing of neurons, is believed to play a significant role in neuronal excitability and is potently inhibited by these AEDs at free plasma concentrations found in patients treated for epilepsy.³ Oxcarbazepine, a structural analog of carbamazepine that is reduced to a monohydroxy-derivative, also blocks sustained repetitive firing.¹¹ And based on this mechanistic profile, oxcarbazepine has been found to be effective only against partial seizures, similar to that of phenytoin and carbamazepine.¹²

Though structurally unrelated to phenytoin and carbamazepine, lamotrigine can also block sodium channels in a voltage-, frequency-, and use-dependent manner.¹³ However, lamotrigine possesses a broader spectrum of clinical activity than either phenytoin or carbamazepine, demonstrating efficacy against several forms of generalized seizures (especially absence seizures), in addition to partial seizures.¹⁴ The mechanistic basis for this difference remained unclear until recently. Lamotrigine was found to enhance activation of the hyperpolarization-activated cation channel (HCN channel), responsible for the so-called I_h or h-current.¹⁵ HCN channels are highly expressed in neuronal dendrites in both thalamus and hippocampus, are activated by hyperpolarization, and tend to stabilize the neuronal membrane potential against both hyperpolarizing and depolarizing inputs.^16,17

Postsynaptic GABA_A receptors are widely regarded as relevant toward the clinical effects of AEDs such as barbiturates and benzodiazepines. Binding of either benzodiazepines or barbiturates to their respective recognition sites on the GABA_A receptor results in enhanced chloride influx and, hence, membrane hyperpolarization.¹⁸ Benzodiazepines increase the frequency of GABA_A receptor channel openings, whereas barbiturates prolong the mean duration of these openings.¹⁹

Other agents have been shown to affect the GABAergic system as well. Vigabatrin is an irreversible inhibitor of the major degradative enzyme for GABA (i.e., GABA-transaminase),²⁰ and tiagabine is a potent and selective blocker of the GABA transporter, that which functions to reuptake GABA from the synaptic cleft.²¹ As predicted from their primary actions, both vigabatrin and tiagabine increase synaptic levels of GABA. Topiramate also increases the GABA_A receptor open duration and burst frequency in an allosteric manner,⁵ and felbamate enhances GABA_A receptor-mediated currents in hippocampal and neocortical through a barbiturate-like action.²² Valproate can evoke a wide variety of biochemical and neurophysiological changes in multiple neurotransmitter systems, but despite numerous studies, its precise mechanisms of action remain a mystery.⁸ Nevertheless, much of the evidence points to valproate’s effects on enhancing GABAergic transmission by enhancing biosynthesis and blocking degradation of GABA, resulting in elevated brain GABA levels.²³ At therapeutic serum levels, valproate also inhibits sustained repetitive firing of cultured mouse spinal cord and neocortical neurons, implicating actions on voltage-gated sodium channels as well.²⁴

Gabapentin and pregabalin are structural analogues of GABA, and although it had been predicted that these AEDs might act on the GABAergic system, they do not interact with either GABA_A or GABA_B receptors, and they do not affect GABA reuptake, synthesis, or metabolism.²⁵ Gabapentin and pregabalin bind uniquely to the α₂δ-1 and α₂δ-2 auxiliary subunits of the high-voltage activated L-type calcium channel.²⁵ It is believed that this interaction is important in decreasing presynaptic neurotransmitter release at both inhibitory and excitatory glutamatergic synapses. Interestingly, gabapentin was also found to enhance h-currents in hippocampal neurons²⁶ similar to the activity of lamotrigine (see earlier discussion).

Felbamate is the first pharmacological agent to both potentiate GABA_A receptor-mediated responses and inhibit NMDA receptor-mediated responses within the same drug concentration range.²² These dual actions are believed to contribute in a synergistic way to protect against seizure activity. Felbamate, like many other AEDs, can also block sustained repetitive firing of neurons, attributable to pathologic firing through voltage-gated sodium channels.²⁷ Similarly, topiramate blocks voltage-gated sodium conductances, but more important, it inhibits AMPA and kainate (specifically, GluR5) receptors.²⁸

Zonisamide is a broad-spectrum agent that has a unique mechanistic profile.²⁹ In cultured spinal cord neurons, zonisamide decreased sustained repetitive firing of action potentials, consistent with actions on voltage-gated sodium channels, and in cultured neurons from rat cerebral cortex, zonisamide blocked low-threshold T-type calcium currents, which predicts efficacy against generalized spike-wave epilepsies—specifically, absence seizures.²⁹

Levetiracetam, one of the newer AEDs, has broadened our conceptual understanding of relevant mechanisms of AED action. As noted earlier, unlike traditional AEDs, levetiracetam failed both MES and PTZ seizure threshold tests, yet had a profound effect in retarding amygdala kindling in rats.^6,9 Levetiracetam’s principal molecular target, originally identified as a specific high-affinity neuronal binding site, was recently demonstrated to be a specific synaptic vesicle protein, SV2A, which is involved in neurotransmitter release.³⁰ The functional consequences of such an interaction, although novel and intriguing, remain unclear.

Potassium channels represent an extremely diverse family of ion channels and generally decrease neuronal excitation by causing membrane hyperpolarization. As such, potassium channels represent a natural target for AED development. None of the currently available AEDs is believed to act primarily to enhance potassium channel activity, but recent studies implicate these channels—at least in part—in the anticonvulsant action of a number of AEDs, including phenytoin, carbamazepine, topiramate, levetiracetam, and possibly lamotrigine and zonisamide.^3–6 In contrast to these AEDs, retigabine, an investigational compound with broad efficacy in animal seizure models, acts primarily through enhanced activation of KCNQ2 and KCNQ3 potassium channels.^31–33 This molecular action is especially intriguing because a rare form of inherited epilepsy—benign familial neonatal convulsions—has been linked to mutations in genes encoding these potassium channel subunits.^34,35

Anticonvulsants known to be clinically effective against absence seizures (e.g., ethosuximide and valproate) can block a subtype of voltage-gated calcium channel known as the “low-threshold” or T-type calcium channel.^3,5 Although the role of T-type calcium currents in the genesis of absence seizures has been somewhat controversial,^36,37 the bulk of evidence supports the involvement of these channels.³⁸ Lamotrigine’s actions on h-currents may also contribute an antiabsence effect, as HCN channels are densely expressed in the thalamus and are critical regulators of pacemaker activity.^16,17 However, although it is appealing to think of absence seizures as simply a by-product of T-type calcium channel and/or h-channel dysfunction, the actual pathophysiological mechanisms are much more complex.³⁹

Finally, several AEDs act either principally or in part by inhibiting certain carbonic anhydrase isoforms.⁴⁰ These include acetazolamide, topiramate, and zonisamide. Carbonic anhydrase inhibitors have been used in epileptic patients for almost 50 years, but it is unclear how clinical effects are achieved with these drugs.⁴¹

Pharmacodynamic Interactions Influencing Efficacy

Often in clinical practice, two or more AEDs are combined in an attempt to achieve either seizure reduction or freedom. A number of prospective studies in patients with newly diagnosed epilepsy have shown that the majority will fully respond to trials of two to three AEDs, administered either as monotherapy or in combination.^42,43 Thus, when two or more AEDs are combined, an improvement in clinical response may be interpreted as either a positive pharmacodynamic interaction and/or infra-additive toxicity. More often than not, combination AED therapy has resulted in pharmacokinetic and pharmacodynamic toxicity and not necessarily significantly improved seizure control.⁴⁴ As shown in Figure 18-1, plasma concentrations need to be measured at the time of event to determine whether an interaction is pharmacokinetic or pharmacodynamic.⁴⁵

Figure 18–1 Schematic representation of an excitatory synapse in the central nervous system and the putative major sites of action of various antiepileptic drugs (AEDs). Plus or minus signs denote activation/potentiation or inhibition, respectively. NMDA, N-methyl-D-aspartate; AMPA, α-amino-3-hydroxy-5-methyl-4-isoxazoleproprionic acid; HCN, hyperpolarization-activated cyclic nucleotide-regulated cation channels; SV2A, synaptic vesicle protein 2A.

Modified with permission from Rho JM, Sankar R. The pharmacologic basis of antiepileptic drug action. Epilepsia. 1999;40:1471-1483.

Overall, it has been extremely challenging to determine whether a particular AED combination produces a beneficial effect due to additive or synergistic effects stemming from enhanced activity at their molecular sites of action. This is because a number of factors determine—dependently and independently—clinical effects of AEDs, including: (1) the variable and often unpredictable pharmacokinetic interactions that ultimately influence delivery of the drugs to their brain targets; (2) the narrow therapeutic indices of most AEDs, which can predispose the patient to loss of seizure control, toxicity, or perhaps, rarely, an improved response; (3) the chronicity of AED treatment, that can result in induction of various forms of drug tolerance and other adaptive changes; (4) the high incidence of AED cotherapy with nonepilepsy drugs; and (5) evolution of the epileptic state, especially in infants and children.

Despite these limitations, investigators have turned to animal models to study the impact of AED combination therapy. Even in animals, however, despite over a hundred published studies addressing greater than 500 AED interactions, we do not yet have a clear understanding of the pharmacodynamic properties of specific AED combinations.⁴⁶ Concomitant pharmacokinetic changes of AED cotherapy have not been properly addressed in the majority of preclinical studies to date, and most have been conducted in rodents using acute seizure models, and thus may not be wholly relevant to the chronic epileptic condition. Nevertheless, a couple of intriguing approaches have been taken in preclinical studies.