IT WAS an impromptu straw poll of his patients that made Canadian neurologist Don Weaver change his career. Four years ago he had asked 100 patients visiting his clinic at Queen’s University in Kingston, Ontario, what would be their dream epilepsy drug, if they could design one to do whatever they wanted. “I was expecting the majority of people to tell me about fewer side effects, and once-per-day dosing,” he recalls.
In fact more than three-quarters of his patients said their dream drug would have stopped them developing epilepsy in the first place. To date, no such therapy exists. All existing epilepsy medicines are anticonvulsants, designed to suppress the seizures that are the disease’s symptoms. They are not always completely effective, and often have nasty side effects. Nothing is licensed to treat the disease’s root cause, and in the past the amount of research in this area has been dwarfed by efforts to develop better anticonvulsants.
To Weaver, that is all wrong, so he decided to switch his focus from the clinic to the lab, to try and develop that dream drug. Over the past few years a growing number of researchers have joined the quest to create a drug that can stop epilepsy in its tracks.
Advertisement
Many people with epilepsy would have been ideal candidates for preventive treatment. It may often be thought of as a disease that people are born with, but about half of patients only acquire it after some kind of damage to the brain. Although sometimes the cause is unknown, it is often a blow to the head, a brain infection, stroke, tumour or a childhood fever that brings on convulsions, though of course most people escape this consequence. What Weaver and like-minded researchers want to do is develop drugs to be given soon after the damage, saving people from a lifetime of taking anticonvulsants.
The researchers have set themselves a tough task, as mystery still surrounds what exactly happens during the development of epilepsy. In some cases, the process – which is known as “epileptogenesis” – progresses rapidly, taking only a few weeks. In others, the first seizure doesn’t strike until more than a decade after the damage was sustained. We have a few clues about some of the changes at the cellular and biochemical level, but it is unclear which are causes of the condition, and which the effects, and the competing theories are at times the subject of fierce dispute.
Despite our relative ignorance, Weaver and other scientists have reached the stage of testing potential agents in animals, and there are even a few early-stage human trials. It may not be too far-fetched to hope that one day, neurologists will have their dream drug that rescues patients from a lifetime of epilepsy.
So what is known about epilepsy? In the brain, neurons are in a state of continual activity, with electrical impulses travelling along their long, strand-like projections, or “axons”, in response to messages from neighbouring neurons. The cells communicate at junctions known as synapses by releasing chemical signals called neurotransmitters. Some of the signals are “excitatory”: in other words, they trigger an impulse in the next neuron. Others are “inhibitory”: they suppress another neuron’s firing. Most neurons connect with many others, perhaps hundreds, and most excitatory circuits are accompanied by inhibitory ones, to keep the firing under control.
A seizure occurs when the balance between excitatory and inhibitory signals is disturbed, and neuronal firing starts to escalate. The frenzy of electrical activity can occur in just one small area, or it can spread through the brain. In most cases the seizure comes to an end spontaneously, but if this doesn’t happen – a condition known as status epilepticus – untreated patients can die.
Anyone can suffer a seizure under certain circumstances. Lack of sleep, for example, too much alcohol, or certain drugs can reduce inhibition or raise excitation in the brain, making a fit more likely – lowering the “seizure threshold”. The 3 per cent of the western population who have epilepsy at some point in their lives have a permanently reduced seizure threshold. The key question is what happens during those hours, weeks or months after a brain injury that results in a lowering of the seizure threshold.
Investigations have focused on two broad areas: the structural changes in the brain, and alterations to the brain’s biochemistry. As well as animal research, a major source of information about the structural changes is brain tissue removed from people with severe epilepsy to get rid of the source of the seizures. Although such surgery seems a drastic step, for some patients it is considered worth the risks.
Most research has focused on temporal lobe epilepsy – in which seizures originate in the temporal lobes, which are on the side of the brain – as is the most common acquired form of the disease. The hippocampus, in the middle of the temporal lobe, seems particularly susceptible to damage, and one of the hotspots in the hippocampus is an area called the dentate gyrus. The role that cells in this area play is one of the most hotly debated issues in epileptogenesis, and advocates of the two main competing theories have been duking it out since the early 1990s.
Neurons in the dentate gyrus known as granule cells are known to receive excitatory inputs from outside the hippocampus, and have long axons known as mossy fibres, which contribute to the hippocampus output (see Diagram). The granule cells are kept in check by inhibitory neurons within the hippocampus known as basket cells.
In a paper published in Science in 1987, Bob Sloviter of the University of Arizona reported that, in a rat model at least, the neurons that excite the inhibitory basket cells seem to die, leading to the basket cells becoming quiescent. “The cells that normally activate inhibitory neurons are among the most vulnerable to injury,” Sloviter says. According to this “dormant basket cell hypothesis”, the lack of inhibition leads to the granule cells becoming hyperexcitable.
The competing theory arises from observations that humans with epilepsy have more mossy fibre growth and axon branching than usual. According to the “mossy fibre hypothesis”, they form synapses with other granule cells, setting up a positive-feedback loop that can lead to their activity spiralling out of control. “You can get an explosive effect with just a few connections,” says Ed Dudek, a neuroscientist at Colorado State University.
But Sloviter reckons the aberrant mossy fibres actually innervate the dormant basket cells in an attempt to reactivate the inhibitory circuit. “We think mossy fibre sprouting is a protective mechanism,” he says.
Between the competing camps are other researchers who think both theories may have some truth to them.
Tipping the balance
Most of the studies on which these theories are based have been done on hippocampal tissue. But Sloviter points out that the brain is full of inhibitory basket cells that may become dormant, prompting hyperexcitability in other brain areas. Over the past 10 years axon sprouting similar to that of the mossy fibres has been observed in the cerebral cortex, which may lead to other forms of epilepsy.
Another possible – though by no means undisputed – player in the development of epilepsy is neurogenesis, the birth of new neurons. Granule cells in the dentate gyrus are one of the few types of neuron that continue to divide in adulthood, and seizures have been shown to boost cell division in the dentate gyrus in rats. “Some of the new neurons might go into abnormal regions and tip the balance towards excitability,” suggests Dan Lowenstein of the University of California, San Francisco, who carried out the research.
So much for changes in the structure of the brain; what about biochemical changes? It has long been known that excessive neuronal firing leads to the release of damagingly high levels of the neurotransmitter glutamate, which in turn leads to cell death. This process, known as excitotoxicity, makes the damaged area of the brain more prone to further seizures, though it is unclear why. Whatever the mechanism, excitotoxicity may be the underlying reason why some epilepsy patients have more frequent seizures as time goes on.
One of the ways of studying the role of excitotoxicity in epilepsy involves giving rats or other animals a small electric shock or dose of toxic chemical at regular intervals. At first this has no noticeable effect, but over time it lowers their seizure threshold until after about two weeks the same small stimulus induces seizures – an effect known as kindling (as in lighting a fire). But the animals never experience spontaneous seizures, so many researchers now consider this a poor model for human epileptogenesis and have turned to other approaches. For example, in what is known as the status-epilepticus model, animals are given a severe shock or high drug dose that immediately induces a long seizure, after which most animals go on to have spontaneous ones.
Kindling research has, however, given us some clues about what goes on in people. For example, the biochemical changes in the brains of kindled animals fit the pattern of too much excitatory signalling and not enough inhibition. The main excitatory neurotransmitter is glutamate, while the main inhibitory one is another amino acid called GABA (gamma-aminobutyric acid). Some studies have suggested that kindling produces more synapse receptors for glutamate, while others point to it producing fewer receptors for GABA. Researchers have also seen reduced levels of a cell membrane ion transporter called KCC2, which is essential for GABA signalling.
A possible intermediary is a growth factor of the nervous system called brain-derived neurotrophic factor (BDNF), which is known to increase axon branching. Mice with only one functional copy of the BDNF gene appear to be less susceptible than normal mice to attempts to induce kindling. In other animal experiments, chemicals that deactivate BDNF have also been found to slow kindling. In animals, levels of BDNF have been found to increase tenfold after a brain injury, and the number of BDNF receptors also increases.
And in the past few years a whole new type of cell has been implicated in seizures: glial cells, traditionally seen as a support network for neurons. Glia maintain levels of ions and neurotransmitters, so disrupting this balance could make nearby neurons hyperexcitable. This theory, however, is still only preliminary.
Which of the changes are cause and which effect is still unclear, as is how events at the biochemical level tie in with those at the cellular or anatomical level. “The big question is how a fleeting experience can leave a lasting effect on the brain,” says James McNamara, a neuroscientist at Duke University in Durham, North Carolina. A plausible chain of events is that brain injury leads to excitotoxicity and cell death, prompting an imbalance between excitation and inhibition, resulting eventually in the cellular and anatomical changes that cause epilepsy.
The complexity of events during epileptogenesis presents drug developers with a problem. Which process should they target? “Is it cell death or axon growth?” asks Asla Pitkänen, an epilepsy researcher at the University of Kuopio in Finland. “We know BDNF is important,” says McNamara. “But we don’t know if it is necessary.”
Nonetheless, the potential market for drugs that block epileptogenesis is enticing enough for researchers and drug companies to take a gamble, using the clues they already have. “Ten to 15 per cent of people with head trauma go on to develop seizures, so millions could benefit from such an approach,” says Francine Gervais of Neurochem, a Canadian biotech firm based in St Laurent, Quebec. And drugs that block excitotoxicity could have wider benefits, by preventing other problems sometimes seen after head injuries, such as deficits in attention, memory and speech.
Golden hour
BDNF is one obvious starting point, and McNamara’s group is planning to find chemicals that interfere with the BDNF pathway – though none are yet known.
Another possible avenue is to use anticonvulsant drugs, as these generally work by reducing neuron excitation or increasing inhibition. What’s more, they are already given to patients with severe head trauma to try to prevent seizures immediately after the injury. However, clinical trials by Nancy Temkin at the University of Washington in Seattle of the anticonvulsants phenytoin and valproate have not proved successful. While they prevented seizures in the short-term, they had no long-term effect in preventing epilepsy from developing.
This does not surprise Pitkänen. “I don’t really believe anticonvulsants will prevent epileptogenesis,” she says. “They are designed to stop seizures, and the mechanism of seizures is different to epileptogenesis. It’s like treating cardiac infarction with something used to treat headaches.”
But anticonvulsants are by no means out of the game. Temkin plans to test another one, magnesium sulphate, and the pharmaceutical company Johnson & Johnson says that its anticonvulsant topiramate seems to have some effect in animals.
It could be that these drugs would have a stronger effect if given sooner after injury. In Temkin’s trials, the time limit was 24 hours, and Larry Benardo, a neurologist at the State University of New York in Brooklyn suggests that this might be too long to wait. “Damage from excitotoxicity occurs quickly, so it’s important to intervene quickly,” he says. As with existing medicines to limit bleeding in the brain after stroke, drugs to prevent epilepsy should be given as soon as possible, preferably within the “golden hour” immediately following injury.
This idea is supported by lab studies to measure the excitability of slices of rat brain. Benardo has found that valproate slows epileptogenesis, but only if given within 30 minutes of injury, and he now plans to retest its effectiveness in humans when given within one hour of injury.
Weaver, whose patient survey sparked his career change, and who is now at Dalhousie University in Halifax, Nova Scotia, is researching a class of compounds that are known to both stimulate the GABA system and suppress glutamate. “We speculate both GABA and glutamate-responsive neurons are affected during epileptogenesis,” he says. “So a comprehensive approach should target both neurotransmitters.”
Weaver’s primary candidate is the chemical messenger beta-alanine, which prevents breakdown of GABA and activates one class of GABA receptor. It also attaches to a sub-type of glutamate receptor without activating it, so blocking binding by glutamate itself. Weaver is working with Neurochem to carry out animal tests of beta-alanine, as well as chemically similar compounds that have better biochemical properties. In a test of their most promising compound only one out of five rats had seizures, compared with all five controls.
How long it is before the work of Weaver and other scientists bears fruit is hard to predict. The mechanisms of epileptogenesis are still far from clear, but basic science is helping to disperse the mists.
And increased recognition of the importance of epileptogenesis research is helping to point the way to new kinds of drugs. “People are coming to grips with the idea that there has to be something better than anticonvulsants,” Weaver says.
