EPILEPSY is bewildering. The sudden and violent nature of a seizure is disturbing to witness, let alone experience. Small wonder, then, that people down the ages have blamed the seizures on demonic possession. It wasn鈥檛 until around 2400 years ago that Hippocrates, the father of modern medicine, first reasoned that seizures must have a natural, physical cause. His explanation 鈥� an imbalance among the four humours, blood, black bile, yellow bile and phlegm 鈥� may seem quaint, but the causes of epilepsy still remain a puzzle.
Epilepsy looks as if it should be a genetic disease because it occurs when brain cells misbehave in ways that you would expect to be controlled by genes. What鈥檚 more, childhood epilepsy often runs in families, and geneticists have even identified some of the genes involved. But the case for inheritance is undermined by many isolated instances of the disease that look identical to familial epilepsy, and by the fact that familial epilepsy can take different forms. Then there are other anomalies. Why do your chances of developing epilepsy drop as you get older? Why does the disease tend to be more severe and affect more of the brain when the first seizure comes early in life?
It was with such questions in mind that the US National Institutes of Health decided to hold a meeting on the genetics of epilepsy a few years ago. I was invited to participate: I am not a neurobiologist, but the organisers wanted to draw on my expertise as a geneticist, in the hope I might suggest better ways to design studies of families with epilepsy. They were unprepared for my response. Instead of recommending bigger studies or analysis of more genes, I suggested rethinking the whole notion of what we mean by the word 鈥済enetic鈥� in relation to epilepsy. I believe the disease is genetic, but not necessarily inherited.
Advertisement
The resistance I encountered was not totally unexpected. At the time, my thinking was unorthodox, even among geneticists. But, even then, there was some evidence for the idea I presented. And recent findings by a team from Harvard University offer additional support. I am becoming increasingly convinced that my view is right. Rethinking our notions of genetics could have far-reaching benefits. It could lead to better treatments and new insight into a variety of neurological conditions and other kinds of diseases that pose similar puzzles to epilepsy.
Around 3 per cent of people in the western world are affected by epilepsy. About half of these cases arise from damage to the brain, such as head trauma. The other cases are either in families with a history of the disease or people who seem to be born with a subtle predisposition towards seizures. The mechanism that triggers seizures is still a matter for debate (see New 杏吧原创, 8 May, p 36), but we do know that they happen when many interconnected neurons begin sending frantic messages to each other, disrupting their ability to carry organised information and leading to a firestorm of uncontrolled activity. This can be confined to a localised area of the brain or be widespread. The effects are similarly variable, causing mild disorientation, or 鈥渁bsence鈥�, in some cases and in others leading to a full-blown fit, which may include uncontrolled eye or muscle movements and breathing difficulties.
What is 鈥済enetic鈥�?
In recent years, specialists have come up with a variety of explanations for the anomalies associated with epilepsy. Perhaps each family has a different affected gene; perhaps different lifestyles account for different forms of the disease in the same family; perhaps isolated cases are due to environmental effects that somehow mimic genetic ones; and so on. But if each case needs a unique explanation, that doesn鈥檛 get us very far towards understanding the disease, and is even less useful in suggesting treatments. That is why I am urging investigators to consider another possibility.
Perhaps we can get out of this muddle by reconsidering what we mean by 鈥済enetic鈥�. The word is mainly used in two ways. The first meaning is mechanistic: genes code for specific proteins, each with a particular function in our bodies. In the brain, for example, genes code for the proteins that make the neurotransmitters that control the firing of neurons. So in epilepsy, the lack of control of neural signals is 鈥済enetic鈥�.
The second use of the word refers to inheritance. All genes vary and these variants are passed down the generations. An undesirable mutation in a gene that codes for a protein that is part of the ion channels in the cell membranes of neurons, for example, could cause them to misfire in a cascade that produces a seizure. If that variant gene were passed from parent to child, the resulting epilepsy would be 鈥済enetic鈥�.
But that definition cannot account for non-familial epilepsy, and neither concept adequately explains why cases can vary so much even in the same family. Perhaps a third, much-neglected, use of the term 鈥済enetic鈥� may be more enlightening. We tend to think of mutations arising in the egg or sperm cells and being passed on to offspring, but they can arise in any cell of your body, throughout your lifetime. As cells divide and multiply these genetic changes will be passed down to their descendant cells. So, somatic mutations tend to end up in patches of tissue, creating localised and variable genetic abnormalities: just what you find in many forms of epilepsy.
鈥淲e tend to think of mutations as arising in the egg or sperm, but they can arise in any cell, throughout your life鈥�
But hold on. Surely one of the noteworthy peculiarities of neurons is that they do not divide once the brain has formed. A somatic mutation in a neuron might kill that cell or make it misbehave, but that would only be a tiny loss among the billions of neurons in the brain. How could it do more harm than that?
The answer lies in how the brain develops and how it works. The brain, like all organs, develops from a few initial cells. Early on in embryonic development, cells that will form the left and right hemispheres separate. Then, step by step, the local regions and cell types differentiate, migrate and interconnect. Copies of any somatic mutation will appear in parts of the brain that descend from the originally mutated cell. The earlier a mutation occurs, the larger the number of cells that will inherit it and the more widespread its effect will be. Mutations occurring later in development will be more localised, affecting perhaps a single hemisphere, or one highly focussed region.
If somatic mutation played a role in epilepsy, that would explain why some fits are severe, affecting most of the brain and much of the body, while others are confined to just one part of the brain, producing only mild absences. And if mutations occurring later in embryonic brain development manifest their effects later, that would explain why late-onset epilepsies are usually more localised or less severe. Somatic mutation is also consistent with the observation that people with no family history of epilepsy can often have the same symptoms as those whose epilepsy is inherited from a parent. Genes that code for neurological functions, such as those producing neurotransmitters, are expressed in neurons in specific parts of the brain, so a mutation in such a gene can affect those parts of the brain regardless of whether the mutation is inherited or occurs during development. On the other hand, somatic mutation may also explain why epilepsy can take different forms among members of the same family. Each may have inherited the same mutation, but epilepsy might arise only in cells that also carry additional and varied somatic mutations.
Extended reach
Even more variability arises from the fact that mutations occurring late in development, and so carried by just a small fraction of neurons, may be able to extend their reach. That鈥檚 because uncontrolled firing of cells in a local area can excite neighbouring normal cells, setting off a chain reaction of impulses 鈥� the electrical 鈥渂rainstorm鈥� of an epileptic seizure. Pre-existing somatic mutations might even help explain why some people develop epilepsy after a head injury, fever or other trauma. Such events seem to trigger misfiring in neurons that are vulnerable to seizures, perhaps because of somatic mutations that had created no previous problems. But once involved in a seizure, neurons might then be modified, 鈥渂urning in鈥� a tendency to misbehave occasionally in the future, or becoming even more damaged.
At the National Institutes of Health meeting I predicted that somatic mutations would be implicated not just in epilepsy, but also in other neurological diseases that seem to be caused by misbehaving neurons. I supported my claim with the arguments outlined above. But recently, Bruce Yanker and colleagues at Harvard University reported a new type of direct evidence for this kind of phenomenon by comparing brain tissue from 30 individuals who died at various ages from 26 to 106. They extracted messenger RNA 鈥� the template for making proteins from genes expressed in a given cell 鈥� from the frontal part of the cerebral cortex of each individual and found that the amount of somatic mutation in genes affecting the function of brain cells increases with age.
This shows somatic mutation is not restricted to embryonic brain development but can also occur in non-dividing neurons of the adult brain. Such mutations will be in single cells scattered throughout the brain rather than in clusters of cells, but the authors suggest that enough of these mutations accumulate over time to contribute to Alzheimer鈥檚 and other diseases. Although Alzheimer鈥檚 is very different from epilepsy at the cell level, it shares many of the same perplexing patterns of variability (Nature, vol 429, p 883).
鈥淐opies of any somatic mutation will appear in the parts of the brain that descend from the originally mutated cell鈥�
If my idea is correct, what are the implications for future research into epilepsy? Because somatic mutations are not transmitted from parent to offspring, we will need to start looking at regions of the brain where the genetic changes responsible for seizures actually have their effect. We can then identify genes that are normally active in these areas and look for the effects of mutations, whether inherited or somatic. Localised somatic mutations may also provide 鈥渕olecular tags鈥� for the cells that carry them, so researchers can single them out from normal cells and target them for treatment.
But there is much more. Somatic mutations affect all cells, not just neurons, and all organisms. By extending our definition of 鈥済enetics鈥� to include what happens across generations of cells within an organism鈥檚 lifetime, we could develop different approaches to treat a whole variety of diseases, not to mention a deeper understanding of some of the weirder, but normal, manifestations of nature.
If this all sounds a bit abstract or revolutionary, remember that in one area of medicine at least, the idea of somatic mutations is already firmly in place. While most mutated cells either die or go unnoticed, surrounded by millions of normal cells, sometimes the genetic change causes them to divide rapidly and multiply out of control. The result is a somatic mutant with which we are all too familiar: a cancer.