Waves of unconstrained neural firing seem to be related to signal transfers between connections between neurons (synapses) in the brain. Since signal molecules are basically genetic (that is, not dietary or due to exogenous chemicals), one would expect epilepsy to be basically 'genetic' in its basic causal physiology. Indeed, some epilepsies have long been known to result from mutations in single genes, occur in Mendelian fashion in families, involve neural signaling molecules, and some of these genes are known to cause epileptic symptoms when mutated in laboratory mice. But most epilepsies are sporadic, not obviously familial, and have resisted attempts to map the responsible genes.
This was the general picture in around 2004, when I was asked to give a presentation to a meeting at NIH, of people working on epilepsy, to discuss what I saw as the general genetic causal landscape at the time. This was early in the age of GWAS and other genomewide approaches.
I don't remember the details, but in my presentation I gave a general picture and issues in genetic mapping, and relevant evolutionary genetics. But I also wanted to make a contribution beyond just a review, and I had thought about the genetics, physiology, and epidemiological patterns of the epilepsies. I suggested what I thought would be relevant, based on previous work I had done over the years on the genetics of cancer and aging. Basically, I hypothesized that many epilepsies might, like many cancers, be due not to inherited mutations but to somatic mutations--that is, mutations in relevant genes that occurred in neural cells during embryonic development or later in life, rather than being inherited from parents.
I found considerable interest in this idea, which I thought (and still think) could be a positive, potentially innovative way to understand the biology and epidemiology of epilepsies. Some interest was shown, and a leading epilepsy neurobiologist offered to work on a paper about this idea and how one might test for it. But for whatever reason, he lost interest before any paper could be published. Partly perhaps, that was because of the difficulty of testing the idea, which would have required identifying the specific misfiring neurons in epilepsy victims and, when they died, sequencing those areas looking for mutations not found in neighboring normal neurons.
After my potential collaborator had more important things to do (or, probably, more important people to please) and dropped the idea of doing a collaborative review paper, I published my ideas in Trends in Genetics in 2005 (paywalled; if you're interested, email me for a pdf). With some helpful discussion with a then-junior colleague, Dan Burgess (now working at Roche, I believe), we thought about ways in which, depending on embryonic or postnatal age, somatic mutation in neural cells or their precursors, could generate epileptic effects. The idea is shown in this figure, drawn in collaboration with Dan, and taken from the paper.
|Possible Epileptogenic somatic mutation scenarios. From Weiss, TiGs, 2005; drawn with Dan Burgess|
Alternatively, a single later-occurring mutation might make the affected neuron too ready to fire under some conditions, and that could entrain firing in the often thousands of other neurons with which it synapses, and their neurons in turn In a sense, such an episode could 'burn in' the synaptic connection in this set of neurons, all genetically normal except for the mutant triggering cell, making the set vulnerable to later episodes triggered by the original offending cell. The location in the brain of the 'parent' aberrant cell(s) and its or their synaptic network would determine the side and part(s) of the brain that were affected by the seizures.
For somatic mutation to have detectable effect at the organismal level, there must be some sort of phenotype amplification, such that even a single or few aberrantly firing cells, that might on their own not be particularly noticeable would entrain enough other cells to cause a seizure. Cancers are now extensively studied for somatic mutation, because of their exuberant growth and histological characteristics. One transformed cell leads to a large number of descendant cells--a tumor.
As noted above, aberrantly behaved neurons even if not individually detectable could, in principle at least, entrain so many other otherwise normal cells in a synaptic network, as to amplify the effect to make it noticeable to the person as a whole and to clinicians. Thus in a variant of the precedent of cancer, one mutant transformed neural cell early in the embryo may be unnoticeable in itself, but after millions of cell divisions it can lead to a clone of misfiring cells.
With this sort of thing in mind, one can naturally also ask about a potential role for somatic mutation, arising in developing neural cells after fertilization in the embryo's later life, in other aspects of brain function--even including traits like intelligence, behavior, learning ability, memory, and personality, or other pathologies such as schizophrenia? Behaviors and especially behavior genetics are highly contentious areas. Some would like these traits and abilities to be due entirely to environments, while others are fervid in their belief that you are what you inherited in your genome--what you are is inborn. That was the view, of course, of the eugenics movement and it's still around today in the belief system of many.
We've posted on this general topic before. The argument for inherency has always been assumed to be about germ-line genotypes, that is, inheritance from parent to offspring. But could similar effects arise by somatic mutation--or, rather, if they can be inherited how could they not in some instances be due to somatic mutation?
Testing for somatic mutation in regard to brain function, even specific traits like epilepsy, has been prohibitively difficult, because specific somatic mutations would need to be identified systematically in specific brain areas or subsets of neurons. Detecting and characterizing somatic mutation in traits like epilepsy will be challenging and may not quickly lead to therapy, but it could at least illuminate etiology and mechanisms.
At least, the idea that I laid out (and others may have as well, though not that I know of), hasn't been tested. But that may be changing, as we'll discuss tomorrow.
[Since first being posted, this has been edited, twice, for clarity and to correct inapt phrasing]