Guilt by association: SCN1A in Temporal Lobe Epilepsy

GWAS. Genome-wide association studies investigate the association of common genetic variants with disease in large patient samples. While this approach has been very successful in many other diseases, the results in epilepsy research have been less convincing. Given the complexity of epilepsy phenotypes, selection of the right epilepsy phenotype has been an ongoing debate. Now, a recent study in Brain finds an intronic variant of the SCN1A gene that is associated with Temporal Lobe Epilepsy (TLE), the most common epilepsy in man. Interestingly, the association with SCN1A seems to be specific for only a particular subtype of focal epilepsies.

Association studies. In a belated effort of spring cleaning, I threw out a bunch of old papers today that I hadn’t looked at in the last few years. One of the publications that I kept was an editorial entitled “Guilt by association” by David Altshuler and colleagues. I have taken the liberty of recycling the title of the 2000 publication for this post. Back in 2000, the first common genetic variants associated with human diseases were identified. Since then, much has happened with respect to genetic studies. Novel platforms made it possible to genotype large patient cohorts for common variants and, more recently for large parts of the human genome, if not the entire genome itself. Genome-wide association studies (GWAS) investigate hundreds of thousands of common genetic markers, so-called Single Nucleotide Polymorphisms (SNPs). In a rigorous study design, variants associated with disease will be filtered out of the baseline noise of statistical fluctuation. This strategy, combined with a validation step and subsequent meta-analysis, was applied by Kasperaviciute and colleagues in Temporal Lobe Epilepsy.

A Forest plot of the study by Kasperaviciute and colleagues, demonstrating the risk found in various subcohorts alongside the confidence interval. Cohorts are ordered by size of the cohort and the size of the rectangle indicates the size of the cohort. The diamond symbol indicates the result of the meta-analysis. This Forest plot indicates quite nicely how smaller cohorts may sometimes have seemingly counterintuitive results as in the case of the Italian cohort that has an opposite association signal. However, the size of the cohort was relatively small and therefore, the confidence interval was quite large, safely covering the overall, final association signal. The Forest plot is from the publication by Kasperaviciute and colleagues, which was published under a Creative Commons licence. HS = hippocampal sclerosis; FS = Febrile Seizures

A Forest plot of the study by Kasperaviciute and colleagues, demonstrating the risk found in various subcohorts alongside the confidence interval. Cohorts are ordered by size of the cohort and the size of the rectangle indicates the size of the cohort. The diamond symbol indicates the result of the meta-analysis. This Forest plot indicates quite nicely how smaller cohorts may sometimes have seemingly counterintuitive results as in the case of the Italian cohort that has an opposite association signal. However, the size of the cohort was relatively small and therefore, the confidence interval was quite large, safely covering the overall, final association signal. The Forest plot is from the publication by Kasperaviciute and colleagues, which was published under a Creative Commons licence. HS = hippocampal sclerosis; FS = Febrile Seizures

TLE and its variants. Temporal Lobe Epilepsy is the most common epilepsy in adults. Epilepsies from the temporal lobe might be due to a large range of factors, ranging from monogenic disorders running families to epilepsies due to trauma or injuries. In most cases, however, the underlying cause is unknown. A classical entity within the temporal lobe epilepsies is the presence of hippocampal sclerosis (HS), referring to a scarring of hippocampus. Given the location of the hippocampus, the subform is then referred to as Mesial Temporal Lobe Epilepsy (MTLE) as opposed to subforms of TLE arising from the lateral aspect of the temporal lobe. Many patients with MTLE and HS have a history of febrile seizures. It has been suggested that prolonged Febrile Seizures lead to Temporal Lobe Epilepsy with hippocampal sclerosis, even though there might be a latent period of one or two decades between the initial Febrile Seizure and the onset of the epilepsy.

SCN1A in TLE. Kasperaviciute and colleagues investigated ~1000 patients with MTLE+HS compared to ~7500 controls. The authors identified a strong association with two SNPs within intron 1 (rs7587026) and intron 3 (rs11692675) of the SCN1A gene. Both variants were then confirmed in a second cohort of a comparable size. When the authors broke down the association by phenotype, they realized that the association signal is entirely due to patients with a history of Febrile Seizures, while MTLE+HS without previous Febrile seizures did not shown any association. Intrigued by this finding, the authors also looked at ~500 patients with Febrile Seizures only and other focal epilepsies with preceding Febrile Seizures. There was no association in either group, leaving the triad of MTLE, HS and FS as the only phenotype associated with the SCN1A variants.

Neonatal form absent. What could intronic variants possibly do to the SCN1A gene? Kasperaviciute and colleagues looked at the gene expression in human brain from a tissue bank. Both the rs7587026 and the rs11692675 SNP lead to a markedly reduced expression of a specific SCN1A variant that is considered a neonatal form. This finding suggests that both SNPs might alter the variant of the SCN1A protein that is generated, which might have subtle consequences for the overall membrane excitability.  While it cannot be excluded that both variants are linked with unknown variants with the SCN1A gene that alter the protein, a delay or shift in subunit expression might be an interesting hypothesis to explain the distinct age distribution of Febrile Seizures.

Attributable risk. As with many other GWAS findings, the presence of both SNPs only increases the risk for epilepsy by a factor of 1.3. Properly speaking, these variants have an odds ratio of 1.3, but for many situtations, this can be seen as an approximation of the relative risk. At first glance, an increase by 1.3 does not seem much and seems negligible compared to the strong risk contributed by de novo mutations in epileptic encephalopathies. However, through their overall frequencies, the risk in the population that can be attributed to this SNP is remarkable. The risk variant is present in roughly 30% of the population and 34% of patients. This means that the population-attritutable fraction is 3-4%, which is the fraction of cases that could be prevented if the effect of this SNP could be compensated or treated. However, the SCN1A variants are only relevant on a population level and the risk for an individual patient is limited. This is in contrast to de novo SCN1A mutations in Dravet Syndrome that explain the phenotype in an individual patient, but are rare amongst a broader range of epilepsies.

Relevance for EuroEPINOMICS. The study by  Kasperaviciute and colleagues established two intronic SCN1A variants as risk factors for the triad of MTLE, HS and FS. Also, the study has sufficient power to single out MTLE, HS and FS as the phenotype that drives the association, while a broad range of other epilepsies show no association. Usually, phenotypes cannot clearly delineated in association studies and this clear-cut finding in the study by Kasperaviciute and colleagues is exceptional, emphasizing that we are finally on the right path to delineating phenotypes according to their genetic architecture. It will be interesting to see whether the SCN1A association signal in MTLE, HS and FS adds to the association seen between SNPs within the SCN1A region and Idiopathic Generalized Epilepsy or whether both signals will cancel each other out in a meta-analysis.

13 thoughts on “Guilt by association: SCN1A in Temporal Lobe Epilepsy

  1. Pingback: The genetics of treatment response in newly diagnosed epilepsy | Beyond the Ion Channel

  2. Pingback: Identifying core phenotypes – epilepsy, ID and recurrent microdeletions | Beyond the Ion Channel

  3. Pingback: Epigenetic signatures – profiling the epilepsies beyond genetics | Beyond the Ion Channel

  4. Pingback: From unaffected to Dravet Syndrome – extreme SCN1A phenotypes in a large GEFS+ family | Beyond the Ion Channel

  5. Pingback: Infantile Spasms/Lennox-Gastaut genetics goes transatlantic | Beyond the Ion Channel

  6. Pingback: Copy number variations and the forgotten epilepsy phenotypes | Beyond the Ion Channel

  7. Pingback: Modifier genes in Dravet Syndrome: where to look and how to find them | Beyond the Ion Channel

  8. Pingback: Mining GWAS mountains for missing heritability | Beyond the Ion Channel

  9. Pingback: Papers of the week – next-level mutation classifiers and 3′UTR variants in Dravet Syndrome | Beyond the Ion Channel

  10. Pingback: The familial risk of epilepsy – revisited | Beyond the Ion Channel

  11. Pingback: A polygenic trickle of rare disruptive variants in schizophrenia | Beyond the Ion Channel

  12. Pingback: The OMIM epileptic encephalopathy genes – a 2014 review | Beyond the Ion Channel

  13. Pingback: SCN1A – This is what you need to know in 2014 | Beyond the Ion Channel

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s