Autism and the methylome

Q: When is an identical twin not an identical twin? A: Pretty much all the time (at least according to some people) as our increasing understanding of the complexity of genetics stretches and modifies long-held beliefs about the building blocks of life and their role in our health and wellbeing.

DNA methylation differences? @ Wikipedia  

Today I'm talking about identical (monozygotic) twins - siblings derived from one fertilised egg - and how the science of epigenetics might have some interesting implications for autism research as per the study by Chloe Wong and colleagues* (open-access). I should perhaps also direct you to some other interesting research recently discussed in this area too (see here) which might be relevant.

Regular readers will probably already know about my amateur interest in epigenetics (see here and here) and how the area of the epigenome - those chemical marks which have the ability to influence the expression of the genome - has started to yield some potentially important observations. At times I'll admit to being slightly too over-excited at the possibilities of epigenetics. Subsequently brought back down to earth by more sobering accounts (see here**) but not yet ready to poo-poo the whole science just yet***.

I don't want to rehash the whole epigenetic story in this one post, so instead am going to concentrate on the particular area covered by Wong et al and their analysis of the methylome (yes, another -ome for you) which is concerned with the addition of methyl groups to various regions of the genome and how that subsequently alters the expression of genes. DNA methylation has been a sort of peripheral topic in relation to cases of autism for quite some time now; brought to the forefront by all that folic acid (see here) and MTHFR research (see here) and the availability of those lovely methyl groups. Suffice to say that we are still very much at the beginning when it comes to looking at the relationship between all these elements and very complicated conditions like autism.

Anyhow, after that very long introduction (I am only an amateur science blogger after all), a few details from the Wong paper might be in order, bearing in mind it is open-access:

• Part of the UK TEDS initiative and partly sponsored by Autism Speaks who were involved in that recent environmental epigenetics symposium (see here), researchers looked at 50 monozygotic (MZ) twin pairs. It wasn't just a case of 25 twin pairs where one twin was diagnosed with an autism spectrum disorder (ASD) compared with 25 twin pairs where both were concordant for ASD. No, instead the authors looked at a variety of phenotypic combinations based on concordance/discordance for ASD and various core traits based on the Childhood Autism Symptom Test (CAST) schedule. Supplementary table 1 shows the combinations (see here).
• A genome-wide analysis of DNA methylation was undertaken on blood samples provided by participants. I can't pretend to intimately know all the techniques that were employed so won't profess to do so. What I can glean from the paper is that both DNA methylation differences between MZ twin pairs discordant for ASD and analyses between groups scoring high and low on the various core symptom areas were completed with "the aim of identifying real, biologically relevant within-twin and between group DNA methylation differences".
• Results: "ASD is not associated with systemic differences in global DNA methylation". In other words, within the twins, there is quite a high degree of similarity when it comes to patterns of DNA methylation. This might tie in with other work looking at methylation in cases of autism and other family members (see here). 
• When looking at DNA methylation patterns between twin pairs discordant for ASD, specific sites of the genome however seemed to show some variability as a function of ASD diagnosis or not. The authors list the top 50 "differentially methylated CpG sites" (see here) showing gene and position, with a combination of hyper- and hypo-methylated regions identified. Top of the methylation differences pops was the NFYC promoter which was "consistently hypermethylated in affected individuals" (see here for some additional papers on this gene). Hypermethylation by the way, generally means gene silencing as per its function when it comes to transposons (think HERVs). 
• Various other data are presented based on either syndrome or trait specific differences across the twins. I'm not going to go through all of these because, well because that's called plagiarism. I will draw your attention to one particular finding which might be important as they identified "one MZ twin pair, concordant for a very severe autistic phenotype, that appear to represent epigenetic outliers at multiple CpG sites across the genome". I'm immediately drawn back to my autism or autisms post and that all-important phenotypic variability as potentially being relevant here too. That and the tie-up with more structural changes to the genome as per the mention of CNVs and hotspots.

I know I've gone on a bit in the post but this is potentially a very important paper. Not only does it put the epigenome, or at least one part of the epigenome, firmly on the autism research map, but it offers something of a partial explanation for [some of] that 'missing heritability' which was talked about not so many years ago (see here).

The fact also that methylation patterns might be variable both intra-twins and intra-ASD is also important; suggesting that as with more traditional genomic findings in relation to autism, there isn't going to be just one epigenomic factor affecting risk or presentation, but rather a plethora of sites which are hyper- or hypo-methylated, potentially also linked to (affecting?) more structural changes to the genome in cases of ASD.

Don't get me wrong, the same questions remain as are seen in other areas of biological functioning with autism in mind: the heterogeneity, the reliance on diagnosis by observation and note-taking, the (elevated) risk of comorbidities, etc. All of which cloud the waters of association. Distinct however from the question of whether there is anything that can be done when structural changes to the genome are observed (outside of gene therapy for example), when talking about methylation, one speculates that this might be something that lends itself to pharmacotherapy as per the already use of DNA methyltransferase inhibitors for example. Yes, going back to that folic acid-autism link work, the possibility also that environment might help shape gene function and some new light on other external factors (noting that I am not advocating anything at the current time).

So endth the lesson for today.

----------

* Wong CC. et al. Methylomic analysis of monozygotic twins discordant for autism spectrum disorder and related behavioural traits. Molecular Psychiatry. April 2013.

** Ptashne M. Epigenetics: core misconcept. PNAS. April 2013.

*** Fraga MF. et al. Epigenetic differences arise during the lifetime of monozygotic twins. PNAS. 2005; 102: 10604-10609.

----------

Wong, C., Meaburn, E., Ronald, A., Price, T., Jeffries, A., Schalkwyk, L., Plomin, R., & Mill, J. (2013). Methylomic analysis of monozygotic twins discordant for autism spectrum disorder and related behavioural traits Molecular Psychiatry DOI: 10.1038/mp.2013.41

Twins and autism: a response

A very short post this one, I promise.

Following the revelations yesterday that the genetic contribution to autism 'might' have been inflated, there has been quite a lot of scientific and media interest in this whole debate. I say media interest, but here in the UK we have heard diddly squat so far about the revelations that identical twins appear to show a not so high rate of autism than fraternal twins. I don't know why.

Anyway, I would like to draw attention to this opinion piece published in the Archives of General Psychiatry (online) by Dr Peter Szatmari of McMaster University and his take on the implications of the Hallmayer paper. Szatmari is no stranger to autism research having published extensively with papers such as this one and this one.

He is fairly candid in his opinions on the 'game changer' that was the Hallmayer paper and the implications for the early concordance studies such as that from Folstein and Rutter in 1977, asking "where did the heritability go?". Whilst I wouldn't like to give a definitive answer to this complicated question, I would perhaps suggest that we should start by looking at the participant numbers included in the Folstein paper - 21 twin pairs. That's all. Forty-two participants. That is pretty much what the whole 'genetics are king' argument was originally based on. I know that this study has been replicated in other participant groups but be under no disillusion that the 1977 paper cast the die.

Many, many people have been talking about the variable relationship between genes and environment in autism down the years. I'd like to think that once the dust has settled and further corroborative work completed (part of which should be appearing in the journal Pediatrics soon), autism research can finally move out of its 'obsession' with purely genetic models of autism and start moving itself into that brave new world in which genes work in unison with environment. At least so no more parents enter a Doctors office to be told blanket that 'autism is genetic' and nothing else.

Genes overestimated, environment underestimated in autism?

Only a short post this one in response to an article which has just appeared in the Archives of General Psychiatry titled: Genetic heritability and shared environmental factors among twin pairs with autism. A copy of the full-text paper can be found here.

For many years, genes have perhaps received the lion's share of attention when it comes to autism and autism research. Studies conducted in the late to mid 1990's and early noughties estimated a significant genetic loading for autism based on studies of the rates of autism in monozygotic and dizygotic twin pairs; that is twins derived from one egg and those derived from separate eggs. A few choice papers detailing such observations can be found here and here.

Whilst genes are important for autism (and nearly every other condition), there has been a bit of a sea-change in recent years away from notions of an 'autism gene' or genes to something a little more complex with regards to genes and environment. If you don't believe me, have a look at my various posts on CNVs and mutations studied with autism in mind.

This recent paper from Hallmayer and colleagues, corroborates the gene-environment interaction and indeed, perhaps swings the pendulum from genes to more of an environmental effect in terms of risk of developing autism. The collaborative study is a detailed one based in our old autism research favourite place California, where data from a very well-defined autism group, autism and broader autism using ADI-R and ADOS, were examined based on their monozygotic or dizygotic membership in order to ascertain whether autism rates were higher in identical twins than fraternal twins.

The study and findings: of the 202 twin pairs looked at (N=404), 242 twins reached the criteria for autism spectrum disorder (60%), of whom 171 twins met the stricter criteria for autism. Based on these data, 192 twin pairs were submitted for genetic analysis to ascertain zygosity, that is whether they were identical or fraternal twins based on various markers. Fifty-four pairs were found to be monozygotic (28%) and 138 were dizygotic (72%). There are other findings based on the breakdown of specific diagnosis and gender, but I perhaps need a little longer to go into those.

What this means is that if autism was a strongly genetic condition, one would expect that there would be a lot more twins with autism derived from the same egg, the same genetic material, than those derived from separate eggs. The fact that those from separate eggs showed a high-ish rate of autism implies that whilst genes might be important, there is perhaps also a significant contribution from some environmental factor/s in connection to a diagnosis of autism or autism spectrum disorder.

The authors do caution about the findings and issues such as their applicability to all diagnoses of autism across the various geographies and ethnicities. I have posted entries about some of these potential environmental factors previously (here and here for example) but as yet no-one has been able to specifically tie down what they might be and the specific temporal window of exposure. I do wonder also if there might be some influence of time in terms of the results found; is the autism of today, the same as the autism of yester-year?

Replication is the next stage of the process, replication in the same geographical area (or thereabouts) and replication overseas, like here in Northern Europe where many of the original genetic studies on autism were carried out. One would perhaps also like to see a little more information about whether the 'type' of autism might also show any effect and also covering the issue of the broader autism phenotype.

For now though, the sea-change continues as environment steps up to share centre-stage alongside genes in the very complicated world of autism research. The question is: will as much money be directed to the study of environmental factors and autism as has been spent on looking at genes?