Thursday, 5 March 2015

Autism, heritability and 'proof of principle' genomic biomarkers

JAMA Psychiatry published a number of interesting articles recently, some of which have grabbed media headlines. "Autism is largely down to genes, twin study suggests" went the BBC headline covering the paper by Emma Colvert and colleagues [1] who, based on an analysis of twin pairs as part of TEDS (Twins Early Development Study), concluded that: "The liability to ASD [autism spectrum disorder] and a more broadly defined high-level autism trait phenotype in this large population-based twin sample derives primarily from additive genetic and, to a lesser extent, nonshared environmental effects." 

The paper from Tiziano Pramparo and colleagues [2] has so far garnered rather less media attention with their "proof-of-principle study" suggesting that "genomic biomarkers with very good sensitivity and specificity for boys with ASD in general pediatric settings can be identified." Said results were based on examination of "leukocyte RNA expression levels" and found some interesting differences between children diagnosed with autism vs. asymptomatic controls including functions relevant to the immune system and inflammatory processes among other things. Go figure.

The Colvert paper covers one of the more heated debates when it comes to autism: the relative contributions of genetics vs. environment to autism. I've covered this issues quite a bit on this blog, perhaps most recently when discussing the paper from Sven Sandin and colleagues [3] and their results leading to press releases stating that: "Environment as important as genes in autism, study finds." The Colvert results seem to have something slightly different to say, following a well trodden path in autism research of results and counter-results swinging pendulum style (see here).

I don't want to trawl over every detail of the Colvert study but it strikes me that there are a few important things to say about the findings within the context of both genes vs. environment and also the growing move towards the plural 'autisms'. First, is their reliance on looking at twin pairs and in particular, some who were monozygotic (MZ) twins and others who were dizygotic (DZ) combined with an analysis of assessed autistic traits in said twins. In effect, authors were comparing twin pairs - MZ vs. DZ - for how well they matched in autism symptoms terms as a function of their degree of genetic similarity. They didn't actually look at the genes potentially involved in autism in this study, which as we have found out from the paper by Ryan Yuen and colleagues [4], are likely to be pretty complex and containing "substantial genetic heterogeneity" even within sibling pairs (see here for my take on this). And yes, I know 'siblings' are not necessarily the same as 'twins'...

There are also some implicit statements in the study of twins. We assume that they are genetically identical (at least MZ twins). Unfortunately, more and more science is realising that sharing the same genes is not necessarily the same as sharing the same gene functions. One word: epigenetics, and as we've seen even with autism in mind, how issues such as DNA methylation mean twins (identical twins) are not necessarily as identical as you might imagine (see here) and how this might explain at least some of the missing heritability noted in such studies. The value-added bit to the study by Chloe Wong and colleagues [5] looking at the methylome with autism in mind was that they too relied on data from TEDS.

"The novel aspect of this study was the inclusion of twins regardless of whether they had a clinical diagnosis. This enabled us to get a more accurate picture of how influential a child’s environmental experiences and their genetic makeup is on ASD, as well as on subtler expressions of autistic skills and behaviours." A quote from one of the study authors also gives us something to ponder. I assume the 'subtler expressions' means the broader autism phenotype (BAP) and the idea that crossing the diagnostic threshold of autism (or ASD) means crossing a blurred barrier where the traits of autism are also present in milder, less pathological forms not necessarily meeting the diagnostic criteria that we've assigned for the condition. Whilst this is a strength of the Colvert paper over other research in this area, it does rather mean that the spotlight is on the 'trait phenotype' of autism. And 'fractionable' autistic traits have been a focus of other research by some of the authors [6] on the Colvert paper.

I'd also like to think that although the Colvert results are important from the point of view that there are potentially shared genes (or even shared epigenomic issues) at work when it comes to autism / autistic traits, this does not mean that such genes are on their own 'causative' of autism. "Some parents are concerned whether things like high pollution might be causing autism” is another quote from another of the study authors who seems to be downplaying the possibility that such environmental factors might play some hand in some autism. I'm similarly guarded about the idea that something like air pollution might 'correlate' with some autism but as we've discovered over recent years, one doesn't talk about environment without also mentioning the idea of genetic 'fragility' to certain environmental issues (see here) based on the preliminary findings from Heather Volk and colleagues [7] for example. That also genes which might predispose to autism may also predispose to other conditions/states as per the Pramparo paper talking about immune function and inflammatory processes is also worth reiterating.

I guess what I'm trying to say is that of course genes are going to be involved in autism. Even those cases of autism where onset is linked to something like infection (see here and see here) or has a regressive element to it (see here), there has to be some genetic involvement. Genes however, don't typically act in isolation from either maturation or the environment they find themselves in. They're dynamic, switching on and off in various tissues in response to all-manner of different variables. Structural genetics, that is looking for the presence of mutations or different variants, is still important to autism research (as per the BCKDK gene work) of that there is no doubt, although even there the processes dictating mutation are likely to be complex. But within the wide - very wide - spectrum included under the diagnosis of autism, I'd be inclined to suggest that science shouldn't yet be ready to give up the idea that environment (however this is interpreted) might yet hold some important clues about some autism...


[1] Colvert E. et al. Heritability of Autism Spectrum Disorder in a UK Population-Based Twin Sample. JAMA Psychiatry. 2015 Mar 4.

[2] Pramparo T. et al. Prediction of Autism by Translation and Immune/Inflammation Coexpressed Genes in Toddlers From Pediatric Community Practices. JAMA Psychiatry. 2015 Mar 4.

[3] Sandin S. et al. The familial risk of autism. JAMA. 2014 May 7;311(17):1770-7.

[4] Yuen RK. et al. Whole-genome sequencing of quartet families with autism spectrum disorder. Nat Med. 2015 Feb;21(2):185-91.

[5] Wong CC. et al. Methylomic analysis of monozygotic twins discordant for autism spectrum disorder and related behavioural traits. Mol Psychiatry. 2014 Apr;19(4):495-503.

[6] Brunsdon VE. & Happé F. Exploring the 'fractionation' of autism at the cognitive level. Autism. 2014 Jan;18(1):17-30.

[7] Volk HE. et al. Autism spectrum disorder: interaction of air pollution with the MET receptor tyrosine kinase gene. Epidemiology. 2014 Jan;25(1):44-7.

---------- Colvert, E., Tick, B., McEwen, F., Stewart, C., Curran, S., Woodhouse, E., Gillan, N., Hallett, V., Lietz, S., Garnett, T., Ronald, A., Plomin, R., Rijsdijk, F., Happé, F., & Bolton, P. (2015). Heritability of Autism Spectrum Disorder in a UK Population-Based Twin Sample JAMA Psychiatry DOI: 10.1001/jamapsychiatry.2014.3028 Pramparo, T., Pierce, K., Lombardo, M., Carter Barnes, C., Marinero, S., Ahrens-Barbeau, C., Murray, S., Lopez, L., Xu, R., & Courchesne, E. (2015). Prediction of Autism by Translation and Immune/Inflammation Coexpressed Genes in Toddlers From Pediatric Community Practices JAMA Psychiatry DOI: 10.1001/jamapsychiatry.2014.3008