Thursday, 12 April 2012

A macroepigenetic approach to autism?

Summary and review papers are always excellent fodder for a blog post because of the ideas and hypotheses they can generate based on a synthesis of lots of different research across different areas. Everyone strives to be 'lateral thinkers' during such papers in an attempt to fit together evidence pertinent to their viewpoint. As to how many actually succeed in making that magical connection is a slightly different story.

One recent review paper which caught my eye was this one by Dufault and colleagues* (full-text) which looked at some of the evidence for the role of diet and toxic substances in the environment as being involved in the increasing numbers of cases of autism in the United States (see the 2012 CDC figures here). I tread carefully with this line of thinking bearing in mind the headline which was covered as part of my post on obese mums and risk of autism - "Obesity in pregnancy causing rise in autism cases" - er, not exactly that simple.

I was drawn to the Dufault paper on several levels. Mention of the word 'epigenetics' (I'm sure you've grown tired of me describing this as changes to gene function over changes to the genome) was a pretty big attraction. As a sidelines, epigenetics is something which the ALSPAC are also working up to now some of their cohort has hit the magical age of 21 years old (see here) alongside some interest from people like David Barker, he of the Barker hypothesis.

I digress.

Reference also to the possible effects of diet and environment on cases of autism was also a big pull from the Dufault paper. Richard Deth (pronounced Deeth) on the authorship list was also a draw given some of his past research on methylation and autism (including reference to the D4 dopamine receptor) and the hyperhomocysteinemia findings which have frequented this blog from time to time.

The latest paper is full-text but a few pointers:

  • 'Macroepigenetics' is a term introduced by this paper which refers to "... the process of examining food supplies and their impact on body metabolism and gene function along with what is known about environmental exposures across populations".
  • Mercury features quite heavily in this paper and in particular, the mercury toxicity model, detailing the link between dietary elements, deficiencies in certain essential metals like zinc and environmental exposures to heavy metals. I know there is still some heated debate about things like mercury and whether or not exposures to different types of mercury might be implicated in at least some cases of autism. I'm not going to say anything else about this particular discussion for now.
  • Consumption of high fructose corn syrup (HFCS) also gets a mention in the paper. To my mind, this is one of the first times HFCS has been linked to autism outside of a past paper from some of the same authors** (full-text). The theory goes that among other things, HFCS has the propensity to affect levels of calcium and zinc; low levels of zinc imply that certain metal binding proteins such as metallothionein might not function optimally to remove heavy metals such as mercury. At this point I would link to an interesting post by Dr Emily Deans on fructose malabsorption in relation to depression highlighting how, alongside other issues with carbohydrate metabolism, problems with fructose can potentially link to problems with zinc. I hasten to add that Dr Deans' discussions are strictly limited to the suggestion of a link between fructose malabsorption and depression and nothing else.
  • Paraoxonase-1 (PON-1), everyone's favourite organo-phosphate (OP) pesticide detoxifier, also gets a mention. Accepting that most people are exposed to some degree of OP residues, the authors present data on the amount of residue present in food and how crops such as wheat and corn seem to be top of the pops in terms of residue amount. Corn obviously being a required raw material in the process of making HFCS and indeed mercury has also been found in samples of HFCS. The implication being that fructose and/or mercury can affect the expression of PON1 and this might be dose-dependent, which when coupled to data from this study by D'Amelio and colleagues*** (full-text) on the expression of PON1 in US and Italian cohorts, suggests a reduced ability in the US group to metabolise some OPs.
  • Methionine, homocysteine and various compounds in-between are also discussed in the paper with particular reference to oxidative stress and DNA / histone methylation. Some interesting facts: PON1 turns out to be pretty essential for reducing homocysteine levels and high homocysteine levels are associated with the under-methylation - hypomethylation - of DNA. This might, just might, tie into these findings suggestive of an inverse relationship between homocysteine levels and PON1 arylesterase activity in a group of children with autism.

I hope you are still here after reading all that. It's taken me a few reads of the paper to start to come to grips with all the various strands of evidence incorporated and the overview presented here is not intended to be a definitive one. I suppose the main conclusion is that various factors might be implicated in producing an epigenetic response in autism including pesticide exposures, heavy metals and dietary components like HFCS. The sum of these factors being greater than the individual factors acting alone. 

I admit that I'm not totally convinced by all the issues discussed in this paper as being universally applicable to all cases of autism. Here in the UK, as in Italy and other parts of Europe, HFCS is still only tentatively findings its way into the shopping baskets of this nation of shopkeepers. Having said that, I have seen nothing that suggests autism has the same aetiological roots across all nations as indeed this paper asserts differences. Added to the fact there is some pretty convincing evidence that for example, zinc deficiency can be comorbid to cases of autism as per a post from a few months back, I have an open mind to the concepts detailed in the paper.

The authors suggest that their review of gene-environment interactions with a particular slant on autism in the United States might offer a kind of template for other individuals or groups to conduct and report similar papers for other potential factors and in other parts of the world. I would champion this notion that we do need to start looking at more complicated patterns of potential risk factors for autism, whether based on endophenotypes, geography or other factors, with a focus on how environment and genes might be working variably, but synergistically.

To finish, I have been getting a little too carried away with the Beautiful South lately and so to conclude, a Song for Whoever (the video is not the real song video but the sound was the best quality).

* Dufault R. et al. A macroepigenetic approach to identify factors responsible for the autism epidemic in the United States. Clinical Epigenetics. April 2012.
DOI: 10.1186/1868-7083-4-6

** Dufault R. et al. Mercury exposure, nutritional deficiencies and metabolic disruptions may affect learning in children. Behavioral & Brain Functions. 2009; 5: 44.
DOI: 10.1186/1744-9081-5-44

** D'Amelio M. et al. Paraoxonase gene variants are associated with autism in North America, but not in Italy: possible regional specificity in gene-environment interactions. Molecular Psychiatry. 2005; 10: 1006-1016
DOI: 10.1038/


  1. There is one more possibility that has not surfaced in the debate yet and I wonder why. Is it possible that GMO modified proteins create unknown peptides triggering both inflammation and possible epigenetic conditioning. I know it is pure speculation, but the coincidence with the sudden growth in the US that is a major consumer of modified food looks suspicious. After all if mankind has struggled to adapt to seeds, we are in a sense rewinding the process.

  2. Thanks for the comment. At this point in time I don't think anyone can discount this as a factor.