More epigenetics, EN-2 and autism... the plot thickens

I don't mind admitting that I was to some extent 'winging it' with my previous post on epigenetics and Engrailed-2 (EN-2) as a consequence of the findings reported by Jill James and colleagues [1] with autism in mind. Although an avid follower of the science of epigenetics when (cautiously) applied to autism, I am by no means any authority on the subject matter particularly when it comes to the nitty-gritty details. You can probably therefore expect similar things in my latest discussions on yet more work from this research group which appeared recently [2].

I have only one rule. Everybody fights, no one quits.

And so, with that pinch of salt at the ready...

The final conclusion made in the most recent James article boils down to the suggestion that "persistent postnatal overexpression of EN-2 suggests that the closing of this programed developmental window may have been missed in some individuals with autism because of epigenetic abnormalities". That being said I think we have quite a way to come before we can substantiate this finding particularly when the main protagonist in the latest article is something called 5-hydroxymethylcytosine (5-hmC) and results which show "that elevated 5-hmC in the EN-2 promoter is associated with a significant decrease in repressive MeCP2 and histone H3K27me3 that appear to over-ride 5-mC hypermethylation". The H3K27me3 bit comes from their previous findings by the way.

To most readers that probably sounds as complicated as it first did to me so I will try and explain more.

EN-2 as I've talked about in that post on the previous James work, has been linked to cases of autism as per the example of the study by Wang and colleagues [3] linking mutations in this gene to cases of autism. The idea being that mice bred without the gene (the homeobox domain of EN2) show some of the [mouse] signs and symptoms of autism alongside issues with the cerebellum and a reduction in the number of Purkinje cells which have been previously noted in cases of autism [4]. The previous James results in this area reported on hypermethylation of the EN-2 promoter region which would normally equate as gene silencing in epigenetic terms, in line with the more structural genomic issues seen in autism that I've just talked about. But, and it is an important point, when they looked at EN-2 expression and protein levels - function and products of the gene - they actually found that levels were increased in their autism samples despite the methylation mark and its 'stop talking' properties. They noted on that occasion that "transcriptional upregulation by other epigenetic mechanisms predominated over the repressive tendencies of DNA cytosine methylation".

Their latest foray into this area sought to further clarify just what might be going on specifically with EN-2 gene-specific DNA hypermethylation previously reported. To do this they focused on both measuring 5-hmC and also 5-methylcytosine (5-mC) among other things based on the same tissue samples (post-mortem cerebellum samples) detailed in their previous study. 5-hmC is apparently an oxidation product of 5-mC mediated via something called TETs.

What they found, far from answering the question of a discrepancy between epigenetic gene silencing of EN-2 but increased gene function and products, actually makes the whole thing a lot more complicated. So they observed "a significant increase in both 5-mC and 5-hmC in the autism cerebellum relative to the control samples". Further that there was "a significant increase in 5-hmC content within the upstream EN-2 promoter region" and "a highly significant positive correlation... was found between 5-hmC content and EN-2 gene expression in the 5’ promoter CpG island in autism but not in control samples". They note that: "that 5-hmC accumulation is mechanistically related to gene upregulation" something which I think ties into other work hinting at the demethylating role for 5-hmC [5].

Insofar as my mention of MeCP2 and histone H3K27me3 from the latest and previous James reports, I can't really say too much more aside from noting again: "reduced MeCP2-mediated gene repression may have contributed to persistent EN-2 gene overexpression in the autism samples". Actually the authors speculate that MeCP2 binding and histone H3K27 trimethylation might work together in a "repressive" manner but when reduced as they were "may contribute to aberrant overexpression of EN-2 in the autism cerebellum" as per their findings.

I have to say that I struggled with getting my head around these findings and I'd quite understand if readers also struggled with my interpretation of them ("If you can't explain something to a six-year-old/granny, you really don't understand it yourself"). I understand that we don't all walk around with our genes stuck in the 'on' or 'off' position and that particularly during foetal and the early post-natal periods, genes are being switched on and off at a surprising rate for many, many different important reasons. I also understand that DNA methylation is an important part of the whole genes switched on or off thing but not the only way that this process can happen as per the authors mention of chromatin and some previous text in this area [6]. With my very limited knowledge of this area, I am however not yet convinced that we have the full story here; specifically in terms of why the original finding of hypermethylation of the EN-2 promotor region (gene silencing) yet increased expression and protein levels were reported. I wonder if indeed we might be able to learn more from a two-hit approach whereby hypermethylation of only one gene allele leaves the other still working?

Just before I finish I'd like to also draw your attention to another paper which has started to ask similar questions about 5-hmC and might be contrasted with the recent James paper. Zhubi and colleagues [7] (open-access here) looked at 5-hmC with a couple of other potentially important genes linked to cases of autism (RELN and GAD1) in mind. They reported: "a significant increase in TET1 expression and an enrichment in the level of 5-hmC... at the promoters of GAD1 and RELN in ASD when compared with CON [controls]". Further that their data are: "consistent with the hypothesis that an increase of 5-hmC (relative to 5-mC) at specific gene domains enhances the binding of MeCP2 to 5-hmC and reduces expression of the corresponding target genes in ASD [autism spectrum disorder] cerebella".

The plot thickens...

So then to some music... Hey Jude.

----------

[1] James SJ. et al. Complex epigenetic regulation of Engrailed-2 (EN-2) homeobox gene in the autism cerebellum. Translational Psychiatry. 2013; 3: e232.

[2] James SJ. et al. Elevated 5-hydroxymethylcytosine in the Engrailed-2 (EN-2) promoter is associated with increased gene expression and decreased MeCP2 binding in autism cerebellum. Translational Psychiatry. 2014. 7 October.

[3] Wang L. et al. Association of the ENGRAILED 2 (EN2) gene with autism in Chinese Han population. Am J Med Genet B Neuropsychiatr Genet. 2008 Jun 5;147B(4):434-8.

[4] Fatemi SH. et al. Purkinje cell size is reduced in cerebellum of patients with autism. Cell Mol Neurobiol. 2002 Apr;22(2):171-5.

[5] Dahl C. et al. Advances in DNA methylation: 5-hydroxymethylcytosine revisited. Clin Chim Acta. 2011 May 12;412(11-12):831-6.

[6] Lasalle JM. et al. Autism genes keep turning up chromatin. OA Autism. 2013 Jun 19;1(2):14.

[7] Zhubi A. et al. Increased binding of MeCP2 to the GAD1 and RELN promoters may be mediated by an enrichment of 5-hmC in autism spectrum disorder (ASD) cerebellum. Transl Psychiatry. 2014 Jan 21;4:e349.

----------

James SJ, Shpyleva S, Melnyk S, Pavliv O, & Pogribny IP (2014). Elevated 5-hydroxymethylcytosine in the Engrailed-2 (EN-2) promoter is associated with increased gene expression and decreased MeCP2 binding in autism cerebellum. Translational psychiatry, 4 PMID: 25290267