Thursday, 4 April 2013

CNV duplication load, hotspots and autism

Viewers here in the UK might remember the catchphrase of one Michael Barrymore on the show 'Strike It Lucky': "What is a hotspot not? A good spot". It is with a rather different kind of hotspot in mind, that I'm posting about today: genomic hotspots and autism with a specific focus on copy number variants.

An intriguing paper by Santhosh Girirajan and colleagues* (open-access) popped up on my Twitter radar recently discussing copy number variant (CNV) load in relation to autism spectrum disorders. Whilst only being an amateur enthusiast when it comes to all things genes and genomic, I can't offer an expert opinion on what CNVs are, just that fairly similar to single-nucleotide polymorphisms (SNPs), we're talking physical alterations to the genome and in particular gains/losses to segments of DNA (I think!)
A hotspot indeed... @ Wikipedia  

I'll start by saying that this is not the first time that CNVs have cropped up on this blog either with autism in mind (see here) or with other conditions such as ADHD (see here) and intellectual disability (ID) in mind (see here).

Indeed readers who looked at that ID link will see that we are talking about another paper from Girirajan following the previous suggestion that ID might be particularly prone to a high CNV load. Keep that in mind for now.

Thankfully the latest paper has been very nicely covered by a ScienceDaily entry (see here with a sigh of relief) so as to patch over my considerable non-expertise in this area. Without plagiarising the paper or SD entry, the general gist of the work was to initially look at CNV data from over 500 people with autism (n=274) or asymptomatic controls (n=242) derived from the CHARGE initiative (see here), to ascertain exactly what the CNV load was and how it might link into some of the signs and symptoms of autism. There was also a further testing group to confirm "the increased duplication load" based on a further cohort of autism and control cases but I'm not going to bore you with all the details.

If I'm reading this right, there were some interesting findings to take from this study:

  • Children with autism "exhibited a significantly elevated copy number load, represented principally as an increase in duplicated base pairs found in large CNVs". Importantly, this copy number load seemed to include quite a bit of de novoso not passed from parents to offspring.
  • Duplication over deletion seemed to be the important variable for autism cases, which as the authors note "is associated with genomic variants with more modest functional impact". As per that previous CNV work with ID in mind, I think the authors seemed to be suggesting a sort of sliding scale of phenotypes based on CNV profiles: ID at the more severe end of the spectrum (with more deletions also), autism somewhere in the middle and dyslexia at the less severe end of the spectrum in terms of functioning. I could be wrong and I could be over-simplifying the whole thing so accept my apologies if so.
  • Copy number load in autism cases also seemed to show some relationship with certain aspects of behaviour. Significant negative correlations for example, were observed between CNV load and VABS scores in core areas of communication and socialisation. That being said, the correlations were not exactly all that great (p=0.048 and p=0.022 respectively) and should be compared with other gold-standard schedules such as the ADOS that did not turn up anything significant. 
  • The notion of genomic 'hotspots' is also raised as a consequence of the results, suggesting that parts of the autistic genome already under the spotlight might be more susceptible to such CNVs. Those words-of-the-hour DNA methylation get a mention alongside the folate story (see here). Personally and with my non-expertise caveat in full working order, I'm also wondering about those archived portions of viruses called HERVs - human endogenous retroviruses - (see this post and this post) which dot the genome and whether they might be contributory in any way, shape or form to any genomic instability in target areas.

Putting aside the complexity of the genome when it comes to autism and the suggestion that we might want to be rethinking how we view the condition** (I've a post coming up on this paper fairly soon), there are some interesting themes emerging from the Girirajan paper. That for example CNV load and the type of CNV (duplication or deletion) might roughly fit into phenotypic differences between inter-related conditions including autism and ID is definitely something worth pursuing in future work. That also such CNVs might tie into genomic hotspot areas is also an important point particularly when it comes to things like systems biology. Lest we also forget the potential importance of any genomic instability and how, with the DNA methylation point in mind, this *might* be attenuated via changes to the environment (remember SAMe)?

To close Chuck Berry and a song about Johnny B Goode and his guitar-playing skills.


* Girirajan S. et al. Global increases in both common and rare copy number load associated with autism. Hum. Mol. Genet. March 2013.

** Moreno-De-Luca A. et al. Developmental brain dysfunction: revival and expansion of old concepts based on new genetic evidence. The Lancet Neurology. 2013; 12: 406-414.

---------- Girirajan S, Johnson RL, Tassone F, Balciuniene J, Katiyar N, Fox K, Baker C, Srikanth A, Yeoh KH, Khoo SJ, Nauth TB, Hansen R, Ritchie M, Hertz-Picciotto I, Eichler EE, Pessah IN, & Selleck SB (2013). Global increases in both common and rare copy number load associated with autism. Human molecular genetics PMID: 23535821