Viewpoint Expert opinions on trends and controversies in autism research.
Opinion / Viewpoint

Genes and environment are two-way street in autism risk

by  /  21 August 2012

Close contact: Exposure to certain environmental agents, such as chemicals found in plastics, may alter the expression of autism-related genes.

Understanding and treating autism would certainly be much simpler if there were only a single or even a handful of underlying culprits. Instead, autism is clearly falling into the category of complex human genetic disorders that include cancer, cardiovascular disease and autoimmunity. Genes and environmental factors are both involved, but using a genetic or other molecular test to predict an individual’s risk of autism remains frustratingly out of reach.

This is in large part because gene-environment calculations involve more than simple math. It may be tempting to consider genes and the environment as separate paths, but I’d like to point out that the important interactions occur on what is more of a two-way street.

In one direction, genetics can influence environmental susceptibility. In the opposite direction, environmental factors can influence the expression of genes via epigenetic mechanisms, which leave unaltered the underlying DNA sequence. The environment can also damage our DNA through mutations or by triggering genomic instability.

The good news is that much has been learned from studies of genetics and gene expression about how neurons and synapses, which transmit signals between neurons, mature in autism5. But despite these successes, there is still not a single genetic test that can predict more than one to two percent of autism cases. Determining whether a given rare variant causes autism is also often difficult to prove.

Personal risk:

The reality is that everyone enters the world with his or her own personal and complex mixture of genes and environmental influences, so finding a single smoking gun in the environment is unlikely.

An interesting example of the complex gene-environment interplay is the influence of both nutritional factors and genes on the one-carbon metabolism cycle, a pathway that includes the vitamin folate. This cycle generates important methyl groups — molecules that can epigenetically modify both DNA and histones, the proteins that serve as organizing spools for DNA.

Studies also show that environmental chemicals can adversely affect this pathway. Exposing mice during pregnancy to common pollutants such as the plastic component bisphenol A (BPA) or the flame retardant polybrominated diphenyl ether (PBDE) lowers DNA methylation levels17,18. Giving the mice folic acid counteracts the effects of BPA, probably by providing additional methyl donors17

PBDE’s effect on methylation in the brain is also linked to social deficits in mice that model Rett syndrome18. So genes and the environment can combine to modify DNA methylation, resulting in changes to gene expression and altered behavior.

Two-way street:

My team led a study, set to be published later this month in Environmental and Molecular Mutagenesis, that examines the effect of long-term exposure to organic pollutants19.

We looked at the effect of PBDEs and polychlorinated biphenyls, or PCBs, on 107 postmortem brains from individuals who had autism of unknown causes, who had autism with a known genetic causes or who showed typical development. We naively expected to find higher exposure levels in the brains of those who had autism from an unknown genetic cause than in those who had autism from an underlying mutation.

Instead, we found that levels of the historic pollutant PCB-95 are strongly associated with duplication of the 15q11-13 genomic region (dup15q), a common CNV among people with autism. Interestingly, all of the dup15q brain samples are from individuals born after 1980, although the U.S. banned the use of PCBs in 1976. 

Among the samples with PCB-95 exposure, all five cases of 15q duplications are inherited from the mother, raising the possibility that maternal PCB-95 exposure may have predisposed the 15q11-13 region to rearrangement. PCBs are not known mutagens, but there is less DNA methylation in the dup15q brain samples than in controls, suggesting that hypomethylation could lead to instability of this region.

This study certainly raises more questions than it answers about the two-way interaction of genes and environment in autism risk. How do the rare new CNVs and mutations observed in autism arise? Could some of these genetic changes be a result of exposure to environmental toxins? How important are DNA methylation levels in genetic stability

We may need to look back as far as a generation or two in epidemiology studies of gene-environmental interaction to get concrete answers to these questions. Animal and cell culture models of these interactions may help reveal the molecular and genomic mechanisms involved.

Janine LaSalle is professor of medical microbiology and immunology at the University of California, Davis.


1: Malhotra D. and J. Sebat Cell 148, 1223-1241 (2012) PubMed

2: Neale B.M. et al. Nature 485, 242-245 (2012) PubMed

3: O’Roak B.J. et al. Nature 485, 246-250 (2012) PubMed

4: Sanders S.J. et al. Nature 485, 237-241 (2012) PubMed

5: Luo R. et al. Am. J. Hum. Genet. 91, 38-55 (2012) PubMed

6: Landrigan P.J. et al. Environ. Health Perspect. 120, a258-260 (2012) PubMed

7: Hultman C.M. et al. Mol. Psychiatry 16, 1203-1212 (2011) PubMed

8: Shelton J.F. et al. Autism Res. 3, 30-39 (2010) PubMed

9: Turner T. et al. PLoS One 6, e26418 (2011) PubMed

10: Tanne J.H. Brit. Med. J. 344, e2768 (2012) PubMed

11: Krakowiak P. et al. Pediatrics 129, e1121-e1128 (2012) PubMed

12: Brown A.S. Dev. Neurobiol. Epub ahead of print (2012) PubMed

13: Kalkbrenner A.E. et al. Epidemiology 21, 631-641 (2010) PubMed

14: Volk H.E. et al. Environ. Health Perspect. 119, 873-877 (2011) PubMed

15: Schmidt R.J. et al. Am. J. Clin. Nutr. 96, 80-89 (2012) PubMed

16: Schmidt R.J. et al. Epidemiology 22, 476-485 (2011) PubMed

17: Dolinoy D.C. et al. Proc. Natl. Acad. Sci. USA 104, 13056-13061 (2007) PubMed

18: Woods R. et al. Hum. Mol. Genet. 21, 2399-2411 (2012) PubMed

19: Mitchell M.M. et al. Environ. Mol. Mutagen. Epub ahead of print (2012) PubMed

9 responses to “Genes and environment are two-way street in autism risk”

  1. Dorothy Bishop says:

    Thanks for this piece, which puts the complexity of autism etiology into perspective. Overall, you do a great job of explaining how environmental and genetic factors can interact, and your own work on this is suggesting important new ways forward.
    I do have a quibble, though. At times you use words such as ‘frequent’ or ‘common’ without specifying just what this means. The naïve reader could, I think, interpret this to mean that the majority of people with autism are affected.
    So for instance, you say:
    “Copy number variations (CNVs), which are large deletions or duplications of DNA, are frequently found in people with autism or other neuropsychiatric disorders”
    Well, yes, but, as you note elsewhere, their causal significance is not established. The link is to a piece that does explain the complexity of the CNV link, but ultimately notes that a recent study found 277 CNVs in 427 unrelated individuals with autism, but only 27 of these were de novo: the rest also were found in non-autistic parents.
    You then go on to talk of duplication of the 15q11-13 genomic region (dup15q) as “a common CNV among people with autism”. I’d like to have seen some specification of the percentage of people with autism who have this duplication. I’m not an expert in this area, but my impression is that this is ‘common’ in the sense that more than one study has found individuals with a CNV affecting this locus, but not in the sense that if you took a large group of people with autism, you’d find lots of instances.
    I do appreciate that, elsewhere in your article you emphasised the heterogeneity and complexity of genetic findings in this field. I do just have a bit of a bee in my bonnet about the way genetics finidngs on neurodevelopmental disorders are reported, which can, I think, be misleading . Given the the dearth of significant findings from genome-wide association studies, researchers get excited by any replicable assocation, but they often fail to report the effect size. This is really just a plea to be more explicit about frequencies when using words such as ‘common’, see:
    And finally! It’s great that you have provided references, but could you provide titles of the articles as well? I realise they are only one click away, but it would be that bit easier to home in on the relevant sources if titles were provided on the SFARI site.

    • emilysinger says:

      Thanks for your comment. I’m working on a story now on the frequency of 15q duplications in autism – it’s the second most common autism linked copy number variation – and will have some specific numbers in the piece. -Emily Singer, News Editor,

    • Janine LaSalle says:

      Thanks for your post. I agree that the comment about the frequency of CNVs in autism was oversimplified in the interest of keeping the article short and accessible. The use of “frequently found” here refers more to the ability to detect CNVs in autism using technologies that didn’t used to exist as opposed to any single CNV being frequent in a population of individuals with autism. I’m glad to hear that Emily will be doing an article on the frequency of 15q duplications in autism.

      A recent paper that I really like on the topic of CNV frequency in autism is listed below because it didn’t make my previous reference list. This was the first study to look at overall relative burden of large CNVs in autism compared to controls, dyslexia, and intellectual disability. They found a striking difference in the frequency of rare CNVs in autism and ID compared to dyslexia and controls, but ID showed a higher frequency of de novo CNVs compared to autism.

      PLoS Genet. 2011 Nov;7(11):e1002334. Epub 2011 Nov 10.

      Girirajan S, Brkanac Z, Coe BP, Baker C, Vives L, Vu TH, Shafer N, Bernier R, Ferrero GB, Silengo M, Warren ST, Moreno CS, Fichera M, Romano C, Raskind WH, Eichler EE. Relative burden of large CNVs on a range of neurodevelopmental phenotypes. PLoS Genet. 2011 Nov;7(11):e1002334. Epub 2011 Nov 10.

  2. RAJensen says:

    The role of CNV’s, inherited or de novo, in human disease has been found to confer both risk and protective effects. A lesson can be learned from AIDS researchers who have identified a common CNV widely distributed throughout the general population. CNV’s in the gene CCL3L1 which maps to chromosome 17q has been identified in a sub-group of people with and without HIV-1 infection after exposure to HIV-1. Lower CNV’s are associated with enhanced risk for acquiring HIV-1 infection while higher CNV’s are associated with reduced risk for acquiring HIV-1 infection. CCL3L1 CNV’s appear to be benign and HIV-1 risk is the consequence of a GxE interaction.

    To be continued due to word count restrictions


    Liu S, Yao L and Zhu H. CCL3L1 copy number variation and susceptibility to HIV-1 infection: a meta-analysis. Plos One. 2010 Dec 30;5(12):e15778.

  3. RAJensen says:

    Klinefelter Syndrome (KS) is associated with autism risk (Bishop et 2011). KS is one of the most common genetic syndromes with a population prevalence of 1 in 500 – 1,000. KS is not inherited, the origins of KS an extra X chromosome. About half the cases are caused by an XY sperm mutation with the other half caused by an XX egg mutation and the KS genotype is XXY in contrast to the normal sex chromosome genotypes, XY (males) and XX (females).

    McAuliffe’s group recently discovered that increasing levels of exposure to PCB’s, as measured in blood, produced an increased frequency of XY sperm in sub fertile men.

    The role of environmental pathogens in causing reproductive errors (sperm or egg) has been neglected by autism researchers who have focused exclusively on the role of environmental exposures as it affects fetal development in the womb and in early postnatal existence.


    Bishop et al 2011. Autism, language and communication in children with sex chromosome trisomies. Arch Dis Child2011;96:954-959 doi:10.1136/adc.2009.179747.

    McAuliffe ME, Williams PL, Korrick SA, Altshul LM, Perry MJ 2012. Environmental Exposure to Polychlorinated Biphenyls and p,p´-DDE and Sperm Sex-Chromosome Disomy. Environ Health Perspect 120:535-540.

    • Janine LaSalle says:

      Thanks for your post and for raising some important points that I think are important for autism researchers to consider.
      1. Your point about CNVs being protective as well as risk factors is an important one to consider. I have a quadrant diagram that I use in presentations where I try to list the know genetic and environmental risk and protective factors in autism. The “risk” side for both genetic and environmental factors is much longer and I tend to think this is simply because researchers don’t look at protective factors. The only environmental protective factor that I’m aware of is use of prenatal vitamins preconception (Schmidt et al study cited above), and the only genetic protective factor I could come up with was being XX versus XY. If there are others that I am unaware of, please let me know!
      2. Your point about considering reproduction and early post-fertilization events is also important for both genetic and epigenetic reasons. In human, most fertilized eggs do not get implanted and the implantation process definitely involves a survival-of-the-fittest selection process where most chromosomal aneuploidies and other common genetic mistakes are not successfully implanted. I think there is good evidence (like the study you cite) that environmental factors can influence the chromosome content of embryos that are successfully implanted. Also, as an epigeneticist, I am struck by recent data on DNA methylomes showing that embryonic stem cells have the highest level of DNA methylation, and these patterns change dynamically during tissue differentiation. Perhaps this is why folate from prenatal vitamins is protective primarily in preconception and the first month of pregnancy.

      More on these type of considerations are in my recent Point-of-View article:
      Epigenetics. 2011 Jul;6(7):862-9. Epub 2011 Jul 1.
      A genomic point-of-view on environmental factors influencing the human brain methylome.
      LaSalle JM.

  4. Human Genetics Journal says:

    I had gone through your blog.Thanks for this piece, which puts the complexity of autism etiology into perspective. But the face that genes plays an important role than environment.Overall, you do a great job of explaining how environmental and genetic factors can interact, and your own work on this is suggesting important new ways forward.

  5. RAJensen says:

    A final thought on your most informative article on DNA methylation. In Klinefelter Syndrome (KS) the XXY genotype is present in all the boys’ cells caused by an XY sperm mutation or an XX egg mutation. A small percentage of boys with Klinefelter syndrome have the extra X chromosome in only some of the cells in their bodies; in these individuals, the condition is described as mosaic Klinefelter syndrome. Individuals with mosaic KS may have milder signs and symptoms depending on how many cells have an additional X chromosome. It occurs as a random event during cell division early in fetal development. As a result, some of the body’s cells have one X chromosome and one Y chromosome (XY), and other cells have an extra copy of the X chromosome (XXY). The role of environmental risk factors (PCB) that contribute to DNA methylation in mosaic KS and in other conditions such as mosaic Down Syndrome is an important area of investigation and I look forward to your future papers.

  6. RAJensen says:

    PS. The most recent SFARI post on the paternal age effect in autism by Virginia Hughes ‘Father’s age dictates rates of new mutations’ can be seen in Klinefelter Syndrome (KS). Lowe’s group examined the sperm in 38 karyotype normal fathers of KS boys. The frequency of XY sperm significantly increased with advancing paternal age.


    Lowe X, Eskanazi B, Nelson D, Kidd S, Alme A. (2001). Frequency of XY Sperm Increases with Age in Fathers of Boys with Klinefelter Syndrome. Am J Hum Genet. 2001 November; 69(5): 1046–1054.

Leave a Reply

Your email address will not be published. Required fields are marked *