16 April 2012. Spontaneously occurring, protein-altering mutations likely contribute to some cases of autism, according to a trio of the largest-to-date exome sequencing studies of the disorder, published April 4 in Nature. The three independent studies led by Matthew State of Yale University in New Haven, Connecticut; Evan Eichler of the University of Washington in Seattle; and Mark Daly of The Broad Institute of Harvard and MIT in Cambridge, Massachusetts, scoured the protein-coding part of the genome, called the exome, to find non-inherited “de-novo” single base pair mutations in individuals with autism. This fingered some new genes (CHD8, KATNAL2, and SCN2A), but also has cast autism as a heterogeneous disorder caused by glitches in at least many hundreds of different genes.
Though schizophrenia and autism may share some genetic liability (see SRF related news story), the studies hold lessons for schizophrenia, not so much for the particular genes they turn up, but for how to make sense of the contributions of rare variants to disease. Just a few years ago, rare genetic abnormalities like copy number variants evoked a smoking gun for some (see SRF related news story). More recently, sequencing has turned up rare events that seem particularly nasty: protein-altering variants in the exome that were also “de novo”—that is, arising spontaneously and not inherited from parents, and so their phenotypic consequences have not yet been tempered by natural selection. Eichler’s group recently used this exome-de novo approach to identify some potentially causal variants in a small autism sample (O’Roak et al., 2012), and two other studies last year did the same for schizophrenia (see SRF related news story and SRF news story).
But the interpretive obstacle here is that rare events are common—that is, everyone carries some kind of protein-coding variant (Pelak et al., 2010). As the sequencers ratchet up their output, researchers are getting a better estimate of the frequency of these sorts of variants in controls, and the three new studies use these estimates, and other approaches, to gauge a particular variant’s role in autism.
And in schizophrenia, even targeted resequencing of those genes with a long history of association with the disorder does not necessarily produce the goods. A fourth study published online April 3 in Molecular Psychiatry from Patrick Sullivan of the University of North Carolina in Chapel Hill and Richard Gibbs of Baylor University in Houston, Texas, re-sequenced portions of the genes considered to be classic schizophrenia players, and found no variants associated with the disorder.
Trios and quartets
In State’s study, first author Stephan Sanders and colleagues sequenced the exomes of 928 individuals from 238 families in the Simons Simplex Collection (SSC), which consists of “sporadic” cases of autism without a family history of the disorder. The families consisted of an individual with autism, two unaffected parents, and in 200 families, an unaffected sibling—making for an extensive quartet dataset. The researchers identified 125 de-novo, disruptive single nucleotide variants that either introduced a stop codon or altered a splice site; this number was significantly greater than the 87 of such variants found in the unaffected siblings. Restricting this count to only those variants landing in genes expressed in the brain enhanced this difference, with 13 in cases and three in unaffected siblings, suggesting elevated rates of de-novo mutations in autism.
To identify genes likely to confer risk, the researchers looked for variants from unrelated individuals with autism converging on the same gene. Two individuals each carried a stop codon mutations in SCN2A (sodium channel, voltage-gated, type II, alpha subunit), a gene implicated in epilepsy, and by Sanders and colleagues' calculations, this convergence was unlikely to occur by chance. When the researchers combined their data with that from Eichler’s study, they found two more genes that were each hit by two disruptive mutations in two unrelated individuals with autism: KATNAL2 (katanin p60 subunit A-like 2) and CHD8 (chromodomain helicase DNA binding protein 8). Other ways of establishing an association of these events with autism did not pan out for State’s dataset: there was no difference in the number of de-novo mutations per person between cases and controls; mutations in cases were not necessarily more disruptive than those in controls; and the genes hit in autism cases were not particularly enriched for gene sets related to autism or other disorders, nor did they belong to functionally related sets of genes, according to pathway analysis. Despite the convergences onto three genes, the researchers’ models predicted about 1,000 risk-promoting ASD genes.
Eichler’s study surveyed 677 individual exome sequences from a different group of 209 families from the SSC that mostly consisted of parent-child trios, but also included 50 unaffected siblings. First author Brian O’Roak and colleagues found 126 de-novo point mutations that were either stop codons or judged to be “severely” protein altering. Pathway analysis indicated that a substantial 39 percent of the mutations landed in genes belonging to a protein network enriched for autism candidates, and important for β-catenin and chromatin remodeling—key processes in brain development. Four times as many mutations occurred in DNA from the father as from the mother, and the number of mutations correlated with paternal age, consistent with the increased risk of autism in children with older fathers.
The study led by Mark Daly sequenced the exomes of 175 autism trios from five different centers. First author Benjamin Neale and colleagues found 161 point mutations in all, including 101 amino acid changes and 10 premature stop codons. Only 46.3 percent of cases carried such variants, however, suggesting that these kinds of variants couldn’t account for a majority of autism. The researchers calculated a mutation rate of 0.92 de-novo events per person, per exome in their trios, which was slightly (but not significantly) elevated over the rate expected to occur by chance. These findings tempered the researchers’ views on the role of de-novo coding mutations in autism.
Still, the protein-protein interaction and pathway analyses conducted by Neale and colleagues suggested that at least some of these point mutations contribute to autism risk: the genes hit by mutation seemed biologically related, and the gene networks to which they belonged were enriched for autism candidates. When Daly’s group looked specifically at the loss-of-function mutations (nonsense, splice, or frameshift) uncovered by all three studies, they found that multiple mutations of this type at SCN2A, KATNAL2, and CHD8 were unlikely to have occurred by chance. These three genes were further evaluated in exome sequencing of 935 autism cases and 870 controls, which produced three more loss-of-function variants in CHD8 and KATNAL2 in cases, and none in controls. This strongly implicates these two genes as genuine risk factors, but they are two of many as-yet unresolved genes, which the Daly’s group estimates to be in the hundreds, each increasing risk 10-20 times.
The schizophrenia study focused on 10 classic candidate genes: COMT, DAOA, DISC1, DRD2, DRD3, DTNBP1, HTR2A, NRG1, SLC6A3, and SLC6A4. Genomewide association studies do not find evidence for common variants in these genes in schizophrenia, but the researchers hypothesized that these genes may still harbor rare variants. First author JJ Crowley and colleagues re-sequenced selected regions of these genes (amounting to 16.8 kb total) in 727 cases and 733 controls, and identified 782 single nucleotide variants, 587 of which hadn’t been seen before. Of these, 92 were novel protein-altering variants (nonsense, missense, or splice site) or variants observed in more than one case. When the researchers genotyped a second sample of 2,191 cases and 2,659 controls for these variants, none were found to be overrepresented in cases compared to controls with genomewide significance. This suggests to the authors that rare coding variants in these well-known genes are not at work in schizophrenia.
The findings beg further scrutiny of the regions harboring the signals that implicated these genes in the first place. This kind of approach can lead to the vast, uncharted spaces between genes, as reported in a study published April 4 in Science Translational Medicine. The study investigates a gene-poor region of chromosome 5 that harbors a genomewide significant signal in autism GWAS, and finds evidence of involvement of a noncoding RNA (Kerin et al., 2012). The molecule is antisense to the transcripts of the gene encoding moesin (MSN), which encodes a protein regulator of neuronal shape, and it was elevated 12-fold in postmortem brain samples from individuals with autism.
As the pace of sequencing findings increases, this sort of work marks just the beginning. Larger sample sizes will need to confirm these first findings, which need to be interpreted with a better understanding of the full extent of human genetic variation. Together, the studies lurch toward ways of considering the different kinds of evidence for ruling in or ruling out a variant in disease, a process that will be informative for future genetic studies of schizophrenia and other mental disorders.—Michele Solis.
Sanders SJ, Murtha MT, Gupta AR, Murdoch JD, Raubeson MJ, Willsey AJ,
Ercan-Sencicek AG, DiLullo NM, Parikshak NN, Stein JL, Walker MF, Ober GT,
Teran NA, Song Y, El-Fishawy P, Murtha RC, Choi M, Overton JD, Bjornson RD,
Carriero NJ, Meyer KA, Bilguvar K, Mane SM, Sĕstan N, Lifton RP, Günel M,
Roeder K, Geschwind DH, Devlin B, State MW. De novo mutations revealed by whole-exome
sequencing are strongly associated with autism. Nature 2012 April 5. Abstract
O’Roak BJ, Vives L, Girirajan S, Karakoc E, Krumm N, Coe BP, Levy R, Ko A, Lee C,
Smith JD, Turner EH, Stanaway IB, Vernot B, Malig M, Baker C, Reilly B, Akey JM,
Borenstein E, Rieder MJ, Nickerson DA, Bernier R, Shendure J, Eichler EE. Sporadic autism exomes reveal a highly interconnected protein network of de novo mutations. Nature 2012 April 5. Abstract
Neale BM, Kou Y, Liu L, Ma’ayan A, Samocha KE, Sabo A, Lin CF, Stevens C,
Wang LS, Makarov V, Polak P, Yoon S, Maguire J, Crawford EL, Campbell NG,
Geller ET, Valladares O, Schafer C, Liu H, Zhao T, Cai G, Lihm J, Dannenfelser R,
Jabado O, Peralta Z, Nagaswamy U, Muzny D, Reid JG, Newsham I, Wu Y,
Lewis L, Han Y, Voight BF, Lim E, Rossin E, Kirby A, Flannick J,
Fromer M, Shakir K, Fennell T, Garimella K, Banks E, Poplin R, Gabriel S, DePristo M,
Wimbish JR, Boone BE, Levy SE, Betancur C, Sunyaev S, Boerwinkle E,
Buxbaum JD, Cook Jr EH, Devlin B, Gibbs RA, Roeder K, Schellenberg GD,
Sutcliffe JS, Daly MJ. Patterns and rates of exonic de novo mutations in autism spectrum disorders. Nature 2012 April 5. Abstract
Crowley JJ, Hilliard CE, Kim Y, Morgan MB, Lewis LR, Muzny DM, Hawes AC, Sabo A, Wheeler DA, Lieberman JA, Sullivan PF, Gibbs RA. Deep resequencing and association analysis of schizophrenia candidate genes. Mol Psychiatry. 2012 Apr 3. Abstract