Whole-exome sequencing has helped two teams track down rare recessive mutations in autism spectrum disorder that may ultimately help in not only diagnosing the condition, but also understanding its diverse biology.
Researchers from Boston Children's Hospital, Harvard Medical School, the Broad Institute and elsewhere started searching for rare, highly penetrant recessive mutations in ASD-affected consanguineous families.
Their findings, which appeared online today in Neuron, suggest that recessive mutations can contribute to ASD cases with unexpectedly simple genetics, involving alterations to individual genes rather than hundreds, as is sometimes the case.
"It does suggest that some fraction of cases are going to be diagnosable in a relatively more straightforward way than we might have thought," Boston Children's Hospital's Timothy Yu, co-first author on the study, told Clinical Sequencing News.
In particular, Yu and his colleagues found examples of causal mutations for ASD in genes already implicated in several Mendelian diseases, including some neurometabolic conditions. At least some of these newly identified mutations appeared to be milder than those described in the same genes previously, suggesting that some ASD cases stem from alternative presentations of conditions that normally affect multiple organ systems.
On the other hand, an independent Boston-led team also reporting in Neuron detected more severe recessive mutations — homozygous and heterozygous compound loss-of-function mutations that appear to knock out the affected gene — by sequencing coding regions of the genome in nearly 1,000 ASD cases and 1,000 controls. These LoF knockouts were roughly twice as common in the cases when researchers considered genes that typically have low rates of LoF variation.
"What this gives us is a genetic mechanism," Mark Daly, senior author on the second paper and a genetics researcher affiliated with the Massachusetts General Hospital, Broad Institute, and Harvard Medical School, told CSN.
"As much more autism sequencing takes place this year, we now know how to approach this analytically," he added, "and with larger samples we'll certainly start pointing to specific genes and pathways that are going to be insightful."
The mutations identified in each of the new studies varied in their severity and penetrance. For instance, Yu and company saw a range of missense, frameshift, or truncating mutations that appeared to be causal in ASD.
Meanwhile, Daly and his colleagues unearthed severe LoF mutations that are expected to obliterate the function of the affected gene, but also seem to contribute to ASD risk through more complex mechanisms.
Together with genetic features implicated in ASD in the past, the inherited mutation mechanisms identified in the studies may eventually pave the way for the use of whole-exome or whole-genome sequencing to diagnose and better treat some ASD cases in a clinical setting.
For the time being, though, those involved in the current studies are hesitant to put a time frame on how long it will take for such tests to enter the clinical arena.
"We think that exome diagnostics has really fantastic potential in understanding autism," Yu said. But, he cautioned, "we don't know what the ultimate diagnostic yield of this is going to be."
"In terms of applying it in the clinic, we're not quite ready for prime time yet," Yu said.
Daly, too, noted that while more widespread diagnostic sequencing for ASD is on the horizon, there is a ways to go before researchers have a clear understanding of the most informative risk genes and some of their clinical effects.
"We can imagine that over the course of the next year, we'll actually get compelling evidence for a meaningful number of genes that do have some predictive value or diagnostic relevance," he said.
Even so, there are those who have started to dip a toe into the sequencing-based diagnostics pool for autism. In 2011, for instance, Mount Sinai School of Medicine's Genetic Testing Laboratory announced plans to introduce a sequencing-based test for autism targeting rare mutations with large effect sizes (CSN 11/16/2011). And just last week, that center said it was gearing up to try to get approval for whole-exome and whole-genome sequencing tests for autism (CSN 1/16/2013).
On the research side, teams are continuing to define the collection of genes that can contribute to autism and the means by which these genes become altered in affected individuals.
While genetic alterations with large effects have been identified — including mutations linked to autism-related syndromes such as Rett syndrome — genetic studies of ASD have shown that the condition often stems from complex genetic changes involving glitches in a wide range of different genes.
Similarly, while some inherited mutations have been tied to ASD, several recent studies have linked ASD to an elevated rate of spontaneous or de novo mutations not found in either parent (CSN 5/17/2011).
Because ASD risk has been pinned on both heritable and environmental factors, though, authors of the new Neuron studies decided it was time to revisit the inherited side of the equation.
To that end, each of the teams used whole-exome sequencing to look for mutations inherited in a recessive fashion, though their approaches — and the types of mutations they identified — were somewhat different.
For their part, Yu and his colleagues settled on a strategy that included whole-exome and other analyses on individuals from consanguineous families, who are especially likely to inherit two copies of the same rare gene mutations.
"Our group has had a long tradition of looking at and trying to study developmental disorders by looking at consanguineous families," Yu noted.
By finding genes linked to high-penetrance disease in these families, he explained, it's sometimes possible to "crack the biological puzzle" and find common biological processes that tend to be upended in a given condition.
"Even if the particular family structure used to find those genes isn't the most common in the world," Yu said, "we're hoping that the pathways that come out of this — and the general themes and principles — will help us understand what's going on in all comers."
For that study, the researchers started by doing whole-genome linkage analyses and exome sequencing on a few of the consanguineous families enrolled by members of the Homozygosity Mapping Collaborative for Autism, or HMCA, explained Boston Children's Hospital's Chahrour.
"We started out with 160 families and took the three most informative families, which have multiple children with autism spectrum disorders, and we applied our method to look for recessive mutations," Chahrour told CSN.
That arm of the study unearthed rare, recessive mutations in a synaptic plasticity gene, along with Mendelian genes involved in metabolic and/or developmental conditions such as nonketotic hyperglycinemia and rhizomelic chondrodysplasia punctate — prompting speculation that some ASD cases are the result of relatively modest mutations to disease genes that normally affect a wider range of organ systems when mutated.
"The mechanism in our families was that we were seeing milder versions of disease than had been previously described," Yu said.
By the end of the study, the team had generated whole-exome sequence data on individuals from 163 consanguineous families in the HMCA collection.
"We ultimately sequenced everyone in the collection, because it's a valuable resource for the community at large," Yu noted. "But we started with the most informative families."
Given that they had found somewhat milder-than-usual mutations in Mendelian disease genes within the first three families, the researchers went on to screen the larger collection of consanguineous family exomes to look for mutations affecting some 70 known disease genes.
There, the team found five families in which affected individuals had severe mutations affecting known disease genes. And two more cases suspected of stemming from similar sorts of high-penetrance mutations fell out of data for hundreds of parent-child trios and quartets assessed through the Simons Simplex Collection.
From their findings so far, Yu said, it seems feasible to think that it may be more straightforward to diagnose at least some cases of autism than was previously thought.
If a subset of ASD is caused by alterations in individual genes with metabolic roles, for instance, finding ways to routinely identify such cases could open new avenues for treatment in those individuals.
"Even if the yield of [diagnosis by] doing whole-exome [sequencing] is 5 percent, say, if in one in 20 kids we can arrive at a specific syndromic or neurometabolic diagnosis," Yu said, "I would argue that that is important."
"Some of these metabolic conditions — not all of them, but some of them — certainly do have interventions that can be applied," he added.
The results also suggest it should be possible to find additional high-penetrance mutations for ASD in non-consanguineous populations by dramatically increasing the number of individuals tested.
Indeed, in the accompanying Neuron study, Daly and his colleagues successfully unearthed rare, inherited mutations associated with ASD in the general population — work that Yu said "speaks really nicely" to the possibility of finding recessive mutations in non-consanguineous populations.
For that effort, collaborators from the Broad Institute and Baylor College of Medicine used Illumina and/or SOLiD instruments to generate sequences from captured coding regions of the genome for 933 individuals with autism and 869 without.
Using this data, the team searched for patterns consistent with a two-hit pattern of ASD inheritance involving rare LoF variants.
Indeed, the researchers found 91 instances in which homozygous or heterozygous compound LoF mutations were expected to produce complete knockouts of genes that normally have LoF variation rates below 5 percent. Two-thirds of the time, these complete knockouts appeared in ASD cases.
Together with existing exome sequence information for thousands more ASD cases and controls, including members of autism-affected parent-child trios and quartets, their data suggest that such rare knockouts are a factor in some 3 percent of ASD cases, as well as 2 percent of the male-specific cases.
"Given that these rare complete knockouts are not found in a single gene but, like the de novo CNVs and SNVs, are distributed across many different genes, these events would have been missed through previous association or linkage studies," Daly and his co-authors noted.
These LoF knockouts appeared a bit more apt to affect genes that have been described as being brain-expressed in past post-mortem studies, the researchers reported.
Even so, Daly noted that more information is needed to understand how widespread that effect is and whether the mutations coincide with genes expressed in specific parts of the brain.
The group is also interested in learning more about the types of biological processes and pathways impacted to the ASD-related rare gene knockouts they uncovered.
If and when routine sequencing-based tests for diagnosing and classifying autism do come to fruition, Daly predicts that it will likely be more advantageous to sequence at least the entire coding region of an individual's genome, rather than try to target sets of specific ASD-associated genes, if only from a price point of view.
"The reality of the technology, though, is that it's actually far less expensive to run an entire exome panel than it is to do a targeted experiment on a set of 10 or 20 genes," he explained. "Given that, I expect that the tendency will be to do the whole-exome or whole-genome experiment and interpret it through sort-of clinical panel lenses rather than doing a targeted sequencing experiment."
On the whole-genome sequencing side, the availability of such data in the research context is useful for uncovering CNVs or larger structural variants that can contribute to ASD, Yu noted.
At the moment, both he and Daly cautioned that there is still a dearth of information for accurately interpreting patterns in the non-coding portion of genomes, making it impractical to consider whole-genome sequencing as a method for diagnosing autism based on the presence or absence of specific regulatory mutations.
That may change going forward as researchers from different centers continue bringing together their sequencing data in consolidated databases and through efforts such as the National Institutes of Health-funded Autism Sequencing Consortium.
"The whole autism community is working toward centralizing the data and the analysis so that collaboratively we can learn much more than our individual datasets tell us," Daly said.
"I think the potential this year is that rather than looking at datasets of 1,000 cases, we'll move to a much larger, collaborative dataset of at least 5,000 cases," he added. "And that's going to be much more informative with respect to specific genes and pathways [involved in ASD]."