In 2007, CSHL researcher Michael Wigler revealed that somechildren with autism have a small deletion on chromosome 16, affecting 27 genesin a region of our genomes referred to as 16p11.2. The deletion—which causeschildren to inherit only a single copy of the 27-gene cluster—is one of themost common copy number variations (CNVs) associated with autism.
Now, by creating mouse models of autism, CSHL scientist AleaMills and colleagues provided the first functional evidence that inheriting fewercopies of these genes leads to features resembling those used to diagnosechildren with autism.
This brings scientists one step closer to diagnosing autismin children at an early age, when they can most benefit from early interventionprogram such as speech and occupational therapy, and possibly discoveringeffective treatments.
"The idea that this deletion might be causing autism wasexciting," says Mills. "So we asked whether clipping out the same set of genesin mice would have any effect."
After altering the copy number of the region of the mousegenome corresponding to human 16p11.2, researchers placed the mouse and a wildmouse into a cage and observed. The action is shown on a video on the CSHLwebsite and is titled, "Climbing episodes of wild-type mice and mice with16p11.2 deletion."
At the beginning of the 10-hour trial, both mice climb tothe lower part of the cage ceiling, returning to the floor with their hindlimbs. Later, wild-type mice progress until they reach the highest point of theceiling of the cage, returning to the floor from different ceiling areas withtheir forelimbs.
In contrast, 16p11.2 deletion mice lack this progressivebehavior, exhibiting the same behavior as earlier in the trial, researcherssay. This type of behavior mimics that of an autistic child often repeating thesame tasks at a lower level compared to a typical child's attempts to climbhigher and progress.
"We were really blown away with the behavior of these mice,"Mills tells ddn. "The behavioralanalyses setup that Dr. Guy Horev in my group designed was just phenomenal, asit was based on the idea that we need to be completely open-minded in how16p11.2 CNVs might act."
It turns out that "this type of unbiased screen is seldomperformed," she says. "Instead, many researchers go into it looking for whatthey think should be seen. Since the clinical features of autism varywidely—even within the same family carrying the identical CNV—we thought itimperative that we look for everything, whether it was considered standard ornot.
"Of course, we got flack for this approach by some of themore rigid behaviorists, but it really paid off," Mills continues. "We likelywould not have found the interesting behavioral changes if we had approached itwith a preconceived or standard approach. I think our work calls for a new lookat how behavioral analyses in rodents should be done, not just in models usedin autism research, but for other neurodevelopmental/neuropsychiatric models aswell."
Mills says the team also studied the brains of thelaboratory mice. Co-author Mark Henkleman of the Hospital for Sick Children inToronto ran MRI scans on the mice and "found eight different regions of thebrain that were severely affected," Mills says. "Interestingly, one of theseregions of the brain is the hypothalamus."
Previous research has linked the hypothalamus with somerepetitive behaviors that are characteristic of autism.
So is this deletion—a missing second copy of this 27-genecluster—the "cause" of autism?
"It's not that simple," Mills says. "There are a number ofdifferent genetic changes that have been found [to be] associated with autism.There could be a network of interacting players, but I don't think there willbe a single region [of the genome] that is responsible for all cases of autism,from what we're seeing."
Still, putting together pieces of the genetic puzzlesurrounding autism may help researchers understand the disorder better, andthat might lead to new therapies, she says. And the development of mice withautism-like behaviors may help scientists in the process.
"We have taken the very first steps in generating thesemodels, and discovering that 16p11.2 CNVs have a dramatic effect in the brain,"Mills says. "This opens up so many more questions: What gene, or groups ofgenes, is responsible for the myriad phenotypes we see? What are the underlyingmechanisms? Can we fix it? When do the very first signs of autism develop?"
In the meantime, Mills ponders the next step.
"We are currently generating mice with chromosome-engineeredsub-deletions; we can then analyze these for the phenotypes we have found forthe full 16p11.2 deletion, thereby honing in on the causative gene(s)," Millssays. "We are going deeper into the mechanism underlying these phenotypes. Ifwe know this, perhaps we can design ways to fix it. We will determine whetherwe are able to correct the phenotypes of the 16p11.2 CNV mice. We are analyzingthese models for early signs associated with autism, in the hope that thismight be useful for a prognostic indicator for the syndrome."