Many of the genes implicated in autism are unusually long — a feature that leaves them susceptible to breaking during cell division, a new study finds1. These genes may be ‘hot spots’ for mutations that crop up in patches of the brain.
The study identified 27 fragile regions in the DNA of cultured precursors to neurons; 12 of the regions land in long genes linked to autism. Cells can make mistakes when repairing these breaks, sometimes disrupting the very sequence they are trying to restore.
Autism-linked genes — especially those that function at the junctions between neurons, or synapses — are roughly four times larger than the average gene in neurons2. The new findings suggest that long genes are particularly prone to mutations.
“The really striking thing is that almost all of [the fragile genes] were involved in synapse function and neuronal communication,” says lead researcher Frederick Alt, professor of genetics at Harvard University. Alt’s work appeared 11 February in Cell.
The study offers clues about the origin of somatic mutations, which affect only some cells in the brain. Somatic mutations in long autism genes could underlie cases of autism that do not have a known genetic origin, says Alt.
The new study uses a clever technique to show how somatic mutations might arise during development, says Mark Zylka, professor of cell biology and physiology at the University of North Carolina at Chapel Hill. “It’s another way of damaging genes we already know are implicated in autism, but damaging them in only a subset of brain cells,” he says. Zylka led a 2013 study implicating long genes in autism but was not involved in the new work.
Alt and his team worked with the stem cell precursors to neurons. They used a combination of drugs and genetic engineering to induce the DNA breaks and to block the cells from repairing the damage or killing themselves off.
To map the location of these breaks across the genome, they used the gene-editing tool CRISPR to introduce a ‘bait’ break in a region of chromosome 12. As the cell tries to repair this break, it may join this region with breaks already present in the cell, providing a neat way to identify those breaks.
This approach revealed nearly 10,000 DNA breaks distributed across the genome. Alt’s team then used an algorithm to home in on 27 hot spots where the breaks cluster, pointing to regions that are fragile.
All 27 hot spots land in genes, which occupy less than 2 percent of the genome. All but one are roughly 20 times longer than the average mouse gene (the remaining one is 4 to 5 times longer). And 12 hot spots affect genes with ties to autism, such as NRXN1, NRXN3 and CTNND2, which code for synaptic proteins.
“This relationship between genes that are affected in autism and double-stranded breaks is quite interesting,” says Sacha Nelson, professor of biology at Brandeis University in Waltham, Massachusetts, who was not involved in the work.
All but one of the 27 hot spot genes also show signs of being actively transcribed into proteins but are also among the last genes that cells copy during division. This combination of features is probably what makes these particular long genes especially fragile, says Zylka.
Long genes naturally take more time to copy than shorter ones do. When the process starts late, a cell may deploy its protein-making machinery before the copying apparatus is finished, resulting in a physical collision between the two sets of molecules that can damage DNA, Zylka says.
It is unclear how often DNA breaks, and what proportion of breaks persist after the cells’ repair attempts. A 2013 study found somatic mutations in roughly 41 percent of neurons plucked from postmortem brains3. Of the 110 DNA duplications and deletions seen in these neurons, 9 overlap with genes that are hot spots for breaks, says Alt. “This implies to us that there’s a good chance that this does occur normally at some frequency,” he says.
For example, environmental factors that stress the copying machinery, such as folate, could induce DNA breaks, Alt says.
It’s also unclear why genes in neurons are especially long. One theory is that DNA breaks might sometimes be beneficial in generating genetic diversity in neurons4. “If this process is happening and if it is there to generate diversity, it’s still going to have to be fine-tuned and regulated in ways we don’t understand yet,” Alt says. “There’s a lot to be learned.”