Autism has been difficult to characterize genetically. It is probably a common endpoint for many genotypes, and is a multifactorial (“complex”) trait. That is, hundreds of genes contribute risk to different degrees, as do environmental factors. Research reports implicate either dozens of genes in genomewide sweeps, or focus on a few genes that encode proteins that act at synapses, such as the neuroligins and neurexins.
Taking cues from the fact that males with autism outnumber females four to one, females are more severely affected, and siblings of females with autism are more likely to also have the condition than siblings of affected males, a team led by Tychele Turner and Aravinda Chakravarti of Johns Hopkins University School of Medicine searched for candidate genes among 13 “female-enriched multiplex families” — FEMFs – that have two or more girls or women with severe autism. The study was published online March 25 in Nature.
Presumably, causative mutations in the female members of these families would have more severe effects. So identifying genes that stand out in their exomes (the protein-encoding part of the genome) and that make physiological sense – that is, affect the brain – could reveal general steps in the beginnings of autism in the broader population. The researchers describe their approach as “modest numbers of samples of rare extreme phenotypes, in contrast to large numbers of typical cases.”
EXPERIMENTAL RESULTS CONVERGE
The FEMFs indeed revealed 18 candidate genes, four of which emerged as the strongest. The researchers further tested the most likely gene, CTNND2, because it had turned up in other studies. CTNND2 encodes a protein called delta-2 catenin. A series of experiments then led to the following findings:
CLUE 1: Most mutations in humans delete all or part of the gene.
CLUE 2: Knocking out the gene in mice and zebrafish disrupts synapses. Therefore the mutation’s effect is a loss of a normal function, rather than a gain of a new function – and it affects neurons.
CLUE 3: The gene is expressed at 20 times higher level in human fetal brain cells than in human adult brain cells. (This is consistent with the fact that the brain changes that set the stage for autism begin prenatally.)
CLUE 4: The Allen Brain Atlas identified genes with which CTNND2 interacts. They include the usual suspects – proteins that act at synapses or in neural extensions, and in the actin cytoskeleton – but also a new role, chromatin modification. This means that absence of CTNND2 protein would affect many genes, a broad stroke that could paint the many manifestations of autism.
ECHOES OF ALZHEIMER’S AND BRCA1
The new autism study brilliantly uses a handful of unusual families to open a door to the inner workings of autism. Even though the news release calls the FEMF strategy a “novel approach” and “unconventional method,” it actually continues the tradition that first drew me to study genetics – severe or unusual cases that provide insights into disease mechanisms that affect many.
Two examples come to mind: Alzheimer’s disease and breast cancer.
The first recognized case of Alzheimer’s disease was Auguste Deter, who began displaying bizarre behavior when she was in her late forties, in the late 1890s. She would scream piteously for hours, often in the dead of night, and traipse around cocooned in bedsheets, propelled by wild hallucinations and delusions. Auguste also had profound memory loss, unable even to write a simple sentence because she’d forget what had just been asked of her. Yet she had glimpses of self-awareness, saying now and then, “I have lost myself.”
Auguste’s terrified husband took her to the Institution for the Mentally Ill and for Epileptics in Frankfurt, where she came under the care of Dr. Alois Alzheimer in 1901. She died five years later, at age 56. In November of that year, after examining her brain, Dr. Alzheimer gave his now-famous lecture on her condition, which was published in 1911 as “eine eigenartige Erkrankung der Hirnrinde” (“a peculiar disorder of the cerebral cortex”).
Alas, Dr. Alzheimer’s meticulous and vivid description was lost to history, even as increasing lifespan revealed many people with forms of the condition that Auguste Deter had.
In 1996, psychiatrist Konrad Maurer rediscovered Alzheimer’s medical records for Auguste Deter, and published an analysis in The Lancet. A year earlier, a team from the University of Toronto had identified the presenilin 1 gene in some families with early-onset Alzheimer’s disease. Then in 2013, researchers discovered that Auguste Deter had a presenilin 1 mutation. (See Comment below, this paper is incorrect. Thank you reader!)
Even more so than the case of Auguste Deter, the new study on autism using female-enriched families reminded me of the 1990 paper in Science introducing the breast cancer 1 gene, better known as BRCA1. That study sought families enriched for early-onset breast cancer.
Mary Claire King famously trolled for susceptibility genes among 329 members of 23 extended families, who included 146 cases of early-onset breast cancer. For anyone who remembers LOD scores (“logarithm of the odds”), a statistic that shows linkage of a phenotype with a particular part of a chromosome, BRCA1 had a good one – 5.98 – signaling something amiss on chromosome 17. Since then, thousands of women and some men have had BRCA1 tests.
CONNECTING THE DOTS THROUGH TIME
The newfound mutations in CTNND2 that may cause or contribute to autism are rare, as are mutations in presenilin 1 among people with Alzheimer’s disease and mutations in BRCA1 among people with breast cancer. But identifying these genes and their pathogenic variants, in the very few patients who serve as canaries-in-the-coalmine, can illuminate at the molecular level how these diseases begin and develop. And that’s a direct route to treating, or at least slowing or controlling, them.