A couple of years ago, I made a prediction about the direction autism research would take, and the biggest blinking neon signpost pointed straight at CNVs, or copy number variations. The variations are in how many copies of a specific DNA sequence a person has. Yes, these do vary, sometimes leading to nothing of much note but other times leading to significant outcomes. To quote myself from a post from last April about the state of the science in autism research--in which I listed CNVs first--I said:
A disorder or developmental difference with this etiological complexity may come down to something as individually variable as the number of copies of a gene sequence a person carries. Bob as five repeats in one area of his DNA, Sue has 10, Sue has autism, Bob does not. But when Bob has a child, by that time those repeats have doubled to 10 in his germ line (where his sperm comes from), and he passes 10 to his child. His child has autism. His next child gets a sperm that carries only 5 repeats and does not have autism. The neighbors all have between zero and four repeats, and none of them has autism. In other words, that saying that people use a lot with autism, "When you meet one person with autism, you've met one person with autism"? Could be more precise than we think.
Why is copy number variation relevant (or not)? Previous work suggests a link between CNVs and schizophrenia and autism, in a yin-yang relationship in which a deletion was associated with schizophrenia while a duplication was associated with autism. It's important to keep in mind that these associations generally crop up in a small percentage of people with either condition, implying that many more pathways remain to be identified, CNV related or not.
Scraping again from a previous post of mine, here's the basic upshot of variation in copy numbers:
Does it hurt the organism for these duplications to happen? Depends on what they affect. While you could probably glide through life just fine with a duplicated finger, a duplicated head might present a bit of a problem, and DNA has similarly influential and not-so-influential areas. We've already linked a few disorders to sequence duplications. Huntington's traces to having a certain number of triplet repeats of the DNA alphabet in a gene. While these might be totally innocuous somewhere else, if these CAG-CAG-CAG repeats occur in great enough numbers in the sensitive areas, the result can be disease.
Autism isn't Huntington's, which traces straight to this series of CAG repeats. Even in Huntington's, a low number of repeats (10 to 28) leads to no symptoms at all, but after a threshold number (>36), symptoms do show up, and the more repeats, often the earlier and more severe the symptoms are.
For a condition like autism, though, the underlying DNA links will likely be more complex and numerous. Despite this probable complexity, though, researchers have to start somewhere, and a promising beginning point is on chromosome 16. Chromosomes have "arms," a short one on top and a long one on the bottom. For what appears to have been purposes of sheer obfuscation, scientists opted to designate the short arm with the French word "petit" (small) and the long arm with the letter "q" because, well, Q follows P in the alphabet. Got that? The arm of specific interest for autism and other neurobiological conditions is the short one, or 16p. A diagram of this handsome chromosome appears below. As you can see, on that short arm (left), a big yellow arrow points to the short arm just before it ends. That's p11.2, or as it is officially known, 16p11.2.
This designated area doesn't indicate a single gene. Instead, it's a section of the chromosome that contains the sequences for about 25 genes (new work says 27). Deletion of this chromosome has been associated with a syndrome of effects, including delayed language development with a greater delay for expressed language than for received language, learning issues or intellectual disability, social impairments, and some non-specific and variable asymmetry of facial features. These features sound a whole lot like autism.
The thing is, even when duplications or deletions are present, the result varies from person to person. Geneticists use a concept, expressivity, to explain variability among individuals in how the presence of a genetic difference shows up. Expressivity describes the magnitude to which a person carrying the genetic difference expresses it.
Even disorders with solely a genetic basis can vary in how individuals express them. No gene is an island, and cells can modify genes, shutting them up with the equivalent of chemical duct tape on the mouth or using them more than normal. The gene is a code for a protein, and its protein also is no island, interacting with innumerable environmental factors from other proteins to nutrition and metabolism that can influence its work. In other words, having a gene difference is not always a guarantee of how a person will express that difference or how intensely.
So what we have here is one candidate section among many of one candidate chromosome among 23 with possible deletions or duplications that lead to neurobiological differences that may vary in expression from individual to individual. Experts often refer to autism as a spectrum disorder, and this single example already yields a potential spectrum of manifestions. Multiply it by, oh, hundreds of other possible CNVs, and you'll realize why it's been so hard to nail down the causes of autism. But, people gotta start somewhere.
Homing in on 16p11.2, a group of researchers recently developed what we in the biz called an "animal model" of autism. The model here is the mouse, an oft-used animal because of our now well-honed ability to manipulate its genes and how they are expressed. We can even tie gene expression to the expression of a protein from the jellyfish called green fluorescent protein and make mice glow green under appropriate light. But in this case, the researchers produced mice that seem to have autism.
The title of the paper they published about their mice is "Dosage-dependent phenotypes in models of 16p11.2 lesions found in autism." Translated, they made mice that lacked 16p11.2 on one of their two #16 chromosomes or had a duplicated 16p11.2 section (these changes in the chromosome are "lesions") on one of their two #16 chromosomes. I note that even the mice with a deletion still had another normal copy of the chromosome, so their "dose" of the genes in that section was half of normal. Thus, some mice had a half dose of 16p11.2 and some mice had a 1.5x dose (correction) (one normal chromosome and one with a duplication), while still others had the regular dose of one copy of the region on each of the two #16 chromosomes.
The investigators found that a half dose of 16p11.2 (deletion) produced worse outcomes in terms of behavior, survival, and brain structure compared to a 1.5x dose or the regular dose. (Unfortunately, the paper is not open access, so readers cannot download it for themselves unless they pay ten bucks.) Behaviorally, the half-dose mice had sleep problems, showed repetitive and restricted behaviors (classic autism terminology), and were hyperactive. Half of them died just after birth. The 1.5x-dose mice also showed behavioral differences, but not to the extent of the half-dose mice. The graph below gives a quick visual representation of the findings:
What's the use of a mouse model of autism or even of focusing on this one section of chromosome 16? Even though this focus already seems pretty fine-grained, scientists can now home in on the 27 genes located in the 16p11.2 region and start doing functional studies, using mice. To determine how the genes might function in the context of autism, they can knock out specific genes or duplicate them (and thus their protein products) and determine how these changes affect the mice. Studies like this not only will give us insight into the effects of these specific genes but may also point researchers to other genes on other chromosomes that have related functions.
Looking for the causes of autism in our genome isn't like looking for a needle in a haystack. The haystack itself is a pile of needles, each one a candidate for involvement. The current needle of focus is 16p11.2. It's just a place to start because we have to start somewhere, and science can be as much about accidents of existing data as about carefully mapped out plans of action. What are some of the other needles that make up the haystack? This should give you an idea. For any candidate gene or group of genes, researchers can start with a mouse model modified at that point in the genome and go from there. You can expect to hear much more about CNVs and different models mice with autism. Like autism itself, the options form a seemingly endlessly refined spectrum of possibilities.
Top image photo credit: This three-dimensional representation of the mouse brain highlights eight regions (shown with different colors) affected by 16.p11.2 deletion. (Image courtesy of Mills@CSHL)