As with many of the genes in our genome, different people carry different versions of FTO. People with two copies of the "fat" version of the FTO SNPs average about 7 pounds (3 kg) heavier than people with two copies of the "thin" version, and they also tend to eat more calories (1, 2).
Despite being the most consistent hit in these genetic studies, FTO has remained a mystery. As with most obesity-associated genes, it's expressed in the brain and it seems to respond somewhat to nutritional status. Yet its function is difficult to reconcile with a role in weight regulation:
- It's an enzyme that removes methyl groups from RNA, which doesn't immediately suggest a weight-specific function.
- It's not primarily expressed in the brain or in body fat, but in all tissues.
- Most importantly, as far as we know, the different versions of the gene do not result in different tissue levels of FTO, or different activity of the FTO enzyme, so it's hard to understand how they would impact anything at all.
An important thing to keep in mind is that GWAS studies don't usually pinpoint specific genes. Typically, they tell us that obesity risk is associated with variability in a particular region of the genome. If the region corresponds to the location of a single gene, it's a pretty good guess that the gene is the culprit. However, that's not always the case...
A New Understanding of the FTO SNP
A paper recently published in Nature upturns everything we thought we knew about the FTO SNPs (3). Dr. Scott Smemo and colleagues begin by pointing out that the FTO SNPs are associated with a non-coding region of the FTO gene. What this means is that the SNPs are in a region of the gene that gets edited out prior to the construction of the FTO protein. Non-coding regions don't contribute to the sequence of proteins such as FTO, but they do often contain regulatory elements that influence how the gene is expressed, or how it's spliced.
Usually, regulatory elements affect the expression of the closest gene, but sometimes they can act at quite a distance. In the case of the FTO SNPs, Smemo and colleagues showed that they're associated with the expression of a gene called IRX3 that's located millions of base pairs away from it! While the SNPs have no association with FTO expression, they are associated with IRX3 expression levels in the brain. People with the "fat" SNPs have a higher level of IRX3 expression.
As I never tire of repeating, if you find a gene that impacts body fatness, the first place to look for its function is the hypothalamus. The hypothalamus is the part of the brain that regulates body fatness, and most genes that affect body fatness have been linked to hypothalamic function in one way or another.
In this case, all signs were pointing to the brain as the key location of IRX3's effects. To test this hypothesis, Smemo and colleagues made genetically modified mice that lack IRX3. What they found is striking: IRX3 knockout mice are unusually lean, and totally resistant to dietary obesity. While normal mice doubled their fat mass when fed a fattening diet, IRX3 knockout mice didn't gain a significant amount of fat on the same regimen. They further reported that specifically interfering with IRX3 function in the hypothalamus also resulted in leanness, supporting the hypothesis that IRX3 exerts its weight-regulating function primarily in the hypothalamus.
What does IRX3 do? It's a transcription factor, meaning it's a protein that regulates the expression of other genes. It's known to be highly expressed in the brain and to play a role in brain development. We can only speculate for the time being, but it may impact the development of brain circuits that regulate body fatness. Or it may do something else-- no one knows.
The identification of IRX3 as the target of the FTO SNPs is really exciting because it suggests there's still a lot to be learned about how the brain regulates body fatness.