Sometimes genetics is like a world-class whodunnit. Imagine: Uncle Donald, a millionaire with an attitude, is found dead in the library, having bled to death from a gunshot wound. The murder weapon is quickly identified as the army revolver belonging to Sir James, the master of the manor, who also had a motive, as he will be inheriting Uncle Donalds fortune. He is arrested and sent to jail, but over the years uncertainty about his conviction grows: a murder just doesn’t fit his profile (he quit the army because he couldn’t stand blood). Then, two years later, a new inspector comes to town: she is convinced that modern methods in forensics can shed new light on the case. Indeed, it does not take long until she discovers fragments of skin from the neighbour, Madame Guillotine, on the weapon, and, in addition, uncovers that Madame was actually a master target shooter in her youth.
Noone is surprised, Madame is known for her vicious temper and had a violent argument with Uncle Donald on the evening of his demise. There is no direct evidence for her guilt, but enough circumstantial data to sway any jury. Then a new witness comes forward: Dr Robinson, an arctic explorer, has just returned from a lengthy mission, where she was studying polar bears, cut off from the world, unaware of the drama unfolding in her home town. She tells the police that she saw someone pour anti-coagulant into Uncle Donalds wine on the evening of her leaving party in the manor. The culprit was neither Sir James nor Madame Guillotine, but Robert, the gardeners adopted son, who is really Uncle Donalds illegitimate child. Robert has a prior conviction for pushing his former girlfriend down the stairs – so a domestic accident, with Uncle Donald bleeding to death would fit his MO. Sounds complicated? Well, in the molecular genetics world, this is more less what has been happening to the FTO locus.
The murder and suspect #1
Scientists have been looking for genetic contributors to obesity for quite some time. If obesity is the molecular equivalent of Uncle Donalds murder, then back in 2007 multiple groups found (one of) the murder weapon(s): non-coding variants in the 1st intron of the FTO gene [1-3], the molecular equivalent of the old army revolver. It was assumed that since the variant is part of FTO it must regulate FTO gene expression (a.k.a. the gun originally belonged to Sir James, murder suspect #1). Also, engineering mice with reduced or increased expression of FTO resulted in leaner or fatter animals, respectively [4-5]. Everything was good.
New forensic methods and suspect #2
But, as with Sir James, things just didn’t completely add up: essentially, there was no direct evidence that those non-coding variants have any influence on FTO gene expression. And then a paper was published earlier this year by Smemo, et al. (Obesity-associated variants within FTO form long-range functional connections with IRX3), where a bunch of new (and some not so new) techniques were used to shed further light on the “crime” . Using 4C, a method to measure chromosomal interactions in three-dimensional space, the authors found that a distant neighbouring gene, IRX3, has close contacts with the non-coding FTO variant.
Moreover, the authors report multiple lines of evidence that IRX3 is, in general, under long-range regulation, and that several distant tissue-specific long-range Irx3 enhancers exist. Finally, they demonstrate that reduced Irx3 expression in mice results in decreased body weight. So, similarly to the imaginary murder, lots of circumstantial evidence is accumulating: long-range interactions (skin fragments) are linking suspect #2 (IRX3/Madame Guillotine) to the murder weapon; there’s proof of the ability to use the murder weapon (Irx3 can regulated from a distance – Madame Guillotine was a master target shooter); and there’s data for an opportunity to commit the crime (Irx3 reduction results in reduced body fat, Mme Guillotine had an argument with Uncle Donald). Direct proof is still missing, but the Smemo et al paper still built a very convincing case based on circumstantial data. For a while it seemed the mystery had been solved.
A new angle: suspect #3
Enter Dr Robinson, arctic explorer. In the FTO world: Stratigopoulos et al with their article Hypomorphism for RPGRIP1L, a ciliary gene vicinal to the FTO locus, causes increased adiposity in mice . In the paper, the authors follow up on some of their previous work, where they showed that a regulatory element in intron 1 of FTO regulates expression of yet another closeby gene, Rpgrip1l. Moreover, Rpgrip1l is expressed in cilia, and it is known that disorders caused by mutations in ciliary genes are often associated with obesity. In their new paper, the authors investigate the effect of reduced Rpgrip1l expression on obesity, and also try to establish a cellular explanation of the phenotype. Using a variety of different diets they show that Rpgrip1l knockout mice are consistently fatter, and both the mice and fibroblasts from patients with Joubert syndrome (which is caused by biallelic mutations in RPGRIP1L) have altered Leptin signalling, a pathway that is well-known to be involved in obesity. In mice, they also reveal altered gene expression patterns in the hypothalamus.
So, who really commited the murder?
Well, yeah, good question. And no answer. Yet. Ideally, in a good whodunnit, the master detective should now put all the pieces together and fully reconstruct the murder. In good science, we should do the same. Ideally, you would generate mice or human cell lines with knock-out or modification of the putative regulatory element, and a) study the effect on gene expression (of all three genes)+obesity, and b) try to rescue the expression of your candidate gene, to see if this reverts the phenotype. Unfortunately, such a regulatory element-centered experiment is still missing, despite the individual genes having been studied extensively. There might be pragmatic reasons for this: maybe the sequence in the 1st intron is not well enough conserved in mice, maybe noone really knows what tissue to look at and when to look at it to detect gene expression changes, maybe the effects are too subtle in measure…?
And could there be more than one murderer?
Also, envision this scenario: maybe Uncle Donald wasn’t killed by a single person. maybe noone really liked him, and Sir James, Mme Guillotine and Robert teamed up? Robert put anti-coagulant in Uncle Donalds wine, Sir James provided the murder weapon and Mme Guillotine fired the shot – and Uncle Donald died from the wound because his blood didn’t clot. It’s not implausible and it would fit all the evidence.
Could the same happen at the FTO locus? Well, it is relatively well documented that functionally related and co-regulated genes are frequently adjacent or in close vicinity along the linear chromosome , and that there is probably also evolutionary pressure to keep them together [8-10]. There is also evidence that regulatory elements (like that in the 1st intron) can contact and regulate the expression of multiple genes, often across extended genomic distances [11,12]. We and others have shown that such units of co-expression often coincide with topological units in the genome (so-called topological domains or TADs) [13-15]. Keeping this background info in mind, let’s take another look at the FTO locus:
You can clearly see that all three candidate genes are located within the same TAD – possibly a sign for shared regulation?
Also, if you have access to the RPGRIP1L paper, take a look at Figure 1C: the authors conclude that in mice carrying a null allele of Rpgrip1l, only the expression of Rpgrip1l is altered significantly. While this might be true, please notice that Fto expression is also altered. Maybe it’s not a significant change, but it’s a change nevertheless. It’s unclear why such a change would occur in the presence of a null allele of Rpgrip1l (with no alteration of the regulatory element), but the genes could be linked in a network, or maybe misexpression of one gene causes a global imbalance in the regulation.
What does it all mean?
Obviously, this is a matter of interpretation and speculation. I would like to believe that all evidence presented to date is true. In this case, there could be a regulatory variant in the 1st intron of FTO (or multiple overlapping regulatory elements), which regulates all three genes: FTO, SIX3 and RPGRIP1L. So, when there is an alteration of this regulatory element, the expression of all three genes changes. Possibly it’s only a subtle change in gene expression for each individual genes, but by affecting all three genes (that all influence obesity, based on the gene knock-out studies) the net effect might be significant. If there are multiple overlapping regualtory elements, maybe one element regulates expression of one gene in one tissue, another the expression of the second gene in a second tissue, and a third does the same for the third gene. Human gene regulation can be complex. Maybe this will be our next model locus to prove this?
1. Dina, C. et al. (2007) Variation in FTO contributes to childhood obesity and severe adult obesity. Nature Genet 39: 724-726
2. Frayling, TM. et al (2007) A common variant in the FTO gene is associated with body mass index and predisposes to childhood and adult obesity. Science 316: 889-894
3. Scuteri, A. et al (2007) Genome-wide association scan shows genetic variants in the FTO gene are associated with obesity-related traits. PLoS Genet 3: e115
4. Fischer, J. et al (2009) Inactivation of the Fto gene protects from obesity. Nature 458: 894-898.
5. Church, C. et al (2010) Overexpression of Fto leads to increased food intake and results in obesity. Nature Genet 42: 1086-1092
6. Smemo, S. et al (2014) Obesity-associated variants within FTO form long-range functional connections with IRX3. Nature 507: 371-375
7. Stratigopoulos, G. et al (2014) Hypomorphism for RPGRIP1L, a ciliary gene vicinal to the FTO locus, causes increased adiposity in mice. Cell Metab 19:767-79
8. Irimia, M. et al (2012) Extensive conservation of ancient microsynteny across metazoans due to cis-regulatory constraints. Genome Res 22: 2356-2367
9. Michalak, P (2008) Coexpression, coregulation and cofunctionality of neighbouring genes in eukaryotic genomes. Genomics 91: 243-248
10. De, S and Babu, M (2010) Genomic neighbourhood and the regulation of gene expression. Curr Opin Cell Biol 22: 326-333
11. Sanyal, A. et al (2012) The long-range interaction landscape of gene promoters. Nature 489: 109-113.
12. Chepelev, I. et al (2012) Characterization of genome-wide enhancer-promoter interactions reveals co-expression of interacting genes and modes of higher order chromatin organisation. Cell Research 22: 490-503
13. Symmons, O. et al (2014) Functional and topological characteristics of mammalian regulatory domains. Genome Res 24: 390-400
14. Nora, E.P. et al (2012) Spatial partitioning of the regulatory landscape of the X inactivation centre. Nature 485: 381‐385
15. Dixon, J.R. et al. (2012) Topological domains in mammalian genomes identified by analysis of chromatin interactions. Nature 485: 376‐380
Smemo S, Tena JJ, Kim KH, Gamazon ER, Sakabe NJ, Gómez-Marín C, Aneas I, Credidio FL, Sobreira DR, Wasserman NF, Lee JH, Puviindran V, Tam D, Shen M, Son JE, Vakili NA, Sung HK, Naranjo S, Acemel RD, Manzanares M, Nagy A, Cox NJ, Hui CC, Gomez-Skarmeta JL, & Nóbrega MA (2014). Obesity-associated variants within FTO form long-range functional connections with IRX3. Nature, 507 (7492), 371-5 PMID: 24646999