Genetics of feeding behavior in mice: finding the ob gene and the db gene

[Cincinnatus]

There's lots of talk about finding "a gene for aggression" or "a gene for homosexuality" etc. I'm less interested in cases where the results are clearly ambiguous like in most (if not all) human studies. Rather I'd like to know what people think we even mean when we refer to a gene as being the gene "for behavior X".

Some examples:

In Drosophila melanogaster there is a gene called fruitless (fru). When mutated, this gene causes a set of behaviors in fruitflies that we might recognize as similar to homosexuality.

DA Gailey, JC Hall Behavior and Cytogenetics of fruitless in Drosophila melanogaster: Different Courtship Defects Caused by Separate, Closely Linked Lesions - Genetics, 1989 - Genetics Soc America

Fixed action patterns like courtship in Drosophila appear to be highly genetically determined with little environmental effect on how and when they are performed. This is in contrast to most human behaviors which are more influenced by environmental or other factors. Birdsong is another good example of a behavior that is inflenced by genetics but highly environment dependent (in most species anyway) birds typically learn their songs from other birds. However, this process can be disrupted by genetic methods.

Also in Drosophila (and many other organisms) there are the so-called "clock genes". These include period (per), timeless (tim) and doubletime (dbt). Mutations of these genes lead to changes in the animal's rest/activity cycles. That is, they change the animal's behavior.

MW Young THE MOLECULAR CONTROL OF CIRCADIAN BEHAVIORAL RHYTHMS AND THEIR ENTRAINMENT IN DROSOPHILA - Annual Reviews in Biochemistry, 1998 - Annual Reviews

---

The existence of fixed action patterns like Drosophila courtship which are highly complex behaviors that can be disrupted by the mutation of a single gene should be reason enough to consider "the gene for behavior X" talk as worthy of examination.

We can distinguish several different ways in which genes might influence behavior. They migh control the actual instance of the behavior while the animal performs it. Alternatively, they might be active only during development during which time they direct the forming of specialized neural circuits either for performing the behavior of for potentiating the likelihood of performing the behavior. After the animal is mature these genes may no longer be active.

For a particularly interesting review article on this topic check out:

BS Baker, BJ Taylor, JC Hall Are Complex Behaviors Specified by Dedicated Regulatory Genes? Reasoning from Drosophila - Cell, 2001 - Elsevier

---

I'd like to know what other people think about a few questions.

1) Does finding a gene that influences a complex fixed action pattern constitute finding "a gene for a behavior"? Or is this somehow less liberally worrisome?

2) Do you know of any instances where genes have been shown to be causal of the performance of a complex behavior that is not a fixed action pattern? Do you expect that we will find uncontroversial examples of this?

3) What do we really mean when we say that a gene controls a behavior?

 

[Cincinnatus]

Well it's been a while now since I posted this with no replies. It seems no one is interested. Or maybe people are intimidated by citations. Could this explain why I've seen so few posts with citations on this forum? Anyway, I certainly didn't intend for responders to actually look up those citations, I merely included them for reference or in case anyone was interested.

 

[BoomBoom]

OK, well I guess I'll throw in my 2 cents for whatever that's worth (2 cents maybe?)...

While there may be some suggestions through statistical correlations that perhaps some people that may have a certain SNP or something for a certain gene may exhibit some sort of behavior (being very vague here), I think it may be "jumping the gun" for us to assume that gene causes that behavior. The sheer number of factors that go into something (especially as complex as a behavioral trait) makes it unlikely that anyone gene is going to be a "cause" of this trait. It is more likely to be a great number of factors all acting together. Also, behavior is somewhat tough to quantify and qualify.

That said though, I believe in the future as we gain much better understanding of expression networks and the like, we will have a better understanding of the influence of particular genes and how all their interactions may influence these things.

 

[Ygggdrasil]

I don't know much about neurobiology, but here's my two cents:

1) Does finding a gene that influences a complex fixed action pattern constitute finding "a gene for a behavior"? Or is this somehow less liberally worrisome?

I think most simple, genetically-encoded behaviors are controlled by neural circuits, not single genes. Certainly, there are certain genes that, when knocked out, can destroy the circuit and prevent the behavior, but I think the relevant quanta to think about are circuits and not individual genes.

2) Do you know of any instances where genes have been shown to be causal of the performance of a complex behavior that is not a fixed action pattern? Do you expect that we will find uncontroversial examples of this?

I'm not too familiar with the literature, so no, I don't know any examples of this. It could be possible however. Some mutations can hyperactivate a specific gene (e.g. by removing an ihnibitor of that gene) and if such a mutation occurs with a gene critical to a behavioral circuit, the resulting phenotype could be performing that behavior in inappropriate settings (e.g. males courting to both males and females).

3) What do we really mean when we say that a gene controls a behavior?

We found this cool mutant fly/worm/mouse and we want to publish in Nature/Science/Cell (j/k). Probably what most people mean is that the gene is necessary for the correct functioning of the circuit controlling that behavior.

 

[Cincinnatus]

Lets ask a hypothetical question. We assume there is some behavior which we can measure and quantify. Then let's say we find a gene such that mutations there affect this behavior in quantifiable ways.

Maybe one of the examples I gave in the first post above satisfies these assumptions, maybe not.

Either way, we can ask the question: Does it make sense to say that this gene "causes" this behavior?

Rather, when we say that such a gene causes that behavior are we using the same notion of causation that we are using when we say "the gene for beta-galactosidase causes the organism to be able to metabolize lactose" or some other less controversial statement ?

In the case of the gene for beta-galactosidase allowing the metabolism of lactose we seem to have a mechanism that links the gene with the effect. The gene codes for a particular enzyme which takes lactose as its substrate and breaks it into its component simpler sugars. We don't have as well a worked out mechanism for any "behavior gene". We can wonder if such mechanistic explanations will ever be possible for behavior genes.

Though, in the case of fruitless (and probably quite a few others) we know a fair amount about what the gene product is doing (it's a transcription factor active during brain development). Maybe someone thinks that knowing this kind of information constitutes the same kind of mechanistic understanding that we have for lactose metabolism.

 

[LowlyPion]

Rather, when we say that such a gene causes that behavior are we using the same notion of causation that we are using when we say "the gene for beta-galactosidase causes the organism to be able to metabolize lactose" or some other less controversial statement ?

Unfortunately behavior seems inextricably intertwined with environmental factors and sensory presentations. And identifying causal links absent the underlying chemistry looks a little treacherous.

I'd say to the extent that you can identify behavior with chemical causation, say psychosis that can be attributed to say some quantifiable dopamine level aberrations or receptor activity or some physical structure as that, then maybe you can discuss that in terms of the behavior effects that may flow from that, but in reality I think the real causation is with gene expression modifying the chemistry or cellular structure and not the consequential behavior that flows from it.

 

[BoomBoom]

Though, in the case of fruitless (and probably quite a few others) we know a fair amount about what the gene product is doing (it's a transcription factor active during brain development). Maybe someone thinks that knowing this kind of information constitutes the same kind of mechanistic understanding that we have for lactose metabolism.

Yeah, I think that once we understand what all the interactions are and what their roles and functions are (truly understanding the whole network), then we will be able to more accurately predict certain outcomes. Although with behaviors we will also need a firm understanding of neural networks as well.

But, at this point, we are not to that level of understanding (not even close)...so I am always skeptical when I hear a news report stating, "a gene has been found that links ___ to ___".

 

[NoTime]

Having looked thru some recent Science articles, it looks like not only genes but methylation and other related processes have effects on mental processing.

 

[Moonbear]

Actually, the "one gene, one behavior" hypothesis is fairly outdated. There was a bit of it still lingering when I was in grad school, but now it's pretty well recognized that behavioral genetics are more complicated than that, with multiple genes influencing a particular behavior, and multiple behaviors affected by a single gene.

Now, if you read about something like a "gene for aggression," it really means, "ONE OF THE GENES for aggression," and does not exclude that it might also be involved in other behaviors, such as fearfulness, maternal behavior, etc.

 

[Ygggdrasil]

We don't have as well a worked out mechanism for any "behavior gene". We can wonder if such mechanistic explanations will ever be possible for behavior genes.

Here's a neurobio example I do know where researchers have identified a few genes that affect feeding behavior and have a decent understanding of the molecular mechanisms (but unfortunately, the circuits controlling feeding behavior still don't seem to be well understood yet).

Mouse geneticists in the 1950s had identified various inbred strains of mice that were severely obese, primarily because they ate too much. This observation suggested that the mice had a mutation in a gene involved body weight homeostasis. Two of the obesity-causing mutations were named ob and db (for obese and diabetic, two phenotypes of the mutations). When better tools in molecular genetics were available in the 1990s, researchers were able to find the location of the mutations in the genome and sequence the ob and db genes (the "I found the gene for obesity" moment).

Examination of the ob gene showed that it encoded a secreted protein that was primarily made in fat tissue. This result immediately suggested that the product of the ob gene, a protein called leptin, acted as a homeostatic signal to suppress food intake: as mice became fatter, the leptin signal would intensify and tell them to eat less, thus preventing them from becoming too fat. This hypothesis was confirmed by experiments showing that injection of leptin into ob mice could correct the obesity phenotype.

Interestingly, injection of leptin into db mice, another strain of obese mice, did not alleviate the obesity phenotype. However, breeding mice with both the ob and db mutations did not increase the severity of the obesity phenotype. These two results led scientists to believe that the db gene might encode the protein that receives the leptin signal, the leptin receptor. Indeed, the identification and sequencing of the db gene confirmed that it did, in fact, encode a cell-surface receptor protein. Further biochemical studies showed that the db gene product bound leptin.

With this knowledge researchers could look at brain slices and identify regions of the brain that contain leptin receptors, as these areas would likely be involved in receiving the leptin signal and activating neurons involving feeding behavior and body weight homeostasis. These studies and a variety of more sophisticated genetic studies led to the identification of two populations of neurons in the hypothalamus that are thought to be the primary neurons involved in receiving the leptin signal and transducing it to other neurons.


So for the ob and db genes, we do have a very good mechanistic understanding of how these genes affect feeding behavior. Granted, the picture becomes fuzzy when you ask what happen once leptin binds to leptin receptors in the hypothalamus (what proteins are turned on by the activated leptin receptor? to which neurons do the leptin-receiving neurons communicate? where does the signal propagate and which neurons are the effectors of the signal?), but I think this is a good example of how genetic studies can lead to molecular understanding of behavior. Of course, in this case the behavior (feeding) was very easily quantified, which definitely helped. With more complex behavior that cannot be so easily quantified (e.g. aggression, maleness, to name a few mentioned in this thread), the genetics won't be as easy.