The house mouse Mus musculus is an incredible research tool in the biomedical sciences, due to its ease of care and its ability to be genetically manipulated. Although mice aren't humans, they resemble us closely in many ways, including how insulin signaling works. Genetic manipulation of mice allows researchers to identify biological mechanisms and cause-effect relationships in a very precise manner. One way of doing this is to create "knockout" mice that lack a specific gene, in an attempt to determine that gene's importance in a particular process. Another way is to create transgenic mice that express a gene of interest, often modified in some way. A third method is to use an extraordinary (but now common) tool called "Cre-lox" recombination (1), which allows us to delete or add a single gene in a specific tissue or cell type.
Studying the relationship between obesity and insulin resistance is challenging, because the two typically travel together, confounding efforts to determine which is the cause and which is the effect of the other (or neither). Some have proposed the hypothesis that high levels of circulating insulin promote body fat accumulation*. To truly address this question, we need to consider targeted experiments that increase circulating insulin over long periods of time without altering a number of other factors throughout the body. This is where mice come in. Scientists are able to perform precise genetic interventions in mice that increase circulating insulin over a long period of time. These mice should gain fat mass if the hypothesis is correct.
I attended an interesting talk this week by Dr. John S. Parks, who studies lipid metabolism at Wake Forest University. He described data from a mouse missing a protein called ABCA1 specifically from the liver (generated using Cre-lox recombination). For reasons that I won't get into here, when fed an obesity-promoting diet, these mice develop heightened insulin resistance in the liver that results in substantially increased circulating insulin relative to normal mice fed the same diet. This suggests that insulin resistance in the liver can lead to whole-body insulin resistance, which is interesting. However, these mice do not become fatter than normal mice; to the contrary, at older ages they weigh slightly less.
Another interesting model is the liver-specific insulin receptor knockout (LIRKO) mouse. These mice lack the insulin receptor specifically in the liver. Although they have more than ten times as much circulating insulin as normal mice, their leptin sensitivity and body fatness are normal (2). Chris Masterjohn commented on this a while back (3).
Then there's the liver-specific IGF-1 knockout (LID) mouse. They maintain the same body weight as normal mice over time, despite having nearly four times more circulating insulin (4, 5).
Another model is the LIKK mouse, which has a modest overexpression of the inflammatory gene NF-kB in the liver specifically. This mouse has elevated circulating insulin yet has a "normal overall appearance, body weight and food intake" (6).
In an interesting 2002 study, researchers orally administered a small molecule insulin mimetic (drug that mimics insulin action) to mice and placed them on an obesity-promoting diet. Mice that received this treatment gained less fat than mice that were not given the drug over the course of six weeks (7), consistent with insulin's ability to constrain fat mass by acting in the brain (8).
These studies suggest that insulin sensitivity in the liver is important for whole-body insulin function, and that low-grade inflammation can impair insulin signaling in the liver, just as it does in other organs. Together with studies showing that preventing the increase in circulating insulin that occurs on fattening diets has no impact on the rate of fat gain in rodents or dogs (9), this suggests that high circulating insulin per se is neither necessary nor sufficient to cause body fat accumulation.
* This does have some hypothetical basis, which centers around the proposal that high circulating insulin causes leptin resistance in the brain. In cultured fat cells, insulin exposure increases SOCS3 expression (10), raising the possibility that the same could happen in other cell types. The reason this is relevant is because in the hypothalamus, SOCS3 is one of the main negative regulators of leptin signaling during the development of obesity. However, I'm fairly certain that insulin has not been shown to suppress leptin signaling in the hypothalamus, and the studies reported above would suggest it does not do so under normal circumstances.