Published online 28 May 2009 | Nature | doi:10.1038/news.2009.523


Flaws found in mouse model of diabetes

Missing protein could limit the use of mice in studies of the disease.

MouseTesting prospective diabetes drugs on mice might give misleading results.Getty / Melina Souza Photography

A protein found in humans — but not in mice — could be important for regulating blood sugar and may play a role in diabetes, researchers have found.

The protein, clathrin isoform CHC22, is produced by muscle and fat cells, and is activated by insulin in response to rising levels of glucose in the blood. In a paper published today in Science1, researchers report that CHC22 helps to store glucose transporters in membrane-bound structures called vesicles. When insulin levels rise, these vesicles move to the cell membrane, where the transporters are then able to take up sugar from the blood and provide a source of energy for the cells.

In patients with type-II diabetes, this system can go awry when insulin's ability to stimulate the uptake of glucose from the blood is compromised. The result: high blood-sugar levels and energy-starved cells.

Yet mice have apparently forged a lifestyle without the benefits of CHC22: although the protein is found in many mammals, mice do not make it. The finding calls into question the use of mice to study some aspects of diabetes, researchers say.

"I certainly would not say that all the work that was done in mice is wrong, and I'm not ready to say we should throw out the mouse models," says Frances Brodsky, a cell biologist at the University of California, San Francisco, and lead author on the study. "I am ready to say that these models are missing a component."

Crucial difference

Brodksy and her colleagues found that reducing CHC22 expression in human-cell cultures using a technique called RNA interference led to a reduction in the number of glucose transporters and the loss of all glucose transporter storage vesicles.

Yet when the team expressed human CHC22 in transgenic mice, the mice developed some hallmarks of diabetes, including high blood sugar and reduced responses to insulin.

The results suggest that the introduction of CHC22 may have disturbed normal sugar responses in these mice. One possibility, says Brodsky, is that mice may not need CHC22 because they do not rely as much on muscle cells to clear glucose from the blood. In humans, skeletal muscle cells do up to 90% of this job, but in mice, the liver does the majority of the work.


The findings highlight the important, but often overlooked, role of vesicle trafficking in diabetes, says Jonathan Bogan, a cell biologist of the Yale University School of Medicine in New Haven, Connecticut. "It's an important pathway to figure out, because it does look like there's some disruption of this pathway in diabetics," he says. Brodsky notes that several genomewide scans in search of genetic changes linked to diabetes have turned up a few hits in the same region of the genome as the gene that encodes CHC22.

The absence of the protein in mice does not invalidate previous work using mouse models, agrees Bogan, but could affect drug development. "The question is, if you're a drug company, do you have to use a human cell line to test your drug on?" he says. "If your target is some upstream molecule involved in insulin signalling, it may not matter. But if it's more downstream, then it could be important." 

  • References

    1. Vassilopoulos, S. et al. Science 324, 1192–1196 (2009).
Commenting is now closed.