Scientists and others have been opening their homes to research animals after the studies conclude, with legislation in some states now mandating adoption.
Researchers should pay closer attention to the diets they use to study obesity in mice, experts advise.
November 23, 2015|
WIKIMEDIA, SCHUMINWEBHigh-fat diets are ubiquitous in rodent research. Scientists feed mice high-fat diets to study the impact of nutrition on diabetes, cancer, and heart disease, and to better understand obesity itself. Companies including Purina, BioServ, Research Diets, Envigo (formerly Harlan Teklad), and CLEA Japan sell standardized high-fat and control diets, making it simple to include high-fat diets in study designs.
But less often do researchers stop to ask what’s actually in the high-fat and control diets they are using, and how rodent nutritional research corresponds to human nutritional research. In recent years, for instance, studies have indicated that fat is far from the lone culprit in human weight gain, and that carbohydrates play a hefty role. “Mice are a good model for human obesity . . . but it seems to me that, for diet, the mouse literature has gone off track,” said Craig Warden, who studies the genetics of body fat accumulation at the University of California, Davis.
What’s in a high-fat diet?
When high-fat mouse diets originated, feeding animals fat seemed like an obvious way to make them gain weight. “What likely happened is that we assumed back in the day that high-fat, high-calorie foods were the problem for human obesity,” said Matthew Ricci, science director of Research Diets.
Upping fat intake worked. When certain mouse strains are fed diets high in fat, they gain weight rapidly. But high-fat diets still contain other components, and Warden explained that this could also influence the health of animals subsisting on them. “The typical diet most commonly used by investigators in the U.S. is a diet that is both high fat and high sugar,” he said.
To Ricci’s mind, worries about the contents of rodent diets can be assuaged by performing properly controlled studies. “Is a high-fat defined diet we’re feeding the mouse really comparable to what a human is eating?” he asked. “Of course not. It’s completely different. However, it’s very controlled. This is why we do animal studies. We control the heck out of them.”
Companies selling high-fat mouse diets often provide “purified” control diets that differ from high-fat diets only in their fat and carbohydrate content. These diets allow researchers to feed control and experimental mice foods with identical caloric content, replacing lost fat with increased sugar in control diets.
Despite the availability of these purified control diets, researchers often simply compare mice fed a high-fat diet to control mice fed ordinary chow. “When you’re going from a purified diet to chow, you’ve just changed absolutely everything,” said Ricci. “In that chow, there’s corn and wheat and soy and oats and animal fat from sometimes sources you don’t know; there are phytoestrogens, heavy metals; there could be pesticides.”
Chow diets may be more or less healthy than high-fat diets in myriad ways. They contain relatively high fiber and carbohydrates and relatively low fat. Carbohydrates are more likely to come from beans and grains and less likely to come from pure sucrose.
Ricci recalled one of his customers trying to determine how a high-fat diet affects colon health. The researchers fed some of their mice a high-fat diet and some chow. The chow-fed animals had large colons, while the colons of the animals on the high-fat diet shrunk. But when the researchers went back and tested the effects of feeding the animals a defined control diet that only differed in relative fat-to-carbohydrate ratio, the colons of all of the animals shrunk. Eventually, they showed that it was not fat, but lack of soluble fiber, that appeared to be responsible for the shrinking colons. Both the high-fat diet and the purified control diet were low in soluble fiber.
Control is key
Warden pointed out that proper controls are lacking in an alarming number of studies. In 2007, he and his colleague Janis Fisler, also of UC Davis, surveyed of the literature and found that, of 35 papers that involved feeding mice a high-fat diet, only five used control diets that were identical save for their fat-versus-carbohydrate ratio.
Warden is currently working on a new survey. He said that more researchers are using proper control diets, but that he still finds many papers comparing high-fat diets to chow.
But Justin Grobe, who studies metabolism at the University of Iowa Carver College of Medicine, pointed out that even purified control diets can generate confusing results. “I’m loading up the control with sugar,” he said. “When I get my results back, am I actually studying a control for a high-fat diet or am I actually studying or what’s happening on a high-sugar diet?”
At the very least, researchers must clearly state what they have fed their animals. “I routinely read the methods sections of papers labeled as ‘high-fat diet,’” Warden wrote in an email. “More often than not, the authors do not provide information about diet composition in the paper and, even worse, do not provide enough information [for me] to determine diet composition by myself.”
Even if researchers performed perfectly controlled studies on rodents, they would still need to ask whether rodents respond to food and dietary changes the same way that humans do.
There is, in fact, some evidence that rodents differ from humans in their responses to carbohydrates and fat. In one study, Richard Feinman, a cell biologist at SUNY Downstate Medical Center in Brooklyn, New York, and his colleagues fed Black 6 mice a no-carbohydrate diet. Protein contributed a fairly standard 20 percent of calories to the diet, while fat contributed a whopping 80 percent of calories. Even though the two groups of mice consumed similar numbers of calories, the mice on this no-carb, high-fat diet gained more weight than mice fed on standard chow, which contained 58 percent carbohydrates, 13.5 percent fat, and 28.5 percent protein.
In humans, meanwhile, many low-carb diets seem to lead to weight loss. Mice generally subsist on a very low-fat diet, and high-fat diets may affect the animals differently than humans. “A laboratory mouse lives on carbohydrates,” says Feinman. “Their ‘normal’ diet is not like a normal diet for humans.”
Clarification (November 24): The ninth paragraph of this article has been updated to reflect that both the high-fat diet and the purified control diet examined in the study were low in soluble fiber. It was previously written that both diets were low in fiber.
November 23, 2015
will just replacing Fat calories with sugar calories in the control diet be sufficient or is there more to it?
November 24, 2015
I'd like to help answer this but I'm not quite sure what you mean. When designing a low fat control diet, usually the fat calories that are removed are replaced by carbohydrate calories, though this can be in the form corn starch (it doesn't have to be sucrose). Since low fat diets that are high in sucrose can induce metabolic derangements in rodents, sucrose is generally not something that's ideal to have in a low fat control diet.
November 24, 2015
Thank you, Ms. Yandell, and Dr. Ricci, for this very important article highlighting the serious flaws that are so widespread in animal nutrition studies. It is frustrating to read headlines proclaiming that high-fat diets cause health problems in mice, only to discover that the methodology behind the headlines renders the results completely meaningless.
Just last week I wrote a critique of a study that suffered from this very same flaw: http://www.diagnosisdiet.com/high-fat-diets-and-depression/.
The study, published in the British Journal of Pharmacology, was entitled High fat diet-induced metabolic disorders impairs serotonergic function and anxiety-like behaviours in mice. It concluded that high-fat diets in mice cause type 2 diabetes, depression, and anxiety.
It turned out that the high-fat diet (made by Research Diets) had been compared to a low-fat chow from another manufacturer (SAFE Diets), and that the carbohydrate quality, as well as almost every single ingredient in the two diets, was completely different, just as Dr. Ricci points out. Given the multiple variables, it was impossible to know what was responsible for the results, although my guess is that it was the high amount of refined carbohydrate and sugar in the high-fat diet.
December 7, 2015
Thanks Dr. Ede and I just read your blog post on high fat diets and depression. I wish more manuscript reviewers would pay attention to the details of the diets as you did, since poor diet controls really only invalidate the study. The answser is a continuous education effort, and I applaud you for that.
December 7, 2015
One other comment I'd like to make - the final study cited in Kate's article (the one by Feinman) unfortunately compares data from mice fed a high fat (no carbohydrate) purified diet to those fed a low fat, grain-based chow diet. So as Craig Warden and I pointed out earlier in the article, this really makes interpreting the data impossible. The diets are just too different from each other.
So the next time you are reading an animal study with an experimental and a control diet, remember that your confidence in their conclusions will be based in part on how well matched the diets were.
May 3, 2016
Hello, thank you for this article. It addresses most of the things I wondered about while reading about diet induced obesity and gut inflammation in mice.
I was wondering one more thing though; do rodents have the same inflammation pathways as humans in the sence that linolic acid and arachidonic acid mostly trigger the formation of pro-inflammatory prostaglandines & leucotrienes and the omega 3 family the anti-inflammatory prostaglandines and leucotrienes and do the diverse saturated fats have similar effects on inflammation as in humans as well? I hardly ever see the type of fat that was used specified in the articles. So either it is not important or it can be added to the list of unfortunate flaws.