December 7, 2009

Rodent Research Blocks Progress in Treating Chronic Disease

A study described in last week's Science Daily points to a problem that has been troubling me about diabetes research: the way that research relies on rodent models that are built on false assumptions about the human diseases they are supposed to model.

You can read about the study describing severe problems with the mouse model for Muscular Dystrophy here:

Science Daily: Mice Holding Back Muscular Dystrophy Research?

It explains that
... two major features of a key DMD [Duchenne muscular dystrophy] gene are present in most mammals, including humans, but are specifically absent in mice and rats, calling into question the use of the mouse as the principal model animal for studying DMD.
This is bad news for people who have been waiting for scientists to find treatments for DMD but this kind of problem is not limited to research on DMD.

A huge amount of "diabetes" research is conducted in db/db mice, ob/ob mice, the New Zealand Obese (NZO) mouse, and Goto-Kakisaki Wistar Rats. A small group of highly respected scientists have built their careers around their knowledge of these particular mice and rats and sent their students out into the academic world where they have spread the use of these rodent models.

The problem is that these rodents became models for human Type 2 diabetes because of their phenotype--i.e. symptoms--matched what researchers 25 years ago thought was the phenotype of Type 2 human diabetes.

They are very fat. They become fat eating high fat diets. The blood sugar of these rodents worsens when they eat fat. Low carb diets are very damaging to these rodent models. In short what we have here is a model of outdated assumptions about Type 2 Diabetes that have not been proven true in humans with the disorder.

Humans with Type 2 diabetes may be fat or of normal weight. Most become fat eating high carbohydrate diets. Human blood sugars usually improve when they eat low carb diets.

The reason that the rodent model does such a poor job matching human experience becomes clear when we look at the genotype of these rodent diabetes models. Their broken genes are not the broken genes found in humans with Type 2 Diabetes.

The ob/ob mouse, for example, is morbidly obese as are many humans with Type 2 diabetes. However, the ob/ob mouse is obese because it has a genetic flaw that keeps it from producing leptin. This makes this mouse catastrophically hungry. But it turns out that after 6 years of hunting for leptin deficient humans, researchers found a grand total of three human families in the entire world who have the same genetic flaw as the ob mouse. Everyone else on earth who is very fat has some other genetic issue causing their morbid obesity.

You would think this discovery would have put an end to the use of the ob/ob mouse in diabetes related research, but you would be wrong. The scientists who have built their careers on their expertise in using this one mouse don't give up easily.

A Google Scholar search for studies conducted after 2005 where ob/ob mice are used as a model for diabetes turns up hundreds of studies.

The db/db mouse is another popular mouse model for diabetes. It is described thus:
"Diabetes (db) is an autosomal recessive mutation located in the midportion of mouse chromosome 4 that results in profound obesity with hyperphagia [compulsive overeating], increased metabolic efficiency, and insulin resistance.
The only problem with this model is that the genes that make the db/db mouse seem to resemble a diabetic human are NOT the genes that have been found in diabetic humans.

This is the problem with all the rodent models of Type 2 diabetes. No one questions that these rodents might have insulin resistance or that they might overeat. But over the past decade the dramatic drop in price of genetic screening techniques has come up with a long list of the most common genes associated with human diabetes. In European populations the list includes genes like TCF7L2, HNF4-a, PTPN, SHIP2, ENPP1, PPARG, FTO, KCNJ11, NOTCh3, WFS1, CDKAL1, IGF2BP2, SLC30A8, JAZF1, and HHEX. Failure of any of these genes can produce diabetes, but the exact phenotype of that diabetes will vary because what is broken is different.

People from non-European ethnic groups have been found to have entirely different sets of diabetic genes, like the UCP2 polymorphism found in Pima Indians and the three Calpain-10 gene polymorphisms found associated with diabetes in Mexicans. As a result the "natural history" of their "Type 2" diabetes is quite different than that of Europeans and they develop different patterns of complications than do people of European backgrounds.

"Diabetic" mice and rats do not have any of these human genetic flaws but other, rodenty gene flaws that give them rodent "diabetes." This does not mean it isn't useful to explore their physiologies, only that we must always keep in mind that they are different from those of humans diagnosed with Type 2 Diabetes--just as individuals diagnosed with "Type 2 Diabetes" can differ dramatically from each other.

The chances that any finding in mice might cross over and apply to humans is very slim. When these studies are reported to the media they should not be reported, as they are now, as if they applied to humans.

Beyond this issue, though, lies the problem pinpointed by the Muscular Dystrophy researchers. The mouse is very different from the human being, physiologically and genetically, in ways that researchers have not probed because, face it, the people who know the most about how to study mice are precisely those whose careers come to a crashing end if mouse research ever goes out of fashion.

We don't know enough, yet, about the genetic differences between rodents and people to know how the genes expressed in the livers of mice differ from those expressed in humans. Until we do, it is pointless to study what pattern of macronutrient intake causes fatty liver in mice. We don't know enough about the many genes involved in the regulation of glucose in the pancreas or brains of mice either. The Muscular Dystrophy research discovered that the genes expressed in the brains of people were significantly different from those of mice.

Glucose regulation in any species involves a complex symphony of interacting feedback loops involving gene expression in the brain, liver, pancreas, muscle and gut. The chances are very high that, because of significant differences between human organs and rodent organs, the findings drawn from rodent research--which form a much more significant part of "what everyone knows about diabetes" in the medical world have led us into dangerously wrong blind alleys.

The value of the low fat/high carb diet for people with diabetes was supported by immense amounts of rodent research. Unfortunately, what that research proved was that fat is a problem for rodents whose diabetes is caused by mutations unrelated to those found in humans which disorder metabolic processes quite different from those that occur in humans.

It is possible that there are proteins in humans that are significantly involved in the production of human forms of Type 2 diabetes that mice don't have or that they don't use the way humans use them. By concentrating so heavily on rodent research, researchers may be missing these vital proteins and the impact of the genes that code for them.

But mice are cheap, live brief lifespans, and can be "sacrificed" without causing massive protests. The students of the original mouse expert researchers have grown into a huge grant-gobbling research establishment that thrives on doing mouse research. It isn't going to go away soon.

Another study published just last week shows that another entrenched group is hoping to get their share of the lucrative diabetes research dollar. Lo and behold, they've created a Fruit Fly Model of Type 1 Diabetes. You can read about it here:

Science Daily: Diabetic Flies: Fruit Fly Model Helps Unravel Genetics of Human Diabetes

As described in the article:
"These mutant flies show symptoms that look very similar to human diabetes," explains Dr. Pick. "They have the hallmark characteristic which is elevated blood sugar levels. They are also lethargic and appear to be breaking down their fat tissue to get energy, even while they are eating -- a situation in which normal animals would be storing fat, not breaking it down."
I have friends who are highly distinguished fruit fly gene researchers and I have immense respect for the work they do, but they are doing basic science looking, for example, at the genes that produce neurons to better understand what a neuron really is.

I am left scratching my head to understand how a fly that lacks blood vessels, a liver, or pancreas can be a "model" for human autoimmune diabetes. A fly with high blood sugar is not a fly with "Type 1 Diabetes." And by the same token, a fat fly that eats too much may be an interesting fly that may teach us something new about fly physiology, but it is not a "Type 2 diabetic" fly.

Diabetes is a set of symptoms produced as the end result of failures at various points in a complex organism with its own evolutionary history. It is not a single condition. It is not caused by a single failure of one gene that can be explored at the level of basic science.

Until doctors and the public understand this better, we will continue to be subjected to "medical truths" derived entirely from the same kind of rodent research that has already set back the treatment of the vague collection of symptoms called "Type 2 Diabetes" for decades.



Scott said...

You forgot another rodent model, the Zucker (fatty) Rat being used in some obesity studies, but increasingly in some rodent studies in type 2 diabetes research. There are also problems in using the NOD mouse in type 1 research. Regardless, there are some very real problems about rodent animal models in all medical research that scientists are slowly beginning to acknowledge are unlikely to translate into human success. For example, researchers have already proven that the cytoarchitechture of rodent beta cells is quite different from human beta cells, but I feel this really is likely the tip of the iceberg with all of this study of murine animal models. There are other problems as well. For example, rodents will eat whatever is put in front of them, while larger animal models are a bit different in that larger animal models more closely mirror human behavior in that larger animals don't necessarily eat when fed. Unfortunately, this really comes down to money. Rodent models are cheaper than larger animal models, but again, the issue is that larger animal models, especially primates, have differences that researchers are only beginning to acknowledge, yet far too many hours of study are dedicated to animal models. We need to ask why is science so slow to change, and what will change these well-established behaviors?

Gretchen said...

I think the problem is not that researchers use rodents, or fruit flies, or flatworms, but that the science press doesn't understand what that means.

The science press disseminates a lot of incorrect information, not just about rodent results.

The scientists themselves also exaggerate the importance of their work to try to boost their reputations so they can get more grants.

But we can't do experiments in humans with no reason. We can't biopsy 10,000 humans in the hopes of finding something interesting. We can't do gene transfers in humans except in exceptional cases. We certainly can't do gene knockout experiments in humans!

And important genes seem to be conserved throughout the animal kingdom, which is why people are still working with Caenorhabditis elegans, which uses insulin/IGF-1 homologs to control growth and reproduction.

The way the system should work is that you try something with rodents or whatever (primates are now too expensive, and animal rights people put pressure on people not to use them at all). If you find a result, you then see if you can reproduce it in humans.

Think of leptin. Yes, there are only a handful of humans who lack leptin. But there are apparently zillions who have leptin resistance, and this may turn out to be more important than insulin resistance.

Without the mice, the leptin would never have been discovered.

Yes, mice are different. But until something better comes along (rabbits and dogs are also quite different), we need the mouse research.

We just need to emphasize the differences.

Jenny said...

Actually, I've tracked several of the stories posted on Science Daily to their source and discovered that SD just reproduces, word for word, the press releases that are sent out by the organizations for whom the researchers work.

So in the case of stories cited in SD the problem is NOT the media spin, it's what was in the press release that the researchers allowed to go out.

Steve Parker, M.D. said...

I understand your frustration, Jenny.

The December 3 issue of New England Journal of Medicine has an article - "A Look at the Low-Carb Diet" - that sounded interesting and educational.

Turns out it's simply about research done in mice. Not just your average mouse either, but a strain deficient in apolipoprotein E, which are particularly susceptible to atherosclerosis when fed a high-fat
Western diet.

The whole point seems to have been to cast aspersions on high-fat low-carb diets eaten by humans. To scare us into thinking that WE'LL develop atherosclerosis.


Jenny said...

I saw that idiotic mouse "low carb" study but there is so much extremely bad research of that nature I rarely discuss it unless there's something particularly egregious about it that leads to discussing a more general principle.

One does wish one could breed a TE/TE mouse that provides a model for tenured faculty. This mouse can only learn to run one maze and is very slow to adapt to changes in its environment, though it defends its territory very aggressively.

Stargazey said...

Jenny, this is completely off-topic, but I need to thank you. Last year you recommended the Escali digital scale as a possible Christmas gift. I bought one and it has been a marvellous addition to my kitchen. (It also helps me figure out proper postage on overweight letters, LOL!) Thanks for the suggestion!

Gretchen said...

I agree that SD simply reproduces press releases, as I've done what you did: tracked them and found the same story on multiple sites.

But I still blame SD and the other science news sources like Eurekalert.

When I worked for a small daily newspaper, we did NOT print press releases as they came in. We edited out the garbage and if necessary checked the facts.

We understood that press releases are a game. The writers try to get as much "free advertising" as possible and we tried to eliminate as much as we could.

So many diabetes press releases go like this:

1. Amazing breakthrough announced, citing a new drug or a new gene product or whatever without actually naming it.

2. List of scientists who worked on the research.

3. Something about the research institution.

4. Description of type 1 vs type 2 diabetes.

5. What the research actually showed revealed only much later.

Jenny said...


Your newspaper had a staff. SD is probably run by a very few people who input hundreds of press releases a week into their database.

SD provides a very useful service to those of us who follow health news but it would take a huge investment in staff to vet those hundreds of studies.

Tthe site is not likely to be earning the kinds of revenue that would make that possible.

It often takes me a couple hours to track down the information I need to be able to comment intelligently on a single study.

Doing that for the hundreds of studies featured weekly would be impossible without a huge budget and paid staffers.

It would be nice if SD made it clearer that they are passing through unedited press releases rather than reporting on stories.

Drs. Cynthia and David said...

Similar problems exist using mice and rats as models for humans in exercise physiology too. But rodents are inappropriate as human models because they have almost all fast twitch muscle, so they don't burn much fat for energy, and they store very little glycogen in their muscles (instead relying on their liver stores), so they burn through what little muscle glycogen they have very quickly. Relying on rodent studies, researchers concluded that humans must carb load because they might run out of glycogen as quickly as rodents do, and you can't have that! So the whole field of sports nutrition advises constant and immediate carb replenishment after exercise. Things are starting to change, but only slowly.


Jonah said...

I think rats are inappropriate for studies of anything having to do with weight because rats typically get heavier throughout their lifespans while humans tend to lose weight after middle age- which means that making rats skinny prolongs their lives in a way NOT analagous to humans. Humans need fat in ways that rats (and especially lab rats) don't.
I also think that animal experimentation in general has a problem when it comes to modeling the destruction of beta cells because of the way that the islet cells of humans are structured differently. Rats and mice and rabbits and dogs and cats and cows have differentiated islet cells- the beta cells all together, the alpha cells altogether- whereas us humans have them mixed, which I strongly suspect has an impact on what happens when there's an antibody attack on any part of the islet cells.

Lucas Tafur said...

I dont know why they still use mouse models instead of using guinea pigs.

Jenny said...

Lucas, When the only tool you have is a hammer, everything looks like a nail.

There are quite a few senior scientists who have built whole careers on work with specific breeds of mice and they've created a self-replicating empire with their students. Take away the mice and they probably have no careers.