Just when you think you know everything there is to know about a model you run tests on, out comes some whippersnapper to upset the scientific apple cart. It makes you wonder how this could have gone on for so long, but as the article says, the researchers became “ingrained” to a certain way of doing research, which I think is also a problem with climate science. – Anthony
Tests in Mice Misled Researchers on 3 Diseases, Study Says
By GINA KOLATA, NYT
For decades, mice have been the species of choice in the study of human diseases. But now, researchers report stunning evidence that the mouse model has been totally misleading for at least three major killers — sepsis, burns and trauma. As a result, years and billions of dollars have been wasted following false leads, they say.
The study does not mean that mice are useless models for all human diseases. But, its authors said, it does raise troubling questions about diseases like the ones in the study that involve the immune system, including cancer and heart disease.
“Our article raises at least the possibility that a parallel situation may be present,” said Dr. H. Shaw Warren, a sepsis researcher at Massachusetts General Hospital and a lead author of the new study.
The paper, published Monday in the Proceedings of the National Academy of Sciences, helps explain why every one of nearly 150 drugs tested at huge expense in patients with sepsis has failed. The drug tests all were based on studies in mice. And mice, it turns out, have a disease that looks like sepsis in humans, but is very different from the human disease.
…
“They were so used to doing mouse studies that they thought that was how you validate things,” he said. “They are so ingrained in trying to cure mice that they forget we are trying to cure humans.”
“That started us thinking,” he continued. “Is it the same in the mouse or not?”
The group decided to look, expecting to find some similarities. But when the data were analyzed, there were none at all.
“We were kind of blown away,” Dr. Davis said.
Full story here
h/t to Indur Goklany
Discover more from Watts Up With That?
Subscribe to get the latest posts sent to your email.
Mice !! How about Rats – the best developed animal in terms of resistance to the widest range of substances, able to survive in the most hostile of climates and the most hostile of environmental loads.
Ever wondered why they are the preferred animal for manufacturers testing the toxicity of chemicals and then extrapolating the results to humans prior to licensing ?? Has always seemed to me that it is precisely because rats are so resistant that they have been so useful as test subjects to ‘big pharma’.
Brilliant responses, DocMartyn and rgb; I just want to add that not only do we realise that “Mice are not people”, but we also know that “People are not people”. For example, some people’s antibodies are broadly cross-reactive against multiple strains of flu, while others are not sufficiently reactive against the virus they have just been infected with. Some of us die from things others have for breakfast.
Sorry previous post should have been Re Doug Huffman.
This article got me thinking about what evolutionary advantage there might be in being sensitive to bacteria? Like, why don’t we have mouse type resistance to bacteria? (“Mice need a million times more bacteria in their blood than what would kill a person”)
Perhaps it’s helped move our populations towards certain environments, out of the slop to richer paddocks.
DaveA says:
> This article got me thinking about what evolutionary advantage there might be in being sensitive to bacteria?
Probably none for us, but for the symbiotic bacteria we carry, the evolutionary advantage is presumably that they can jump ship when they sense that their host is in trouble. It may just be that by killing the host quickly and in a particular way, their chances of being acquired by the next host are better than starving to death with the present host.
> Like, why don’t we have mouse type resistance to bacteria? (“Mice need a million times more bacteria in their blood than what would kill a person”)
I would question whether that’s really true, and for what kind of bacteria. The most common enterobacteria that we and mice share, E. coli and P. aeruginosa, can kill us equally well, and they don’t even need to be in the blood to achieve that; they can simply let the blood out.
Josualdo says:February 12, 2013 at 4:42 am That mice and men are not the same was known.
Precisely, the concept is of mice and men, not of mice or men.
Biogerontologist says:
February 12, 2013 at 6:52 am
[re the question whether or not cell lifespan is the key question of aging: ]
“This is still under debate in the ageing field. The causes of human ageing are not understood at this time.”
—
This is a strange statement.
Of course, with a complex system like a living organism, we never completely understand anything, but we can have some useful, if limited, degree of understanding. Otherwise, there would be no vaccination, no drugs, no vascular surgery.
So, what could a partial understanding of aging look like?
Let us first clarify the concept of “cell lifespan”. The way it was used here relates to a cell’s generational limit — the number of times a cell is permitted to divide; past this limit, the offspring cells will no longer be viable.
Another relevant, more literal interpretion of the term “cell lifespan” relates to the maximum time a cell may survive without dividing. Why is this relevant? Because it applies to most cells in the brain, which are already formed at the time of birth, or shortly thereafter, and then have to last us through our entire lifetime.
It is important to understand that the number of cell divisions that go on in our organs is vastly different. In the brain, there isn’t much going on at all, whereas for example the bone marrow produces about 500 kilograms each of red and white blood cells over our lifetime, and potentially quite a bit more in long-serving blood donors. Nevertheless, most people do not die because they run out of blood cells, or skin cells, or intestinal epithelial cells, which also regenerate at a very fast clip. So, clearly, the cell generational limit does not as such limit life expectancy.
It is also interesting to note that, as biogerontologist states, a generational limit applies in human cells but not in cells from mice. Nevertheless, mice live only something like 2 years or so, even if not eaten by a cat. So there.
Now, if it is not the generational limit that limits life expectancy, what is?
What do people actually die from? Which organs do give out most commonly? It is interesting to note that the brain is much higher on this list than the bone marrow. What limits the life span of a brain cell? Broadly speaking, a lifetime of collecting waste. Sure, there are mechanisms for removing debris, but they work imperfectly, and in an aging brain such as mine, a lot of accumulated debris will always be observable. In a rapidly dividing tissue, intracellular waste will be diluted with each successive cell division; in long-lived cells, such dilution does not occur. Therefore, the tissues with low regeneration rates are the hard ones to optimize for longevity, not the ones with high turnover.
Now, back once more to the question why humans, but not mice, developed a strict generational limit for cell divsion. This generational limit is one of many safeguards against uncontrolled cell growth, aka tumors. A mouse will most likely be eaten, or otherwise die of dementia before it gets a tumor. Man has a generational limit to help him grow old; mouse hasn’t, because its chance of growing old is negligible anyway.
“So bottom line. We know mice aren’t human. We use them as a mammalian model. For somethings they are fantastic, for other things less so. ”
The problem is that they’re being used as a model without trying to verify that they’re a GOOD model prior to proceeding with the research. It’s not that mice are worthless, it’s that scientists got lazy and assumed that the mice would be a good model for what they’re researching rather than putting in the effort to find out before wasting everyone’s time and money.
DaveA says:
February 12, 2013 at 7:14 am
This article got me thinking about what evolutionary advantage there might be in being sensitive to bacteria? Like, why don’t we have mouse type resistance to bacteria? (“Mice need a million times more bacteria in their blood than what would kill a person”)
—
We actually do have competitive resistance to bacteria. What kills us in septicemia is the very power of our immune system. In most cases, however, the vigilance of our immune system will prevent dangerous numbers of bacteria to accumulate in our system.
The real point to this study is not, as many have said, that “mice are not men” — no one ever thought they were. But they did think that mice acted as valid research proxy for men in many areas. Careers were spent (along with lots of money) using mice as proxies (they call them models, but that will stir up too much confusion here) in sepsis, burn and trauma research.
The use of mice as proxy for this research apparently had not been validated — which is what this study set out to do. It was a negative result — the proxy is NOT valid — results in these areas of mice do not apply to men. They knew this — the drugs developed on mice for sepsis (150 of them!) failed to work in humans. Question: Why wasn’t this study done after the first five or six drugs didn’t work?
This will stir up a lot of trouble in medical research. They should have been validating the use of mice as a proxy for each specific use as a separate study — like validating models and methods in Climate Research…It simply MUST be done or years of work and billions of dollars will be wasted following false leads.
@Kip Hansen: you say, “They should have been validating the use of mice as a proxy for each specific use as a separate study”
They should, no doubt. The problem is in what “they” were, or were not. For drug companies, following false leads is normal business. They call it “screening”. They are using their own funds for that, and they have discovered stuff in the past by random trial and error. They also have the money to do the kind of validation you are talking about, and I bet they do, as a rule.
For researchers working off public grants, validating someone else’s theory or experimental result has become all but impossible.
“Boss, can I spend a couple days validating the results that German lab published in Nature last month?”
“Yes, please do, by all means. We must be sure we’re doing the right thing.”
“Boss, can I send a P.O. for these supplies and equipment and spend 3-4 months validating the results that German lab published in Nature last month?”
“You’re kidding me? My grant officer is not going to authorise it. And these results have been peer-reviewed and double-checked, and cross-validated. We can’t waste our time like that.”
http://www.smbc-comics.com/index.php?db=comics&id=2881 (about a week ago)
This bring to mind “The Invisible Gorilla” study. This was talked about on NPR just yesterday. A test group was asked to watch a video. It had two teams of children. One group in white t-shirts, the other in black. The observer test subjects were asked to count how many times the team in white t-shirts passed a ball. During the video a gorilla comes out, gestures at the camera, then walks off screen. After the test the observers were asked if they saw the gorilla. 87% did not. They had become so focused on seeing what they expected that the outrageous passed them by completely unnoticed.
Gene Selkov says:
February 12, 2013 at 7:03 am
Brilliant responses, DocMartyn and rgb; I just want to add that not only do we realise that “Mice are not people”, but we also know that “People are not people”. For example, some people’s antibodies are broadly cross-reactive against multiple strains of flu, while others are not sufficiently reactive against the virus they have just been infected with. Some of us die from things others have for breakfast.
This has always been my concern. The medical trials bring together a group that seems to have been chosen for a political poll – similar lifestyles and occupations. Then they carry out a climatologist level statistical analysis of the effects of diet, of drugs, of natural things such as sunshine. All the while they totally disregard genotypes which they must realize will affect the result. If you trial sunscreens and 75% of the group are dark skinned and 25% are fair skinned red heads it will not be a surprise to have 25% outliers in the trial – but this will be hidden by primitive statistics.
So yes as Gene Selkov says “People are not people” in medical trials.
“kcrucible says:
The problem is that they’re being used as a model without trying to verify that they’re a GOOD model prior to proceeding with the research. It’s not that mice are worthless, it’s that scientists got lazy and assumed that the mice would be a good model for what they’re researching rather than putting in the effort to find out before wasting everyone’s time and money.”
You cannot do work on sepsis in humans: that’s it. If you are a clinical scientist you may be allowed to observe the team who are treating the patient, but you are not going to get any research done.
Same with burns and trauma. Even Ph.D. students who volunteer for everything are not going to have a 20 pound weight dropped 12 inches onto their skull or have a iron on maximum temperature applied to their back for 30 seconds.
I have friends who work on sepsis, most of them use rats. If any of you want to volunteer to have the scientists inject you in you abdominal cavity with excreta, follow progression and then remove you organs, just drop an email address.
Yes, biomedical research is one hell of a mess.
It is not just mice. We have plant prejudice to deal with.
CB2
Medical Marijuana prohibition is a crime against humanity and a violation of the religious precept – heal the sick.
Pass it on.
DocMartyn, I think that you are huffing and puffing a bit here. A point well made further up in the comments is that observation led to the conclusion that sepsis is not simply a by-product of wounds. Observation of non-wound related sepsis does not require people to subject themselves to trauma. People with this kind of sepsis have to be treated, and not all of them will be treated the same way, for all sorts of reasons. Detailed observation of those conditions and treatment outcomes is a perfectly valid way of gathering data.
Indeed, it is almost certainly a better way of gathering data than doing nasty things to meeces.
Why is that a problem for you? In medical research, all sorts of things are out of bounds for double-blind tests for ethical reasons. We just have to rely on observation and analysis. It’s not perfect, but that’s just how it is.
Empirical research has a lot going for it, you know.
rgbatduke says:
February 12, 2013 at 6:31 am
“It is almost as satisfying to shoot down a widespread but incorrect belief as it is to find something positive, and in the long run, our brains work by BOTH building neural connections AND brutally pruning back the ones that don’t work. Without the pruning, you get a teenage adolescent (literally — proliferation has happened but not the pruning) — and as those of us who have raised or been them well know, teenagers have a sort of brilliant and self-destructive madness until that pruning takes place.”
It is more satisfying. Finding something positive involves making a promise to keep working even without reward.
My oh my. How wonderful it would be if every adult recognized that he/she had been a teenager and, even better, raised a teenager.
Powerful metaphor for the role of criticism in science.
johanna, OK lets play it your way.
Work out how to organize a double bind trial to treat sepsis in humans, without an animal model. First off you need to recruit hospitals, and so need a lot of clinicians to sign up.
You need written procedures for the ER staff to recognize a patient falling with in you studies parameters, then the fun starts,begin by getting informed consent from patients coming into the ER .
Now treatment. What you going to do? Come on tell me.
johanna says:
February 12, 2013 at 1:55 pm
“DocMartyn, I think that you are huffing and puffing a bit here. A point well made further up in the comments is that observation led to the conclusion that sepsis is not simply a by-product of wounds. Observation of non-wound related sepsis does not require people to subject themselves to trauma. People with this kind of sepsis have to be treated, and not all of them will be treated the same way, for all sorts of reasons. Detailed observation of those conditions and treatment outcomes is a perfectly valid way of gathering data.”
We are seeing too many cases of people in intensive care dying from an immune system that has turned against them. Physicians are scratching their heads. Research can and must be done.
Those who are chastising the assorted scientists about using the “murine”
(mouse) model should consider what other avenues these scientists could
use for testing.
Dogs? Cats? Spider or rhesus monkeys? Chimpanzees? All of these
animals get PETA and their ilk riled up more than the poor little rodents.
To use some of these other mammals is simply asking for a nuisance
lawsuit (or three) from an animal-rights group; a raid to ‘free’ the animals–
and destroy hundreds of thousands of dollars of lab equipment; ongoing
pickets and other harassment; ‘hacktivism’ against any of your computers
connected to the Internet; and assorted other unpleasantness.
Computer modeling? Please. We here on this Web site have frequently
derided climate computer models as being inadequate to the task. How
much more complicated is a living organism with multiple feedback
loops devoted to trying to maintain homeostasis!
So how about suggesting an alternative? DocMartyn and
rgbatduke put forth disclaimers up-thread stating
that the assorted biomedical researchers & doctors know that the murine
model is flawed–but the most practical game in town. Suggest a better alternative,
bearing in mind the cautions I voice above.
Asking someone to put up an alternative because your system doesn’t work is one of the oldest wheezes of the CAGW crowd. It is not up to me to develop a model for biomedical research. What I said was that observation of existing cases of non-wound related sepsis is a valuable source of data – and given the limitations of mice, quite possibly the best source of data we have.
I detect a defensive note in these querulous comments which we have heard before in other spheres of ‘science.’
When I worked for Hercules back in the ’80’s I spent some time in the company of one of our toxicologists regarding one of our products. He related how he had become a target of the professional muckraker, Jack Anderson. Seems the good doctor wrote a scathing objection to a negative report on on another of our products based on a mouse study. Jack thought the company was throwing around its ill-gotten weight, hammering at a poor, defenseless three-letter agency. The question was all about the delta on the incidence of a particular cancer in a particulare strain of mice, said agency finding a significant increase (exact figures escape me). Our toxicologist noticed that there was a significant time lapse between sacrificing and necropsy of the control group (earlier) and test group (later). Mortality from cancer for this particular strain of mice was 100%. In time, they would all die of this particular cancer, at a predictible rate. When you graphed the mortality vs time for the control group and test group against the known parameters for the strain, there was no difference!! The agency had “found” a difference by delaying the examination of the test group. Not that Jack bothered asking the good doctor before publishing.
So what, now we know the murine model for sepsis is flat out wrong, let’s continue using it??? Sure, I’ll give you money for that </sarc. Ye gods and little fishes!!
The murine model for sepsis is an adequate one. You can make them sick with Chron’s just like humans; they will haemorrhage like humans; their blood will be infected just like human, with the same kind of stuff; their immune response will be largely similar.
That does not mean it is a good model for drug testing. Drugs are very fiddly.
But, the only efficient “drug” that actually prevents sepsis (PEG lavage) in those situations where it is known to occur with a very high probability has been tested both in mice and in humans, with similar results. A model is only good as long as you understand what it does.
Unfortunately, there are no drugs to treat an ongoing sepsis. But mice are very helpful in learning about its causes.
Performing an intervention on a mouse model of a human disease tells you how the intervention works in that model.
Many of us use strains of mice with a much modified immune system to grow human tumors and to then treat those tumors. Now we know that the tumor growing in the mouse is different than the same tumor growing in a human. We then give chemotherapy and see what happens. Now we then assume that what happens in the mouse is ‘pretty much’ the same as in the human. It isn’t the same, but if the mouse liver turns to mush you know that isn’t going into a person.
What the hell do you people think we can do?
Seriously?
We DO NOT KNOW a better system to guess what is going to happen in people than a combination of tissue culture and animal experiments.
This is the bottom line. Animal experiments give us our best guess.
We are not lying to you, the mice models are not perfect. in some diseases they are damned good, other, suck. Anything to do with aging brain, no models work. Immune system, models can be iffy. Human brains and human immune systems are quite removed from other animals. Makes work hard, slow going and mostly wrong in a strict model=human sense.
So we know all this. I am not lying to you that the models are temperamental. Sometimes fantastic, sometimes completely crap.
We could work on primates instead, but the costs of research would be two orders of magnitude higher and the outcome only slightly improved.
However, i will tell you what will not work. Looking at sick people. Physicians have a duty of care and they have to threat their patients to the best of their ability; under threat of lawsuit. So they treat their patients. Different patients have different doctors and different preexisting conditions and present with a different cocktail of drugs for their hemorrhoids, premature balding, hormone replacement therapy, pain pills for soft tissue injury and copper bracelet.
Patients present at different ages, heights, ethnicity and different drug histories; give me mice that are almost clonal any day.
So don’t shot the pianist’s as they are doing their best.
WE KNOW MICE ARE MODELS; they are bound to be bad in some disease states.
In the case of gerontology and auto-immune diseases mice don’t mimic humans well. We don’t have any good aging models and even chimps have a different immune system to us. Chimps cost a small fortune and have to be looked after in retirement, and as the HIV/AIDS people will tell you, don’t mimic our immune system either.
@ur momisugly Doc (hopefully not mine) Martyn:
“However, i will tell you what will not work. Looking at sick people.”
Holy crap.