Diabetes ResearchOn this page go to:
Von Mering and Minkowski
The orthodox view, the current paradigm of the biomedical community is that the great advances made in the understanding and treatment of diabetes, particularly the insulin treatment that considerably increased duration and quality of life of diabetic patients from the 1920s onwards, not only were greatly aided, but indeed would not have been possible without animal experimentation.
This is exemplified by what the biologist Sir Peter Medawar said in "The Pissing Evile", in D. Pyke (ed.), The Threat and the Glory:
"So far as insulin is concerned, it was only by experimentation on dogs that it came to be learnt that removal of something manufactured by the pancreas caused diabetes... In the continuing debate between experimentalists and champions of the rights of animals, the discovery of insulin remains a shining example of the benefactions experimental animals have conferred upon man."
Doctor and vet Ray and Jean Greek, in their book on the scientific limitations of animal experimentation Sacred Cows and Golden Geese, call insulin "the animal experimenters' poster drug".
Let's look at the history of medical knowledge of diabetes, its treatment and the discovery of insulin.
There are two moments in the history of diabetes which conventional medical storiography considers as breakthroughs: they are associated with the names of von Mering and Minkowski, and Banting and Best.
Isaac Asimov, in A Short History of Biology, writes that the suspicion that pancreas was somehow linked with diabetes arose because in 1893 two German physiologists, Joseph von Mering and Oskar Minkowski, had removed the pancreas from dogs, observing the rapid onset of a severe form of diabetes. This is generally the conventional historical account.
In fact over a century earlier, in 1788, Thomas Cawley performed an autopsy on a patient who had died of diabetes and discovered stones and signs of tissue damage in the pancreas, which led him to be the first to suggest a connection between diabetes and the pancreas. He also described patients showing pancreatic lesions who developed diabetes. Subsequently many other physicians described similar degenerative and pathological changes in the pancreas of diabetics well into the 20th century.
After Cawley, there was a series of similar observations of clinical origin or based on autopsies, among which: in 1833 Bright found pancreatic cancer in a diabetic patient; in 1869 Langerhans discovered the Islands (or Islets) of Langerhams, the part of the pancreas whose beta-cells produce insulin which is affected in diabetes; in 1870 Bouchardat, who replaced the practice of giving extra sugar to diabetic patients, to compensate for sugar loss in the urine, with the then-revolutionary therapy of dietary change and exercise, observed the link between diabetes and pancreatic lesions; in 1895, Hansemann reviewed the literature and found many cases of diabetics showing pancreatic lesions, but nevertheless reached the conclusion, based on dog experiments, that diabetes and the pancreas have no association.
So, before Minkowski, it was already known that many diseases of the pancreas, including pancreatic cancer and an inflammation of the pancreas called pancreatitis, produce diabetic symptoms, and the association between diabetes and the pancreas was already established. Then why did von Mering and Minkowski depancreatization of dogs receive all the credit for this knowledge?
The reason why the biomedical community considers Minkowski's the decisive discovery is because it was the result of an experiment, i.e. a manipulation of nature to test a theory, rather than an observation.
The idea behind the preference of the experiment over the observation is that in an experiment, being something that the researcher creates rather than simply witnesses, the variables are more strictly controlled and so results should be more reliable.
The Bench Scientists Position
The bench scientists position is that nothing can replace experiments as ways to test hypotheses because in well-designed experiments, as opposed to observations, comparisons, clinical and epidemiological studies, the subjects are chosen to fit strict criteria, the environment is rigidly controlled, and so the control over the relevant variables is more stringent, indeed total, reaching a level of control over variables that is impossible to attain otherwise. Since we can't ethically perform experiments on humans, their position continues, that leaves us with no alternative but to experiment on nonhuman animals, the closest we can get. On the other hand, we can have in vitro experiments, but they cannot replace animal experimentation because we need to experiment on intact systems.
This position is well exemplified by the view expressed by medical researcher and pharmacologist William Paton in the book Man and Mouse:
"It was the introduction of deliberate intervention into the study of natural phenomena, as opposed to observation, which so inspired the investigators of the seventeenth century onwards, and made them talk of 'the way of experiment'. Galileo, Torricelli, Gilbert, Boyle, Hooke, and their peers showed what a wealth of new knowledge could be gained if, instead of just observing natural events, you set up your own test to make events happen under your own control, choosing materials, procedures, and comparisons."
Interestingly, Galileo, the creator of the experimental method, said that observations can do as well as experiments, if the latter are difficult to perform.
Here is an example of the way in which this way of thinking is applied to diabetes research:
"There was research into diabetes as early as the 18th Century. In 1788 by [sic] Thomas Cawley observed, that the pancreas was damaged in patients who had died from diabeates [sic]. However, at this time there was no way of telling whether this was cause or effect: he could equally well have noted changes in the kidney, eye, blood vessels or nerves of such patients. None of these is the cause of diabetes: they are all effects of the disease. In 1889 Joseph von Mering and Oskar Minkowski showed that removing the pancreas from the dog [sic] produced diabetes. This was the first demonstration that there was an anti-diabetic factor produced by the pancreas which enabled the body to use sugars in the blood properly."
There are several things about this statement which are incorrect or incomplete. First of all, Cawley's observation referred to was not a single one: observations of damaged pancreas found in diabetics autopsies were repeated during over a century. Second, at the time there was plenty of clinical evidence of patients with pancreas diseases, like pancreas cancer, or with pancreatic lesions who subsequently developed diabetes, so it was an easy inference that damaged pancreas was the cause, and not the effect, of diabetes.
In addition, the comment about changes in the kidney, eye, blood vessels or nerves of diabetic patients is not valid in that not all people with diabetes develop the same complications - some develop more than one, some none at all - whereas all diabetics have pancreas problems. Which explains why Cawley and other medical scientists of the time arrived at the right conclusion about the relationship of cause and effect between diseased pancreas and diabetes based solely on human autopsies and clinical observations.
Furthermore, if it is absolute certainty that we are after, I cannot see how removing the pancreas from a dog to produce diabetes can make us surer that a damaged pancreas causes diabetes in humans than all the weight of clinical evidence mentioned in the above paragraph. First, we have a different species; second, we have a completely different mechanism: in human patients diabetes is not caused by removal of the pancreas.
The notion that only experiments, whose character is that of an intervention rather than a simple, passive observation of "nature's experiments", can be relied on for a rigorous testing of hypotheses derives from physics and is transferred from there to biology through a school of thought called "reductionism". Scientists who believe in the primacy of experiments in biology may not be entirely aware that they are embracing a form of reductionism.
Reductionism is a 19th-century philosophy of biology, based on the belief that all life phenomena, processes and events can be eventually reduced to the laws of physics and chemistry. Reductionism is closely associated with classical determinism, the idea that there is no chance, that, given a certain cause, a certain effect will inevitably follow: its purest expression is the theory of Kant-Laplace, from the late 18th century, whose causality principle which says that, if we know all the initial conditions of a closed system, and we know the laws governing it, we can predict with perfect accuracy all the future events of the system.
Reductionists also believe that laws of nature are universal, i.e. not restricted in time and space and having no exception. Same cause, same effect. For example, all hydrogen molecules are the same, and they are the same in and outside the lab. So it is hardly surprising that numerous scientists embrace determinism, since deterministic, universal laws enable researchers to predict future events if they know the initial conditions.
In other words, scientists who profess that animal experiments cannot be replaced are saying, whether knowingly or not, that animals, including us, are like clocks, machines that can be eventually understood in purely physico-chemical terms, more complex than a clock but not qualitatively different.
The fact is that both reductionism and determinism are philosophies which have been amply surpassed by developments, not just in philosophy but even more in science.
Determinism, developed in the 1700s, was the philosophy behind the classical mechanics of Galileo and Newton. Modern physics, with its quantum mechanics, has rejected the principle of causality of Kant-Laplace when it introduced Heisenberg's uncertainty principle, which states that we cannot determine at the same time both the position and the velocity of an electron.
So, the illusion of a tight control on a system allowing an exact prediction of future events went out of window.
Even closer to home, biology, from the time in which the theory in support of animal experiments was formulated by the French physiologist Claude Bernard in the mid-1800s, has taken a completely different path. What happened to it? The theory of evolution by natural selection happened, which not coincidentally Claude Bernard always rejected because it was so much in contrast with his mechanistic ideas.
Reductionism is now showing itself to be less and less acceptable in view of the new biological theories and concepts, in contemporary evolutionary biology. Experimental physiology has distanced itself more and more from the rest of current biology. The current physiologists' paradigm does not grasp the full implications of evolution for the practice of physiology. This is not surprising, since the founder of that paradigm, Claude Bernard, rejected the theory of evolution. This paradigm is still today associated with biological reductionism, which is not accepted by most evolutionary biologists any more.
A major evolutionary biologist, a creator of the New Synthesis, Ernst Mayr, says in his Toward a New Philosophy of Biology: Observations of an Evolutionist:
"The changes in the physical sciences involve, among other things, a rejection of the strict determinism of classical physics. Scientists now recognize that most physical laws are not universal but are rather statistical in nature, and that prediction therefore can only be probabilistic in most cases. They have also realized that stochastic processes [non-deterministic, because a system's future state is determined not only by the process's predictable actions but also by a random element] operate throughout the universe, at every level, from subatomic particles to weather systems, to ocean currents, to galaxies. In the study of these processes, observation has been elevated to the status of a valid scientific method wherever the experiment is difficult or impossible to perform, as in meterology and cosmology." [Emphasis added]
In fact, what served to confirm one of the most important theories in all history of science, Einstein's general relativity theory, was a mere observation of a total solar eclipse in May 1919 by Sir Arthur Eddington. In astronomy, observations are often a way to test hypotheses and this has not prevented it from developing as an advanced science.
This strenuous effort by biomedical experimentalists to create conditions permitting total control over the relevant variables, which they consider the supreme advantage of experiments over observations, is directed to a goal which appears less and less attainable in the light of the new scientific developments and ideas: a strict, deterministic experiment in a non-deterministic, statistical, probabilistic world.
Anyway, even if experiments on animals could achieve the type of rigour their creators are after, this would still leave them in a difficult predicament. First of all, what animal species to choose to experiment on, if the purpose of the test is of medical nature? This in itself is an arbitrary choice, that reduces the level of rigour. Researchers usually say that they choose animal species which they know are similar to humans in the relevant aspects under investigation.
How do they know that? One source of their knowledge, since this is a comparison, must be the humans themselves. But, by their own admission, biomedical scientists do not consider clinical and epidemiological studies an adequate source of knowledge, only a "second best". So, according to their own view, they don't know that a particular species is similar to humans.
They may try to overcome this obstacle by experimenting on more than one species. But even more damning, after they have, for example, tested a drug on several lab species, they may have found that in a certain species the drug under study was harmful, in another it was innocuous but useless, in another it was effective and harmless, and in still another it was effective but with severe side effects. This inter-specific variety of results is a common occurrence.
So, the problem now is: which species' experimental results to choose to form predictions on what will happen when the drug is administered to humans?
Now it gets interesting because we can see that, the more rigorous and controlled the conditions are during an experiment, the more these conditions distance themselves from the reality to which they should eventually refer and apply.
The more rigour and control we find in animal experiments, the more of them we lose in the process in the process of interpreting the results and extrapolating them to humans. Researchers are realizing that more and more, and because of that we are witnessing a shift to a greater use of epidemiological studies on humans, from the field of the health effects of diets to the studies on the toxicity of organophosphates.
Even the previously greatest supporters of animal tests, like the USA Environmental Protection Agency (EPA) and National Institutes of Health (NIH), have publicly acknowledged their limitations and moved on from the use of lab animals to other research tecniques.
Unfortunately, it's two steps forward and one step back. In Europe, the REACH (Registration, Evaluation, Authorisation and Restriction of Chemicals) EU directive, which imposes the testing of something like 170,000 chemical substances and products on animals in the forthcoming (possibly a hundred) years, is going against the tide of events. But on this, on another page.
Not only other animal species differ profoundly from the human species, but lab animals, being imprisoned in an artificial environment all their lives, are kept in conditions of extreme stress, which does not apply to humans. Add to this the fact that human diseases are not to be found in other species, and that the aetiology which defines human illnesses is not what artificially caused in lab animals only similar symptoms, and you can see how all the rigour that researchers strived for in the laboratory evaporates, disappears and gets lost outside of it.
Scientists sometimes see that too, as Adams, Whisnant and Wiebers in "Animal models of stroke: are they relevant to human disease?" in Stroke:
"Molinari suggested we modify measurement of human outcomes to conform to animal research so that we can reproduce the results of animal research in humans. We suggest the converse, namely that there is a great need to design experimental projects and therapeutic interventions in such a way that they have relevance and practical application to the human condition." [Emphasis added]
Besides, to use results of experiments on animals as a way to confirm a hypothesis about humans requires an extrapolation, that is an additional level of hypothesis (a theoretical layer) which is not required in the case of observations of humans. We are getting further and further from the declared goal of stringency and control ober the whole process.
Whereas it is entirely plausible to assume that two hydrogen atoms are identical and will behave identically in and outside the laboratory, when it comes to living beings, extremely more complex systems than pure physico-chemical entities, this kind of simplistic assumptions become inadequate.
After Darwin, and after the revolutionary developments in the physics of the early 20th century, the idea that the laws of nature are deterministic and that we can ignore the role of chance in natural processes has no scientific ground. But the view of the primacy of experiments which is still behind the current biomedical paradigm depends on that very idea.