Critique of Pharmacokinetics and Metabolism Studies in Mice, Rabbits, and Dogs Performed at Huntingdon Research Center

Murry J Cohen, M.D.; Stephen R. Kaufman, M.D.; and Jerry Vlasak, M.D.
September 2001

In an attempt to shed light on the pharmacokinetics and metabolism of sucralose in humans, the Huntingdon Research Center team of John, Wood, and Hawkins measured these physiological processes in mice,[1] rabbits,[2] and dogs.[3] They chose mice "to assist in the evaluation of results from toxicology studies [in laboratory animals] and the extrapolation to man."[1] Rabbits were used because "information derived from this study is helpful in the evaluation of results from toxicology studies of sucralose and in extrapolation of the animal data to man.'[2] (sic) Finally, dogs were included in the study "to assist in the evaluation of results from toxicology studies and to extrapolate the findings to man."[3]

These studies, like studies attempting to determine the effects of chemicals on humans by researching mice, are highly flawed. They typically generate much data, but none that can be reliably used to understand human pharmacokinetics. Predictably, in order to interpret animal data, researchers need human data for comparison: "Co-chromatographic and autoradiographic analyses of both mouse and human urine samples were conducted to compare the urinary components of the two species."[1] Also, characteristically, interspecies differences are found. For example, humans and mice metabolize sucralose differently, with metabolites found in one but not the other species. One human metabolite only "co-chromatographed with a trace component in mouse urine, detectable only by radiochromatogram scanning," while a "mouse urine component…did not co-chromatograph with any human urinary component."[1] The rabbit data were influenced by the fact that rabbits eat their feces - 30-80 percent of the total daily fecal excreta[2] - and by the finding of a "notable difference of a relatively prolonged excretion phase [in rabbits] compared with the other species tested." The dog data revealed different sucralose metabolites in dog urine compared to human urine, with one metabolite "present in only trace amounts in dog urine but…more prominent in human urine."[3] The researchers concluded in the mouse paper, "Overall, the metabolism of oral doses of sucralose in the mouse is similar to that in man following oral administration of 14C-sucralose" [emphasis added]. One may assume, since the researchers published the papers in a scientific journal that deals with human toxicology, that they believe the same of their rabbit and dog results.

While we discuss the use of mice to learn about human physiology, many of our points apply equally well to the use of rabbits and dogs. There are many problems with using mice to learn about pharmacokinetics in humans. Differences between species in metabolism and distribution prevent generalization of results. A toxic substance remains in the bloodstream and organs of humans longer than in rodents since humans and large animals metabolize and distribute substances much slower than do small animals.[4] Also, excretion factors similarly prevent generalization. Since humans excrete compounds more slowly than mice, humans can develop a greater accumulation over many years with only a very small daily intake.[5] Because humans are less closely related genetically to one another than are laboratory mice, there is a much greater variation in response to a given chemical in humans than in mice.[6] Mice are also subject to contamination from various laboratory conditions, including chemical contamination from laboratory floors and researchers' gloves,[7] viral contamination with LDH-virus,[8] and chemicals found in the bedding used. In fact, shavings from cedar trees used in bedded contained a carcinogen.[9] Many other factors can introduce uncertainty or error into mice tests, including inbred strains of animals; stressful living conditions; unusual routes of test compound administration; enhancement of the susceptibility of the animals by deliberately suppressing immune mechanisms; contamination in the compounds; exposure to high doses of a chemical; microbial diseases; chemically polluted feed, water, air, and bedding; infestation by vermin; cannibalism; and diet.[10], [11]

One measure of the accuracy of testing chemicals in mice is the overlap between the effects of a putative chemical in mice and humans. Chemicals are often tested in mice for both carcinogenicity and anti-cancer chemotheraputic value. Only slightly more than one-third of compounds that caused cancer in humans also did so in mice. In a National Cancer Institute study, only 7 of 19 known human carcinogens caused cancer in rodents given high doses over their lifetimes, using the Standard National Cancer Institute Protocol.[12] Likewise, anti-cancer agents that are effective in mice are ineffective in humans. Between the mid-1950s and the mid-1980s, the National Cancer Institute screened 400,000 putative anti-neoplastic chemicals on leukemic mice,[13] but the few compounds effective against mouse leukemia had very little effect on human tumors.[14]

The conclusion of John et al. that mouse metabolism of sucralose is similar to human is astonishing. "Similarity" is simply not sufficient to allow one to draw conclusions about human metabolism from mouse studies. This is implicit in the experimenters using human data - at least in the mouse and dog studies - to understand and interpret their findings in these species. But, already having human data, of what value are the findings of animal experiments?

The standard recommendation ends one paper: "Further investigations would be necessary to confirm the identity of the mouse metabolites."1 However, all the possible subsequent animal studies that could be perfored will not and cannot overcome the problem of species dissimilarity and the impossibility to extrapolate the data from another species to humans with reasonable confidence. Testing substances in animals to understand toxicity, carcinogenicity, mutagenicity, teratogenicity, or any other effect in humans is wasteful.

Noted toxicologist and animal testing advocate David Rall has noted "the hazardous process of attempting to extrapolate data for laboratory test systems to man" and concludes that species similarity breaks down in the realm of metabolism. Since "Man is among the most heterogeneous systems on earth, [I]f sufficient effort were exerted, one could find that metabolically some man would exhibit a pattern like a rat, another man would exhibit characteristics of a dog and so forth."[15]