150 • Alternatives to Animal Use in Research, Testing, and Education 
TESTING METHODS 
Toxicology as a science began in the 16th cen- 
tury and has advanced with the growth of the 
chemical, pesticide, drug, and cosmetic industries. 
The concept of protecting the public from harm- 
ful effects of chemicals dates back to laws of an- 
cient civilizations that made it illegal to adulterate 
the food supply (25). The importance of toxicol- 
ogy to public health has received considerable at- 
tention in the United States since the 1930s. Pub- 
lic awareness of the value of toxicological testing 
has also been furthered by disasters such as Mina- 
mata disease (methyl mercury poisoning in Japan), 
the thalidomide tragedy, and, more recently, the 
development of cancer in those exposed to diethyl - 
stilbestrol (DES) in utero. 
Designing a Test 
There are two approaches to toxicology— mech- 
anistic and descriptive— and these affect the de- 
sign of experiments and the choice of biological 
end points to be measured. Mechanistic toxicol- 
ogy focuses on the chemical processes by which 
a toxic effect occurs and relies heavily on the tech- 
niques of physiology, biochemistry, and analyti- 
cal chemistry to monitor these processes. A sim- 
ple example of this approach would be a series 
of experiments showing that a certain substance 
is metabolized in the liver, that one of the by- 
products of metabolism happens to be a potent 
liver carcinogen, and that liver cancer typically 
follows administration of that substance. Mech- 
anistic tests are custom-designed and are closely 
related to research. They can contribute greatly 
to the design and interpretation of descriptive tests . 
Mechanistic toxicology plays a major role in the 
development of methodologies that could replace 
whole -animal testing. 
Descriptive toxicology deals with phenomena 
above the molecular level and may rely heavily 
on the techniques of pathology, statistics, physi- 
ology, and pharmacology, e.g., the evaluation of 
changes in the appearance of an organ or its con- 
stituent cells, the presence of tumors, or signs of 
irritation. This approach does not necessarily re- 
quire an understanding of the mechanisms by 
which toxic effects occur, although if mechanis- 
tic information were available, it would be used. 
In terms of the test substance and species in the 
preceding hypothetical case, descriptive toxicol- 
ogy would show that a certain substance causes 
liver cancer in a particular species within a cer- 
tain time. It might also show the approximate rela- 
tionship between the substance dose and the inci- 
dence of the liver cancer. Regulatory schemes 
requiring testing most often rely on descriptive 
toxicology. 
Mechanistic toxicology provides an approach to 
extrapolation from one species to another based 
on known similarities and differences in physiol- 
ogy. The closer the test animal is biologically to 
humans or the greater the number of species in 
which the effect is detected, the more likely it will 
occur in humans as well. The reliability of extrap- 
olations from descriptive experiments is greatly 
enhanced when mechanistic information is also 
used. Similarly, the use of mechanistic informa- 
tion in the design of descriptive tests contributes 
greatly to the reasonableness of any later extrapola- 
tion to humans if human toxicity data are lacking. 
Most state-of-the-art toxicological tests require 
whole animals. Although in vitro alternatives are 
being developed (see ch. 8), different end points 
would be measured. For example, whole animals 
will probably continue to be needed to look for 
effects in previously unknown target organs, to 
evaluate effects that represent an interaction of 
multiple organ systems, to monitor metabolism and 
pharmacokinetics, or to evaluate healing or dimin- 
ished responsiveness to the toxic substance. Thus, 
whole -animal use is unlikely to stop entirely in the 
foreseeable future. 
Choice of Species and Strain 
In 1904, the Food and Drug Administration (FDA) 
was still using human employees to test food pre- 
servatives (e.g., boric acid, salicylic acid, their de- 
rivatives, and formaldehyde) for toxicity (25). Use 
of animals remained limited until a few decades 
ago, when breeding technology provided large 
numbers of animals with carefully controlled 
genetic characteristics, thus allowing toxic effects 
to be more easily detected than had previously 
been the case . Animal use has grown with increas- 
