Twelve

TOXICITY TEST PROCEDURES FOR
CHEMICAL SUBSTANCES

1. SUBSTANCES TO BE TESTED AND NUMBER OF ANIMALS USED

The previous chapter considered the production and testing of "biologicals" - substances which are derived from living matter. The great majority of substances mentioned in this and the following chapter are not organic in origin but are manufactured by the chemist. Drugs, including analgesics; cosmetics; household products; and substances used in agriculture (e.g. pesticides) and industry, are subject to the toxicity tests to be discussed.

A substance is defined as toxic if, when eaten, inhaled or absorbed through the skin, it causes by its chemical action either damage to structure or disturbance of function, or both. Safe and effective dosages have to be estimated for all drugs, and safe concentrations for household products. There is also a difference in toxicity to be calculated for the ingredients of a drug under any and all possible circumstances and for the drug product under reasonable conditions of use. Drug manufacturers are very concerned about such things, and so are certain government agencies such as the Food and Drug Administration (with respect to food additives, drugs, cosmetics), the U.S. Consumer Product Safety Commission (hazardous substances: irritants, corrosives, inflammable and "highly toxic") and the Environmental Protection Agency (insecticides, fungicides, rodenticides).

In all, there are in the U.S. up to two million such substances already in distribution, with 13,000 substances listed in the toxic substances list of the National Institute for Occupational Safety and Health and 600-1,000 new substances entering the market every year. (Eagleton, T., 1977).

Animals are extensively used in testing all these substances for toxicity and very large numbers are involved. According to the Institute for the Study of Animal Problems, dealers now estimate the current annual demand for all types of research and testing at 100,000,000 animals: 50 million mice, 20 million rats and about 30 million other animals, including about 200,000 cats and 450,000 dogs. (Rowan, A.,1979a, p.8). It is not known how many of these are involved in acute and chronic toxicity testing because American official bodies like the National Research Council have failed to produce statistics on the subject, but a British group of scientists, the Committee for Information and Animal Research, estimated that acute toxicity testing - including the notorious LD/50 tests - consumed one-fifth (about one million) of the animals used in Great Britain for experimental purposes in 1975. (CRAE,1977, p.25).

One-fifth of the estimated 100 million animals currently used in the U.S. would amount to 20 million.


2. TEST PROCEDURES

Acute Toxicity

Acute toxicity tests are a "search for untoward reactions at a high dose level." Research facilities, particularly pharmaceutical ones, frequently mention these tests in their Annual Reports to the U.S. Department of Agriculture under the Animal Welfare Act (cf.p.241). They can be found uner column D of the Report, which is for experiments "involving pain or distress without the use of appropriate anesthetic, analgesic or tranquilizer." Thus Procter and Gamble:

"Acute Oral Toxicity Studies - 48 dogs. These are product safety tests designed to determine potential local corrosive action or systemic toxic effects after ingestion of a product. They are a necessary part of a total oral toxicity program submitted to the FDA to insure the safety of a product prior to marketing. It is necessary that the animals receive no other medication since their use may interfere in the evaluation of the test. The dogs are given a specific dosage of a product by gavage [forced feeding by stomach tube] and are observed for four hours after dosing and daily thereafter for two weeks. Most Products cause only transient emesis or diarrhea, although occasional signs of systemic toxicity, e.g., neurologic dysfunction, are observed." (USDA/APHIS,1976, Procter & Gamble Co.).

Bristol Laboratories:

"During 1973 approximately 30 dogs and one squirrel monkey received acute doses of a compound which resulted in convulsions and eventual death." (USDA/APHIS,1973, Bristol Labs.).

Tulane University School of medicine (Laboratory of Environmental Medicine):

"241 rabbits. Skin and eye irritation test and dermal LD/50s ["lethal dose in 50% of animals tested"] were performed according to Federal Hazardous Substances Act procedures .... We did cause some pain and distress because we were studying the acute skin and eye irritation potential of various chemicals. If we were to use drugs, anesthetics, or analgesics to relieve the pain and distress we would be defeating the very purpose of the study." (USDA/APHIS,1976,Tulane Univ.).

Column D requires a "brief explanation" of why pain-relieving drugs were omitted. Those who drafted the Animal Welfare Act presumably expected that this explanation would serve as a check on painful experiments performed without anesthetics or analgesics. However, in the vast majority of cases, facilities reporting such experiments circumvent this by referring merely to the government agency regulating the test substances.

In the 1970 amendments to the Animal Welfare Act the secretary of DHEW is told to promulgate standards for the use of pain-relieving drugs. This he has failed to do, so buck-passing references to a whole potpourri of government regulations can be used to justify the suffering inflicted on many thousands of animals.

As a reminder of what the animals have to undergo, here is a partial list of the common signs of toxicity which may be observed in acute toxicity tests in rodents: behavior: unusual vocalization, restlessness; twitch, tremor, paralysis, convulsions; muscle tone: rigidity, flaccidity; salivation, lachrymation; difficulty in breathing; diarrhea, constipation, flatulence; swelling of the sex organs and breasts; skin eruptions; discharge, including hemorrhage, from the eye, nose or mouth; abnormal posture, emaciation. These are all signs of distress which can be observed; what the animals feel can only be imagined.


The LD/50 Test

This is an acute toxicity test in which a number of animals are given different amounts of a substance to determine what dose will kill 50% of them over a predetermined period of time. If the substance is a gas, then the lethal concentration (LC/50) in air which will produce the same effect is determined.

The test is principally used to give a rough indication of the toxicity of a to produce, as a statutory requirement, a figure for comparison with the toxicity of known compounds. Thus the warning notices which appear on packaging under the Federal Insecticide Fungicide and Rodenticide Act are based on the LD/50: "DANGER-POISON" means that 50% of the test animals were killed by a dose of 50mg. or less per kg. of body weight; "WARNING," by a dose of 51 to 500mg./kg.; "CAUTION" by 501 to 5000mg./kg.
1 However, even when the test is not mandated by government regulations, it is frequently used by manufacturers of drugs and consumer products on the supposition that it is a protection against proceedings for negligence.

Variables Impair the Accuracy of LD/50 Tests

Recently, the LD/50 has been under increasing fire from scientists as well as humanitarians. At first glance, the test seems clear-cut since its end-point is an event as definite as death. But, as I have reiterated in this book, tests on animals are beset by numerous variables. Thus, while the actual death is definite, the moment it occurs is on a sliding scale: some die soon, some later. Furthermore, there are differences in susceptibility between species: penicillin is highly toxic in guinea Pigs but generally nontoxic in humans. Some drugs, when tested by the LD/50, appear to be relatively nontoxic, even when given in large doses, but when administered in small doses over a longer period of time, prove to be killers. The value of the LD/50 may also he influenced by the age of the animal, its sex, state of nutrition and bedding material, also by the temperature and humidity, the time of the day or year the test substance is administered and by the method of administration. J.K. Morrison and co-authors, writing in Modern Trends in Toxicology, state:

"It cannot be denied, however, that a reproducible LD/50 determination can he achieved provided that experiments are carried out under stringently controlled conditions guaranteeing uniformity over all these numerous factors which contribute to variations. But we are not convinced that such effort and experience are either economically or scientifically justified, since the resultant information contributes little to the design of future toxicity trials in animals, or to the assessment or safety in man." (Morrison,J.,1968,p.12).


Stuffed to Death

There is no justification for using the LD/50 to test relatively nontoxic substances such as cosmetics and food additives. The forced feeding of huge quantities of these may cause the stomach to rupture or may kill the animal through the physical insult to its system: this bears no relevance to the actual toxicity of the substance. It is claimed that this grotesque practice is being superseded by the "limit test", in which "a dose of the substance at one percent of the animal's body weight (equivalent to the ingestion of 1/4 lb. of face cream by a one-year old child) is administered to a group of animals. If this produces little or no toxic reaction, then the substance is presumed to be safe and no further testing is required." (Rowan.A., 1979b, p.5). R.P. Giovacchini, whose tests at Gillette Laboratories are cited by the Food and Drug Administration as setting standards for "adequately substantiating" cosmetics, is another who cautions against LD/50s if there is a possibility of death resulting from shock caused by "the sheer volume of material ingested." To avoid this, he also recommends one percent of body weight as the maximum dose. (Giovacchini, R.,1972, p.364).

Modifications of LD/50 Test Suggested by Experimenters

Considering that the standard LD/50 may use from 60 to 100 animals, it is important to press for at least a reduction of this wastage pending the phasing out of the test itself. A suggestive comment comes from P.S. Rogers, managing Director of Hazleton Laboratories (Europe), one of the largest contract oratories in the world:

"The test itself is usually run in two parts, involving a dose ranging test with two animals per group, followed by a repeat experiment using larger groups of animals over a narrower dose range .... The LD/50 is a crude measure of toxicity. There is really little scientific justification for the test because reproducibility is not good, it can even vary from day to day, and the results are dependent on the animal strain used. Sufficient information can usually be obtained from the dose ranging sector of the test." [Emphasis added]. (CRAE,1977,p.18).

Battelle's Pacific Northwest Laboratories have made some attempt to moderate the rigors of an LD/50-type experiment on ten beagles who received oral doses of a radionuclide which caused gastrointestinal lesions, diarrhea and death.

"To minimize pain the animals were sacrificed when clinical signs indicated death was imminent. To minimize the number of animals required for the experiments, the dose groups were completed in sequence so that data from one dose group could be used to plan sequential dose groups." (USDA/APHIS,1976,Battelle Memorial Inst.).

The above discussion was largely concerned with acute toxicity testing.

This section concludes with several paragraphs on more prolonged - subacute and chronic - test procedures.


Subacute Toxicity

In general this defines the biological activity of a compound, an estimate of the "no-effect" dosage, and the maximum tolerated dosage. The experiment usually covers a three month period. The animals generally used are, first, the albino rat, ten of each sex in each of six groups. one group receives no dosage (controls), the other five receive gradually decreasing dosage - starting from the maximum tolerated dosage which has been previously determined by a "range-finding" test.

Concurrently, dogs are used, of, known species (usually beagles) and colony-raised, two males and two females on each of three dosage levels, plus a control group. The route of administration of the tested substances is as close to the proposed human usage as possible. Frequently in feeding experiments the animals will refuse the test material and then the stomach-tube is necessary. Biochemical tests are carried out both before and during the experiment; at the conclusion, the animals are killed and their organs and tissues studied. (Assn. of Food and Drug Officials, 1979, p.26).

Chronic Toxicity

Chronic toxicity tests are used to determine if a small dosage, apparently harmless on subacute testing, is toxic on long exposure. The Food and Drug Administration generally runs these chronic tests for two years. Over this length of time, larger groups of rats and dogs than used in the subacute tests are given graduated dosage so as to determine a "no-effect" level and a toxic level of dosage. At the end the animals are killed and autopsied. Sometimes much larger numbers of animals are used to be sure that there is no toxic effect. The FDA cites an instance where the use of 24 animals showed no effect at a certain dosage, whereas testing 200 rats at that dosage revealed an increase in liver weight among enough of them to be significant. (Ibid., p.36).

3. COMPARISON OF TESTS IN INTACT ANIMALS AND CELL CULTURES

With chemical entities numbered in the millions and hundreds of new ones entering the home, the farm and the environment every year, the attempt to test all these toxicologically becomes overwhelming. The effect on humans is what we need to know, but animals have been the traditional stand-ins. However, as P.C. Rofe points out in an article reviewing the growing use of tissue and cell culture in toxicology,

"Toxic manifestations in the whole animal, be they change, in metabolic patterns or alterations in functional efficiency of specific organs, are still secondary to changes occurring within the cell ... the facility of being able to observe the reaction of tissues from man under the same conditions as those of experimental animals is a unique advantage."

Many systems using either animal or human cell or organ culture, as well as plant material and microorganisms, have been created with emphasis on those which are rapid, inexpensive and can discriminate between those chemicals whose properties represent a high toxicity risk and those which are relatively innocuous. (Rofe, P.,1971).

In Vivo vs. In vitro: Comparing Tests on Animals and Animal Cell Cultures

While there is no difficulty in proving that many chemicals are toxic to cells in culture, it is more difficult yet essential to demonstrate to what degree the cellular response is related to the whole animal's response to the toxic compounds. Only if the effect on the cells in culture can be shown to correlate closely with the animal reaction - indeed can predict the toxic effects in animals of other compounds of similar structure - are in vitro tests on cells likely to supersede the in vivo ones in animals. Or, if not supersede, then at least screen out the substances which are acutely toxic and thus spare animals some of the LD/50type tests described on p.203-208.

As an effort in this direction, the work of Dr. Robert I. Christian at the University of Cincinnati is interesting. According to a report in the Christian Science Monitor of Mar. 10,1978, he has tested 30 chemicals on mammalian cell cultures and compared his findings for them with tests of the same chemicals in the whole animal. Toxicologically, Christian found that the results in the two methods of testing agreed very well. (Salisbury, D., 1978).

Of course, if questions are asked about the transport of a toxic compound in the body, about its metabolism or transformation into less or more toxic forms, its chronicity, etc., a cell culture test which merely records the toxicity of the dose in terms of inhibition of cell growth will be of little use. One must ask appropriate questions to get information of value from tissue culture.

Autian and Dillingham asked the following question: given a series of methyl- and halogen- substituted alcohols, each of which is known to kill by a certain dose 50% of a group of mice (LD/50) in seven days, can a similar effect be produced in cell culture: i.e. 50% inhibition of cell growth (ID/50), with high correlation between the two procedures?

They used a mouse fibroblast L cell culture. When the ID/50s were calculated for the alcohols, giving the intrinsic toxicity (Ti) of each, it was found that the correlation between the Ti values and the LD/50s - between in vitro and in vivo - was poor: there was a 164-fold variation. To reduce the gap, it was necessary to introduce another factor into the equation: the degree to which each of the alcohols separated either into the aqueous or into the lipid (oily) compartment of the cell culture. Expressed in terms of "lipophilicity," the most lipophilic alcohol was 60 times more "oil-loving" than the least. When the equation was corrected for this, the 164-fold variation above was reduced to 4-fold, which, allowing for variations due to statistical error, biotransformation and metabolism of the compounds, is considered excellent. (Dillingham,E.,1973).

As the investigators point out, their alcohol study, although it provides meaningful information with respect to structure-activity relationships and supports the basic equivalence of in vitro and animal response,

"...Only suggests the direction for the development of good predictive systems. Without question ... animal toxicity parameters other than the LD/50 will be needed to deal with different classes of biological responses. Tissue culture parameters other than growth inhibition will undoubtedly be needed."

They add that, as the list of chemicals tested in this manner grows, mathematicians will have data from which to predict the toxicity of new compounds and to determine what component parts of the molecule are contributing to the overall toxicity. This is similar to the role of mathematicians described on p. 247.


In Vitro vs. In Vitro: Comparing Tests on Human and Animal Cell Cultures

An excellent review of toxicity testing in cell cultures by Professor Roland M. Nardone of Catholic University of America describes many of the advantages - and limitations - of the in vitro method. But Nardone has some critical comments.

"The development of in vitro testing to date has been, at worst, almost happenstance, and, at best, without taking full advantage of the technical and theoretical advances of recent years. Such amorphous growth is wasteful, inexcusable, and has led to the publication of much research of questionable value."

To correct this, he suggests rigorous standardization of test procedures, including the biological material used - cells, tissues and organs, media and culture vessels.

"Serious consideration," he continues, "should be given to the establishment of the toxic response of one or more well-characterized 'reference' cell lines to diverse, ultrapure reference chemicals drawn from several classes such as metals, acetone, and National Cancer Institute standard carcinogens. The response of such a test system could then serve as a yardstick or standard against which other cells can be compared, against which other toxicants can be compared, and for the continued assessment of the reproducibility of a particular protocol." (Nardone, R., 1977).

To study the subject of interspecies variations, P.C. Rofe has proposed that a range of adult human tissues be exposed to the action of selected compounds. "These compounds would be chosen for their contrasting toxicological effects and for the amount of data available about them. A selection of tissues from the principal laboratory animals would also be exposed under the same conditions." At the conclusion of the exposure, a sample of each tissue would be rapidly frozen and subjected to chemical and microscopic analysis, with emphasis on obtaining metabolic data. Since the liver is the main site of metabolism of foreign compounds, its study in culture would be of particular importance, and liver metabolites formed after exposing it to the test compounds could be added to the culture fluid of the other tissues. "From these studies it would be possible to design a routine procedure, the results of which would be available in the early stages of a major investigation, and would indicate not only the most appropriate species to use but also the most vulnerable tissues." (Rofe, P.,1971, p.683).

In other words, these investigators wish to standardize the use of tissue and cell cultures in toxicology. They would compare the chemical reactions of human and animal cells -under the same conditions - to a variety of compounds and their metabolites. From this would emerge routine in vitro procedures which would he reproducible, and the standardization would hopefully reduce the variables responsible for inconclusive experimental results and much wastage of animals.


NOTES

1. Comment by W.N. Scott Universities Federation for Animal Welfare): "My chief objection is to people who continue to stuff animals with inert drugs until the substance almost comes out of their ears. Assessing toxicities of the order 5000 mg./kg. is not killing the animal with the toxicity of the drug, but by distension of the stomach or something like that." (CRAE, 1977, p.30).
Back to text


Chapter Thirteen | Back to Table of Contents