TESTING FOR LONG-TERM TOXICITY 829 Table V Strain Differences in Induction of Liver Cancer by Ethionine • Duration of Ethionine Livcr Cancer Rat Strain Sex (Months) (%) Carworth Farms Wistar M 5 86 Holtzman M 5 0 Carworth Farms Wistar M and F 7.5 100 Holtzman M 7.5 60 Holtzman F 7.5 25 Fischer M 7.5 100 Fischer F 7.5 90 Data from Farber (17). Ethionine was fed at 0.25% of the diet. Table VI Mammary Cancer in Female Rats After Single Dose of 7,12-Dimethylbenz(a)anthracene" Dose Per Cent with Median Induction Strain (mg) Mammary Cancer Time (Days) Long-E vans 20 16 82 Marshall 30 0 ... Sprague-Dawley 20 100 41 Wistar 45 100 50 August 30 90 90 Chester Beatty 45 100 50 Data from Boyland and Sydnor (18) and Sydnor et al. (19). In most stains of rats, oral administration of a single large dose of 7,12- dimethylbenz(a)anthracene to 40-50 day old females represents an over- powering carcinogenic potential at a sensitive time period. Usually 100% of these animals develop mammary cancer in 25-80 days. How- ever, some strains such as the Long-Evans or the Marshall hardly re- sponded or did not respond at all (18, 19) (Table VI). Thus, in planning long-term toxicity (or carcinogenicity) tests, it is important to use strains of animals which respond to carcinogens and are not totally resistant, but which are not so extremely responsive that misleading results are obtained, due to their high spontaneous tumor incidence. In our usual practical, this is accomplished by including a positive control (known carcinogen) in each test series. If the animals do not respond to the appropriate known carcinogenic agent, the results ot• the test series are then in doubt. The route ot• administration may also influence the site at which tumors appear. Although most of the polycyclic hydrocarbon carcino-

830 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS gens such as benzo (a)pyrene, dimethylbenz(a)anthracene, or 3-methyl- cholanthrene cause tumors when painted on mouse skin, feeding these materials to infant mice can lead to hepatomas and numerous other tumors (9). However, in adult mice feeding of the hydrocarbons usually leads to relatively few tumors, generally of the forestomach, since much of the carcinogen passes through the gastrointestinal tract unchanged (3). A more striking example is that oral administration of dibutylnitrosamine to rats caused mostly liver cancer while repeated subcutaneous injection led to bladder tumors in all rats (20, 21) (Table VII). The solvent or vehicle employed for administration by gavage or skin painting can alter the response dramatically. An example is the work of Bingham and Falk (22) who found that if dodecane, noncar- cinogenic by itself, were added as a solvent, the skin tumor response of mice to benzo(a)pyrene was increased tremendously (Fig. 1). The fact that diet may also radically affect the outcome of experi- ments in carcinogenicity was emphasized almost 30 years ago by Table VII Carcinogenic Effect of Nitrosodibutylamine in Rats • Tumor Incidence Dose ( % ) Average Induction Route (mg/kg) Frequency Liver Bladder Time (Days) Oral 10-75 Daily 58 35 150-540 Subcutaneous 200-400 Weekly 10 100 208-334 Data from Druckrey et al. (20, 21). • 2 x 16 • •.) C.) 2 Xt(• 2 / • In Decahn I tn Decalin + Dodecane 50 SKIN TUMOR INCIDENCE (%) I O0 Figure 1. Skin tumor incidence in mice after repeated cutaneous application of low concen- trations of benzo(a)pyrene (BP) in decalin or n-dodecane and decalin. Data from Bingham and Falk (22)