Toxicologic Pathology

Livers of male Sprague-Dawley rats were evaluated for foci and nodules 90 days and 16 months after one or two oral doses of 1,2-dibromoethane (DBE). Rats (190 g) were given the following oral treatments. Controls received corn oil (2.5 ml/kg) at 0 and 24 hr. DBE × 1 received corn oil at 0 hr and 75 mg DBE/kg at 24 hr. DBE × 2 received 75 mg DBE/kg at 0 and 24 hr. All rats underwent a two-thirds partial hepatectomy at 28-29 hr, a one-third partial hepatectomy at 90 days, and were given 0.05% phenobarbital in drinking water for 4 months beginning at 1 yr. DBE was given to the DBE × 2 group twice within 24 hr because the compound is a hepatocyte mitogen. At 90 days, no appreciable changes were evident in any group. At 16 months, the incidence of nodules in the DBE × 2 group (25 of 41) was double that of the DBE × 1 group (12 of 38) (p < 0.01) and triple that of the control group (6 of 34) (p < 0,0001). Both the DBE × 1 and the DBE × 2 groups had higher incidences of eosinophilic foci, and also higher numbers, larger sizes, and larger areas of γ-glutamyltranspeptidase-positive foci than the control group. These results demonstrate for the first time that hepatocyte foci and nodules are initiated by limited exposure of animals to DBE.

Hyperbaric oxygen (HPO) was administered to rats (100% O₂ at 2.8 atm for 90 min) immediately or 1 hr after severe carbon tetrachloride (CCl₄) intoxication in order to study the mechanisms of protection against hepatocellular injury by hyperoxia. Slight to moderate hepatocellular injury was observed, particularly by morphologic criteria, 4 hr after CCl₄ intoxication. Little cell death was observed; 24 hr after CCl₄, 20% of the untreated animals died. In the survivors, the following typical changes occurred in the liver: extensive hepatocellular swelling, vacuolization and necrosis; severe ultrastructural alterations; binding of CCl₄ to microsomal lipids; elevation of lipid peroxidation products (conjugated dienes); little decrease in cytochrome b5 and severe decrease in cytochrome P-450 levels. Serum (ransaminase (alanine aminotransferase and aspar-tate aminotransferase) levels were elevated. Immediate treatment with HPO prevented the mortality and markedly decreased the hepatocellular necrosis 24 hr after intoxication. Immediate HPO treatment did not lower the levels of free CCl₄ in the liver. However, the rise in lipid peroxidation products caused by CCl₄ intoxication at 4 hr was reduced. Delayed treatment with HPO (1 hr after CCl₄) prevented the mortality but was less effective in preventing necrosis. Some hepatocellular protection was still demonstrable. In particular, the rise in lipid peroxidation products was reduced.

Hyperoxia protects hepatocytes against CCl₄ toxicity. The rapid decline in protective effect within 60 min of intoxication suggests that hyperoxia inhibits CCl₄ activation and/or damage from molecular intermediates. Hyperoxia has little effect on the progression of sublethal injury to cell death in the livers of CCl₄-intoxicated rats.

Studies were carried out on microsomes isolated from the highly differentiated (slow-growing) Morris hepatoma 9618A, on microsomes and plasma membranes from the poorly differentiated (fast-growing) Morris hepatoma 3924A, and rat liver used as control. The lipid composition (phospho-lipid and cholesterol content, degree of fatty acid unsaturation) and peroxidation of such membranes has been correlated with the order and fluidity of the membrane bilayer. The results indicate that substrate availability is the rate-limiting step in microsomal and plasma membrane lipid peroxidation of hepatoma 3924A. From diphenylhexatriene fluorescence depolarization measurements it appears that the changes in lipid composition cause an increase in the order of the lipid bilayer on going from the control to hepatoma 9618A and 3924A microsomes, while fluidity is virtually unchanged. Conversely, for similar chemical changes, in plasma membranes from hepatoma 3924A the order is nearly the same and there is a decrease in fluidity. The changes in the above parameters of tumor membranes might be partly related to the loss of protective enzymes against oxygen radicals. This is supported by the observation that inhibition of liver superoxide dismutase and glutathione reductase, by treatment of rats with diethyldithiocarbamate and chloroethyl nitrosourea, respectively, renders the microsomal membranes more resistant to lipid peroxidation in vitro.

Favism is an acute hemolysis occurring in glucose-6-phosphate dehydrogenase (G6PD)-deficient (Mediterranean variant) individuals after intake of fava beans. Divicine (D), 2,6-diamino-4,5-dihydroxypyrimidine, is present in high amounts in the beans, and is suspected to play a role in hemolysis. Its mechanism of action was studied in a cell-free system and in G6PD (Mediterranean variant)-deficient red cells (RBC). Upon hydrolysis of the inactive β-glucoside vicine, reduced divicine is formed. Oxygen acts as a one- or two-electron acceptor; superoxide anion and hydrogen peroxide are formed, respectivley, together with the semiquinoid free-radical form of D. This free radical gives an electron spin resonance (ESR) signal, which is similar to that of the alloxan free radical. Added reduced glutathione (GSH) is rapidly oxidized with a stoichiometry of one to one, and the ESR signal is abolished. Additional GSH is oxidized by hydrogen peroxide and by a slow redox cycle which continuously regenerates oxidized D. The fast-direct and the slow-indirect oxidation result in nonstoichiometric oxidation of GSH.

D added to G6PD-deficient RBC rapidly oxidizes GSH with an end point kinetics and a stoichiometry of one to one. Hydrogen peroxide and superoxide anion are scavenged in the RBC and no redox cycling is taking place. No GSH is regenerated even after long incubation periods. After the primary event, i.e., oxidation of GSH and —SH groups, a number of metabolic, rheologic, and membrane modifications, together with increased erythrophagocytosis take place in G6PD-deficient, D-treated RBC only. There is almost complete correspondence between D effects and RBC changes during the favic crisis. Among these effects, increased erythrophagocylosis, membrane cross-bonding, and formation of Heinz bodies seem to be causally related with the hemolytic process.

In vitro metabolic trends were assessed for 100 genotoxic agents detected in two bacterial test systems, i.e., the Ames reversion test and a liquid DNA-repair test in Escherichia coli. Seventy-five compounds were found to undergo a more or less pronounced decrease of genotoxicity, in at least one of these models, in the presence of rat liver homogenates or in other metabolic systems (up to 40 different preparations, from various sources, for chromium compounds). A number of these deactivable compounds are reported in the literature to yield negative or equivocal results in animal carcinogenicity assays, which may imply the existence of metabolically regulated thresholds in the initiation of cancer. Several examples are provided to support this hypothesis.

The in vitro treatment with a pharmacologic agent (N-acetylcysteine) enhanced detoxification mechanisms, either by stimulating enzyme activities promoting glutathione formation in liver homogenates or by reacting itself with direct-acting mutagens and with the genotoxic metabolites of procarcinogens.

Rats were treated with N-nitrosomorpholin and subsequently with phenobarbital. After several weeks of phenobarbital treatment, foci of altered cells and tumors developed in the livers of the animals. In liver tumors, the activity of glutathione S-transferase was higher than in the surrounding tissue and remained elevated even after the withdrawal of the promoter. Like the glutathio-one S-transferase from normal liver, tumor glutathione S-transferase activity is composed of several, predominantly basic, isoenzymes with an apparent molecular weight of approximately 45,000.

Gangliosides are ubiquitous plasma membrane components whose structural characteristics make it possible to establish multiple interactions with the pericellular microenvironment. Several receptorial activities of gangliosides have already been recognized and the possible involvement of gangliosides in growth control, adhesion, differentiation, and immunologic recognition is at present under study. The study of the role of gangliosides in cell growth control has been approached in our laboratory by investigating whether the ganglioside changes found in SV40-transformed Balb/c3T3 cells (SV3T3 cells) revert to a normal pattern in a variant of SV3T3 cells with growth properties similar to those of normal parental cells, the concanavalin A-selected SV3T3 revertant cells. These latter cells showed the same reduction of the more complex gangliosides as found in SV3T3 cells, while their amount of II³NeuAC-LacCer (G_M3) was greatly increased compared to that found in either normal or transformed 3T3 cells. In order to study the role of gangliosides in the adhesion process, we analyzed the ganglioside structure of the so-called substrate-attached material, a cell surface structure involved in the adhesion of cells to substrate, from cultures of Balb/c3T3, SV3T3, and concanavalin A-selected SV3T3 revertant cells and from cultures of a system of transformed cells with different metastatic potential: the nonmetastatic B77-3T3 and the highly metastatic AA6 cells. Compared to normal cells, all the transformed cells contained smaller quantities of gangliosides in their substrate-attached material. In both normal and transformed cell lines, the gangliosides of the substrate-attached material, compared to those of detached cells, were found to be enriched by IV³NeuAC, II³NeuAc-GgOse₄Cer (G_D1a); the only exception were the AA6 cells which contained rather low levels of G_D1a in both detached cells and substrate-attached material. When ganglioside compositions of the substrate-attached material, detached cells, and plasma membranes were compared in Balb/c3T3, SV3T3, and concanavalin A-selected SV3T3 revertant cells, a prevalence of G_D1a was noted in the substrate-attached material and not in the plasma membranes. Moreover, the level of G_D1a was scarcely modified in the substrate-attached material from cultures of a neuroblastoma cell line after growth in media containing chelating agents, a treatment which is known to modify the composition of cellular gangliosides.

The overall indications emerging from these findings suggest a role for G_M3 in growth control and a role for complex gangliosides in the adhesion process.

Ethanol in amounts equivalent to a man drinking a pint of beer has a dramatic effect on the metabolism and distribution of nitrosamines in rats. It prevents the first pass clearance of dimethylnitrosamine and thus exposes the extrahepatic organs to oral doses of this carcinogen. By selectively inhibiting metabolism in liver and kidney, ethanol increases the amount of diethylnitrosamine activated in the esophagus between 1.8- and 4.6-fold. It is suggested that there may be a link between these observations and the increase in human esophageal cancer which is associated with alcohol consumption.

To study the interactions between chemical carcinogens and oncogenic retroviruses, BALB/Mo mice which carry the Moloney murine leukemia virus (M-MuLV) as an endogenous virus, and conventional (M-MuLV-free) BALB/c mice, as well as their Bc1 (M-MuLV+ or M-MuLV−) hybrids were injected neonatally with a single dose of urethane. BALB/Mo and V+ Bc1 mice showed accelerated lymphoma development; similar results were obtained in BALB/Mo mice receiving one or two doses of urethane transplacentally. Lung adenomas developed with shorter latency and higher incidence in BALB/Mo mice given urethane at birth; however, significant differences in the incidence of lung adenomas in BALB/Mo mice were found only in two experiments.

Additional short-term experiments were carried out to investigate the mechanism of the higher susceptibility to sister chromatid exchange induction observed in BALB/Mo lymphocytes. It was found that BALB/Mo spleen lymphocytes incubated with cordycepin, an antiviral antibiotic, with or without mitomycin C treatment, showed reduction in both M-MuLV synthesis and sister chromatid exchange frequency, and the latter values were similar to those seen in control cultures.

These data suggest that the integration of M-MuLV proviral DNA into the host genome is per se not sufficient to increase the susceptibility to carcinogenic stimuli, but that other events, such as viral gene expression and amplification, are most likely required for the chemical-viral synergistic effect to occur.

Several studies are reviewed dealing with the mechanisms which regulate the cell cycle progression in normal and cancer cells. Using Yoshida AH 130 ascites tumor cells, it has been found that the G₁-S transition of these cells is impaired by specific inhibitors of the electron flow through the respiratory chain (antimycin A), although respiratory ATP can be replaced by glycolytic ATP. The above transition can be also inhibited by the addition of physiologic substrates, mainly pyruvate, by a mechanism which appears linked to a modification of the cellular redox state and can be totally reversed by adding adenine to the culture medium. Adenine equally removes the block produced by antimycin A, pointing out a respiration-linked step of purine metabolism restricting the cell recruitment into S. A substantial protection of this step against the inhibitory effects of pyruvate and antimycin A has been obtained by the addition of folate and tetrahydrofolate, suggesting that the respiration-linked limiting step of tumor cell cycling involves folate metabolism and its connection to purine synthesis. The biologic relevance of these findings is stressed by the fact that pyruvate addition also inhibits the proliferation of concanavalin A-stimulated lymphocytes as well as of bone marrow hemopoietic cells in the presence of colony-stimulating factors. On the other hand, pyruvate only slightly affects the growth kinetics of malignant lymphoblasts and of Friend erythroleukemia cells either in the absence or in the presence of the differentiation inducer dimethylsulfoxide.

Different protocols of chemically induced hepatocarcinogenesis were applied to Wislar rats under identical experimental conditions. The following conclusions may be drawn after an analytic comparison of these results. Various chemical carcinogens show different carcinogenic capacities. Diethylnitrosamine is more potent than N-nitrosomorpholine which is more active than 2-acetylaminofluorene. A short-term exposure to such carcinogens is sufficient to initiate but not necessarily to complete the carcinogenic process. It can be promoted to completion by either a noncarcinogenic promoter or a carcinogen. From a systemic point of view, it appears that, as in the skin models, two-step protocols are not always equivalent to protocols using the same agent during the whole treatment. Moreover, the results observed with a multistep protocol indicate that during the initiating phase the carcinogen plays a selective role distinct both from a pure initiating role and from the promoting effect. The results obtained lead to the conclusion that the distinction between initiation and promotion remains purely operational as it still does not correspond to the nature of well-established biologic processes.

Mutagenesis and neoplastic transformation assays on mammalian cells in culture have been extensively used for quantitative estimates of the activity of carcinogens, in spite of the limitations that such in vitro systems have when compared with in vivo systems for tumor induction. In order to assess the validity of these correlations, a series of studies was undertaken in our laboratory with the BALB/3T3 Cl A31-1-1 mouse embryo cell line. Different carcinogens were found to induce dose-dependent frequencies of transformation, including the direct-acting alkylating agent N-methyI-N'-nitro-N-nitrosoguanidine (MNNG) and carcinogens that were metabolically activated by these cells through different pathways (benzo[a]pyrene, 3-methylcholanthrene, aflatoxin B₁, and benzidine). Their respective level of activity on a molar basis was different from that obtained in standard Salmonella + S9 mutagenesis tests. Studies currently underway indicate the possibility of lowering the serum content in the medium considerably, thereby reducing a major variable in the assay. Methods were established for the induction of ouabain-resistant (oua^r) mutants in these cells. Studies were conducted by applying 30-min MNNG exposures to cells that were synchronized by serum deprivation followed by serum-induced release from growth block. While maximal induction of mutants occurred in the S phase, the transformation frequency remained constant for treatments in G₁ and early or late S. In subsequent studies, cytotoxicity, alkali-labile DNA lesions, oua^r mutations, and neoplastic transformation were analyzed concurrently in this cell line after cells were exposed to two concentrations of MNNG and the exposures were protracted for different time periods (30, 60, 90, 120, and 240 min; 24, 48, and 72 hr). A marked temporal dissociation was found in the exposure times required to induce maximal frequencies of mutations and of transformation. Cytotoxicity increased for periods up to 100-200 min; mutations reached a maximal induction level after a much shorter exposure time (30-60 min); DNA damage detected by alkaline elution was already maximal by 30 min. Transformation frequencies, however, reached maximal levels only after exposure periods 1-3 hr longer than those required for maximal mutation. The ratio of transformation to oua^r mutation frequencies was 3.7 for short treatment times (30-60 min), but it increased to more than 20 for exposure times of 240 min or longer. These studies support the hypothesis that a single gene mutational event is not sufficient to account for the expression of neoplastic transformation.

One important early contribution to the control of chemical carcinogenesis is provided by the enzyme pattern responsible for the generation and disposition of reactive metabolites. Especially well studied is the important group of enzymes responsible for the control of reactive epoxides. Many natural as well as man-made foreign compounds, including Pharmaceuticals, possess olefinic or aromatic double bonds. Such compounds can be transformed to epoxides by microsomal monooxygenases present in very many mammalian organs. By virtue of their electrophilic reactivity such epoxides may spontaneously react with nucleophilic centers in the cell and thus covalently bind to DNA, RNA, and protein. Such alterations of critical cellular macromolecules may disturb the normal biochemistry of the cell and lead to cytotoxic, allergenic, mutagenic, and/or carcinogenic effects. Whether such effects will be manifested depends on one hand on the chemical reactivity as well as other properties (geometry, lipophilicity) of the epoxide in question. On the other hand, enzymes controlling the concentration of such epoxides are another important contributing factor. Several microsomal monooxygenases exist differing in activity and substrate specificity. With respect to large substrates, some monooxygenases preferentially attack at one specific site different from that attacked by others. Some of these pathways lead to reactive products, others are detoxification pathways. Moreover, enzymes metabolizing such epoxides represent a further determining factor. These enzymes include epoxide hydrolases and glutathione transferases. These enzymes do not play a pure inactivating role, but can in some cases also act as coactivating enzymes. Enzymes involved in biosynthesis and further metabolism of epoxides differ in quantity and sometimes also in substrate specificity between organs, developmental stages, sexes, and animal species. They therefore represent one important contributing factor to differences in susceptibilities between species and individuals.

Although both the epidemiologic and experimental studies have led to the identification of chemical carcinogens, the limitations in epidemiologic approaches and the need for primary prevention of cancer require a greater reliance on experimental studies. Long-term carcinogenicity studies in experimental animals have been instrumental in identifying chemicals with carcinogenic activity, and, in some cases, the experimental evidence has preceded the epidemiologic evidence (for 4-aminobiphenyl, aflatoxin B₁, diethylstilbestrol, melphalan, mustard gas, and vinyl chloride). A better understanding of the multistage process of carcinogenesis and the findings from various short-term tests available more recently may provide a more solid basis for extrapolating experimental findings to man.

The starting point of this study is the observation that succinate, the well-known Krebs cycle intermediate, strongly inhibits the incorporation of amino acids into protein of tissue slices. The results presented in this paper show that this somewhat peculiar succinate effect, which is present in regenerating liver and in a well-differentiated hepatoma and absent in two anaplastic hepatomas, is well-marked in all stages of hepatic carcinogenesis by N,N'-dimethyl-4-aminoazobenzene. A working hypothesis is that the inhibition by succinate of the amino acid incorporation into protein is mediated by a shift of the redox level in the cell toward a more reduced condition.