THE IN OVO CARCINOGENEICITY ASSAY (IOCA): A REVIEW OF AN EXPERIMENTAL APPROACH FOR RESEARCH ON CARCINOGENESIS AND CARCINOGENICITY TESTING

Harald Enzmann¹ and Klaus D. Brunnemann²

1Bayer AG, Institute of Toxicology, 42096 Wuppertal, Germany and ²American Health Foundation, Valhalla, NY 10595, USA

Received 7/21/97 Accepted 11/14/97

3. CARCINOGEN-INDUCED CELLULAR CHANGES IN EMBRYONIC AVIAN LIVER

3.1 Induction of adenoid formations by carcinogens

Exposure of avian embryonic liver to the hepatocarcinogenic nitrosamines N-nitrosomorpholine (NNM) and diethylnitrosamine (DEN) resulted in the occurrence of tubular (adenoid, ductular) formations instead of the physiological trabecular organization of the embryonic liver (10, 11). Extracellular deposits of a brownish substance (bile pigment) were occasionally observed in these cell populations. Neither basal membranes nor fibrillary structures were detectable at the luminar cell pole. After high doses of the nitrosamines, the whole liver was composed of a mixture of small basophilic hepatocytes and large acidophilic hepatocytes including cells with a ground glass appearance. A similar mixture of large acidophilic hepatocytes and small basophilic hepatocytes has long been shown in the liver after exposure of rats to the very same nitrosamines (12). The occurrence of tubular structures is a typical histological feature of adenoid hepatocellular tumors (13), but may also be observed in severely damaged human liver (14,15).

3.2 Carcinogen-induced hepatocellular nuclear enlargement

The cell and nuclear size of the hepatocytes in embryonic liver is severely increased after exposure to chemical carcinogens. This effect on the nuclear size is dose dependent (table 1) and occurs at doses lower than the doses required for the induction of the tubular formations or preneoplastic focal lesions (see below). Single hepatocytes with unusually large nuclei (figure 1) are observed after exposure to low doses that fail to elicit cytotoxicity in the embryonic liver. At low doses, only a small minority of hepatocytes exhibit enlarged nuclei after exposure to low doses of carcinogens whereas the vast majority of nuclei are apparently normal. Therefore, the 95% quantile or the 99% quantile rather than the arithmetic mean or the median of a sample is the most appropriate means of assessing the exposure-induced nuclear size differences.

Figure 1: Effect of diethylnitrosamine on nuclear size: obviously enlarged hepatocyte nucleus (center) after exposure to 0.4 mg of diethylnitrosamine. Hematoxylin eosin stain, objective magnification 40-fold.

3.3 Quantification of carcinogen-induced nuclear enlargement

The measurement of cell and nuclear size may be performed using flow cytometry (16). Most hepatotoxins, however, cause qualitatively and quantitatively different damage to periportal and perivenous hepatocytes (17,18) and cell suspensions are not suitable for the discrimination of periportal and perivenous hepatocytes. In addition, cell suspensions do not allow an exact distinction between cells of the carcinogen-induced focal lesions (see below) and the surrounding extrafocal tissue. Therefore, we preferred to measure nuclear size in the tissue sections. There were no differences in nuclear size between periportal and perivenous hepatocytes in untreated control embryos (table 2).

It is worth to consider that thickness of the section may be a crucial point in the planimetric measurement of nuclear areas. The measured areas of peripherally cut nuclei (small areas) strongly depends on the thickness of the section. Focusing the microscope for the maximal nuclear area will have the consequence that with an infinite increase in section thickness no peripherally cut nuclei, but exclusively the maximal (central) area of nuclei is measured. On the other hand, the ideally central profiles will give the maximal area values both in infinitely thin and in infinitely thick sections. Therefore, whereas small profiles may be underrepresented in the measured data and their number may be biased by the thickness of the section, the maximal area values will hardly be biased by variations in thickness of the section, if enough profiles per sample are measured.

3.4 Significance of the carcinogen-induced nuclear enlargement

The size of liver cell nuclei may reflect various physiological and pathological processes. In rodents, distinct alterations in the ploidy pattern are characteristic during postnatal development (19). Alterations of the nuclear size of parenchymal cells have frequently been observed during various stages of hepatocarcinogenesis. Various authors reported the occurrence of enlarged or hyperploid nuclei at early stages of the carcinogenic process (20-24), effects of the carcinogenic treatment on the number of binucleated hepatocytes (24-27) or proposed in vitro test systems based on the carcinogen-induced nuclear enlargement (28).

At least some of these carcinogen-induced phenomena seemed also to be linked to liver regeneration. Regeneration after partial hepatectomy (29-31) or after toxic injury (32) results in a shift towards higher ploidy levels and a reduced number of binucleated hepatocytes. Consequently, the occurrence of enlarged nuclei during chemically induced hepatocarcinogenesis was regarded as to be due to the combination of regenerative and toxic effects (33). More recent experiments, however, showed that enlarged nuclei occurred after low doses of carcinogens that failed to induce other signs of nonspecific toxic effects (34). Nuclei with unusually high ploidy level were observed in LEC rats with hereditary hepatitis (35) before these rats developed preneoplastic liver cell foci and hepatocellular carcinomas without any deliberate exposure to carcinogens (36). According to Clawson and coworkers an enlargement of nuclei was also observed after, low, and essentially non-toxic doses of mutagenic carcinogens (37). Similarly, effects on the nuclear size were observed in the IOCA even after small doses that did not induce foci of altered hepatocytes and did not induce nonspecific toxic effects, such as cell death, loss of glycogen or fatty change. This observation is not compatible with the assumption, that the enlarged nuclei merely reflect a compensatory regeneration due to toxic cell loss. Both the occurrence of aneuploid or hyperploid nuclei caused by nonspecific effects on the mitotic spindle (38) and a carcinogen-mediated acceleration of the naturally occurring polyploidization program (39) may account for the observed effects.

These findings suggest that the occurrence of enlarged nuclei at early stages of experimental carcinogenesis, induced by low doses of chemical carcinogens indicate cellular changes, that may be involved in the carcinogenic process.