It was not until people lived long enough to die of cancer that an association between exposure to certain chemicals and development of cancer was noted. In 1761, John Hill identified the use of snuff as the probable cause of an increased incidence of cancers of the nasal cavity and in 1775, Percival Pott noted the early appearance and high incidence of scrotal cancer in chimney sweeps who had poor personal hygiene. Yamagiwa and Ichikawa were the first to induce cancers in animals in 1915 by the application of coal tar to the ears of rabbits. In 1935 Sasaki and Yoshida observed liver cancers in rats treated with o-aminoazotoluene. Since then, the skin and liver have become the major organs studied for experimental induction of malignant tumours.

These model systems have been analysed extensively in regard to the nature of the cellular changes that take place during development of cancer. One of the most compelling questions regarding the origin of cancer is whether mature cells in an organ dedifferentiate or whether cancer arises from transformation of pluripotent ‘stem’ cells. This question will be addressed from the standpoint of the sequence of cellular changes that occur in the liver during the induction of cancer of the liver by chemicals.

The cellular changes that occur during hepatocarcinogenesis provide evidence for both dedifferentiation and stem cell models of the origin of cancer, depending on the emphasis placed on importance of the different cell lineages seen. The sequence of cellular changes that supports the dedifferentiation theory of hepatocarcinogenesis consists of the appearance of small ‘foci’ containing a few large, mature hepatocytes that are more basophilic than normal hepatocytes and that express abnormal amounts of certain enzymes, such as increased &ggr;-glutamyl transpeptidase, epoxide hydrolase and glutathione S-transferase, and decreased levels of ATPase and &bgr;-glucuronidase. ‘Nodules’ then appear: these are composed of large hepatocytes that push aside the adjacent liver. Over time most of these nodules disappear or ‘remodel’; however, one or a few hepatocellular carcinomas eventually develop, at a time when most of the nodules are no longer detectable. This may take as long as 2 years. The apparent progression from foci to nodules to hepatocellular cancer has led to the concept that hepatocellular carcinomas arise through the action of chemicals on mature hepatocytes, eventually leading to dedifferentiation. This hypothesis is supported by the fact that chemical hepatocarcinogens are metabolized to active forms by adult hepatocytes, and that primary hepatocellular carcinomas induced by chemicals are often seen in remodelling nodules.

Although the dedifferentiation model has received most of the experimental attention over the last 30 years, there is increasing evidence to suggest that there is actually a different cellular lineage of chemically induced primary hepatocellular carcinoma. Stem cells are defined as multipotent cells that divide to produce one daughter cell that stays as a stem cell while the other daughter cell expresses a differentiated phenotype. Tissue stem cells are determined for differentiation to a specific cell type. The liver stem cell as represented by the oval cell has the capacity to differentiate into duct cells or hepatocytes. Small oval shaped cells appear at the beginning of the liver plate and extend progressively into the liver: many of these have characteristics of bile ducts, but also express liver proteins. At the same time that the foci and nodular changes described above are occurring there is also proliferation of these cells, which usually first appear in the periportal area before nodules are seen. These oval cells may arise from a putative periportal liver stem cell or from transitional duct cells. The identification of a true liver stem cell as the origin of hepatocellular carcinomas is complicated by the observation that chemical hepatocarcinogens also induce early proliferation of fat storing Ito cells, which are derived from mesenchyme and line the liver sinusoids. However, since the earliest proliferating periportal cells do not have the ultrastructural appearance of Ito cells, they are probably not precursors of primary hepatocellular carcinoma. Proliferating Ito cells do produce transforming growth factor &bgr;, however, and this may influence the differentiation of the proliferating oval cells or nodular cells.

The transitional duct cells are connected to the terminal bile duct (canal of Hering), which consists of two or three ductule cells surrounded by a basement membrane that connects larger bile ducts on one side and to transitional duct cells on the other. The transitional duct cells do not abut basement membranes; rather, they form a tight junction with hepatocytes and border one side of a bile canaliculus with a hepatocyte. Carcinogens may act directly on terminal duct cells or on adjacent stem cells that differentiate into terminal duct cells, or both. In any case, proliferation of periportal cells is seen very early after exposure to chemical carcinogens, and is followed by massive proliferation of terminal duct cells, both of which may be identified as oval cells. Monoclonal antibodies that recognize epitopes on different cells that appear during chemical hepatocarcinogenesis can be used to identify cells phenotypically belonging to the oval cell lineage in persistent nodules, in remodelling nodules, and between the nodules, as well as in carcinomas. Cells intermediate between oval cells and hepatocytes have been identified, not only using these monoclonal antibodies, but also by morphological and autoradiographic criteria. Differentiation of oval cells into hepatocytes also occurs within neoplastic nodules.

The possible origin of liver cancer from a stem cell casts serious doubt on the dedifferentiation theory of cancer. The liver has been used as an example of an organ in which there is normally very little cellular proliferation, but which is a common site of primary tumours, at least in Africa and southeast Asia. In general, the incidence of neoplasms is much higher in organs that have a high cellular turnover, such as blood-forming cells, skin, and gastrointestinal epithelial cells, than in organs with a low turnover rate, such as brain or bone. The presence of liver stem cells has been questioned as there are very few dividing cells in the normal liver; liver regeneration following chemical injury or partial hepatectomy occurs by division of mature hepatocytes. However, oval cell proliferation may be seen after severe liver injury, and there also appears to be a regular turnover of liver cells, with continued production of hepatocytes and bile duct cells arising at the portal zone of the liver and ‘streaming’ to the central zone. Liver-like stem cells have also been identified in the pancreas. Exposure of the liver to hepatocarcinogens results in proliferation of the liver stem cell population, with eventual malignant transformation. Although the nature of the malignant transformation event remains unknown, concentration of efforts upon this lineage should be more rewarding than continued study of foci and nodular cells.

Bucher NLR, Malt RA. Regeneration of liver and kidney. Boston: Little Brown and Co, 1971.

Dunsford HA, Karnasuta C, Hunt JM, Sell S. Different lineages of chemically induced hepatocellular carcinoma in rats defined by monoclonal antibodies. Cancer Res 1989; 49:4894 - 900.

Everts RP, Nagy P, Marsden E, Thorgeirsson SS. A precursor-product relationship between oval cells and hepatocytes in rat liver. Carcinogenesis 1989; 8:1737 - 40.

Evarts RP, Nakatsukasa H, Marsden ER, Hsia CC, Dunsford HA, Thorgeirsson SS. Cellular and molecular changes in the early stages of chemical hepatocarcinogenesis in the rat. Cancer Res 1990; 50:3439 - 44.

Farber E. Similarities in the sequence of early histologic changes induced in the liver of rats by ethionine, 2-acetylaminofluorene and 3-methyl - 4-dimethylaminoazobenzene. Cancer Res 1956; 16:142 - 8.

Farber E, Sarma DSR. Biology of disease: hepatocarcinogenesis: a dynamic cellular perspective. Lab Invest 1987; 56:4 - 22.

Goldsworthy TL, Hanigan MH, Pitot HC. Models of hepatocarcinogenesis in the rat: contrasts and comparisons. CRC Crit Rev Toxicol 1986; 17:61 - 89.

Grisham JW, Porta EA. Origin and fate of proliferating hepatic ductal cells in the rat: electron microscopic and autoradiographic studies. Exp Mol Pathol 1964; 2:242 - 61.

Hill J. Cautions against the immoderate use of snuff. London: Baldwin and Jackson 1761.

Opie EL. The pathogenesis of tumors in the liver produced by butter yellow. J Exp Med 1944; 80:231 - 46.

Pierce GB, Shikes R, Fink LM. Cancer: A Problem of Developmental Biology. Engelwood Cliffs: Prentice-Hall, Inc, 1978:84.

Pott P. Chirurgical Observations. London: Hawkes, Clarke and Collins, 1775: 48 - 63.

Sasaki Y, Yoshida T. Experimentelle erzeugung des Lebercarcinoms durch futterung mit o-Amidoazotoluol. Virchow's Arch (A) 1935; 295:175 - 200.

Scarpelli DG. Multipotent developmental capacity of cells in the adult animal. Lab Invest 1985; 52:331 - 3.

Sell S. Perspectives in cancer research: is there a liver stem cell? Cancer Res, 1990; 50: 3811 - 15.

Sell S, Leffert HL. An evaluation of cellular lineages in the pathogenesis of experimental hepatocellular carcinoma. Hepatology 1982; 2:77 - 86.

Sell S, Salman J. Light and electron microscopic autoradiographic analysis of proliferating cells during the early stages of chemical carcinogenesis in the rat induced by feeding N - 2-fluorenylacetamide in a choline deficient diet. Am J Pathol 1984; 114:287 - 300.

Teebor G. Sequential analysis of liver carcinogenesis. In: Becker FF, ed. Cancer: A Comprehensive Treatise. Vol 1. New York: Plenum 1975:345 - 52.

Tounier I, Legres L, Schoevaert D, Feldman G, Bernuau D. Cellular analysis of alpha-fetoprotein gene activation during carbon tetrachloride and d-galactosamine-induced acute liver injury in rats. Lab Invest 1988; 59:657 - 65.

Zajick G, Ariel I, Arber N. The streaming liver III. Littoral cells accompany the streaming hepatocyte. Liver 1988; 8:213 - 18.