INDUCTION OF FOCI OF ALTERED, y-GLUTAMYLTRANSPEPTIDASE- POSITIVE HEPATOCYTES IN CARCINOGEN-TREATED RATS FED A CHOLINE-DEFICIENT DIET

Three recent lawsuits are focusing public attention on the environmental and occupational health effects of the world's largest and fastest growing manufacturing sector-the $150 billion semiconductor industry. The suits allege that exposure to toxic chemicals in semiconductor manufacturing plants led to adverse health effects such as miscarriage and cancer among workers. To manufacture computer components, the semiconductor industry uses large amounts of hazardous chemicals including hydrochloric acid, toxic metals and gases, and volatile solvents. Little is known about the long-term health consequences of exposure to chemicals by semiconductor workers. According to industry critics, the semiconductor industry also adversely impacts the environment, causing groundwater and air pollution and generating toxic waste as a by-product of the semiconductor manufacturing process. In contrast, the U.S. Bureau of Statistics shows the semiconductor industry as having a worker illness rate of about one-third of the average of all manufacturers, and advocates defend the industry, pointing to recent research collaborations and product replacement as proof that semiconductor manufacturers adequately protect both their employees and the environment.

IT HAS BEEN SHOWN conclusively in recent years that the process of chemical carcinogenesis in the liver consists of at least 2 basic, probably multi-step, stage3: initiation and promotion (Peraino et al., 1973;Pitot et al., 1978). The initial interaction of the carcinogen with target cells generates a small population of initiated cells with the acquired property or properties of evolving eventually into neoplastic cells. Establishment of the latter, however, depends upon the intervention of factor(s) which promote specifically the proliferation and further evolution of initiated cells. In the case of hepatocarcinoma induction, a sequential emergence in the liver occurs, as a consequence of initiation and promotion, of successfully smaller populations of hepatocytes displaying phenotypic changes, through which the entire process evolves from normality to preneoplastic lesions and malignancy (Farber, 1976). Among the induced phenotypic alterations are changes in enzyme activities so characteristic as to represent reliable histochemical markers for the identification of the newly emerging populations of liver cells. Thus, islands of hepatocytes deficient in ATPase and/or glucose-6-phosphatase activities have been shown to appear within a short time after exposure of experimental animals to several chemical hepatocarcinogens (Scherer et al., 1972). Scherer & Emmelot (1975, 1976 have devised an efficient and reproducible pro-cedure for the induction of these islands, and have used it to study and define several of their properties. In the procedure, a single dose of diethylnitrosamine (DEN) a potent hepatocarcinogen, is administered to rats 20 h after partial hepatectomy (PH). XWithin 8 weeks, numerous microscopic islands of altered hepatocytes are formed which, after histochemical staining, can be easily identified, quantitated and sized. The number and size of the islands have been shown to be proportional to the dose of DEN administered, the enzyme deficiencies to persist in the island cells for many months, and the islands to be associated with the ultimate appearance of hepatocellular carcinomas, even though they can be induced also by single, low, noncarcinogenic doses of DEN.
More recently, a different procedure has been described by Solt & Farber (1976). In this, a single, large dose of DEN (or of other chemical carcinogens) is administered to intact rats in order to initiate liver cells. Proliferation of the latter is then selectively stimulated by application of a "selection pressure" consisting of 2 components, PH, and administration of 2-acetylaminofluorene (AAF). The "selection pressure" is predicated on the fact that initiated liver cells also acquire resistance to the cytotoxic and cytostatic effects of chemical carcinogens such as AAF, whilst uninitiated cells remain sensitive to those effects (Solt et al., 1977). Thus, the intense and generalized stimulus for cell growth provided by PH is applied only to initiated cells. Indeed, within 10 days after PH, numerous foci and minute but visible nodules of altered hepatocytes are formed in rats previously injected with DEN. The rapidity with which the foci emerge, and the size they attain, is most probably due to the fact that Solt & Farber's procedure evokes not only the steps of the initiation stage of chemical hepatocarcinogenesis, but also stimulates or heightens early steps of the promotion stage. The elicited foci and nodules show a high activity of the enzyme y-glutamyl-transpeptidase (E.C. 2.3.7.2, (GrT) in contrast to normal hepatocytes and hepatocytes surrounding the foci and nodules, which express no histochemical activity of the enzyme. Identification, quantitation and sizing of the foci and minute nodules can thus be performed easily after histochemical staining of GGT. Even though the procedure suffers from some shortcomings, primarily due to the use of 2 different carcinogens, it has already proved to be a verv useful tool in the analysis of the early steps and cellular events of chemical hepatocarcinogenesis (Cayama et al., 1978;Laishes & Farber, 1978).
Feeding a choline-deficient (CD) diet to rats markedly enhances hepatoma induction by chemical carcinogens (Lombardi & Shinozuka, 1978;Shinozuka et al., 1978a, b). In a previous study  we used Solt & Farber's procedure to investigate wl-hether the CD diet modifies the initiation stage and/or the promotion stage of liver carcinogenesis. Whereas the diet appeared to have no effect on initiation, evidence was obtained that it has a strong promoting action. Indeed, after administration of a single initiator dose of DEN, a selection pressure consisting of feeding a CD diet containing AAF was found to promote the proliferation of initiated cells, and their evolution to foci or altered hepatocytes, as effectively as the selective pressure used by Solt & Farber (1976). In other words, the diet could effectively replace PH as one of the selection pressures. In this paper we report the results of further studies showing that feeding a CD diet alone, without any other manipulation of the animals, can efficiently promote the proliferation of initiated liver cells and their evolution to foci of altered hepatocytes. controls. Food and water were given ad libitumn. The animals were maintained on laboratory chow (Purina, Ralston Purina Co., St Louis, MO) for at least one week before the beginning of the experiments. Semisynthetic, semipurified CD and choline-supplemented (CS) diets were prepared as previously described (Shinozuka et al., 1978c). DEN (Aldrich Chemicals, Madison, WI) was dissolved in a 0.90o NaCl solution and was administered i.p. PHs were performed according to the method of Higgins & Anderson (1931). The animals were fasted overnight before being killed, and anaesthetized with pentobarbital (60 mg/kg body wt). A blood sample was withdrawn from the abdominal aorta, and the liver was resected and weighed. Serum aspartic aminotransferase (E.C.2.6.1.1, SGOT), plasma GGT and ox-foetoprotein (AFP) and liver AFP wAere determined as previously indicated (Shinozuka et al., 1978c;Sell & Gord, 1973). Blocks of liver tissue wvere fixed in Stieve's solution, and sections were stained with haematoxylineosin (HE) for light-microscopic examination. Histochemical localization of GGT was performed by the method of Rutenberg et al. (1969). Blocks of liver were quickly frozen in dry ice and 5,um cryostat sections were cut and fixed in acetone. The sections were incubated in freshly prepared medium containing y-glutamyl-4-methoxy-2-naphthylamide as substrate, and fast blue-BBN as coupling agents. The sections were counterstained writh haematoxylin. Control sections incubated in a similar medium without substrate were uniformly negative. All reagents were obtained from Sigma Chemical Corp., St Louis, MO. Foci of GGT+ hepatocytes with a diameter larger than 125 Htm were counted, and their number per cm2 of sections determined. The surface area of the sections was measured with the aid of a planimeter, and the size of the foci with the aid of an AOocular micrometer. The number of foci per liver was calculated as indicated by Scherer et al. (1972). Immunofluoreseence staining of liver sections for AFP and albumin was performed as previously described (Sell, 1978). Differences betw!een the means were evaluated statistically by Student's t test and regarded as significant if P < 0 05. Experimental dosiqns. Five experiments Nere performed, and the basic design of the experiments is presented diagrammatically in Fig. 1. Experiments 1-4 were designed to test whether feeding the CD diet alone, without subjecting the animals to other manipulations or feeding of AAF, could promote the proliferation of initiated cells, and their evolution to foci of altered, GGT+ hepatocytes. DEN, 150 or 30 mg/kg body w%t, was administered to intact animals in Expts 1 and 2, respectively. In Expts 3 and 4, 30 mg or 15 mg DEN/kg of body weight, respectively, were administered 18 h after PH. Then, all animals were fed laboratory chow for 1 week, in order to allow recovery of the liver from any toxic effect of the carcinogen. At this point, the animals in each experiment were divided into 2 groups, one of which was fed a plain CS diet, and the other a plain CD diet. Subgroups of 3-4 rats were killed after 2-8 weeks of feeding the CS or CD diet.
Expt 5 was designed to test whether the CD diet could substitute for AAF as a selective agent in Solt and Farber's procedure, in which PH follows DEN exposure. DEN, 30 mg/kg body wt, was administered to intact rats. After being fed laboratory chow for 1 week, the animals were divided into 2 groups, one of' which was fed a plain CS diet, and the other a plain CD diet. One week later, PH was performed on all animals, and 3-4 rats were killed at the end of the 2nd and 4th week of feeding the CS or CD diet.

RESULTS
The results of all experiments are shown in Table I. Within each experiment there was no statistically significant difference in body weight of rats fed the CS or CD diet. However, the liver weight of rats fed the CD diet was consistently about double that of rats fed the CS diet. This finding is most probably accounted for by the accumulation of fat that occurs in the liver of rats on a CD diet (Lombardi, 1971).
The hepatocYtes in the foci were arranged irregtularly wTithout a distinct sinusoidal platterin and in plat,es more than one cell thick, had slightly basophilic cyt,oplasm, contained an occasional small droplet of fat atnd displayed frequent mitosis (Fig. 2). Foci of hepatocvtes identical to these also (levelopedI in the liver of the rats in Expts 2 annd 4 after 4 wxeeks of feeding the CD diet. Iti general, the nutimber of foci increased over the perio(l of feeding the diet. However, gross alterattions of the surface of the liver were present in none of the animals. In rats administered 15 mg DEN/kg after PH (Expt 4) and killed after 8 weeks on the CD diet, a slight periportal fibrosis was present, in addition to many foci of basophilic, fat-free hepatocytes. Oval cells and/ or intermediate cells were frequently seen scattered throughout the parenchyma and often at, the periphery of the foci of basophilic, fat-free hepatocytes (Fig. 3). In Expts 1, 2, 4 and 5, the liver of rats on the CS diet showed a well preserved lobular architecture and fatty change, necrosis, inflammation or fibrosis were never seen. However, foci of basophilic hepatocytes similar to those observed in the liver of rats fed the CD diet were observed in rats fed the CS diet in Exp. 3.

Histochenmistry
After histochemical staining of liver sections, foci of CGxUxT+ hepatocytes were readily seen in rats fed the CD diet but, in those fed the CS diet, only in Expt 3 (Fig. 4). In rats fed the CD diet, the hepatocytes in the foci, in addition to the GGT positivity, also showed little or no fatty change. The latter property indicates that the GGT+ foci correspond to the foci of basophilic hepatocytes observed in the HE-stained sections. Data on the quantitation and diameter of the GGT+ foci are presented in Table I. Significantly more foci developed when the single, initiator dose of DEN was administered to the rats 18 h after PH (Expts 3 and 4) than when it was administered to intact rats (Expts 1 and 2). Actually few foci developed in the liver of rats fed the CS diet only in Expt 3, in which the animals received 30 mg DEN/ kg after PH. After administration of 30 mg DEN/kg to intact rats, performance of PH 1 week after the rats had been on the CD diet slightly increased the number of foci formed (Expt 5 vs Expt 2). In general, the number as well as the size of the foci increased with increasing time on the CD diet. Bile-duct and ductular cells, and oval and/or intermediate cells were also intensely GGT+. No foci of GGT+ hepatocytes have been seen so far to develop in the liver of rats subjected to no other experimental manipulation than feeding either the CS or the CD diet for up to 6 months.

IDIS( U SS1ON
The results of the present study show clearly that, after administration to rats of a single initiator dose of a chemical hepatocarcinogen, feeding the animals with a (ID diet efficiently and reproducibly promotes in a relatively short period of time the proliferation of initiated liver cells, anid their evolution to foci of altered hepatocvtes. In Solt & Farber's procedure (1976), such a promoting action is achieved via 2 effects, a growth stimulus for initiated cells provided by PH or other mitogenic agents, and selective inhibition of the proliferation of uninitiated cells by exposing the animals to AAF. It is evident, therefore, that similar effects can also be induced by a diet devoid of choline without any further manipulation of the animals. Indeed, in another study  a CD diet containing AAF was found to have as strong a promoting action as PH and exposure to AAF; that is, the CD diet could effectively r eplace PH in the original procedure of Solt and Farber. The results of Expts 2 and 5 of the present paper show that the diet can also replace AAF, but less efficiently, since the nuimber of foci/cm2 of liver sectioni in Expt 5 was only slightlv more than that in Expt 2 (Table I). This fact may explain why 2-4 weeks on the CD diet are requiired to induce foci of altered hepatocytes, in contrast to the 7-10 days required in Solt and Farber's procedure.
Unlike the CD diet, the CS diet has no promoting action. Indeed, no foci, or only a minimum number of foci of altered hepatocytes were seen in the present study in rats injected with a single dose of DEN, and then fed the CS diet for various periods of time. Irrespective of the diet fed, the number of foci induced in the liver was greater when an initiator dose of 30 mg DEN/kg body weight was administered 1 8 h after P1 rather than to intact rats (Table I, Expts 2 and 3). This finding is consistent with that of other studies (Scherer & Emmelot, 1976) and may be explained on the basis of a recent report by Cayama et al. (1978) that, the ntumber of liver cells initiated by a single dose of a chemical carcinogen is maximal when the chemical is administered 18 h after PH. An increase in number as well as in size of the induced foci was found to occur with increasing duiration on the diets. The increase in number may be due either to an increasing fraction of initiated cells being promoted, with time, to proliferate and evolve or, more probably, to more foci attaining the minimum scoring diameter of 125 /m. The increase in diameter is undoubtedly due to cell proliferation. Indeed, in HEstained sections, hepatocytes in mitosis were frequently seen in the foci (Fig. 2). Despite no histological evidence of hepatocyte necrosis, SGOT levels in rats fed the CD diet were higher than in rats fed the CS diet (Table II). A low rate of liver cell death could represent one possible mechanism underlying the promoting action of the diet . On the whole, levels of plasma GUGT showed good correlation with the number of foci of (GGT hepatocytes in the livers. Therefore the sertum enzyme levels could be used to monitor the emergence, number and further evolution of the foci, were it not that GGT is also present in, and possibly secreted bv, bile ductular and duct cells and oval and/or intermediate cells. On the other hand, no correlation was found to exist between serum AFP levels and number of foci in the liver and, by immunofluorescence staining, the cells in the foci were found to be consistently AFP-. The sefindings are consistent with those made and discussed in another study . In the present experiments, the foci of altered hepatocytes that developed in the liver of rats on the CD diet had properties very similar to those described by Solt & Farber (1976) and Solt et al. (1977) for the foci induced with their procedure, such as arrangement of the hepatocytes in plates more than one cell thick, basophilia of the cytoplasm, GGCT positivity, and a high mitotic rate. Administration of a single initiator dose of 30 mg DEN/kg 18 h after PH, followed by feeding the CD diet for 4 weeks, led to the formation of the largest number of foci, 27/cm2 of liver section, or 2 5 x 104/liver (Table I, Expt 3). These numbers are certainly adequate for carrying out kinetic and other types of study on the origin, development and further evolution of the foci. Large numbers might be induced, if needed, by administration of a larger dose of DEN, since it has been shown that there is a direct relationship between the size of the initial dose of DEN and the number of foci formed in the liver (Solt & Farber, 1976;Shinozuka et al., 1979). It is evident, therefore, that administration of a single initiator dose of a chemical carcinogen, 18 h after PH, followed by feeding laboratory chow for 1 week and the CD diet for varying lengths of time (depending mostly upon the size of the initiator dose and the number of foci/ liver required) constitutes a new and efficient procedure for the induction of foci of altered hepatocytes in rat liver. The niew procedure appears to have at least one major advantage over the original procedure of Solt and Farber, namely, that it requires the use of a simpler selection pressure, one avoiding the need for application of a strong mitogenic stimulus such as PH or other mitogenic chemicals and, more importantly, the use of AAF, a second chemical hepatocarcinogen. Another advantage, of secondary importance, is the availability of a new marker for the identification of the foci, namely, the absolute or relative absence of fatty change. In the experiments reported in this paper, as well as in those described elsewhere , identification of foci of basophilic hepatocytes in HE-stained sections of livers from rats on the CD diet was consistently made easy by the fact that the basophilic hepatocytes had no or very little fatty change, in contrast to the surrounding hepatocytes (Fig. 1). The same property was shown in histochemically stained sections by the hepatocytes of the GGT+ foci, indicating the identity of the foci. The new procedure, used alone or in conjuinction with that of Solt and Farber, may prove to be a valuable, alternative or additional tool in studies concerned with the analysis of the early steps and cellular events of chemical hepatocarcinogenesis, and the mechanism(s) underlying promotion of liver carcinogenesis.