Enhancement by thyroxine of gastric carcinogenesis induced by N-methyl-N'-nitro-N-nitrosoguanidine in Wistar rats.

The affects of L-thyroxine (T4) on the incidence and histology of gastric cancers induced by N-methyl-N'-nitro-N-nitrosoguanidine (MNNG), and on the labelling index of gastric mucosal epithelial cells were investigated in Wistar rats. After oral treatment with MNNG for 25 weeks, the rats received s.c. injections of T4 (0.2 microgram kg-1) in depot form every other day until the end of the experiment in Week 52. This long-term treatment with T4 significantly increased the incidence of gastric cancers in Week 52. However, it did not influence the histological type of the gastric cancers. It also caused significant increases in the labelling indices of the fundic and antral epithelial cells. These findings indicate that T4 enhances the development of gastric cancers, and that its effect may be related to its effect in increasing proliferation of gastric epithelial cells.

Summary The affects of L-thyroxine (T4) on the incidence and histology of gastric cancers induced by N-methyl-N'-nitro-N-nitrosoguanidine (MNNG), and on the labelling index of gastric mucosal epithelial cells were investigated in Wistar rats. After oral treatment with MNNG for 25 weeks, the rats received s.c. injections of T4 (0.2 jug kg-') in depot form every other day until the end of the experiment in Week 52. This long-term treatment with T4 significantly increased the incidence of gastric cancers in Week 52. However, it did not influence the histological type of the gastric cancers. It also caused significant increases in the labelling indices of the fundic and antral epithelial cells. These findings indicate that T4 enhances the development of gastric cancers, and that its effect may be related to its effect in increasing proliferation of gastric epithelial cells.
Thyroid hormones have important regulatory roles in the morphology and biochemistry of gastrointestinal mucosal cells (Hernandez et al., 1988). Long-term T4 administration increased the mitotic activity of the fundic 'stem' cells of the stomach and basal and secretagogue-stimulated acid secretion via its effect on parietal cell mass (Adeniyi & Olowookorun, 1989). We recently found that prolonged administration of T4 significantly increased the incidence of rat colon tumours induced by azoxymethane (Iishi et al., 1992). These findings suggest that T4 might influence gastric carcinogenesis. Therefore, in the present work, we examined the effect of treatment of rats with T4 on the development of gastric cancers in rats.

Animals
Sixty 6-week-old male Wistar rats (SLC, Shizuoka, Japan) were used in this study. They were housed in stainless steel suspended wire mesh cages, under controlled environmental conditions of 12 h light and 12 h darkness, 30-50% humidity, and 21-22'C. Each rat was given standard laboratory chow (Oriental Yeast, Tokyo, Japan) at 60 kcal per day.

Experimental design
The animals were given drinking water containing MNNG (25 ytg ml-'; Aldrich, Milwaukee, WI) for 25 weeks. The stock solution of MNNG was prepared at 1 mg ml1' in deionised water and kept in a cool, dark place and renewed every week. Just before administration to rats it was diluted to 25 tg ml-' with tap water. Rats were given 20 ml per day of MNNG solution each, supplied from bottles covered with aluminium foil to prevent photolysis of MNNG, and the solution was renewed every other day. Safety precautions were taken in use of MNNG. From Week 26, the rats were given normal tap water ad libitum and were randomly divided into two groups of 30 rats each. Group 1 was given s.c. injections of the vehicle, plain olive oil only, while Group 2 was given s.c. injections of T4 in depot form (Sigma, St. Louis, MO, USA; 0.2 jig kg-' body weight) in olive oil until Week 52. Injections were given at various sites every other day in a volume of 1 ml kg-' body weight between 2 and 3 p.m. Tissue sampling Animals that survived for more than 49 weeks were included in effective numbers, because the first tumour of the glandular stomach was found in a rat in Group 2 that died in Week 49. All surviving animals were killed at the end of the experiment in Week 52. All rats were autopsied, and the stomach and other organs were carefully examined. The stomach was opened along the greater curvature, pinned flat on a cork mat, and fixed in Zamboni's solution (Stefanini et al., 1967) for histological examination. The fixed stomach was cut into longitudinal, 3 mm wide strips. The specimens were embedded in paraffin, and 5 gm thick serial sections were stained with hematoxylin and eosin. Sections were examined without knowledge of which group they were from.

Classification ofgastric cancers
Histologically, adenocarcinomas were defined as lesions in which neoplastic cells had penetrated the muscularis mucosae to the submucosa or deeper layers. Adenocarcinomas were classified as very well-differentiated, well-differentiated, and poorly differentiated, as reported previously (Tatsuta et al., 1988b). Very well-differentiated adenocarcinoma: cancers showing atypical glandular structure with a tubular or papillary pattern and an arrangement of cells comparable to that enclosing normal gastrointestinal crypts (Figure la). Welldifferentiated adenocarcinoma: common type, less differentiated glands consisting of disorderly masses of atypical cells containing a small amount of intracellular mucin (Figure lb); mucinous carcinoma, active mucin secretion, often resulting in mucinous nodules with a large amount of extracellular mucin, with only a few isolated groups of tumour cells ( Figure Ic). Poorly differentiated adenocarcinoma: anaplastic carcinoma, highly anaplastic cells scattered individually with no typical glandular or tubular differentiation; signet-ring cell carcinoma, tumour cells with a large amount of intracellular mucin, giving the cells a signet-ring appearance ( Figure  Id).
Measurement of labelling index of gastric mucosa Five rats in each group were killed in experimental Weeks 30 and 52 to determine the labelling index of the gastric mucosa with an immunohistochemical analysis kit (Becton-Dickinson, Mountain View, CA) for assay of bromodeoxyuridine (BrdU) incorporation (Gratzner, 1982;Morstyn et al., 1983). For this purpose, the rats were starved for 12 h and then received s.c. injection of either 1 ml kg-' of olive oil (Group 1) or 0.2 tLg kg-' of T4 (Group 2). One hour later, Correspondence: H. lishi. the animals received an i.p. injection of BrdU (20 mg kg-'), and after another hour were killed with ether. The stomach was removed and fixed in 70% ethanol for 4 h. Sections of 3 ltm thickness were immersed in 2 N HCI solution for 30 min at room temperature, and then in 0.1 M Na2B407 to neutralise the acid. The sections were then stained with anti-BrdU monoclonal antibody (diluted 1:100) for 2 h at room temperature, washed, treated with biotin-conjugated horse anti-mouse antibody (at 1:200 dilution) for 30 min, and stained with avidin-biotin-peroxidase complex for 30 min. The reaction product was localised with 3,3'-diaminobenzidine tetrahydrochloride. Cells containing BrdU were identified by the presence of dark pigment over their nucleus.
The labelling index of the gastric mucosa were determined by counting BrdU-labelled and unlabelled cells in the zone of proliferating cells (Eastwood & Quimby, 1983) without knowledge of which group the sample was from. The zone of proliferating cells in the fundic mucosa was defined as a 250-iLm rectangular area between the highest and lowest labelled cells in well-oriented sections. Ten such rectangular areas in each rat were examined. In the antral mucosa, all cells below the highest labelled cells in each pit-gland column were regarded as being within the zone of proliferating cells. In each rat, 100 well-oriented columns of pits and glands were examined, and the labelling index was calculated as the number of BrdU-labelled cells/total number of cells within the zone of proliferating cells.
Measurements of serum T4 and T3 Serum T4 and triiodothyronine (T3) were measured in Weeks 30 and 52. For this purpose, five rats in each group were kept for 12 h without food, and then received an s.c. injection of either 1 ml kg-' of olive oil (Group 1) or 0.2 tLg kgof T4 (Group 2). Two hours later, they were anesthetised with ether and blood was obtained by cardiac puncture. The serum was separated and stored at -20°C. Within 1 week, the serum T4 and T3 were assayed with commercial radioimmunoassay kits (Gamma Coat T4 RIA kit and Gamma Coat T3 RIA kit). Statistical analysis Incidence and distribution of the different histological types of gastric cancers were analysed by the chi-square test or Fisher's exact probability test (Siegel, 1956). Other results were all analysed by one-way analysis of variance with Dunn's multiple comparison (Snedecor & Cochran, 1967;Miller, 1966). Data are shown as mean ± s.e.m. 'Significant' indicates a calculated P value of less than 0.05.

Results
Incidence and histological type ofgastric cancer The body weight, food consumption and incidence, numbers and histological types of gastric cancers in each group are summarised in Table I. In Week 52, animals that received T4 (Group 2) had significantly lower body weights than untreated rats. Table I also shows that there were no significant differences between the food consumption in the two groups in Weeks 30 and 45. In Group 1 (olive oil), four gastric cancers were found in four (20%) of 20 rats examined. In Group 2 (T4), 12 gastric cancers were found in 11 (55%) of 20 rats examined. The incidence of gastric cancers in Group 2 was significantly greater than that in Group 1. All the tumours induced in the glandular stomach were identified histologically as adenocarcinomas. Almost all were very welldifferentiated, and neither mucinous carcinomas or poorly differentiated cancers were found in this series. There was no significant difference in the histological types of adenocarcinomas in the two groups: all cancers in Group 1 were very well-differentiated, while in Group 2 very well-differentiated adenocarcinomas were found in 11 (92%) of 12 tumours and the other was well-differentiated. All cancers were found in the antral mucosa, and no metastases were detected at the macroscopic and/or microscopic level.
Labelling index of gastric mucosa and serum 14 and T3 levels showed significantly higher labelling indices and increased number of cells in the zone of proliferating cells in both the fundic and antral mucosa than Group 1 (olive oil). Table III shows that at both times, administration of T4 significantly increased the serum levels of T4 and T3.

Discussion
In the present work, we found that T4 enhanced gastric carcinogenesis induced by MNNG in Wistar rats. Long-term s.c. administration of T4 in depot form significantly increased the incidence of gastric cancer, but had no influence on their histological type at autopsy in Week 52.   The mechanism of this effect is not known, but three possible explanations may be considered. One is an immunomodulatory role of T4. Thyroid hormones have immunostimulatory (Pierpaoli et al., 1969;Bachman & Marshaly, 1986) and immunosuppressive (Gupta et al., 1983) effects on the lymphocyte population. As observed after other endocrine treatments, e.g., with estrogen or cortisone in mice (Milisauskas et al., 1983;Hochman & Cudkowicz, 1979), the inhibition of natural killer cell activity caused by long-term T4 administration might be ascribed to the induction of suppressor cells.
A second possibility is an influence of T4 on the secretion and/or synthesis of regulatory peptides such as epidermal growth factor (EGF) and somatostatin. Walker et al. (1981) found that 5 and 10 days treatments with T4 significantly increased the EGF concentration in the submaxillary gland of mice. EGF is a well characterised polypeptide that exhibits mitogenic effects on a wide range of cell types after binding to specific transmembrane receptors (Cohen, 1983). As EGF can stimulate mucosal growth throughout the gastrointestinal tract, it has been suggested as playing a role in gastrointestinal carcinogenesis. Yasui et al. (1990) found that prolonged s.c. administration of EGF significantly increased the incidence of gastric cancers induced by MNNG. A high level of T4 was shown to increase the secretion and hypothalamic content of somatostatin in vivo and in vitro (Berelowitz et al., 1980). We found previously that prolonged s.c. administration of somatostatin significantly increased the incidence and number of gastric cancers (Tatsuta et al., 1989).
A third possibility is an acceleration of cell proliferation by T4. de Launoit and Kiss (1989) found that T4 dramatically stimulated the cell division by MXT (mouse) and MCF-7 (human) mammary cancer cell lines. In mammary tumorigenesis, T4 may exert its stimulatory effect at the level of the nuclear chromatin receptors (Oppenheimer, 1979). In the present work, we found that administration of T4 significantly increased the number of cells in the zone of proliferating cells and the labelling index of the antral and fundic epithelial cells. Adeniyi and Olowookorun (1989) found that chronic T4 administration significantly increased the mucosal thickness and volume, the parietal cell count per unit mass of the glandular stomach. However, they did not examine the effects of T4 on the antral mucosa. Recently, we examined the effects of long-term administration of T4 for 20 weeks on the fundic and antral mucosa in rats without MNNG pretreatment and found that T4 significantly increased the labelling indices in both the antral (25.0± 1.1 vs 16.6 1.4%, P<0.01) and the fundic (17.8 1.1 vs 10.4 0.5%, P <0.001) mucosa, compared to those in control group. However, we observed neither significant differences between the average heights of the antral and fundic mucosa in the two groups nor mucosal dysplasia h d associated with T4 treatment in non-carcinogen treated rats.
The present study showed that T4 administration increased the yield of gastric cancers after MNNG treatment, presumably by acting as a non-genotoxic growth stimulator for the mucosal cells. Watanabe et al. (1992) found that oral administration of NaCl after MNNG pretreatment significantly increased the incidence of gastric cancers and that it also significantly increased the height of the pyloric mucosa. Many investigators (Takahashi et al., 1986;Tatsuta et al., 1988a;Kobori et al., 1984) have indicated that enhanc-ing effects on gastric carcinogenesis may be related to stimulation of cell proliferation and elongation of the mucosa. These findings suggest that a non-genotoxic component increases tumour yield when administered subsequent to a genotoxic agent.