Experimental induction of ovarian Sertoli cell tumors in rats by N-nitrosoureas.

Spontaneous ovarian tumors are very rare in ACI, Wistar, F344 and Donryu rats; the few neoplasms found are of the granulosa/theca cell type. Ovarian tumors were also rare in these strains of rats when given high doses of N-alkyl-N-nitrosoureas continuously in the drinking water for their life-span; however, relatively high incidences of Sertoli cell tumors or Sertoli cell tumors mixed with granulosa cell tumors were induced in Donryu rats after administration of either a 400 ppm N-ethyl-N-nitrosourea solution in the drinking water for 4 weeks or as a single dose of 200 mg N-propyl-N-nitrosourea per kg body weight by stomach tube. Typical Sertoli cell tumors consisted of solid areas showing tubular formation. The tubules were lined by tall, columnar cells, with abundant, faintly eosinophilic, often vacuolated cytoplasm, and basally oriented, round nuclei, resembling seminiferous tubules in the testes. In some cases, Sertoli cell tumor elements were found mixed with areas of granulosa cells. The induction of ovarian Sertoli cell tumors in Donryu rats by low doses of nitrosoureas may provide a useful model for these tumors in man. ImagesPLATE 1.PLATE 2.PLATE 3.PLATE 4.PLATE 5.PLATE 6.PLATE 7.PLATE 8.PLATE 9.PLATE 10.PLATE 11.PLATE 12.PLATE 13.PLATE 14.PLATE 15.PLATE 16.


Introduction
There is a need for good animal models for human ovarian tumors. Spontaneous ovarian tumors are rare in rats, and there are few useful methods of experimental induction of ovarian tumors in this species (1). This paper describes the experimental induction ofovarian Sertoli cell tumors by N-alkyl-Nnitrosoureas as a potential animal model of human ovarian tumors.

Materials and Methods
Four strains (ACI/N, Slc:Wistar, F344/DuCrj, and Donyru) of female rats were used for this study. The housing conditions of the animals and occurrence of spontaneous tumors in all strains except Donryu rats have been described in previous papers (2)(3)(4)(5).
Two additional experiments were carried out using Donryu rats. In the first experiment, (Experiment 1), forty 10-week-old rats were administered a 400 ppm solution of ENU in the drinking water for 4 weeks. In the second experiment (Experiment 2), 80 rats were given a single dose of 200 mg PNU per kg body weight by stomach tube. After the treatment, the animals in both experiments were divided into two subgroups. One-half of the rats in each experiment were then maintained on a standard diet, and the other half were administered a diet containing 0.25% phenylbutazone for 2 years. One hundred animals were also maintained as untreated controls. Observations were continued until week 110, when all survivors were sacrificed. The purpose of these two experiments was to investigate the promoting potential of phenylbutazone in rats. A short report has been published on these experiments (21).
Ovarian Tumors in Treated Rats Table 1 summarizes incidences and types of ovarian tumors in rats given N-alkyl-N-nitrosoureas continuously in the drinking water for a life-span and lists other target organs. All these chemicals were strongly carcinogenic; the target organs differed with the compound and strain of rats used. Incidences of ovarian tumors, however, were very low in all experiments; only a few Sertoli cell tumors were observed in Donryu rats given B-BNU or ANU, and in F344 rats given low doses of ENU. Table 2 shows the incidences and histological types of ovarian tumors observed in Donryu rats given a 400 ppm ENU solution in the drinking water for 4 weeks (Experiment 1) or a single dose of 200 mg PNU per kg body weight by stomach tube (Experiment 2). In this table, the numbers of rats in each subgroup were totaled within each experiment, because no modifying activity of phenylbutazone for induction of ovarian tumors was observed. Ovarian tumors were induced at incidences of about 50% in both experiments. Mean survival times of rats with ovarian tumors were 52.7 weeks in Experiment 1 and 78.6 weeks in Experiment 2. In contrast, only two tumors were found in 96 untreated control rats, and the mean survival time of tumor-bear-ing animals was 109.5 weeks. Many other neoplasms, such as tumors of the uterus, kidney, mammary gland, and intestine, and leukemias were also observed in both experiments.
The majority of the induced ovarian tumors were unilateral. Tumors were round, oval or irregular in shape, and of various sizes. Their cut surfaces were usually solid, although occasional cyst formation and areas of hemorrhage were evident. On histological examination, most of the ovarian tumors induced were found to be Sertoli cell tumors which consisted of solid areas of tubular formation. The tubules were lined by tall, columnar cells, with abundant, faintly eosinophilic, often vacuolated cytoplasm and basally oriented, round nuclei. These tubules were separated by thin fibrovascular stroma and resembled seminiferous tubules in the testis (Plates 5-10). There was variation in the size and form of the tubules. In some cases, hyaline bodies of varying sizes, which were positive for PAS, were observed in the cytoplasm of tumor cells or in the lumen of tubules (Plate 11). Mitotic figures were usually rare. No histological findings suggesting that these tubules had arisen from downgrowth of the surface epithelium were found in any case. In cases diagnosed as Sertoli cell tumors mixed with granulosa cell tumors, histological patterns ranging from typical tubular structures to those more characteristic of granulosa cell tumors were observed within single tumor nodules, and differentiating between Sertoli cell tumors, granulosa cell tumors, and mixed tumors was sometimes difficult (Plates 12- aIn all experiments, rats were administered solutions of these chemicals daily in their drinking water for the life-spans. bNumber of rats with ovarian tumors/number of rats treated.  15). In most tumors, proliferation of Leydig-cell or theca-cell elements was not apparent.

Discussion
Generally, spontaneous ovarian tumors are uncommon in rats (1). Of the few spontaneous ovarian tumors arising in F344 rats, the most common type was granulosaltheca cell tumors (22,23). Our results were in good agreement with the findings in other laboratories. Recently, however, Gregson et al. (24) reported that the most commonly observed ovarian tumors in Sprague-Dawley rats were tubular adenomas arising from downgrowth of the surface epithelium. The occurrence of tubular structures in ovarian tumors of rats and mice is well documented (1,(25)(26)(27)(28)(29); however, the classification and nomenclature of tubular adenomas in experimental animals is complicated and controversial. Tubular adenomas ofmice are generally accepted as arising from downgrowth of the surface epithelium (28). It is possible that Sertoli cell tumors, which are sex cord tumors, may have been included in the tubular adenoma category reported by others because tubular structures resembling seminiferous tubules were apparent in some reported cases.
Methods such as transplantation of the ovary, hormonal treatment, irradiation, or administration of chemical carcinogens have been used in attempts to induce ovarian tumors in rats. For example, granulosatheca cell tumors can be induced in rats by the transplantation of the ovary to various ectopic sites such as the spleen (30) and also by intrasplenic transplantation of the ovary coupled with 7,12-dimethylbenzanthracene (DMBA) treatment (31). Granulosa/theca cell tumors appear to be hormonally active, because in rats with these tumors, hyperplastic proliferation of the uterus, vagina, or mammary gland is observed.
There is little direct evidence that hormones are responsible for the induction of tumors. There is also little experimental evidence that ovarian tumors can be induced in rats by irradiation, although this is the oldest and most effective method for induction of ovarian tu-mors in mice (32). Polycycic hydrocarbons such as DMBA and 3-methylcholanthrene have mainly been used for chemical induction of ovarian tumors. Although they result in good tumor yields in mice, there are only a few reports of success in rats. In one report, Kato et al. (33) succeeded in inducing ovarian tumors in rats by treatment with DMBA. Most of these DMBA-induced tumors were adenocarcinomas that were considered to originate from the surface epithelium (Plate 16), suggesting a useful experimental model for human ovarian carcinomas. Sertoli cell tumors of the ovary are quite rare both in human beings and rats, and there has been no previous report of their successful high yield, other than the recent report by Stoica et al. (34). Our study suggests a useful experimental animal model for Sertoli cell tumors.
In several ovarian tumors observed in the present study, histological patterns ranging from typical tubular structures to those more characteristic of granulosa cell tumors were observed within single tumor nodules. Sertoli cell differentiation has been seen in mouse granulosa cell tumors (28), and Knowles (35) also described mixtures of both elements in rat ovarian tumors induced by ENU coupled with radiation treatment. There is a degree of architectural and cytological overlap between these two neoplasms in human beings (36). Mixed tumors observed in the present study indicate that Sertoli and granulosa cell elements are homologous, with a common origin from the sex cord.
Recently, Knowles (35,37) reported low incidences of ovarian tumors, which showed tubular formation and were interpreted as of Sertoli cell type, in Harwell Mouth Tumor rats given ENU together with radiation.
Stoica et al. (34) also reported induction of Sertoli cell tumors of the ovary of Sprague-Dawley and BD-IV rats by intraperitoneal on transplacental treatment with ENU. These data and our present results show conclusively that such nitrosoureas as ENU and PNU can induce Sertoli cell tumors in some strains of rats. Incidences of ovarian tumors were very low when high doses of these nitrosoureas were given to rats in the drinking water for a life-span (Table 1). High doses of N-nitroso compounds show a pronounced organotropism in rats. Development of other tumors such as ovarian tumors is masked by the elimination of animals early in the study because of selective induction of specific tumors with short latent periods. For example, as shown in Table 1, continuous oral administration of ENU at doses of 400 to 100 ppm rapidly and selectively induced erythroleukemias in Donryu rats, and no ovarian tumors were detected in a total of 104 treated rats (12). The fact that the amounts of the chemicals given in the present two experiments described in this paper were low might explain the high rate of induction of ovarian tumors. Another important factor may be variation in susceptibility of different strains of rats. Only two Sertoli cell tumors were found in a total of 203 F344 rats given low doses of ENU (9), (Table 1). Schmahl and Habs (38) reported that the location of tumors induced by N-nitroso compounds depends not only on the chemical structure of individual carcinogens, but also on the administration route, strains of rats used, or dose of the compounds. Our results are consistent with their opinion.
The possible hormonal function of Sertoli cell tumors observed in the present study is still uncertain and is the focus of ongoing research in our laboratory.