Induction of Sarcomas by a Single Subcutaneous Injection of 7,12-Dimethylbenz[a]anthracene into Neonatal Male Sprague–Dawley Rats: Histopathological and Immunohistochemical Analyses

Abstract

Animal experiments have shown that carcinogenicity of chemicals is higher in fetal or neonatal periods than adult. We investigated sensitivities to a carcinogen in peri-neonatal rats with a model of sarcomas-induction by a subcutaneous injection of chemo-carcinogen that has rarely done in neonatal rats. Neonatal male SD rats were injected with 7,12-DMBA 10, 100, and 500 μg, which resulted in sarcomas-induction in 0, 62, and 94% of rats. Male SD rats were injected with DMBA 500 μg at 0, 3, 7, 14, and 21 days, which resulted in sarcomas-induction in 94, 70, 64, 50, and 44% of rats. Although the induced sarcomas were occasionally in mixed morphological feature as previous reports for sarcomas of rat, each was immunohistochemically in almost monotonous pattern, and classification was feasible. The incidence of rhabdomyosarcomas was higher in rats neonatally injected with a higher dose of DMBA than a lower dose, and in rats injected at peri-neonatal periods than later periods. In histological observations for the site of injection before overt sarcomas develop, clusters of atypical mesenchymal cells emerged as previous studies, but also those were immunohistochemically differentiated into rhabdomyocytes and other mesenchymal cells. We consider these findings may contribute a little to elucidation of process of sarcomas-induction in rats.

Introduction

Chemical carcinogenesis is a major cause of malignant neoplasm, and polycyclic aromatic hydrocarbons (PAHs) are major chemical carcinogens contained in cigarette smoke, burnt foods, exhaust fumes of vehicles, etc. Although the epidemiologic data do not show apparent increase of pediatric cancer incidence in advanced nations in recent years (Linet et al., 1999), many animal experiments have shown that the toxicity of chemicals is much higher (Bulay and Wattenberg, 1971; Vesselinovitch et al., 1979; McClain et al., 2001) and the development of tumor is faster (Flammang et al., 1997) in fetal or neonatal periods than in adolescent or adult, and transplacental or neonatal exposure of carcinogen may enhance development of tumors afterwards (Mandybur et al., 1978; Everson et al., 1980; Hasegawa et al., 1995; Waalkes et al., 2004).

We had at first tried to investigate the influence of exposure of carcinogen at infant period upon incidence of breast cancer that is epidemiologically increasing in Japan, and various doses of 7,12-Dimethylbenz[a]anthracene (DMBA), a representative compound of PAHs, were injected subcutaneously into neonatal female SD rats. The result was that, in the group in which DMBA 100 μg was subcutaneously injected at 0 days of age, incidence of breast tumor was slightly higher compared to the control group, but 3 rats over effective 7 gave rise to overt sarcomas at the site of injection. We concluded that this model was not suitable to investigate the influence of carcinogens-exposure at infant period upon incidence of breast tumor in later life, but it was suggested that neonatal period might be a high sensitivity period for sarcomas-induction at the site of injection of the rats.

The purposes of experiments for sarcomas-induction may be generalized as follows: assessment of sensitivity of various states of host; assessment of carcinogenicity of various kinds of carcinogens; investigation of intervention or therapy to developmental process: classification of the induced sarcomas: observation of prognosis; histopathological observation for developmental process, etc. The model for sarcomas-induction by subcutaneous implantation of chemo-carcinogens has been used in many animal experiments since 1934 (Berenblum, 1949; Roussy) for many advantages: simplicity of technique, easiness to trace the implanted points, simplicity of statistical analysis because of appearance of almost only 1 tumor in each animal, etc. But not many studies until now has investigated incidence of sarcomas in rats exposed to a carcinogen at neonatal period. Also there are few papers that show classification and frequency of the induced sarcomas on the basis of immunohistochemical findings or investigate developmental process of the sarcomas. We thought that those works would contribute to elucidation of process of sarcomas-induction in experimental models.

In this paper, we work on experiments in which sarcomas were induced in rats injected with DMBA at neonatal and infant periods and classified with morphological and immunohistochemical findings according to a recent WHO classification of soft tissue tumors of humans (Fletcher et al., 2002), and we observe morphologically the site of neonatal DMBA-injection at regular intervals when most of the rats did not give rise to overt sarcomas to get clues to understand the developmental process of sarcomas.

Materials and Methods

Animals and Maintenance

Inbred male Sprague–Dawley (SD) rats maintained at the Division of Laboratory Animal Science, Research Center for Life Science Resources, Kagoshima University, were given a commercial diet (CE-2, CLEA Japan, Inc.) and tap water ad libitum. The room was maintained at 23 ± 2°C, 55 ± 10% relative humidity, with a 12-hour light and 12-hour dark cycle. DMBA was purchased from Wako Pure Chemical Industries, Ltd., Japan. Rats were euthanized at necropsy by intraperitoneal injection of pentobarbital 120 mg/kg (Nembutal, Dainippon Pharmaceutical Co. Ltd., Japan), whereas most of the rats in debilitation interpreted by significant weight loss and low activity were euthanized by diethyl ether inhalation (Diethyl Ether, Nacalai Tesque Inc., Japan). Experiments were carried out in accordance with the Guide for Animal Experimentation, Faculty of Medicine, Kagoshima University, and approved by the committee of animal experimentation of Kagoshima University as category C of the Categories of Biomedical Experiments Based on Increasing Ethical Concerns for Nonhuman Species (1987).

Experiment 1. [Purpose: to examine the incidence of sarcomas in rats neonatally injected with various doses of DMBA]

Various doses of DMBA (Group I, 0 μg; Group II, 1 μg; Group III, 10 μg; Group IV, 100 μg; Group V, 500 μg; Group VI, 1000 μg) dissolved in 0.05 ml sesame oil were injected into subcutis below panniculus carnosus at the center back of 0-day-old male SD rats. After the injection of DMBA, the back was observed weekly until necropsy. The latent period was defined as the interval between the day of injection and the day when a nodular lesion became recognizable by palpation at first. When the lesion reached more than 20 mm in diameter, the rat was necropsied. The rat without a tumor was necropsied at day 250. All of the rats were euthanized at the necropsy. The hair on the back was shaved and the tissues were widely excised including skin, subcutaneous tissue, fascia of dorsal skeletal muscles and adjacent organs in case of tumor invasion. The abdomen and the thorax including their cavities were also inspected, but there were no remarkable findings in them except occasional invasions or metastases of the histologically same tumor as the back. Excised tissues were given single or plural longitudinal split lines for good fixation, and stuck to filter paper to preserve shape. The tissues were fixed in phosphate-buffered 10% formalin (pH 7.2) for 24 hours at room temperature, and then sliced with serial, transverse cuts about 3-mm thickness. Representative pieces of the tissues were thoroughly sampled to make paraffin blocks. Hematoxylin and eosin stained sections (H&E, 4 μm thick) were made for all blocks, and immunohistochemistry was performed for representative sections. Transmission electron microscopy was conducted for representative examples of the sarcomas.

Experiment 2. [Purpose: to examine the incidence of sarcomas in rats injected with DMBA at various days of infancy]

DMBA 500 μg dissolved in 0.05 ml sesame oil was injected into subcutis below panniculus carnosus at the center back of male SD rats of various ages (Group V, 0 days; Group VII, 3 days; Group VIII, 7 days; Group IX, 14 days; Group X, 21 days of age). After the injection of DMBA, the back was observed weekly until necropsy. When a palpable tumor reached more than 20 mm in diameter, the rat was necropsied. The rat without a tumor was necropsied at day 250. Definition of the latent period, the procedure of necropsy, the findings of abdomen and thorax, and the processing of excised tissues were the same as Experiment 1.

Experiment 3. [Purpose: to observe the process by which sarcomas develop]

DMBA 500 μg dissolved in 0.05 ml sesame oil for experimental group, or only 0.05 ml sesame oil for control group, were injected into subcutis below panniculus carnosus at the center back of 0-day-old male SD rats. Five or more rats were necropsied at each day 7, 14, 21, 49, 63, 77, 91, 105, and 119. The procedure of necropsy and the processing of excised tissues were almost same as in Experiments 1 and 2, but the thickness of serial, transverse slices after formalin-fixation was almost equally 2 mm.

Immunohistochemistry

Immunohistochemistry for α-SMA, Desmin, MyoD1, S-100, Myoglobin, ED1, Vimentin and AE1/AE3 was performed in all of the sarcomas in Experiments 1 and 2. Immunohistochemistry for α-SMA, Desmin, MyoD1, S-100, Myoglobin and Ki-67 was performed in all of the lesions consisting of a cluster of atypical mesenchymal cells in Experiment 3. After the blocking of endogenous peroxidase activity, deparaffinized sections (4 μm thick) were pretreated by microwaving (800W) for 10 minutes in 10 mM citrate buffer (pH 6.0), except for AE1/AE3, which was pre-treated μy microwaving (800W) for 10 minutes in 50 μM EDTA buffer (pH 8.0), S-100 which was pretreated by incubating with 0.1% trypsin dissolved in 0.1% CaCl₂(pH 7.6) for 30 minutes and myoglobin, which was not pretreated at all. After cooling, the sections were incubated with primary antibodies overnight at 4°C. The primary antibodies used and dilution rates were as follows; α-SMA (clone; αsm-1, Novocastra, UK, ×500), Desmin (clone; D33, DAKO, USA, ×10), MyoD1 (clone; 5.8A, DAKO, USA, ×200), S-100 (rabbit anti-cow, DAKO, ×5000), Myoglobin (rabbit anti-human, DAKO, USA, ×3000), ED1 (clone; ED1, Serotec, UK, ×800), Vimentin (clone; V9, diluted, DAKO, USA, ×20), AE1/AE3 (clone; AE1 and AE3, DAKO, USA, ×2000). Immunovisualization was performed with avidin-biotin complex (Elite ABC kit, Vector Laboratories, USA), except for S-100 with EnVision + System Labelled Polymer-HRP, (DAKO, USA). For chromogen, 3,3′-diaminobenzidine tetrachloride (0.5 mg/ml) containing 0.02% hydrogen peroxide (30% w/v) was used. Positive and negative controls were set up with various tissues of rats (skeletal muscle, connective tissue, lymph nodes, liver, spleen, etc.).

Ultrastructure

For transmission electron microscopy, types of representative tumors were chosen, cut into 1-mm-sized cubes, and immersed in Karnovsky’s fixative containing 2.5% glutaraldehyde and 2% paraformaldehyde dissolved in 0.1 M PBS (pH 7.4) for 24 hours at 4°C. The cubes were then immersed in 2% OsO₄ for 2 hours, then in 4% uranyl acetate for 1 hour, then dehydrated and embedded in Epoxy resin. Ultrathin sections (about 60 nm thick) were cut and observed with an electron microscope (H-7000; Hitachi, Japan).

Classification of Tumors

We examined the tumors in Experiments 1 and 2 histologically and tentatively classified according to WHO classification of soft tissue tumors of humans (Fletcher et al., 2002). The tumors could be divided into two categories at first: rhabdomyosarcomas and other sarcomas. The sarcomas in which more than 5% of cells showed nuclear staining for MyoD1, or more than 10% showed cytoplasmic staining for myoglobin were classified as rhabdomyosarcomas. Figures 1A, 1B and 1C show a typical pattern of the sarcoma classified as pleomorphic rhabdomyosarcoma with relatively well differentiation, and Figures 1D, 1E and 1F show a typical pattern of the sarcoma classified as embryonal rhabdomyosarcoma with relatively poor differentiation. Desmin was well-stained for pleomorphic rhabdomyosarcomas, but was negative for embryonal rhabdomyosarcomas with poor differentiation. Electron micrograph Figure 1G shows numerous myofilaments with Z bands in the cytoplasm of an embryonal rhabdomyosarcoma. Sarcomas in which cells were entirely negative for MyoD1 and Myoglobin were classified as other sarcomas. There was no sarcoma in which 5% or less of cells expressed MyoD1 and 10% or less expressed Myoglobin. The sarcomas consisting of pleomorphic or storiform pattern with scattered positive staining for α-SMA and negative staining for Desmin, Myoglobin and MyoD1 were classified as fibrohistiocytic tumors: of this category, sarcomas with inconspicuous myxoid areas were classified as pleomorphic malignant fibrous histiocytoma (Figure 2A); of this category, sarcomas accompanied with prominent (>50%) myxoid area were classified as myxofibrosarcoma (Figure 2B). A myxofibrosarcoma of Figure 2B showed negative staining for MyoD1 (Figure 2C) and no striated muscle fiber was observed in the cytoplasm of a cell resembling myofibroblast in the tumor (Figure 2D). Sarcomas consisting of spindle cells with eosinophilic fibers and fascicular growth pattern with diffuse positive staining for Desmin, scattered or diffusely positive staining for α-SMA and negative staining for Myoglobin and MyoD1 were classified as leiomyosarcoma (Figure 2E). Sarcomas consisting of spindle cells with herringbone growth pattern, negative staining for Desmin, Myoglobin and MyoD1 and almost negative staining for α-SMA were classified as adult fibrosarcoma. Sarcomas consisting of small round cells with fine chromatin and scant cytoplasm with a diffuse growth pattern around the relatively plentiful vascular spaces (Figure 2F) were placed in unclassified sarcoma group. These sarcomas showed diffuse positive staining for S-100 (Figure 2G) and negative staining for α-SMA, Desmin, Myoglobin, MyoD1 and ED1 (rat homologue of human CD68). We speculate from the morphology and immunohistochemical pattern that these tumors might have primitive neural differentiation, but we could not classify these tumors confidently because we could not establish immunohistochemistry of NSE, CD56, CD99, CD34 and Chromogranin A for any tissue of rats. In other sarcomas category, main cellular lineages seem to consist of myofibroblasts, fibroblasts, histiocytes and smooth muscle cells except in the unspecified sarcomas. Apparent composite sarcomas with components of rhabdomyosarcomas and other sarcomas were not observed. No sarcoma was positive for AE1/AE3, and all of the sarcomas were strongly positive for Vimentin. We could not establish LCA (CD45) staining in any tissue of rats, but no atypical lymphoid cells were observed at all, therefore we consider that there is no possibility that lymphomas occurred in our experiments. There was no mesenchymal tumor with a predominantly benign component.

Criteria to Call Cells Myofibroblast-Like Cells

The cells we call myofibroblast-like cells were spindle cells with architecture of random or loosely fascicular arrangement, and a cytoplasm visible but less abundant and less elongated than smooth muscle, as the description of AFIP to myofibroblasts (Table 5–2 of Tumors of Soft Tissues, 3rd series of AFIP, Kempson et al. 2001). These cells emerged conspicuously in the tissues at Experiment 3. Immunohistochemically, almost half of these cells were positive for α-SMA, and the stain for α-SMA in positive cells was typically intense at the periphery of the cytoplasm (tram-track pattern) as the description of Enzinger and Weiss’s Soft Tissue Tumors to myofibroblast (Weiss and Goldblum, 2001). (Figure 7L). Desmin positive cells were scarce. We consider that strict distinction in various kinds of mesenchymal cells such as myofibroblasts, fibroblasts, histiocytes or even Schwann cells except myoblasts is difficult in routine morphological and immunohistochemical observation, and that, any kind of these cells may exist in some rate in a cluster of myofibroblast-like cells, although myofibroblasts may occupy predominantly.

Statistics

The incidence of the sarcomas was examined with Fisher’s exact test, and the latent period was examined with the Student’s t-test (2-tail test). The cut-off for significance was taken as p = 0.05.

Results

Experiment 1. (to observe the incidence of sarcomas in rats neonatally injected with various doses of DMBA)

In Group I (0 μg of DMBA), II (1 μg), III (10 μg), IV (100 μg) and V (500 μg), sarcomas were induced in 0, 0, 0, 62, and 94% of effective rats, respectively (Table 2). Subtypes of the sarcomas are shown in Table 1. Body weights and survival rates are shown in Tables 3 and 4 and Figures 3 and 4. Approximately half of the sarcomas were rhabdomyosarcomas. The other half consisted of other sarcomas. Apparent composite sarcomas with components of a rhabdomyosarcoma and another sarcoma were not observed. The incidence of sarcomas overall increased significantly with the dose of DMBA. The latent period for sarcomas overall was significantly shorter in Group V than IV. The incidence of rhabdomyosarcomas was higher in Group V than I, II, III, and IV (Table 2).

Experiment 2. (to observe the incidence of sarcomas in rats injected with DMBA 500 μg at various days of infancy)

In Group V (injected at 0 days of age), VII (3 days), VIII (7 days), IX (14 days), and X (21 days), sarcomas were induced in 94, 70, 64, 50, and 44% of effective rats, respectively (Table 5). Subtypes of the sarcomas are shown in Table 1. Body weights and survival rates are shown in Tables 6 and 7 and Figures 5 and 6. As well as Experiment 1, approximately half of the sarcomas were rhabdomyosarcomas, and the other half consisted of other sarcomas. Apparent composite sarcomas with components of a rhabdomyosarcoma and another sarcoma were not observed. The incidence of sarcomas overall was significantly higher in Group V than IX and X. The latent period for sarcomas overall was significantly shorter in Group V than IX and X. The incidence of rhabdomyosarcomas in rats injected at peri-neonatal periods (the sum of Group V and VII) was significantly higher than in rats injected at later periods (the sum of Group VIII, IX and X) (Table 5).

Experiment 3. (to observe the process by which sarcomas develop)

In experimental group, a cystic lesion containing oil was observed at the site of injection beneath the panniculus carnosus in all of the rats at day 7. At day 14, the disappearance of skeletal muscle and connective tissue with deposits of fibrinoid materials in lower part of the panniculus carnosus was also noted in most of the rats (Figure 7A). At this early stage, mass-forming lesions, such as proliferation of myofibroblast-like cells rhabdomyoblast-like cells, fibrosis or granulation, were not observed. The cyst continued thereafter in most of the rats, but the oily content turned serous or bloody after day 21. At day 21, a proliferative lesion of bland spindle cells mainly consisting of rhabdomyoblast- and myofibroblast-like cells was observed near the cyst in some of the rats (Figures 7B, 7C and 7D). Although the cellularity was high in some parts, monotonous proliferation of atypical cells was not observed at this stage. At day 49 and day 63, fibrosis progressed in most of the rats. Mild nuclear enlargement of rhabdomyocytes of the panniculus carnosus was seen in some of the rats at this stage, but the rhabdomyocytes were preserving polarity and lacking nuclear hyperchromasia and pleomorphism. On the other hand, proliferation of other mesenchymal cells (mainly myofibroblast-like cells) around the cyst was observed in some of the rats at day 49 and day 63. At this stage, the other mesenchymal cells were lacking nuclear hyperchromasia and a monotonous pattern of proliferation.

At day 77 or later, an atypical lesion sized in diameter 1 mm to 15 mm consisting of a cluster of atypical rhabdomyocyte-like cells became evident near the panniculus carnosus in some of the rats. The incidence of this lesion is shown in Table 8. This lesion showed an increased N/C ratio, nuclear hyperchromasia, high cellular density and disordered polarity (Figures 7E and 7F). The atypical cells in this lesion was immunohistochemically positive for MyoD1 (Figure 7G), Desmin and Myoglobin Ki-67-positive rate was high compared to surrounding tissues (Figure 7H).

On the other hand, at day 77 or later, an atypical lesion sized in diameter 1 mm to 10 mm consisting of a cluster of other atypical mesenchymal cells (mainly atypical myofibroblast-like cells) became evident near the cyst in some of the rats. The incidence of this lesion is shown in Table 8. This lesion showed an increased N/C ratio, nuclear hyperchromasia, high cellular density, front-formation at the margin and monotonous pattern of proliferation (Figures 7I and 7J). Almost half of the other atypical cells in this lesion was immunohistochemically positive for α-SMA (Figure 7K), and most of the stain for α-SMA was intense at the periphery of the cytoplasm similar to the tram-track pattern of myofibroblasts (Figure 7L). In this atypical lesion, scarce cells were positive for Desmin, and no cells were positive for Myoglobin, MyoD1 and S-100. Ki-67-positive rate was high compared to surrounding tissues (Figure 7M).

All of the clusters of atypical rhabdomyocyte-like cells were located near the panniculus carnosus, while all of the clusters of other atypical mesenchymal cells (mainly atypical myofibroblast-like cells) were located contagious to the cysts, as shown in Figure 8.

An atypical composite-like lesion consisting of a cluster of atypical rhabdomyocytes and a cluster of other atypical mesenchymal cells (mainly myofibroblast-like cells) was observed in 2 rats necropsied at day 77.

Besides the above-mentioned lesions, there were no such atypical lesions as a cluster consisting of epithelial cells, smooth muscle cells of arrector pili or vascular wall, peripheral nerve cells, etc.

In control rats injected with sesame oil at 0 days of age, small foci of inflammation, fibrosis or encapsulated small cystic lesions (1 to 4 mm in size) containing serous fluid were seen in a few rats, but neither proliferative nor atypical lesions were seen at any day.

Discussion

In animal experimental models for sarcomas-induction by subcutaneous implantation of carcinogens, PAHs have been representative carcinogens, and rat has been one of the most sensitive species (Berenblum, 1949). The investigations of sarcomas induced by exposure to foreign materials also have some relevance in a standpoint of preventive medicine: many kinds of materials are being developed for implantation into human bodies for medical purposes; a minority of recipients of prosthetic materials gives rise to sarcoma still now (Keel et al., 2001).

In our Experiment 2, the incidence of sarcomas was highest in the group in which DMBA was injected during the neonatal period. In other studies using mature rats, relatively high doses of PAHs were required to induce a high incidence of sarcomas compared to our Experiment 1: single subcutaneous injection of approximately 8 mg of DMBA dissolved in 0.4 ml lanolin into white rats whose weights ranged 85 to 290 gm was reported to induce tumor in 100% (16/16) (Davenport et al., 1941); single subcutaneous injection of DMBA 1 mg dissolved in 0.5 ml tricapryline into albino rats (ages and weights were not mentioned) was reported to induce tumor in 84% (16/19) (Berenblum, 1949). From these facts, we speculate that sensitivity to chemo-carcinogens injected into subcutaneous tissue of born rats probably shows peak at peri-neonatal period if otherwise unmanipulated. Several factors that may contribute to high sensitivity of neonatal rats are as follows: state of cell cycle, expressed proteins, protooncogenes and tumor suppressor genes; low activity of phagocytosis and metabolism-related enzymes; a lower degree of looseness of subcutaneous tissues which might prevent carcinogens from diffusing, etc. Most of these factors have not been fully investigated yet.

The result of our Experiment 1 was similar to the result of an experiment of Toth in which single subcutaneous injection of DMBA 10, 100, 1000 μg dissolved in 0.05 ml of tri-n-caprylin into neonatal Lewis rats resulted in sarcomas-induction in 8% (3/37), 17% (8/46), and 75% (9/12) at the site of injection, while survival rates were 63, 41, and 10% (Toth and Shubik, 1963). Toth also pointed out in the paper that, in newborn Swiss mice, subcutaneous injection of DMBA gave rise to a high incidence of malignant lymphomas and lung adenomas but only to very few sarcomas at the site of injection. Neonatal mice models seem more suitable for screening compounds for carcinogenic activity because of the high sensitivity with a lower dose of carcinogen: single subcutaneous injection of DMBA 30 μg dissolved in 0.015 ml 1% aqueous gelatin into neonatal Swiss mice resulted in induction of malignant lymphoma in 32% (8/25) (Pietra et al., 1959); single subcutaneous injection of DMBA 10 μg dissolved in 0.02 ml 3% aqueous gelatin into neonatal BALB/c mice resulted in induction of lung tumors in 100% (26/26) (Walters, 1966). But, it is generally thought that genetic and viral background is significantly different between mice and rats; for example, suspicious mechanism of lymphomas-induction in mice is activation of a latent virus that has been supported by demonstration of a filterable leukemogenic agent obtained from both radiation and chemically induced lymphomas (Toth and Shubik, 1963). Therefore we consider that any result of experiments using rats have a peculiar value independent of any result of experiments using mice.

We classified the induced sarcomas in Experiment 1 and 2 with morphological and immunohistochemical findings. Classification of sarcomas of rats was scarcely done in the past. Thomas tried to classify their induced sarcomas of rats with morphological, electron microscopic and immunohistochemical methods and reported that most of the sarcomas displayed mixed histological pattern, that is, undifferentiated, spindle cell, fibrosarcomatous and rhabdomyosarcomatous areas seen simultaneously (Thomas et al., 1977). In our Experiment 1 and 2, all of the sarcomas did not show immunohistochemically various or colliding pattern in a tumor as for expression of MyoD1, Myoglobin, Desmin, α-SMA and S-100 even though they showed a somewhat mixed morphological pattern. Therefore, we consider classification of the sarcomas of our experiments was feasible. But some problems remain in application of a classification for human sarcomas to sarcomas of rats at present: as the mention of Thomas, the morphology of sarcomas of rats is occasionally varied and not completely consistent with histological pattern of human sarcomas; the number of commercial antigens with cross-reactivity to rat tissues seems small.

In spite of these unfavorable circumstances for classification, it can be said that the incidence of rhabdomyosarcomas was higher in rats neonatally injected with higher dose of DMBA compared to rats injected with lower dose in Experiment 1. The incidence of rhabdomyosarcomas was also higher in rats injected at peri-neonatal periods than rats injected at later periods in Experiment 2. Also in Group V of Experiments 1 and 2, the incidence of rhabdomyosarcomas was higher than the models using mature rats (Tokiwa et al., 1987; Nikitin et al., 1993; Stefanou et al., 1998). These results remind us a difference of sensitivities between rhabdomyosarcomas and other sarcomas with the difference of dose of carcinogens or the difference of age of hosts, and the possibility of a high sensitivity to detect skeletal muscle differentiation in the immunohistochemical method that we used.

We could not confirm at present whether the clusters of atypical mesenchymal cells in Experiment 3 were early neoplastic lesions with the potential to develop into a malignant neoplasm or merely atypical regenerative lesions, because the morphological features of preneoplastic stages of sarcomas are still not well defined (Guillou et al., 1998), and transplantation may be necessary for strict verification of neoplasia. However, we consider that the clusters of atypical mesenchymal cells in Experiment 3 acted as soils or seeds for rhabdomyosarcomas or other sarcomas considering the previous reports of animal experiments that proliferation of atypical fibroblasts or mesenchymal cells (Vasiliev, 1959; Thomas et al., 1977; Nikitin et al., 1993), irregular proliferation of capillaries (Richter et al., 1999), foci of cells with the ultrastructural features of a rhabdomyosarcoma (Westwood et al., 1979) emerge on early days after exposure to chemo-carcinogens before overt sarcomas appear.

The different location between the clusters of atypical rhabdomyocyte-like cells and the clusters of other atypical mesenchymal cells in Experiment 3, as shown in Figure 8, also reminded us the possibility that different neoplastic lesions may be derived from different original cells. We suppose from the location that most of the clusters consisting of atypical rhabdomyocyte-like cells may be induced from regenerating rhabdomyocytes of the panniculus carnosus while most of the clusters consisting of other atypical mesenchymal cells may be induced from other mesenchymal cells near the cystic lesions where stimuli of the carcinogen seem to be most intense. However, we can not affirm the origin of rhabdomyosarcomas or other sarcomas at present because mesenchymal tissues are reported to contain pluripotent mesenchymal stem cells capable of forming skeletal muscle, fibroblasts or smooth muscle, etc. (Toma et al., 2001; Tholpady et al., 2003).

We consider that the resolution of the original tissues of individual sarcomas and the resolution of the possibility of histology-transition in the process of sarcomas-development in an animal experimental model would contribute to many fields including human pathology. We also consider that the findings of our experiments suggest a little about the nature of subcutaneous tissue of neonatal rats and the process of sarcomas-induction.