JOURNAL OF PATHOLOGY, VOL. 157: 263-274 (1989)

INDUCTION OF SOFT TISSUE TUMOURS IN F344 RATS BY SUBCUTANEOUS,

INTRAMUSCULAR, INTRA-ARTICULAR, AND RETROPERITONEAL INJECTION OF NICKEL

SULPHIDE (Ni,S,) MASASHI SHIBATA, KEISUKE IZUMI, NOBUYA SANO, AKIKO AKAGI AND HISASHI OTSUKA

Second Department of Pathology, School of Medicine, The University of Tokushima, Japan

Received 1.5 July 1988 Accepted 26 July 1988

SUMMARY The carcinogenicity of nickel sulphide (Ni,S,) injected into subcutaneous (s.c.), intramuscular (i.m.), or retroperito-

neal intrafat (i.f.) tissue, or the intra-articular space (i.a.) of male F344 rats was studied. Rats were given a single injection of 0.5 mg of Ni,S, and were observed for 48 weeks. Malignant soft tissue tumours were induced in 18/19 rats (95 per cent) by S.C. injection, 19/20 rats (95 per cent) by i.m. injection, in 16/19 rats (84 per cent) by i.a. injection, and in 9/20 rats (45 per cent) by i.f. injection of Ni,S,. The i.f. injection of Ni,S, resulted in a lower tumour incidence and the appearance of tumours 10 weeks later than its injection by other routes. The tumours were examined histologically, ultrastructurally, and immunohistochemically with antibodies against desmin, vimentin, and cytokeratin. The 62 tumours induced by injection of Ni,S, by different routes were identified as rhabdomyosarcomas (RMS, 3 9 , malignant fibrous histiocytomas (MFH, 18), fibrosarcomas (FS, 5 ) , and unclassified sarcomas (4). All 19 tumours induced by i.m. injection of Ni,S, were rhabdomyosarcomas; those induced by S.C. or i.f. injection were mainly MFHs. However, a number of RMSs were also found in groups that received i.a., s.c., and i.f. injections; five FSs also developed in these groups. Four sarcomas induced by S.C. and i.a. injections were not classified. No synovial sarcoma developed.

KEY WoRDS-Nickel sulphide, rhabdomyosarcoma, malignant fibrous histiocytoma, ultrastructure, immunohisto- chemistry.

INTRODUCTION

Since Gilman reported a high incidence of sarco- mas in rats after a single injection of Ni,S,,' many investigators have used this simple method of carcinogenesis. Histological examination has shown that the sarcomas induced at the site of intramuscular injection of Ni,S, have been mainly rhabdomyosarcomas, but some undiffer- entiated sarcomas, fibrosarcomas, liposarcomas, neurofibrosarcomas, hemangiosarcomas, or leio- myosarcomas have also been Rhabdo- myosarcoma has also occurred after intrate~ticular~ or intrarenal' injection of Ni,S,.

Addressee for correspondence: H. Otsuka, Second Depart- ment of Pathology, School of Medicine, The University of Tokushima, Kuramoto-cho 3-18-15, Tokushima 770, Japan.

0022-341 7/89/030263-12 $06.00 0 1989 by John Wiley & Sons, Ltd.

There are also many recent reports of induction of soft tissue sarcomas in various sites in rodents by injection of various other chemicals such as dehydroretronecine,6 20-methylcholanthrene,' 4- hydroxyamino)quinoline 1 9,lO-dimethyl- 1,2-ben~anthracene,'&'~ dinitropyrenes, and dini- trofl~oranthenes.'~

The diagnosis and classification of soft tissue sarcomas are frequently difficult, particularly when these tumours are poorly differentiated. Rhabdo- myosarcomas exhibiting pleomophic morphology are not always easy to differentiate from malignant fibrous histiocytomas and more detailed identifica- tion of tumour cell differentiation in various ways is needed to determine the origin of these tumours. Recently, the immunoperoxidase technique has been shown to be useful for identifying several

264 M. SHIBATA ET AL.

muscular proteins found during myogenic differen- tiation of human'"" and experimental'' sarcomas. Cytokeratin, a marker of epithelial differentiation, has also been found to be useful in the diagnosis of human synovial sarcomas."-"

The present study was designed to investigate the tumourigenicity of Ni,S, following its injection into four sites (intramuscular, subcutaneous, intra- fat, and intra-articular space) in F344 rats. We examined which mesenchymal components were susceptible to the carcinogen and what types of sarcoma developed in different tissues. Electron microscopy and immunohistochemical techniques were used to aid the histological analysis of the induced t umours.

MATERIALS AND METHODS

Carcinogen Nickel sulphide (Ni,S,, average particle size 3-

5nm) was obtained from INCO Ltd. (Toronto, Ontario, Canada) by the courtesy of Dr. Shigeto Yamashiro (Department of Biomedical Sciences, University of Guelph, Guelph, Ontario, Canada).

Animals One hundred and twenty male Fischer 344 rats

(Charles River Japan Inc., Atsugi) 7-8 weeks old, weighing 136-198 g at the time of Ni,S, injection, were used. The rats were housed in plastic cages (2 or 3 rats/cage) and given a pellet diet (Oriental MF, Oriental Yeast Co., Tokyo) and tap water ad libitum.

Injection procedure The animals were randomly divided into eight

groups. The route of administration of Ni,S,, the number of rats, and the dose of Ni,S, injected are given in Table I. For i.m., s.c., i.a. injections, the rats were anaesthetized with ether; 5mg of Ni,S, powder suspended in 0.25 ml of aqueous procain G penicillin was injected into the interscapular subcu- taneous tissue (s.c) or gastrocnemius muscle (i.m.). for i.a. injection, 5 mg of Ni,S, in 0.1 ml of the same vehicle was injected into the intra-articular space of the knee joint (i.a.), using a 1 ml sterile tuberculin syringe with a 25-gauge needle. For injec- tion into retroperitoneal fatty tissue (if.), the rats were anaesthetized by intramuscular injection of ketamin hydrochloride into the gluteal muscle, and

the abdomen was opened by a midline incision under sterile conditions. The intestines were protected, and a needle was inserted into the retro- peritoneal fatty tissue and 5 mg of Ni,S, in 0.25 ml of procain G penicillin injected. The abdominal muscles were sutured with silk thread and the skin incision was closed by surgical clips. Control rats were treated identically expect that no nickel sulphide was injected. Animals were examined weekly for tumour development and were killed when tumours were more than 3cm in diameter. Rats without palpable tumours were observed for 48 weeks after injection and then killed.

Light microscopy

Several pieces of primary and metastatic tumours, when present, including their largest cut surface, were fixed in 10 per cent neutral buffered formalin and embedded in paraffin. All sections were cut at 4 ,um thickness and stained with haema- toxylin and eosin (H&E). Selected sections were stained with periodic acid-Schiff (PAS), Azan, phosphotungstic acid-haematoxylin, Sudan I11 or alcian blue, or by silver impregnation for reticulin fibres.

Electron microscopy

Small pieces (about 1 mm') of tumour tissue were fixed in 2.5 per cent glutaraldehyde/O.l M phosphate buffer (pH 7.4) and then 1 .O per cent OsO, in the same buffer. These specimens were dehydrated in a graded acetone series and embedded in Epon. Ultrathin sections were cut with an LKB Ultrotome 4802A and examined with a Hitachi H-300 electron microscope after staining with uranyl acetate and lead citrate.

Immunohistochemistry Tests for desmin, vimentin, and cytokeratin were

done by the avidin biotinylated horseradish com- plex (ABC) method on sections of tissue fixed in 10 per cent buffered formalin and embedded in paraffin. Anti-desmin antibody (rabbit IgG against chicken gizzard desmin) was given by Dr Jun-ichi Abe (First Department of Pathology, School of Medicine, University of Tokushima). Mono- clonal anti-vimentin antibody and monoclonal anticytokeratin antibody (PKK-1) were obtained from Dako (California, U.S.A.) and Labsystems (Helsinki, Finland), respectively.

SOFT TISSUE TUMOURS INDUCED BY Ni3S2 265

Table I-Experimental groups, incidences and sizes of primary tumours, and metastases

No. Amount of Observ. Effect. No. rats Size of of Ni,S, period No. with tumour

Group Route rats (mg)* (week) rats tumour (cm)? Metastases

I

I'

TI

11'

111

111'

IV

IV'

Subcutaneous (interscapular)

Intramuscular (gastrocnemius)

Intra-articular (knee)

Intrafat (retroperitoneum)

20

10

20

10

20

10

20

10

48

48

48

48

48

48

48

48

19

10

20

10

19

10

20

7

18

0

19

0

16

0

9

0

*Suspended in 0.25 or 0.1 ml of procain G penicillin. tMean & standard deviation. $Two in lungs. @ix in lungs; two in lungs and rctroperitoneum. IIOne in lungs and generalized lymphnodes; one in lungs and retroperitoneum; two in retroperitoneum.

RESULTS The results of oncogenesis tests in each group,

including the tumour incidence, average diameter of the primary tumour, and metastases in each group, are summarized in Table I. No tumour developed at the injection site in any control rats (groups 1', 11', III', and IV) within 48 weeks of injection. The tumour incidence in rats which received Ni,S, by s.c., i.m., and i.a. routes (groups I, 11, and 111) was 95-84 per cent, whereas after i.f. injection of Ni,S, (group IV) it was only 45 per cent (9/20), which was significantly lower (P<0.01 for group IV versus groups I and 11; P < 0.025 for group IV versus group 111 k2-test). The cumulative incidence of tumours in each group is illustrated in Fig. 1. After s.c., i.m., and i.a. injections, the first tumours appeared in or at about week 30. The cumulative incidence of tumours was highest in the S.C. group between weeks 34 and 41, but was almost the same in these three groups in week 48. However, in the i.f. group (IV), the first tumour appeared 10 weeks later than in the other groups and the tumour incidence was also lower.

Histological diagnosis of tumour was based on light microscopic findings with reference to ultra- structural and immunohistochemical findings. All

tumours were sarcomas and were classified as rhabdomyosarcomas (RMS, well or poorly differ- entiated types), malignant fibrous histiocytomas (MFH), fibrosarcomas (FS), and unclassified sarcomas.

The histological classification of the tumours in the four groups is shown in Table 11. RMSs were found in all groups; all 19 tumours in the i.m. group and half of those in the i.a. group were RMSs. MFHs were mainly seen in the S.C. and the i.f. groups, but a few were also observed in group 111 (La.). Uncommonly, FSs and unclassified sarcomas were found in a few rats.

Rhabdomyosarcoma (RMS) RMSs developed in rats in all experimental

groups. As in group 11, the RMSs in groups I, 111, and IV always involved adjacent muscles, which were skin muscles or back muscles (s.c. group), periarticular muscles (La. group), and lumbar muscles (i.f. group). RMSs were classified into two groups (well and poorly differentiated) on the basis of the predominant cell types.

Well differentiated RMS-Most of these tumours exhibited cellular pleomorphism. They contained

266

1 00

x v

m 50

e II)

M. SHIBATA ET AL.

- S.C. - i.m. H i.a.

M i.f.

Weeks after start of experiment

Fig, 1-Cumulative sarcoma incidence in male Fischer rats after s.c., im., La., or i.f. injection of 5 mg of Nils,

Table 11-Histopathological classification of the tumours

Total No. Histological type Group oftumours RMS (Well) (Poor) MFH FS Unclassified

I (S.C.) 18 5 (2) (3) 10 1 2 I1 (i.m.) 19 19 ( 5 ) (14) 0 0 0 I11 (i.a.) 16 8 (2) (6) 3 3 2 IV (i.f.) 9 3 (1) ( 2 ) 5 1 0

62 35 (10) ( 2 5 ) 18 5 4 ~ ~ - - _ - -

RMS: Rhabdomyosarcoma; well: well differentiated rhabdomyosarcoma; poor: poorly differen- tiated rhabdomyosarcoma; MFH: malignant fibrous histiocytoma; FS: fibrosarcoma.

strap-shaped cells with two or more nuclei arranged in a row and/or small, round cells with one nucleus or giant cells with several nuclei (Fig. 2a). The cytoplasm of these cells was fibrillar or granular and showed varying degrees of eosinophilia. Cross- striations were rarely seen, except in one RMS in the retroperitoneum which contained many such cells. Desmin was detectable in almost all tumour cells (Fig. 2b). Many tumour cells had a few glycogen granules shown by PAS staining. Ultrastructurally, thin and thick filaments formed myofibrils and sarcomeres (Fig. 3), but their maturity varied in different tumour cells.

Poorly differentiated RMS-These tumours showed diffuse proliferation of round to oval cells (Fig. 4a) or an irregular arrangement of spindle cells. The tumour cells had relatively large nuclei and scant, slightly eosinophilic cytoplasm. Mitotic figures were common. Occasionally, larger cells with single nuclei and more eosinophilic cytoplasm were seen among small immature cells. The former cells had fairly well-formed sarcomeres. Stromal myxoid change was frequently observed. In areas in immature small cells, devoid of any differentiated cells, a diagnosis of RMS was difficult. Desmin- positivity was detectable in all these tumour cells

SOFT TISSUE TUMOURS INDUCED BY Ni3S2 267

Fig. 2-(a) Well differentiated RMS. The tumour tissue consists of strap-shaped cells with many nuclei arranged in rows and small rounded cells with one nucleus or large cells with several nuclei. (b) Well differentiated RMS: Almost all the tumour cells are strongly positive for desmin (lmmunoperoxidase with haematoxylin counterstaining)

except the small immature ones (Fig. 4b), which also had no intracellular PAS-positive granular material. Poorly differentiated myoblasts and immature cells were the main constituents of poorly differentiated regions. The cytoplasmic filaments in poorly differentiated cells varied, sometimes forming dense aggregates simulating Z bands, but not sarcomeres (Fig. 5). Immature cells had many free ribosomes or polysomes, pinocytotic vesicles, a fairly well developed RER, and some mitochondria (Fig. 6). Desmosome-like intercellular junctions were sometimes observed.

MulignantJibrous histiocytornu (MFH) MFHs developed after s.c., i.a., or i.f. injection of

Ni,S,, being commonest after S.C. (10/18) or i.f. (5/9) injection, and were diagnosed on the criteria by Enzinger and Weiss.*' A storiform and/or whorl pattern was predominant (Figs. 7a and 7b) but a hemangiopericytomatous growth pattern with dilated vascular lumina was also seen. Myxoid change of the stroma and inflammatory infiltration of cells, mainly lymphoid cells, of the periphery

of the tumour were additional findings. Neither des- min nor PAS positivity was detected in any tumour cells, even in bizarre tumour cells with abundant cytoplasm resembling differentiated myoblasts.

Ultrastructurally, three major cell types were seen: fibroblast-like cells, histiocyte-like cells, and intermediate forms. A few myofibroblasts, xan- thoma cells, and multinucleated giant cells were also seen occasionally. Fibroblast-like cells, which were characterized by a well-developed RER, a nucleus with heterochromatin, and a prominent nucleolus, were the predominant cell type (Fig. 8). Histiocyte- like cells were characterized by many cell processes on their surface, a well-developed Golgi zone with lysosomal bodies, and a nucleus with peripheral aggregates of chromatin. Intermediate cells had organelles characteristic of both fibroblast-like and histiocyte-like cells. Large bizarre cells had a large nucleus with a pominent nucleolus and abundant cytoplasm containing a well developed RER with a dilated lumen, many mitochondria, and sometimes lipid droplets, residual bodies, and peripherally located microfilament bundles.

268 M. SHIBATA ETAL.

Fig. 3-Well differentiated rhabdomyoblast. Thick and thin filament and 2 bands forming an incomplete sarcomere are seen. Bar = 1 ,u

Fihrosurconiu These tumours had a rather uniform and often

fascicular growth pattern and consisted of fusiform or spindle-shaped cells of very uniform size and shape (Fig. 9). No multinucleated or bizarre giant cells or pleomorphic cells were seen. Various amounts of interstitial collagen fibres were demon- strated by Azan staining, and well-oriented fine reticulin fibres were observed in silver-impregnated sections. Desmin and cytokeratin were not detec- table in these tumour cells.

Uncluss fied surcomu Two sarcomas in group I could not be classified.

One was anaplastic and showed diffuse prolifer- ation of small, round mesenchymal cells. The other contained scattered, small spindle cells with markedly myxoid stroma and showed extensive tissue necrosis. Both these tumours gave positive reactions for vimentin, but not for desmin or PAS- positive granules.

Two tumours in ,-group I11 were unclassified sarcomas which showed diffuse proliferation of round or oval cells of varying size with eosinophilic

cytoplasm and rather distinct cell borders (Fig. 10) and, occasionally, giant cells with both single and multiple nuclei were seen. Extensive stromal myxoid change was observed. Almost all tumour cells gave a positive reaction for vimentin, but negative reactions for desmin and cytokeratin. Unltrastruc- turally, the tumour cell showed many fine cell processes, a well-developed RER, and Golgi appar- atus with adjacent lysosmal bodies and many lipid droplets. These findings seemed to indicate a tend- ency for the cells to differentiate into histiocyte-like cells.

DISCUSSION

In this work, a single injection of Ni,S, into four different soft tissues of rats induced tumours in all sites. A high incidence of sarcomas was observed after intramuscular, subcutaneous, or intra-articu- lar injection of 5 mg of Ni,S,, but a lower incidence of tumours and a longer induction time were found after the same injection into fatty tissue. Based on the dose dependence of the carcinogenic effect of an intramuscular injection of Ni,S, reported by Sunderman,, this phenomenon may be due to

SOFT TISSUE TUMOURS INDUCED BY Ni3S2 269

Fig. 4-(a) poorly differentiated RMS. Round or oval cells have proliferated diffusely. Mitotic figures are common. (b) Poorly differentiated RMS. Round cells are positive for desmin. but small immature cells are desmin-negative. (Irnmunoperoxidase with haemdtoxyh counterstaining)

dilution of the carcinogen by its spread and the low susceptibility of fatty tissue to its carcinogenic effect. In fact, multiple small scars were observed scatted in the retroperitoneal fat and/or inside the fascia of the lumbar muscles in rats with no tumours after i.f. injection of Ni,S,

There are many reports of light and electron microscopic studies on the tumour induced by i.m. injection of Ni,S,;22-25 many of these tumours were considered to be rhabdomyosarcomas. However, it is difficult to diagnose poorly differentiated tumours without obvious myogenic differentiation as rhab- domyosarcomas. Recently, several muscle proteins such as myoglobin,I6 desmin,I4 and Z-p r~ te in ’~ have been used as diagnostic indicators of human rhabdomyosarcoma. Altmannsberger et al. l 7 repor- ted that 24 sarcomas induced by i.m. injection of Ni,S, all gave a positive reaction with an antibody to desmin and were identified histologically as rhabdomyosarcomas. We investigated the value of immunohistochemical demonstration of myo- globin, myosin, desmin, and cathepsin B in sar- comas induced by Ni,S, and found that desmin was the most reliable diagnosis of rhabdomyosarcomas as a marker of myogenic differentiation.26

All tumours induced by i.m. injection of Ni,S, were rhabdomyosarcomas. Because of the simi- larity between diving stem cells in rhabdomyosar- coma and satellite cells, Bruni suggested that muscle satellite cells were progenitors of the rhabdomyosarcomas induced by Ni3S,.27 However, rhabdomyosarcomas developed also in some rats given Ni,S, by s.c., La., or if.; all these tumour involved muscles adjacent to the injection sites. This suggests that the carcinogenic potential of Ni,S, is linked to striated muscle or its progenitor cells. Tissue necrosis, degeneration, and regene- ration of mucles occur as early reactions to i.m. injection of Ni,S,.,* The same tissue reactions in other sites, adjacent to muscle, may be due to leakage or spread of the carcinogen; another possi- bility is that the RMSs which occurred after Ni,S, injection by routes other than the i.m. route may have developed from immature pluripotential mesenchymal cells present in the connective tissue at the injection site, as Damjanov et al. reported that RMSs developed in the testis after local injection of Ni,S,.4

MFHs developed in rats after s.c., i.a., and i.f. injection of Ni,S,, and were the predominant type

270 M. SHIBATA ET AL.

Fig. 5-Poorly differentiated rhabdomyoblast. Numerous intracytoplasmic aggregates simulating a Z band are seen. Bar = 1 pm

Fig. &Immature cell in an RMS. Many free ribosomes, pinocytotic vesicles, and fairly well developed RER are seen. Bar = 1 prn

SOFT TISSUE TUMOURS INDUCED BY NijSZ 27 1

Fig. 7 1 a ) MFH. Elongated tumour cells forming distinct storiform pattern. (b) Pleomorphic area of MFH. The tumour tissue is composed of irregularly arranged spindle and giant cells. Marked infiltration of inflammatory mononuclear cells is observed

Fig. 8-Ultrastructural appearance of an MFH. Many fibroblast-like cells (F) of varying size and some histiocyte- like cells (H) are observed. Bar = 5 pm

212 M. SHIBATA ET AL.

Fig. 9-Fibrosarcoma. Uniform sized, fusiform tumour cells show fascicular growth

of tumour after S.C. or i f . injection of Ni,S,. The light microscopic and ultrastructural features of the tumour cells were similar to those in human MFH. Since the recognition of human MFH as an entity, there have been reports that MFH may be induced in experimental animals by certain carcinogens, e.g., in the subcutis by 4-(hydroxyamino)quinoline 1 dinitropyrenes, and dinitrofluoranthene;

in the testes by Ni3S,;4 and in intra-articular spaces by 9,lO-dimethyl- 1,2-benzanthracene (DMBA)."S~~

Three possible progenitor cells of MFH have been suggested: the h i~ t iocy te ,~~ undifferentiated stem cell,30 and fibrobla~t.~' In this work, ultrastruc- tural studies on experimental MFHs showed that fibroblast-like cells were predominant, but we could not determine the type of cell from which the MFHs actually originated.

Ghadially and Roy first reported experimental inducation of sarcomas by intra-articular injection of DMBA in 1966," and since then there have been several reports of induction of these tumours, but sarcomas induced by i.a. injection of Ni,S, have not been reported previously. Based on gross and light microscopic findings, Ghadially and Roy considered that DMBA-induced sarcomas were of

Fig. l&Unclassified sarcoma induced by i.a. injection of Ni,S, showing diffuse proliferation of round to oval cells in the myxoid stroma

synovial origin." On the other hand, Homma and Wiinsch" and Sakamoto" reported that sarcomas obtained by DMBA treatment had no features of synovial sarcoma and were all MFHs. Our obser- vations show that the sarcomas induced by intra- articular injection of Ni3S, can be classified into four types: rhabdomyosarcomas, MFHs, fibrosarco- mas, and unclassified sarcomas. Of the 16 sarcomas induced, eight were identified as RMSs and only three as MFHs, while five showed no histological characteristics of MFH. None had the character- istics of synovial sarcoma. On electron microscopy, the two unclassified sarcomas were found to show histiocytic differentiation, but their exact origin could not be determined.

ACKNOWLEDGEMENTS

The authors thank Dr Shigeto Yamashiro for kindly giving the Ni,S,, and Dr Jun-ichi Abe for providing anti-desmin antibody. Part of this work was supported by a Grant-in-Aid for Cancer Re- search from the Ministry of Health and Welfare, Japan.

SOFT TISSUE TUMOURS INDUCED BY Ni3S2 273

REFERENCES 1. Gilman JPW. Metal carcinogenesis: 11. A study on

the carcinogenic activity of cobalt, copper, iron, and nickel compounds. Cancer Res 1962; 22: 158-162.

2. Sunderman FW Jr. Carcinogcnicity and anticarcino- genicily of metal compounds. In: Emmelot P. Kriek E, eds. Environmental Carcinogenesis. Amsterdam: Elsevier/North-Holland Biomedical Press, 1979;

3 . Yamashiro S, Gilman JPW, HullandTJ, Abandowitz HM. Nickel sulphide-induced rhabdomyosarcomata in rats. Acta Pathol Jpn 1980; 30: 9-22.

4. Damjanov I, Sunderman FW Jr, Mitchell JM, Allpass PR. Induction of testicular sarcomas in Fischer rats by intratesticular injection of nickel sub- sulfide. Cancer Res 1978; 38 268-276.

5. Sunderman FW Jr, Maenza RM, Hopfer SM, et al. Induction of renal cancers in rats by intrarenal injec- tion of nickel subsulfide. J Environ Pathol Toxicol

6. Allen JR, Hsu IC, Carstens LA. Dehydroretronecine- induced rhabdomyosarcomas in rats. Cancer Res

7. Hatakeyama S. Morphological changes of myofibril in the carcinogenetic course of 20-methylcholanth- rene-induced rhabdomyosarcoma. Acta Pathol Jpn 1981;31: 1029-1043.

8. Konishi Y, Maruyama H, Mii Y, et al. Malignant fibrous histiocytomas induced by 4-(hydroxyamino)- quinoline 1 -oxide in rats. J Natl Cancer Znst 1982; 68:

9. Maruyama H, Mii Y, Emi Y, et al. Experimental studies on malignant fibrous histiocytomas: TI. Ultrastructure of malignant fibrous histiocytomas induced by 4-(hydroxyamino)-quinoline 1 -oxide in rats. Lab Invest 1983; 48: 187-198.

10. Ghadially FN, Roy S. Experimentally produced synovial sarcomas. Cancer 1966; 19 1901-1908.

1 1. Homma W, Wunsch PH. Experimental-induced sarcomas after intra-articular injection of 9-10- dimethyl-1-2-benzanthracene. Arch Orthop Trauma Surg 1983; 102: 111-113.

12. Sakamoto K. Malignant fibrous histiocytoma induced by intra-articular injection of 9,lO-dimethyl- 1,2-benzanthracene in the rat: pathological and enzyme histochemical studies. Cancer 1986; 57:

13. Tokiwa H. Otofuji T, Nakagawa R, et a/. Dinitro derivatives of pyrene and fluoranthene in diesel emission particulates and their tumorigencity in mice and rats. In: Ishinishi N et al. eds. Carcinogenic and Mutagenic Effects of Diesel Exhaust. Amsterdam: Elsevier, 1986; 253-270.

14. Gabbiani G, Kapanci Y, Barazzone P, Franke WW. Immunochemical identification of intermediate-sized filaments in human neoplastic cells: a diagnostic aid for the surgical pathologist. Am J Pathol 1981; 104: 2 0 6 2 16.

165-1 92.

1979; 2: 1511-1527.

1975; 35: 997-1002.

8 59-865.

23 13-2322.

15. Mukai M, Iri H, Torikata C, et al. Immunoperoxi- dase demonstration of a new muscle protein (Z- protein) in myogenic tumours as diagnostic aid. Am J Patholl984; 114: 164-170.

16. Mukai K, Rosai J, Hallaway BE. Localization of myoglobin in normal and neoplastic human skeletal muscle cells using an immunoperoxidase method. Am JSurg Patholl979; 3: 373-376.

17. Altmannsberger M, Weber K, Droste R, Osborn M. Desmin is a specific marker for rhabdomyosarcomas of human and rat origin. Am J Pathol 1985; 118:

18. Miettinen M, Lehto VP, Virtanen I. Keratin in the epithelial-like cells of classical biphasic synovial sarcoma. Virchows Arch [Cell Pathol] 1982; 4 0

19. Salisbury JR, Isaacson PG. Synovial sarcoma: an immunohistochemical study. J Pathol 1985; 147: 49-57.

20. Leader M, Patel J, Collins M, Kristin H. Synovial sarcomas: true carcinosarcomas? Cancer 1987; 59: 20962098.

21. Enzinger FM, Weiss WS. Malignant fibrohistiocytic tumors. In: Soft Tissue Tumors. St Louis, Toronto, London: C.V. Mosby, 1983; 166-198.

22. Bruni C, Rust JN. Fine structure of dividing cells and of nondividing, differentiating cells of nickel sulfide-induced rhabdomyosarcomas. J Natl Cancer Znst 1975; 54: 687-696.

23. Friedmann I, Bird ES. Electron-microscope investi- gation of experimental rhabdomyosarcoma. J Pathol

24. Hildebrand HF, Biserte G. Ultrastructural investi- gation of Ni,S,-induced rhabdomyosarcoma in Wistar rat: comparative study with emphasis on myofibrillar differentiation and ciliar formation. Cancer 1978; 42: 528-554.

25. Yamashiro S, Basrur PK, Gilman JPW, Hulland TJ, Fujimoto Y. Ultrastructural study of Ni,S,- induced tumors in rats. Acta Pathol Jpn 1983; 33: 45- 58.

26. Sano N, Shibata M, Izumi K, Otsuka H. Histopatho- logical and immunohistochemical studies on nickel sulphide-induced tumors in F344 rats. Jpn J Cancer Res 1988; 79 212-221.

27. Bruni C. Mitotic activity of muscle satellite cells dur- ing the early stages of rhabdomyosarcomas induction with nickel subsulfide. In: Mauro A e t al., eds. Muscle Regeneration. New York: Raven Press, 1979; 265- 273.

28. Daniel MR. Strain differences in the response of rats to the injection of nickel sulphide. Br J Cancer 1966;

29. Kauffman SL, Stout AP. Histiocytic tumors (fibrous xanthoma and histiocytoma) in children. Cancer 1961; 14: 469432.

30. Fu YS, Gabbiani G, Kaye GI, Lattes R. Malignant soft tissue tumors of probable histiocytic origin

85-95.

1 57-1 6 1.

1969; 97: 375-382.

2 0 886-895.

274 M. SHIBATA ET AL.

(malignant fibrous histiocytomas): general consider- ations and electron microscopic and tissue culture studies. Cancer 1975; 35: 176-198.

3 1 . Hoffman MA, Dickersin GR. Malignant fibrous histiocytoma: an ultrastructural study of eleven cases. Hum Patholl983; 14: 91 3-922.