Int. J . Cancer: 21, 516-522 (1978)

NEOPLASTIC TRANSFORMATION OF HAMSTER BRAIN CELLS I N VITRO BY POLYOMA VIRUS C. DE M ~ c c o M. , F. TRIPIER, J. HASSOIJN, C. LIPCEY,G. MEYERand M. TOGA Unite de Recherches de Canctrologie exptrimentale ( U . 119 de L'INSERM) 27, Bd. Lei' Roirre, 13009 Marseille, France The transformation of cultivated hamster brain cells by polyoma virus i s reported. The transformed cell line contained polyoma virus-specific nuclear, surface and transplantation antigens. Subcutaneous and intracranial inoculations revealed high tumorigenicity of the cells. Brain-specific S 100 protein was found in these tumors with immuno-peroxidase staining, suggesting that they were of a nervous nature. Both in vivo and in vitro, the cells had glial features as studied by phase contrast, light and electron microscopy. Type-H virus-like particles were found in the tumor cells and might have played a role in the viral transformation.

Brain cells maintained in monolayer culture can undergo neoplastic transformation after infection in vitro by various papovaviruses (Shein, 1967; Santoli et a/., 1975; Tanaka et al., 1976; TixierVidal and de Vitry, 1976). As far as we are aware, there axe only two reports dealing with the transformation of brain cells by polyoma virus (Shein, 1968, 1970). However, no definite proof of the nervous nature of the cells and the viral origin of transformation was given. We describe herein the establishment of a line of hamster brain cells transformed in vitro by polyoma virus. This line, named HCxPy, contains the brainspecific S 100 protein (Moore, 1965) and contains pol yoma virus-specific antigens. MATERIAL AND METHODS

Animals Inbred Syrian hamsters from o u r colony were used. Cells The primary cultures were initiated from newborn hamster cerebral hemispheres. Leptomeninges and pia mater vessels were carefully removed. Cerebral hemispheres were finely minced and dissociated with 0.25% trypsin and the cell suspension seeded into Falcon plastic flasks. The growth medium was PUC N-16 supplemented with 25% fetal calf serum, glucose ( 5 mglml) and glutamine (2 mM). Cultures were incubated at 37" C in a humidified atmosphere containing 5 % CO, and fed at approximately 3-day intervals. Polyoma virus infection of cells After 12 days, when a monolayer of flat polygonal cells was obtained, the cells were trypsinized, washed and infected with POlYOma virus (Toronto strain produced in Primary mouse kidney cultures) at a multiplicity of 500 PFU/cell, for 3 h at 37" C with

stirring. Then, the cells were washed three times and seeded into new plastic flasks. Infected cultures and uninfected cultures, kept as virus free controls, were re-fed every 3 or 4 days and subcultured when they reached confluency. Detection of polyoma virus-specific antigens HCxPy cells were tested at passage 9 for polyomaspecific nuclear antigen (TAg) and at passages 10,24 and 63 for pol yoma-specific surface antigen (SAg) by the indirect immunofluorescence technique (Meyer, 1971). Normal hamster kidney cells and spontaneously transformed hamster brain cells (named " CT ") from our laboratory, were used as negative controls and polyoma virus-transformed hamster fibroblasts (8 Spy) as positive controls. To demonstrate the tumor-specific transplantation antigen (TSTA), adult hamsters were protected by nine consecutive subcutaneous injections of the infected brain cells at 3-day intervals with amounts of cells increasing from 5 x l o 3 to 5 x lo6. One week later they were challenged with 5 times the 50% tumor dose (TD,,) of a polyoma virus-induced hamster sarcoma (CT 54). The control animals were immunized either with polyoma virus or with CT cells, or were unimmunized. After 6 weeks' observation, the number of tumors was recorded. Morphology of the cultured cells The cells were observed at various subculture levels by phase-contrast and electron microscopy. For the latter, the cultured cells were fixed in 2% glutaraldehyde in cacodylate buffer, post-fixed in 1 % osmium tetroxide in the same buffer, dehydrated in alcohol and embedded in Araldite inside the culture flasks. Tumorigenicity assay Newborn or adult hamsters were injected subcutaneously or intracranially with inocula of 5 x loE and loEcells respectively, and then examined twice weekly for the presence of tumors. Tumor fragments were fixed in a solution of 4 % paraformaldehyde0.1 % glutaraldehyde in PBS, embedded in paraffin, and sectioned for histologic examination and immunoperoxidase staining. For electron microscopy, tumor fragments were fixed in 4% glutaraldehyde and embedded in Araldite.

Received: October 10, 1977 and in revised form February 13, 1978. Requests for reprints should be addressed to: Dr. J. Hassoun, Laboratoire de Neuropathologie-Facult6 de Mkdecine, 27, Bd. Jean Moulin, 13005 Marseilles, France.

POLYOMA-TKANSFOKMED BRAIN CELLS

517

FIGURE1 - HCxPy in vifro, passage 42. (a) Note two kinds of cells: small stellate and large binucleated flat cells. Phase contrast, x400; (b) stellate cells showing branched and anastornotic processes. Phase contrast, x 400; (c) ultrastructure of a stellate cell. Finely dispersed chromatin without margination,short ergastoplasmicchannels and microfilaments, x 13,500; ( d ) detail of c: fascicle of 7-9 nm microfilaments, X 33,500.

518

DE MlCCO ET AL. TABLE I1

Detection of the S 100 protein

TUMORIGENICITY OF HCxPy CELLS

S 100 protein was demonstrated on histologic sections of paraffin-embedded tumors with peroxidase-anti-peroxidase (PAP) staining (Sternberger, 1974). Antisera against S 100 protein, kindly supplied by Dr. L. F. Eng and Prof. P. Mandel, were used in 1/25, 1/50, and 1/100 dilutions. The specificity of the staining for brain cells was checked on the following tissues : hamster brain, benign human astrocytoma, hamster kidney and polyoma virus-induced hamster sarcoma. Negative controls were also performed by substituting non-immune rabbit serum for specific antiserum.

Primary tumor

RESULTS

4

1X l O 6

5

5 x lo6 2 x 105

About 120 days after viral infection, during the 4th passage in vitro, several foci of large, refringent cells with increased multilayered growth appeared in the infected cultures. No such alterations occurred in the uninfected controls during a 6-month observation period. Confluent cultures of transformed cells split at a ratio of 1/2 became confluent within 7 days as compared with 15 or 20 days for the uninfected controls. Gradually, the growth rate of transformed cells increased and after the 7th passage, cultures could be divided at a ratio of 1/2 every 48 h. PUC medium was then replaced by Eagle's minimum essential medium (MEM) containing 10% fetal calf serum, glucose and glutamine at the same concentrations as above. This transformed cell line named " Hamster cortex polyoma " (HCxPy) has been maintained for 24 months and is now in the 75th subculture. By phase-contrast microscopy, two kinds of cells were shown to grow simultaneously : small, stellate, triangular or multipolar cells with very long, thin branched processes, and large flat cells which were often multinucleated (Fig. la, b). A cell type intermediate between these two types was occasionally observed. All types displayed the characteristics of transformed cells : high nucleo-cytoplasmic ratio, numerous nucleoli, tridimensional growth of confluent cultures with piling-up of the cells. By electron microscopy, the small, multipolar cells showtd a clear, spherical nucleus with a finely dispersed chromatin and a scanty cytoplasm containing numerous ribosomes, mitochondria, short ergastoplasmic channels and few 7- to 9-nm microfilaments (Fig. l c , d ) . The large cells presented an TABLE I DETECTION OF POLYOMA-SPECIFIC TSTA I N HCxPy CELLS

Immunogenic agent

Number of animals inoculated with sarcomatous cells Number of tumors obtained

virus

HCxPy cells

;is

Unimmunized animals

15

15

15

15

6

6

15

15

Passage No.

Time cei, dose Number of tumors/ of appearance of cumors Number of animals (days)

5x

lo8

20 to 30

2x106 1 x 108

13 21

2

1 x106

10 to 14

3

2 x 108 1 x106

7 to 13 18 10 to 20

1

6

15

10

irregular, dark nucleus and a large cytoplasm with numerous ribosomes and mitochondria but were devoid of filaments. Regarding the polyoma virus-specific antigens, 10% of the HCxPy cells were found to be positive for T Ag, whereas no kidney cells were so. Depending on the subculture, staining for S Ag was positive in 60% to 90% of the HCxPy cells. One per cent of the hamster kidney cells and 2 to 5 % of the spontaneously transformed hamster brain cells (CT) showed non-specific fluorescence. As shown in Table I, HCxPy cells were as effective as polyoma virus in protecting hamsters against challenge by a polyoma-induced transplantable sarcoma whereas CT cells were not. This immunogenic property confirms the presence of a pol yoma-specific transplantation antigen (TSTA) on the HCxPy cells. Tumors appeared 20 to 30 days after subcutaneous injection of 5 x lo6cells from passage 18 into newborn hamsters. Intracranial tumors arose 40 to 60 days after local inoculation of lo6 cells from passage 41 into adult animals. These tumors were serially ti ansplanted in adult hamsters. After five passages in vivo, tumors arose after subcutaneous inoculation of 2 x loKcells (Table 11). Gross examination revealed whitish-grey tumors with areas of necrosis and hemorrhage. There was local invasion of surrounding tissues and sometimes lung metastasis. Microscopic examination showed highly cellular and monomorphic tumors composed of criss-crossed bundles of elongated cells (Fig. 2a). Numerous areas of necrosis surrounded by a rosette-like arrangement of cells were found. Some tumors contained pleomorphic cells with large, eosinophilic, Mallorypositive cytoplasms and bizarre, peripheral nuclei (Fig. 2b). Electron microscopic examination showed elongated cells with numerous tapered processes. Their cytoplasm often contained filaments 7-9 nm wide (Fig. 2c, d ) which were more numerous in the processes where they formed wavy bundles around aggregates of dense bodies and mitochondria. Some cells exhibited clumps of type-H virus particles (Bernhard and Tournier, 1964) inside endoplasmic reticulum channels (Fig. 2e).

POLYOMA-TRANSFORMED BRAIN CELLS

519

FIGURE 2 - HCxPy tumors. (a) Intracerebral tumor infiltrating leptomeninges, cortex and perivascular sheaths, H. and E., x 100; (6) subcutaneous tumor showing elongated or swollen astrocyte-like cells, H. and E., x 100; (c) and ( d )ultrastructure of an intracerebral tumor cell. Presence of numerous fascicles of 7-9 nm microfilaments; c, X 18,000, d, x 53,500; (e) type-H virus-like particles in an ergastoplasmic channel of an intracerebral tumor cell, x 80,000.

520

DE MICCO ET AL.

FIGURE 3 - S 100 immunostaining. (a) Subcutaneous HCxPy tumor stained with PAP technique. S 100 protein is present in most of tumor cells, x 500; (b) and (d), details of PAP stained tumor cells. Note S 100-positivecytoplasmas and processes. Nuclei are negative, X 1,000. (c) negative control: NCxPy tumor reacted with non-immunized rabbit serum, X500.

The PAP staining for S 100 protein was found to be significantly positive in 100% of HCxPy cells, the cytoplasm of which appeared brown with all the dilutions of specific antisera used (Fig. 3a, b, d). The stain could vary slightly from one cell to another, but these differences corresponded to the plane of section of the cells and to the size of the cytoplasms. Hamster kidney and polyoma virus-induced hamster sarcoma were unstained as were HCxPy tumors reacted with non-immune rabbit serum (Fig. 3c). In normal hamster brain, only the perikarya and processes of neuroglial cells were stained. Human neoplastic astrocytes were also positive.

DISCUSSION

Recent works have shown the transforming ability of various papovaviruses for brain cells (Santoli et al., 1975; Tanaka et a/., 1976; Tixier-Vidal and de W r y , 1976). To date; mouse polyoma virus has not shown such properties since, in vivo, intracerebral inoculation induces sarcomas (Rabson and Kirschstein, 1960). The results we report here show that transforrnation of brain cells by polyoma virus is possible in vituo. The latency times for transformation (4 months) are much longer than those required for connective tissue cells: as reported by Stoker and MacPherson

POLYOMA-TRANSFORMED BRAIN CELLS

(1961) and by Montagnier et al. (1966) colonies of transformed cells could be detected on the top of a confluent monolayer 13 to 15 days after plating of pol yoma-infected hamster fibroblasts. This difference could explain the exclusive occurrence of sarcomas after intracerebral inoculation of polyoma virus. Nevertheless, HCxPy cells are indisputably transformed: they are highly tumorigenic in the hamster and have a very low serum requirement, since with a 1 % concentration of serum cultures divided at a ratio of 1/2 they are confluent after 72 h, while normal glial cells need 15% to reach confluence within 10 to 15 days. The persistence of the polyoma virus genome in the transformed cells is proved by the significant presence of polyoma virus-induced antigens : antiserum to S antigen gave a positive reaction with HCxPy cells as well as polyoma virus-transformed hamster mesenchymal cells. There was no reaction with either normal hamster mesenchymal cells or spontaneously transformed brain cells. This type of serum gives no cross-reaction with cells transformed by methylcholanthrene, Rous sarcoma virus or hamster sarcoma virus (unpublished data) but it can sometimes give positive results with some SV40 virus-transformed cells or with embryonic fibroblasts. As Barra et al. (1977) have proposed, this is due to the heterogeneity of S antigen, of which only a part is specific to polyoma virus. The presence of TSTA on HCxPy cells was investigated with an immunogenicity assay. We showed that these cells were able to protect adult hamsters against a grafted polyoma-virus-induced hamster sarcoma. In order to eliminate the activity of some tissue antigens or antigens related to transformation but distinct from polyoma-virus-induced antigens, we used as controls uninfected spontaneously transformed hamster brain cells (CT). Despite their tumorigenicity, these cells did not protect the animals, so we concluded that the immunization obtained with HCxPy cells was specific for the virus. This is supported by the results of Barra et al. (1977) who showed the specificity for polyoma virus of TSTA activity evidenced by an imniunogenicity assay. Thus, extracts of polyoma virus-transformed hamster cells protected hamster and mice against subsequent challenge by polyomavirus-transformed syngeneic cells but were unable to protect hamsters against the challenge by SV40transformed cells. To our knowledge, type-H virus particles have never been reported in neuroglial cells. As an oncogenic potential has been proposed for these particles (C6sarini and de Micco, 1972; Mayo et al., 1973), we suggest that they might have played a role with polyoma virus in the transformation of these cells. The identification of the HCxPy cells was based on morphologic and immunohistochemical criteria. During transformation, the cell morphology became simplified and only two cell types persisted: stellate and large flat cells. Stellate cells with branched processes were reminiscent of the TL, and TL, types described by Rorke et al. (1975) and the 2 n A type described by Ponten and MacIntyre (1968) in dis-

521

sociated cultures of human brain, and identified as glial cells. Similar patterns and interpretation were reported by Maunoury et al. (1972) in dissociated human brain cultures. The ultrastructural characteristics of these stellate cells (finely dispersed chromatin without margination, short ergastoplasmic channels, microfilaments 7-9 nm in diameter) are suggestive of a glial origin (see Markesbery and Lapham, 1974). Large, flat, sometimes multinucleated cells lacked any definite criteria of differentiation by light and electron microscopy. However, phase-contrast observation of various intermediates between these and stellate cells suggested that they might be a n undifferentiated glial precursor. Identical observations have been reported by Lim and Mitsunobu (1974) in brain cultures of rat fetus. The histological appearance of the tumors induced in hamsters by subcutaneous or intracerebral inoculation of HCxPy cells was more often that of poorly differentiated tumors associating spindle and large cells. A great number of experimental virus-induced tumors of the central nervous system possess these characteristics (Walker et a/. 1973; Murao et nl., 1974). More rarely, an obvious astrocytic differentiation could be observed in HCxPy tumors, with the presence of large cells having clear, eccentric nuclei and abundant eosinophilic cytoplasm reminiscent of grade I1 or 111 human astrocytomas. From the ultrastructural point of view, however, the tumors always presented a number of cells possessing bundles of microfilaments, 7-9 nm in diameter, which are classically described in human (Poon et al., 1971) and experimental (Vick and Bigner, 1972) astrocytomas. S 100 protein which was present in HCxPy tumors is considered to be specific to nervous tissue (Moore 1965). For most authors, this protein is localized in neuroglia (Cicero et al., 1970; Ludwin et al., 1976; Matus and Mughall, 1975). S 100 protein has also been reported in neurons (Haglid et al., 1975; Tabuchi and Kirsch, 1975). In our material, the localization of the protein was exclusively glial in normal control hamster brains although it was impossible to determine the astrocytic or oligodendrocytic nature of the cells stained with the PAP technique. All non-nervous tissues were always negative in our material. In conclusion, we feel that there is sufficient morphologic and immunohistochemical evidence to assert that the HCxPy transformed cell line is neuroglial in nature. ACKNOWLEDGEMENTS

We wish to thank Prof. P. Mandel and Dr. L. F. Eng for helpful advice and the generous gift of S 100 antisera. We thank Mrs. N. Beau and Mr. G. Cohen for technical assistance. This work was supported by grants from the GEFLUC-Marseilles (Groupement des entreprises franqaises dans la lutte contre le cancer) and by research contract No. 75.5.054.6. from INSERM (Institut National de la Sante et de la Recherche MCdicale). Dr. M. F. Tripier is " ChargCe de Recherches '' at INSERM.

522

D E MlCCO ET AL.

TRANSFORMATION NEOPLASIQUE I N VITRO DE CELLULES CEREBRALES DE HAMSTER PAR LE VlRUS POLYOME Les auteurs rapportent les resultats d’une experimentation portant sur la transformation in vitro de cellules cerebrales de hamster par le virus polyonie. La lignee cellulaire transformee H C x Py possi.de les antigenes specifiques de ce virus (TAg, SAg, TSTA). Elk niontre une forte Cumorigenicitt5 a p r b inoculation intracerebrale ou sous-cutank. Son origine nerveuse est confirmee par la prksence de la p r o t h e S 100 mise en evidence par les reactions immunohistochimiques. In vivo et in vitro, les cellules prtsentent les caracteristiques morphologiques de la neurologie, aussi bien en contraste de phase qu’en microscopie optique et electronique. Les particules virus-like H presentes dans les cellules tumorales pourraient jouer un rBle adjuvant dans le processus de transformation. REFERENCES

BARRA,Y., ASTIER,A. M., and MEYER,. G . , Isolation of polyoma virus-induced surface antigens in hamster cells : potassium chloride solubilization and differential precipitation. J. nat. Cancer Inst., 58, 259-262 (1977).

by human adenovirus type 12. Acta med, Okayama, 28, 47-58 (1974). PONTBN,J., and MACINTYRE, E. H . , Long term culture of normal and neoplastic human glia. Acta path. nricrobiol. scand., 14, 465-486 (1968). BERNHARD, W., and TOURNIER, P., Infection virale inapparente de cellules de hamster dCcelee par la microscopie POON, T. P., HIRANO, A., and ZIMMERMAN, H. M., Electron electronique. Ann. Inst. Pasteur (Paris13107, 447-452 (1964). microscopic atlas of brain tumors. Grune and Stratton, New York (1971). C~SARIN J.I P., , and DE MICCO,C., Studies on type-H viruslike particles in hamster: their role in oncogenesis. Int. J. RABSON, A. S., and KIRSCHSTEIN, R. L., Intracranial sarcomas Cancer, 10, 174-185 (1972). produced by polyoma-virus in Syrian hamsters. Arch. Path., CICERO, T. J., COWAN,W. M., MOORE,B. W., and SUNTZEFF, 69, 663-671 (1960). RORKE, L. B., GILDEN, D. H . , W R O B L E ~ S K Z., Aand , SANTOLI, V., The cellular localization of the two brain-specific proteins D., Human brain in tissue culture. IV. Morphological S 100 and 14-3-2. Brain Res., 18, 25-34 (1970). characteristics. J. comp. Neurol., 161, 329-340 (1975). HAGLID, K. G., HAMBERCER, A., HANSSON, H. A., HYDEN, H., SANTOLI, D.! WROBLEWSKA, Z.: GILDEN,D. H.: GIRARDI: A ., PERSSON,L., and RONNBACK,L., Immunohistochemical H.: Human brain in tissue culture. 111. localisation of S 100 protein in brain. Nature ( L o n d , ) ,258, and KOPROWSKI, PML-SV40 induced transformation of brain cells and estab748-749 ( 1 975). lishment of pxmanent lines. J . comp. Neurol., 161, 317-328 LIM,R.,and MITSUNOBU, K., Brain cells in culture: morpho(1975). logical transformation by a protein. Science, 185,63-65 (1974). SHEIN,H. M., Neoplastic transformation induced by simian LUDWIN,S . K., KOSEK,J. C., and ENG, L. F., The topovirus 40 in Syrian hamster neuroglial and meningeal cell graphical distribution of S 100 and G F A proteins in the cultures. Arch. ges. Virus Forsch., 22. 122-142 (1967). adult rat brain: an immunohistochemical study using SHEIN:H. M., Neoplastic transformation of hamster astrohorseradish peroxidase-labelled antibodies. J. comp. Neurol., cytes in vitro by simian virus 40 and polyoma virus. Science, 165, 197-207 (1976). 159, 1476-1477 (1968). MARKESEERY, W. R., and LAPHAM, L. W., A correlated light SHEIN, H . M., Neoplastic transformation of hamster astroand electron microscopic study of the early phase of growth cytes and choroid plexus cells in culture by polyoma virus. in vitro of human fetal cerebellar and cerebral cortex. J. J. Neuropafh. exp. Neurof.,29, 70-88 (1970). Neuropath. exp. Neurol., 33, 1 13-1 27 ( I 974). STERNBERGER, L. A., Immunocytochemistry, p. 129, PrenticeMATUS.A,, and MUCHAL, S., lmmunohistochemical localisaHall Inc., Englewood Cliffs (1974). tion of S 100 protein in brain. Nature ( L e n d . ) , 258, 746-748 STOKER,M.? and MACPHERSON, I., Studies on transforma(1975). tion of hamster cells by polyoma virus in vitro. Virology, 14, MAUNOURY, R., VEDRENNE, C., ARNOULT,J., CONSTANS, 359-370 (1961). J. P., and FEBVRE, H., Culture in vitro de tissu glial normal et TABUCHI,K.. and KIRSCH,W. M., Immunocytochemical neoplasique. Croissance, cytologie, ultrastructure. Neurolocalization of S 100 protein in neurons and glia of hamster chirurgie, 18, 101-120 (1972). cerebellum. Brain Res., 92. 175-180 (1975). MAYO,J., LOMBARDO, J. L., KLEIN-SZANTO, A. P. J., CONTI, TANAKA, R., KOPROWSKI, H., and IWASAKI. Y., Malignant C. I., and MOREIRA, J. L., An oncogenic virus carried by transformation of hamster brain cells in vitro by human hamster kidney cells. Cancer Res., 33, 2273-2277 (1973). papovavirus BK. J. nat. Cancer Ins?., 56,671-672 (1976). MEYER,G . , Viral genome and oncogenic transformation: TIXIER-VIDAL, A,, and DE VITRY, F., Ultrastructure and nuclear and plasma membrane events. Advanc. Cancer Res., cytochemical features of SV 40 transformed hypothalamic cell 14, 71-159 (1971). lines. Cell Tissue Res., 171, 39-60 (1 976). MONTAGNIER, L., MACPHERSON, I., and JARRETT, O., An VICK,N. A., and BICNER,D. D., Some structural aspects of epithelioid variant of the BHK 21 hamster fibroblast line and dog brain tumors induced with the Schmidt-Ruppin strain of its transformation by polyoma virus. J. nar. Cancer Inst., 36, the Rous sarcoma virus. In: W. G. Bingham, Jr. (ed)., 503-512 (1966). Progress in experimental tumor research, Vol. 17, pp. 59-73, MOORE,B. W., A soluble protein characteristic of the nervous Karger, Basel (1972). system. Biochem. Biophys. Res. Comm., 19, 739-744 (1965). WALKER, D. L., PADGETT, B. L., ZU RHEIN, G ., and ALBERT, A. E., Human papovavirus (JC); induction of brain tumors MURAO, T., OHMORI, H., SONOBE,H . , MATSUO, K., in hamsters. Science, 181, 674-676 (1973). TSUTSUMI, A,, and OGAWA,K., Brain tumors induced in rats