Journal of the Neurological Sciences, 1976, 30:55-64

55

© Elsevier Scientific Publishing Company, Amsterdam - Printed in The Netherlands

TWO SEPARATE M E M B R A N E - B O U N D A N T I G E N S ON H U M A N G L I O M A CELLS IN TISSUE C U L T U R E D E T E C T E D W I T H SERA F R O M G L I O M A PATIENTS BY I M M U N O F L U O R E S C E N C E

CI. SOLHEID, G. LAURO and G. PALLADINI Department of Neuropathology, Born-Bunge Foundation, Berchem-Antwerp (Belgium) and Cattedra di Patologia Generale, Facolt~ di Scienza and Istituto di Biologia Generale, Facolt& di Medicina e Chirurgia, Roma (Italy)

(Received 16 March, 1976)

SUMMARY Sera from patients with malignant and benign gliomas, as well as sera from healthy donors, were tested by indirect immunofluorescence to detect antibodies against antigens on the membrane of glioblastoma, astrocytoma, reactive perimetastatic glia, normal glia and fibroblasts in tissue culture. Sera from glioblastoma patients reacted with glioblastoma, astrocytoma and reactive glial cells; they were negative on normal glia and on fibroblasts, whereas sera from astrocytoma patients were unreactive. Sera from control patients were positive in 7 out of 15 cases, although some differences were noted in the pattern of reaction. Absorption with astrocytoma powder, with glioblastoma and reactive glial cells indicated that all the positive cell lines expressed an astrocytoma-associated antigen " A " , while only glioblastoma lines and reactive glial line shared a supplementary antigen " G " . Neither of these 2 antigens seemed to be present in significant amount in normal brain, since the positive reactions could not be abolished by absorption with normal brain powder. The relationship between these 2 antigens and the process of increasing malignancy in gliomas is briefly discussed.

INTRODUCTION The recognition of the brain as an "immunologically privileged site" (Medawar 1948) resulted in a considerable interest in central nervous system immunology, esThis work was supported by a grant from the "Fonds voor Geneeskundig Wetenschappelijk Onderzoek" nr. 3.0033.76, and from the "Fonds Baron Kronacker" for brain tumour research. Reprint requests to: Dr. CI. Solheid, Born Bunge Foundation, 59 F. Williotstraat, B-2600 Berchem, Belgium.

56 pecially as certain neurological diseases are now considered to be auto-immune in nature. Nervous system components were found to be antigenic when injected in allogeneic or heterologous hosts. These studies of nervous system specific antigens were extended from the normal brain to the primary nervous system tumours, especially when it became clear that other cancers could express particular antigens: tumourspecific transplantation antigens (Morton, Eilber and Malmgren 1970), onco-fetal antigens (Gold and Freedman 1965) or antigens coded by tumour-viruses (Klein, Clifford Klein, Smith, Minowada, Kourilsky and Burchenal 1967). The search for brain-tumour antigens was paralleled by a search for a possible immune response elicited in the host, at both cellular (Levy, Mahaley and Day 1972; Kumar and Taylor 1973: Brooks, Netsky, Normansell and Horwitz 1972) and humoral levels (Febvre, Maunoury and Constans 1971; Eggers 1972; Kornblith, Dohan, Wood and Whitman 1974). The different papers report variable results as to the existence of turnout-specific antigens in human brain neoplasms: with heteroantisera against normal brain extracts, Delpech, Delpech, C16ment and Laumonier (1972) were not able to demonstrate tumour-specific antigens in glioblastomas; Wickremesinghe and Yates (1971) however, demonstrated normal brain antigens in low malignancy gliomas which progressively disappeared as malignancy increased. With immunofluorescence, Wahlstr~m, Linder and Saksela (1973) demonstrated the presence of normal brain antigens in the cytoplasm of fetal, adult and malignant human gila in tissue culture, using a rabbit antiserum against lyophilisate of normal brain. With a similar antiserum against a glioblastoma lyophilisate, they demonstrated turnout-specific membrane antigens on tissue-cultured glioblastomas (Wahlstr/Sm, Linder, Saksela and Westermark 1974). In autologous human systems, Trouillas (1973) demonstrated "carcino-fetal" glial antigens, common to malignant glioma and fetal glial cells. Using indirect membrane immunofluorescence on living human glial cells in tissue culture, we looked for the presence of anti-glioma antibodies in the serum from glioma and other patients. In this paper, we present evidence that the glioma patients' sera contain antibodies which cross-react with antigens on the membrane of glioma and reactive glial cells in tissue culture. However, we observed a reaction in some non-glioma patients too. Absorption experiments were subsequently undertaken to study the relationship between the antigens on the several lines used, and to attempt to differentiate between reactivity of glioma and non-glioma patients. MATERIALAND METHODS Cell cultures

Cultures were initiated from biopsy specimens obtained at neurosurgery for neoplastic, vascular or traumatic brain disease. The specimens were finely minced with scissors and further dissociated by incubation for approximately 20 min in a solution of 0.12 ~o trypsin (Difco) and 0.02 ~ EDTA. After 3 washings with nutrient mixture (see below), the cells were seeded on glass coverslips in Falcon 35mm Petri dishes and in square glass bottles. The nutrient mixture is Ham's F 10 supplemented with 15 fetal bovine serum (Gibco or Flow) and 1 ~ .antibiotic-antimycotic mixture (Gibco)

Fig. 1. Patterns of positive immunofluorescence: A: fine punctiform pattern, Evans Blue counterstain. x 128; B: more positive, patchy, irregular pattern, Evans Blue counterstain, x 128; C: irregular spots, no counterstain, x 128; D: numerous spots outlining the cell shape, no counterstain, x 128.

58 containing penicillin, streptomycin and fungizone. When the cell monolayers reached confluency, the cells were subcultured by brief treatment with the trypsin-EDTA mixture, followed by pipetting with nutrient mixture, so that the cells became detached. The culture period for the cell lines used in this study varied between 6 months and 3 years; in this period some of them had been frozen in liquid nitrogen. The cell lines used were 6 cultures derived from glioblastomas (T63, T10I,T158, T185, T187, T244), one culture derived from an oligodendroglioma (T54) which after a few passages appeared to be composed of astrocytes, and which we shall refer to as astrocytom a, one culture derived from reactive glia in and around a cerebral metastasis of breast cancer (G253), one culture derived from normal glia obtained at lobectomy for intracerebral hemorrhage (G146) and one culture of fibroblasts from the spinal dura mater (F264). T54 and G146 were shown to possess some astrocytic metabolic features, i.e. the selective vacuolisation under ouabain treatment (Solheid and Palladini 1974). Sera were obtained from neurosurgical patients in the immediate pre- and postoperative periods and from other patients during their observation in the neurological clinic. The sera were diluted 1:1 with phosphate-buffered saline (PBS), divided in small aliquots and frozen until use.

Membrane immunofluorescence Coverslips carrying the test cells were incubated in a moist chamber with 1 drop of test serum for 30 min; they were then rinsed 3 times with PBS and further incubated with 1 drop of goat- or rabbit anti-human I g G conjugated to F I T C (Nordic Pharmaceuticals) for 30 min. After 3 washings, the cultures were fixed and counterstained TABLE 1 REACTIVITY OF THE SERA FROM TUMOUR PATIENTS (GLIOBLASTOMAS, GLIOSARCOMA, MIXED OLIGODENDROGLIOMA, ASTROCYTOMAS) ON CULTURES FROM GLIOBLASTOMAS, ASTROCYTOMA, REACTIVE GLIA, NORMAL GLIA, FIBROBLASTS Reactivity on glioblastoma cultures was tested on 3 cell lines at least (see text). Patients No. 51 60 65 87

Target cells Name

glioblastoma astrocytoma reactive glia normalglia

V.Br... (glioblastoma) -~ Ver...a -t Via... 4Wil... ~-

! 00

Bel...

121 122 108 119

Bay... Jk Wag... -~Chr... (gliosarcoma) -+ V.Ee...(mixed oligodendroglioma) V.d.S... (astrocytoma) Exe...a

40 52

:k

+

:k

Two serum samples tested from same patient.

±

J

fibrobtasts

59 with 4 ~ formaldehyde and 0 . 0 5 ~ Evans Blue in PBS for 10 min. All procedures were carried out at room temperature. The coverslips were then mounted in PBSglycerol 1:1 ; the preparations sealed with nail varnish and examined under a LeitzOrthoplan microscope with Ploem-Opak fluorescence system. Specificity controls of the reaction included incubation without test serum or blocking the reaction with nonfluorescent anti-IgG before addition of the fluorescent antibody. Control results were uniformly negative. Absorption o f test sera Pieces from normal brain and 1 astrocytoma obtained at autopsy were homogenised andlyophilised. Absorption was performed by adding 100 mg lyophilisate per mi of serum overnight at 4 ° C. Afterwards the tubes were centrifuged at 5.000 rpm and the supernatant collected. Some absorptions were performed on cells in monolayer culture: this method was preferred as the first approach since we did not know the possible action of trypsin on the cell surface antigens: 0.3 ml of serum was laid on dense monolayers in 35 mm Petri dishes for 4 hr at 4 °C; afterwards the serum was carefully aspirated with a fine Pasteur pipette and transferred to a "fresh" monolayer. This was repeated until the underlying monolayer appeared negative after addition of the conjugate.

RESULTS The pattern of reactivity was a punctiform or patchy fluorescence irregularly scattered among the cell surface, sometimes outlining the cell shape. The staining intensity was variable and ranged from a few spots to a dense, reticular pattern (Fig. 1). There were no signs of "capping" or endocytosis under our experimental conditions. Tables 1 and 2 show the results obtained with the sera from glioma and non-glioma patients (which we shall refer to as " t u m o u r " and "control" patients). Tests were scored as positive ( ÷ ) when at least 10 ~/o of the cells exhibited more than 2 fluorescent spots by visual impression; weak positive results ( ~ ) refer to a positivity of less than 10 ~ . Most of the sera were tested against an average of 4 glioblastoma lines; when it occurred, a positive reaction was never limited to 1 cell line: the positive sera reacted in a similar way, although some quantitative differences could occur with all glioblastoma lines. The T158 line however showed a weak positive reaction with 1 serum and was negative with all others, and was therefore discarded from further experiments. The results are thus presented as reactivity on glioblastomas, without further specification of the cell lines. As can be seen from Table 1, all but one of the glioblastoma patients showed a positive reaction against glioblastoma, astrocytoma and reactive glial cells in culture. They were negative with normal glia and fibroblasts. Control patients also showed a variable degree of reactivity, not always with the same pattern as the glioblastoma sera, in that they were not always positive on the astrocytoma cells. The sera from astrocytoma patients were negative on all the cell lines. The specificity of the positive reactions was further investigated by absorptions with tissue powders and cultured cells. A first absorption with normal brain lyophilisate

6O 'FABLE-2 REACTIVITY OF NON-TUMOURAL PATIENTS ON THE SAME CULTURES The first 5 sera were tested against the panel of cultures, the others (without number) only against 1 or 2 glioblastoma cultures. Patients

Target cells

No.

Name

77 102

Ama... Koc...

105 107 113

Diagnosis

depression cerebrovascular insufficiency Giu... healthy Lau... healthy Wit... healthy Vande... mult. sclerosis Left... cervical spondylosis Van Be... epilepsy Van Ae... meningioma Deth... aneurysm Poe... headache Brey... cervical arthrosis Van Bra... meningioma Rob... headache Sel... post-concussional syndrome

glioastroblastoma cytoma

reactive glia

--

--

--

+ + + +

+ -+ --

+ -÷ ÷ ~-

normal glia

fibroblasts

m

m

_

--

-+

did n o t change any pattern. A second experiment involved the a b s o r p t i o n o f the positive sera on a s t r o c y t o m a lyophilisate: this resulted in a loss of the reactivity on the a s t r o c y t o m a T54, with little or no change o f the pattern on the glioblastomas or the reactive glia. The only exception was the serum f r o m a mixed oligodendrogliomae p e n d y n o m a (119) whose reactivity remained unchanged. A b s o r p t i o n on glioblastoma or reactive glia monolayers, either directly or after absorption on astrocytoma, b r o u g h t a b o u t an abolition o f all positive reactions. A b s o r p t i o n on fibroblast monolayers had no effect. These absorption experiments are summarized in Table 2. The situation was somewhat different with the control sera: while absorption with glioblastoma or reactive glia abolished all reactions, absorption with a s t r o c y t o m a lyophilisate resulted in a negative pattern on T54 and on some gliobtastomas, with no change on the reactive glia (Table 3). DISCUSSION O u r experiments indicate the presence, in the sera f r o m gtioblastoma and f r o m some control patients, of antibodies which react with antigens on the m e m b r a n e o f h o m o l o g o u s glioblastoma, a s t r o c y t o m a a n d reactive glial cells in tissue culture. A reaction with normal glia and fibroblasts was never observed, I n studies o f other h u m a n cancers, antibodies against h o m o l o g o u s or autotogous m e m b r a n e - b o u n d or cytoplasmic t u m o u r a l antigens were demonstrated by immunofluorescence or c o m p l e m e n t

61 TABLE

3

EFFECT OF THE ABSORPTION OF SERA WITH ASTROCYTOMA POWDER, GLIOBLASTOMA AND REACTIVE GLIAL CELLS IN MONOLAYER: THESE ABSORPTIONS WERE BOTH SUMMARIZED lN THE SAME COLUMN, SINCE THEY GAVE SIMILAR RESULTS

Sera absorbed were the positive sera of Table 1 and the fully tested positive sera of Table 2. Patients

Absorbing tissue astrocytoma powder

glioblastoma or reactive gila monolayer

glioblastoma astrocytoma reactive glia

glioblastoma astrocytoma reactiveglia

~(±) ~ + { (±) + ± q

. -N N

Gliablastomas No. No. No. No. No. No. No.

51 60 66 65 87 100 122

. --

.

.

.

N N

-N N -N N

-N N -N N

+

--

--

--

----

-~ w ~

--

i~

N N -N

N N -N

N N -N

-----

~ + ~ + + ~

Mixed oligodendroglioma N o . 119

+

~

Non-tumoural No. No. No. No.

+ N

102 105 107 113

+(--) -k- ( - - ) +(--) +(±)

( ± ) : refers to a slight diminution of the pattern on some cultures. ( - - ) : refers to negativisation of the pattern on some cultures. = not done.

fixation (Morton and Malmgren 1968; Morton et al. 1970; Whitehead 1973; The, Huiges, Schraffordt Koops, Lambert and Nieuweg 1975). In these studies, as in ours, an important number of control patients' sera reacted with the tumour ceils. Several hypotheses were put forward to explain this positivity: common exposure of patients to a tumour-associated infectious agent (Morton et al. 1970) or acquisition of new antigenic determinants by the tumour cells during the life in vitro, which makes them react with sera from patients with the same tumour, from patients with other tumours and from normal donors (The et al. 1975). If the latter possibility exists in our study, it should be regarded as a phenomenon restricted to neoplastic and reactive glia, since our 2 "normal" cell lines, glial and fibroblastic, remained negative. The distribution of the positivity within the panel of cell lines tested suggests that the antibodies in the positive sera are directed against antigens on glial cells with a strong growing pattern in vivo, neoplastic and reactive. This is not, a priori, surprising since metabolic similarities between "reactive" and neoplastic astrocytes have been suggested (Benda, Takesada and Sweet 1970) but this points out that reactive glial cells behave immunologically differently from normal gila; it must be emphasized, however, that these reactive astrocytes were found in and around a brain metastasis of mammary carcinoma, and

62 it would be interesting in that respect to compare the antigenic structure of such reactive glia with reactive gila in non-neoplastic disease, such as brain abscess. The specificity of our experiments and eventual cross-reactivities were investigated by absorptions with normal brain lyophilisate, astrocytoma lyophilisate, with glioblastoma and reactive glial cultures. The lack of any effect of absorption with normal brain rules out the possibility of auto-immunity against normal brain antigens in our patients. This finding, together with the negativity of all reactions on normal glial cultures, is in contrast with the findings of Wahlstrt~m, Saksela and Troupp (1973) who found a cellular immunity among patients against normal adult and fetal glial cells. Unfortunately, we were not able to study fetal tissues, but our experiments clearly indicate that a humoral immune response against antigens other than adult glia is mounted by the host. Further absorptions were carried out to analyse the various patterns of reactivity. Absorption with astrocytoma lyophilisate was able to abolish the reactivities of glioblastoma sera on astrocytoma cells, leaving unaffected or slightly diminished the reactivity on glioblastoma and reactive glial cells. Absorption with these glioblastoma or reactive glial cells on the contrary removed at once all reactivity. We therefore conclude that the sera of our glioblastoma patients were able to react with 2 different antigens or groups of antigens: the first being expressed on the surface of astrocytoma, glioblastoma and reactive glial cells (which we shall refer to as " A " antigen), the second being shared by glioblastoma and reactive glia ( " G " antigen). Antibodies against the " A " group are removable by absorption with astrocytoma powder and by glioblastoma or reactive glia, those against the " G " group only by treatment with the glioblastoma and reactive glial cells. The situation with the control sera is somewhat different: the pattern of positivity is nearly the same as with glioblastoma sera, but absorption with astrocytoma powder results sometimes in a loss of reaction on the glioblastoma cells (sera 102, 105, 107), indicating, in our working hypothesis, a reactivity chiefly against " A " antigens, also present on these cells. An interesting fact is that the serum samples from astrocytoma patients did not show any evidence of sensitisation. This is perhaps due to the integrity of the bloodbrain barrier in these benign gliomas. During the preparation of this manuscript, Coackham and Lakshmi (1975) described the presence of an antigen on the surface of human astrocytomas in culture, using a heteroantiserum against cultured astrocytoma cells and complement-dependent cytotoxicity as an experimental system. The cytotoxicity could be diminished markedly by absorption with other astrocytoma cells. Unfortunately, these astrocytomas, as well as the one used for immunisation, were of grades II-IV and showed areas ofmalignancy. Following our hypothesis the cells used for immunization, for cytotoxicity tests and for absorptions expressed both " A " and " G " antigens, so that no distinction could become apparent. In human tumour systems, gliomas provide a model of graded malignancy, ranging from benign fibriUary astrocytomas to very anaplastic glioblastoma. Our studies with alloantisera indicate that the evolution towards malignancy is reflected

63 in the antigenic structure of the glioma cells. Astrocytomas express a particular antigen (A) but, probably due to blood-brain barrier integrity, there is not necessarily sensitization of the host. When the cells become more malignant, a new antigen appears (G). These two antigens are able to sensitize glioblastoma patients a n d some of the n o n t u m o u r a l subjects. They are also f o u n d o n reactive perimetastatic glial cells. This antigenic similarity between neoplastic a n d reactive glia was also detected when studying the cellular i m m u n i t y in our patients (Solheid a n d D u m o n , in preparation). F u r t h e r a b s o r p t i o n experiments with the use of solid i m m u n o a b s o r b e n t s are n o w u n d e r t a k e n to differentiate between these 2 antigens. Finally, it should be stressed that the preservation o f the antigenic properties of o u r cell lines is evident, despite m o r p h o l o g i c a l " s i m p l i f i c a t i o n " in vitro ( L u m s d e n 1971). Only 1 culture o u t of the 10 used in this work proved devoid of antigenic d e t e r m i n a n t s d e m o n s t r a b l e with the patients' sera. There should thus be some optimism a b o u t the use of long-term cultures of n o r m a l a n d neoplastic glia in studies of braintumour immunology.

REFERENCES Benda, P., M. Takesada and W. Sweet (1970) Demonstration of an astrocyte-specific cerebroprotein by an immunofluorescence study of human brain tumors, J. Neurosurg., 33 : 281-286. Brooks, W. H., M. G. Netsky, D. E. Normansell and D. A. Horwitz (1972) Depressed cell-mediated immunity in patients with primary intracranial tumors, J. exp. Med., 136: 1631-1647. Coakham, H. B. and M. S. Lakshmi (1975) Tumour-associated surface antigen(s) in human astrocytomas, Oncology, 31 : 233-243. Delpech, B., A. Delpech, J. Cl6ment and R. Laumonier (1972) I~tude immunochimique et immunologique des tumeurs du cerveau humain, Int. J. Cancer, 9: 374-382. Eggers, A. E. (1972) Autoradiographic and fluorescence antibody studies of the human host immune response to gliomas, Neurology (Minneap.), 22: 246--250. Febvre, H., R. Maunoury and J. P. Constans (1971) l~tude immunologique des tumeurs c6r6brales in vitro, Neuro-Chirurgie, 17: 203-207. Gold, P. and S. O. Freedman (1965) Specificcanceroembryonic antigens of the human digestive system, J. exp. Med., 122: 467-481. Klein, G., Clifford, P., E. Klein, R. T. Smith, J. Minowada, F. M. Kourilsky and J. H. Burchenal (1967) Membrane immunofluorescence reactions of Burkitt lymphoma cells from biopsy specimens and tissue cultures, J. nat. Cancer Inst., 39: 1027-1044. Kornblith, P. L., F. C. Dohan Jr., W. C. Wood and B. O. Whitman (1974) Human astrocytoma - Serum-mediated immunologic response, Cancer, 33 : 1512-1519. Kumar, S. and G. Taylor (1973) Specific lymphocytotoxicity and blocking factors in tumours of the central nervous system, Brit. J. Cancer, 28, Suppl. I: 135-141. Levy, N. L., M. S. Mahaley Jr. and E. D. Day (1972) In vitro demonstration of cell-mediated immunity to human brain tumors, Cancer Res., 32: 477-482. Lumsden, C. E. (1971) The study by tissue culture of tumours of the nervous system. In: D. S. Russell and L. J. Rubinstein(Eds.), Pathology of Tumours of the Nervous System, Edward Arnold, London, pp. 334-420. Medawar, P. B. (1948) Immunity to homologous grafted skin, Part 3 (Fate of skin homografts transplanted to brain, to subcutaneous tissues and to anterior chamber of the eye), Brit. J. exp. Path., 29: 58-69. Morton, D. L. and R. A. Malmgren (1968) Human osteosarcomas - - Immunologicevidence suggesting an associated infectious agent, Science, 162: 1279-1281. Morton, D. L., F. R. Eilber and R. A. Malmgren (1970) Immune factors in human cancer - - Malignant melanomas, skeletal and soft tissue sarcomas, Progr. exp. Tumor Res. (Basel), 14: 25-42.

64 Solheid, CI. and G. Palladini (1974) Morphological changes induced by ouabain in normal and neoplastic h u m a n glia in monolayer culture, Acta neuropath. (Berl.), 28 : 253-260. The, T. H., H. A. Huiges, H. Schraffordt Koops, H. B. Lambert and H. O. Nieuweg (1975) Surface antigens on cultured malignant melanoma ceils as detected by membrane immunofluorescencc method with h u m a n sera - - Lack of tumor-specific reactions on melanoma lines. In : W. Hijmans and M. Schaeffer (Eds.), Fifth International Conference on lmmunofluorescence and Related Staining Techniques, New York Academy of Sciences, New York, N.Y., pp. 528-540. Trouillas, P. (1973) lmmunologie et immunoth6rapie des tumeurs c6r6brales --- Etat actuel, Rev. neurol., 128: 23-38. Wahlstr6m, T., E. Linder and E. Saksela (1973) Glia-specific antigens in cell cultures from rabbit brain, human foetal and adult brain, and gliomas, Acta microbiol, stand. Section B, 81: 768-774. Wahlstr6m, T., E. Linder, E. Saksela and B. Westermark (1974) Tumor-specific membrane antigens irt established cell lines from gliomas, Cancer, 34: 32-35. Wahlstr/Sm, T., E. Saksela and H. Troupp (1973) Cell-bound antiglial immunity in patients with malignant tumors of the brain, Cell. lmmunoL, 6: 161-170. Whitehead, R. H., (1973) Fluorescent antibody studies in malignant melanoma, Brit. J. Cancer., 28: 525-529. Wickremesinghe, H. R. and P. O. Yates (1971) Immunological properties of neoplastic neural tissues, Brit. J. Cancer, 25: 711-720.