Tumor-associated neural differentiation antigens detected on C1300 neuroblastoma cells by hybridoma monoclonal autoantibodies

Tumor-associated neural differentiation antigens detected on C1300 neuroblastoma cells by hybridoma monoclonal autoantibodies

CELLULAR 1MMUNOLOGY 68, 75-92 (1982) Tumor-Associated Neural Differentiation Antigens Detected C 1300 Neuroblastoma Cells by Hybridoma Monoclonal A...

5MB Sizes 0 Downloads 6 Views

CELLULAR

1MMUNOLOGY

68, 75-92 (1982)

Tumor-Associated Neural Differentiation Antigens Detected C 1300 Neuroblastoma Cells by Hybridoma Monoclonal Autoantibodies

on

ROBERTO P. REVOLTELLA,* RITA BUSINARO,* GIULIANA LAURO,? AND AMELIA TOESCA* *Laboratory of Cell Biology, CNR. Via G. Romagnosi 18/A. 00196 Rome, and tlnstitute of Pathology, Science, University of Rome, Rome, Italy Received November

25, 1981: accepted December 16, I981

Immune isoantisera and hybridoma monoclonal autoantibodies against syngeneic Cl 300 neuroblastoma (NB) cells were produced from BALB/c mice. Isoantisera were obtained (i) from mice immunized with membrane preparations from cloned NB cells and (ii) from mice bearing NB tumors. After repetitive absorptions on several different syngeneic or allogeneic tumor cell lines and syngeneic normal kidney, liver, spleen, bone marrow, and brain mouse tissue powders, these sera still retained antibodies reacting with tissue-differentiation antigens present on both NB cells and normal nerve sympathetic cells on cryostat whole body sections of neonatal mice. Monoclonal autoantibodies against NB cells were the products of the fusion between plasmacytoma cells and spleen cells from mice bearing syngeneic NB tumors. These anti-NB monoclonal antibodies revealed a restricted spectrum of distinct alloantigenic specificities against syngeneic bone marrow, fetal and adult brain cells, and nerve sympathetic cells present on neonatal rather than adult mice. A mixture of four monoclonal antibodies, recognizing, respectively, an epitope of the Ia complex and three distinctive neuronal-restricted antigens, proved to be a powerful and specific probe for histological immunodiagnosis of neuroblastoma, on cryostat sections of NB tumors, metastases, and tumor-draining lymph nodes.

INTRODUCTION The Cl 300 neuroblastoma (NB Cl 300) is a transplantable tumor of sympathetic nerve cells, which arose spontaneously in the body cavity of a newborn male mouse (I) and was later adapted to grow in continuous culture (2, 3). The cells of several clones of this line retain the capacity to undergo neuronal differentiation in vitro, and have for many years been considered as a model for the growth, differentiation, and function of sympathetic nerve cells (4-6). Evidence of an immune reactivity against NB cells was demonstrated in syngeneic mice with a growing NB tumor, of both the cellular (7- 11) and the humoral ( 12- 14) type. This responseis, however, usually unable to protect the recipient host against the growth of the tumor cells. Conversely, successful protection was achieved in mice preimmunized with glutaraldehyde-fixed NB cells or NB membrane preparation, so that subsequently injected viable NB cells were rejected (15, 16). It has been demonstrated that antibodies reacting with NB cells are produced in these animals; no clear evidence, 75 OOOS-8749/82/050075-18$02.00/O Copyright 0 1982 by Academic Press, Inc. All rights of reproduction in any form reserved.

76

REVOLTELLAETAL. TABLE 1

Absorption of Anti-NBR6 Hybridoma Monoclonal Antibodies from Immune (IM) or Tumor-Bearing (TB) BALB/c Mice by Syngeneic Spleen, Bone Marrow, Brain, Liver, and Kidney Cell Powders” Ratio of cell cultures absorbing serum anti-NB antibodiesb

Cells Lines and clones Neuroblastoma’ Melanoma’ Renal adenocarcinoma Fibrosarcoma Sarcoma S- 180 Testicular tumor Friend leukemia cells’ Clone 745 Clone 3BM-78/35 Fibroblasts L-29 (TK-) LD (TK-) 1DF BH4 (Polyoma) Normal tissue cell suspensions Bone marrow leukocytes Liver, spleen, kidney cells Brain Adult Fetal

Strain of origin

Before absorption

011 011 011 111

TB 316 011 011 011 011 ‘/’

IM 416 011 011 011 011 Ol’

TB 316 011 011 011 01’ 011

DBA/Z DBA/Z

111 ‘I’

‘I1 ‘/’

01’

011

011 01’

GH Cd BALB/c BALB/c

011 011 011 112

011 011 011 ‘I2

011 01’ 01’ 01’

011 011 011 011

BALB/c BALB/c

313 o/3

113 o/3

013 o/3

o/3 o/3

BALB/c BALB/c

113 313

o/3 113

o/3 o/3

o/3 o/3

A, BALB/c (C X DBA) F BALB/c SW CFW BALB/c

IM 416

After absorption

011

0 Diluted antisera (1:lOO) were mixed with equal volume of cell powders (see Materials and Methods) and incubated at 4°C for 2 hr. After centrifugation (14,OOOgfor 8 min), remaining antibody activity was tested by RIA and CF. b Number of positive cell lines or primary cultures/total number of cell lines or cultures tested. c Cells expressing Ia specificity, as determined by RIA using hybridoma monoclonal antibodies.

however, has been provided as to whether the antigenic determinants detected by these sera are unique to NB cells, or are shared by tissues of the animal as well. Xeno- and alloantisera have been used for the serological analysis of NB cells in the search of common tumor-associated antigens (reviews: 17, 18). However, these analyses have been complicated by the necessity to remove the species-specific antibodies by absorption. This paper describes the specificity of (i) serum antibodies from BALB/c mice immunized against membrane preparations of syngeneic cloned NB cells, (ii) serum antibodies from mice with growing NB tumors, and (iii) hybridoma monoclonal autoantibodies obtained by fusion between plasmacytoma cells and spleen leukocytes from NB tumor-bearing mice. In all these cases anti-NB antibodies were produced defining distinct antigenic specificities on NB cells. These antibodies also localized alloantigenic determinants associated with normal mouse tissues, including sympathetic nerve cells, on cryostat sections of neonatal mice.

Cl300 NEUROBLASTOMA-HYBRIDOMA

AUTOANTIBODIES

77

FIG. 1. Detection of antigens on clone NB6R cells with isoantisera from BALB/c mice bearing N B6R neuroblastomas, using the immunoperoxidase assay. The cells incubated with serum preabsorbed on a mixture of liver, kidney, brain, and bone marrow cell powders from 3-month-old BALB/c mice show black precipitates on the surface and cytoplasm (A), while the cells treated with the same dilution of a normal mouse serum remain unlabeled (B).

MATERIALS

AND METHODS

Mice

Inbred BALB/c, A, C57Blj6, and AKR mice were purchased from Jackson Laboratories, Bar Harbor, Maine. Outbred Swiss Webster mice were obtained

78

REVOLTELLA

ET AL.

FIG. 2. Immunoperoxidase staining of normal paravertebral nerve sympathetic cells on a cryostat whole body section of a neonatal (3-day-old) BALB/c mouse exposed to anti-NB isoantibodies obtained from mice bearing NB tumors.

from a local producer. Neonatal and 2- to 3-month-old male and female mice were used in the experiments. Cells A cell line derived from the murine NBC1300 was originally obtained 8 years ago from Dr. F. Jacob, Institut Pasteur, Paris, and several clones (NB/R series) were isolated in agar from it (9). Clone NB6R is highly tumorigenic in strain A mice (TDSo = 3 x IO3cells per mouse). This clone also originates tumors in BALB/ c mice (TDSO= 4 x IO4 cells per mouse). When a mouse is inoculated SCwith lo6 NB6R cells, a solid neuroblastoma tumor originates at the site of the inoculum and the mouse rapidly succumbs to the progressively growing tumor mass. Clones NA and neuro-2A (N-2A) (two hypoxanthine guanine-phosphorybosyl-transferase, HPRT-deficient clones) were originally supplied, respectively, by Dr. F. A. McMorris and Dr. E. Levine of the Wistar Institute of Anatomy and Biology, Philadelphia, Pennsylvania. The fibroblast NCTC clone 29 and clone LD (two

Cl300 NEUROBLASTOMA-HYBRIDOMA

AUTOANTIBODIES

79

4-10

I plate fusion 3-21

100%

4-24

FIG. 3. Specific anti-NB reactivity as detected by complement fixation (CF) assay, of a series of hybridoma antibodies during sequential limiting dilution of positive cells from the moment of fusion.

tyrosine kinase, TK-deficient clones) are derived from subcutaneous tissue of adult C3H mice. Somatic cell hybrids (NBL 5, 18, 21) were derived from the fusion of clone NA (HPRT-) and LD (TK-) in HAT-selective medium and cloned in agar. Fibroblast clone IDF originated from a BALB/c mouse. Fibroblast clone BH4, derived from BALB/c fibroblasts transfected with polyoma virus (19) was obtained

80

.b

I
ET AL.

r--mz.-‘O

; :: 2 UT nc

///

100

6.25

0.39 QO2”None

30

300 2.700 24.300

l/Dilution inhibitor (pg) l/Dilution FIG. 4. Solid-phase radioimmunoassay to determine anti-NB hybridoma antibodies. Plastic wells (96 well) were coated with NB membranes (2.0 rg protein/well), washed, saturated with BSA, and exposed overnight to hybridoma supernatant (1:lO). After washing, rabbit anti-mouse Ig antibody was added to the well followed, after new washings, by ‘251-labeledPA. (A) Specific binding by a positive hybridoma fluid (0) as compared to an unrelated hybridoma culture fluid (0). (B) Specific inhibition of antibody binding by varying concentrations of NB membranes (e) as compared to no inhibition by BALB/c kidney and liver membranes (X). (C) Typical binding curves by different anti-NB hybridoma monoclonal antibodies.

from Dr. C. Basilica, New York University. Mouse Friend-erythroleukemia (FLC) clone 745 was obtained from Dr. C. Friend, The Mount Sinai University, New York; FLC Clone 3BM/78-35 was obtained in this laboratory (20). Both FLC clones derived from DBA/2 mice. The following murine cell lines were originally obtained from the American Type Culture Collection, Rockville, Maryland: fibrosarcoma clone HSDM,C, originated from a SW mouse (ATCC CCL 148); renal adenocarcinoma clone RAG, originated from a BALB/c mouse (ATCC CCL 142); melanoma clone M-3, originated from a (C X DBA) F, mouse (ATCC CCL 53.1), sarcoma clone CCRF S-l 80, originated from a CFW mouse (ATCC CCL 9.1); testicular tumor clone l-10, originated from a BALB/c mouse (ATCC CCL 83). These cell lines were all adapted to grow in bacteriological petri plates in high glucose Dulbecco’s modified Eagle’s medium (DMEM) (Grand Island Biological Co. Rockville, Md.) supplemented with 15% fetal bovine serum (FBS) (Flow Laboratories, Ayrshire, England), 100 U/ml penicillin and 50 rg/ml streptomycin, in a 10% COJair humidified incubator at 37°C. Primary cultures of adherent bone marrow and spleen leukocytes were obtained from BALB/c mice as follows: 2 x lo8 cells (in 20 ml DMEM 15% FBS) were placed in 100 X 20-mm plastic dishes and incubated overnight at 37°C. The nonadherent cells were removed by repeatedly washing the dishes with warmed DMEM. The adherent cells were released by gently scraping the surface with a sterile rubber policeman. Such cells were 90% viable by the erythrosine-B dye exclusion test. Suspension of whole brain cells were obtained from adult or neonatal BALB/c mice. Neural differentiation of NB cells was induced by culturing for 48 hr in DMEM supplemented with 1% FBS, as elsewhere described ( 18).

Cl 300 NEUROBLASTOMA-HYBRIDOMA TABLE Binding in RIA of Anti-NB6R

2

Hybridoma Monoclonal Antibodies Hybrids, and Fibroblasts” Hybridoma

Cells

Fibroblasts Clone LD (TK-)

monoclonal

to Cl 300 NB Clones, ~__ antibodies

1

2

3

4

5

6

I

8

1320 840 1910

3400 1815 3200

2140 1620 2150

5120 6200 5100

2380 3100 1850

5100 8120 5300

6200 1400 1200

8100 2410 3120

7600 3200 4400

2650 9120 1300

5600 11210 3718

1910 6120 310

8120 3400 4120

3200 4618 3120

7400 11150 6120

3100 4250 1710

2400 3200 1420

12500

860

315 1180 1200

750 620 1840

460 120

870 1300 2100

4100 850 3100

1200 1200

2100 300 850

2310 420 1380

121

150

180

205

126

98

250

Clone:

Neuroblastoma Lines NA (HPRT-) N-2A (HPRT-) NB-Cl300 Clones NB-1R NB-6R NB-IOR Somatic cell hybrids NBL-5 NBL-18 NBL-21

81

AUTOANTIBODIES

1100 3240

180

1100 210

680

__-

9

5320 2150

a RIA = 2.0 rg protein of cell membranes (see Materials and Methods)/well were incubated with hybridoma monoclonal antibody (1:50) followed by ‘251-labeled rabbit anti-mouse Ig antibody (specific activity about 10 mCi/mmol, approximately 50,000 cpm/well).

Preparation

of Cell Membranes

Washed cells, at a concentration of not less than 5 X 10’ cells/ml in DMEM without serum, were sonicated for 2 min at 4°C. The disrupted cells were spun at 600g for 20 min at 4°C and the resulting supernatant was then centrifuged at 105,OOOgfor 90 min. The protein concentration in this membrane pellet was determined by the Lowry method (2 1). This method has been successfully used during the solubilization of antigenic determinants from various tumor cells (22) including NB6R cells (7). Similar membrane preparations were also made from mixed kidney, spleen, liver, and brain cells from normal 2-month-old BALB/c mice. This mixture was used in the experiments as control. Antibodies

against NB Cells

Zsoimmune sera. Serum antibodies were obtained in BALB/c mice immunized with membrane preparations of NB6R cells emulsified in complete Freund’s adjuvant (CFA (Difco) as previously described (16). Briefly, adult (3-month-old) mice were inoculated SCat weekly intervals for 4 weeks with NB membrane (0.1 pg protein/mouse) in CFA. After 7 days from the last injection, the mice were bled, the sera were pooled, decomplemented (56°C for 30 min), and kept frozen in aliquots at -20°C until use. This schedule of immunization proved to protect the mice against the challenge with lo6 viable NB6R cells: none of these mice developed a tumor, at least within 70 days of the challenge. On the contrary, five untreated mice developed a tumor and all died within 70 days of the challenge with

82

REVOLTELLAETAL

FIG. 5. Binding of combinations of hybridoma antibodies to NB6R membranes in solid-phase radioimmunoassay. NB membrane-coated wells were incubated with saturating quantities of one hybridoma antibody, washed, and then incubated with a second hybridoma antibody or control plasmacytoma antibody. The quantity of total antibody bound was then measured by incubation with rabbit anti-mouse Ig followed by ‘251-IabeledPA, as described under Materials and Methods. First hybridoma antibody (1) from clone 7, (2) from clone 8, (3) from clone 9.

lo6 viable NB6R cells. Immune sera were also obtained from BALB/c mice with growing NB tumors (see below). Hybridoma monoclonal autoantibodies. Monoclonal antibodies reacting with NBC1300 cells were prepared by a modification of the method of Gefter et al. (23). Briefly, BALB/c mice were inoculated SCwith lo6 viable NB6R cells per mouse. The mice rapidly developed a solid neuroblastoma tumor at the site of the inoculum. Three weeks after cell inoculation, at a time when the palpable tumor was approximatively 1 X 1 cm large, the mice were exsanguinated and the sera were collected and pooled. Spleen cells from these mice were fused with P3 X 63Ag8-SP-2/O cells (a HPRT-deficient MOPC-21-derived cell line producing no immunoglobulins) and were kept in HAT medium in Falcon T-25 flasks for 10 days. Surviving cells were then aliquoted into 96-well microtiter plates (Cooke Laboratories Products) for initial cloning. Hybridoma culture supernatants were screened for antibody activity against NB cells. The hybridomas, initially positive, were subsequently subcloned by limiting cell dilution. Stable clones were then maintained in vitro and as ascites in Pristane-primed BALB/c mice. Antibody Determination Microcomplement fixation test (CF). The method employed was adapted from the original method described by Wasserman and Levine (25). Culture supernatants, sera, and ascites were first heated (56°C for 30 min), cleared by centrifu-

Cl300

NEUROBLASTOMA-HYBRIDOMA

AUTOANTIBODIES

83

FIG. 6. Immunopezoxidase staining of NB6R eels by clone 7 hybridoma monoclonal antibodies, demonstrating a selective plasma membrane localization. FIG. 7. Immunoperoxidase staining of NB6R cells by clone 8 hybridoma antibodies: antibodies were directed against an epitope present on the plasma membrane of undifferentiated cells, as shown by the lack of staining of the processes departing from the cell body (see inset).

gation (14,000g for 5 min), and dialyzed against buffer before testing. Guinea pig serum was used as a source of complement. Solid-phase radioimmunoassay (RZA). Sera, ascites, or predialyzed culture fluids were assayed for anti-NB antibodies by an adaptation (26) of the indirect

84

REVOLTELLA

ET AL.

solid-phase radioimmunoassay of Langone (27). Polyvinylchloride microtiter plates (96 well by Cooke Laboratories Products) were used as the solid absorbing surfaces. The wells were filled with 0.05 ml of the cell membrane preparations (2.0 rg protein/well, in 0.15 M phosphate-buffered saline, pH 7.4, PBS). After 6 hr of incubation, the wells were washed and left at room temperature overnight with 1.0% BSA in PBS, to saturate any free binding site on the plastic surface. After washing, serum, ascites, or predialyzed culture fluids were added and incubated for at least 18 hr. The plates were washed again and an additional incubation was performed with a rabbit anti-mouse immunoglobulin serum (NL-Cappel Laboratories Inc. Cochranville, Pa.). Bound specific antibody was detected by addition of ‘Z51-labeledprotein A (PA, Pharmacia Fine Chemicals, specific activity approximately 100 Ci/mmol, about 50,000 cpm/well). Alternatively, bound antibody was detected after direct incubation with a saturating amount of 12?-labeled rabbit IgG (Fab) anti-mouse Ig. After 1 hr incubation with the labeled ligand, the wells were washed, cut out, and assayed for radioactivity. As a control, the specimens tested were also added to wells coated with only BSA. However, no significant binding was observed with BSA-coated plates. Results are expressed as net binding (specific antibody-control antibody). Antibody

Control

For most of these studies we used (i) normal BALB/c mouse sera; (ii) anti-NB immune sera, preadsorbed on NB cell powders to remove all NB-specific antibodies; (iii) a monoclonal antibody with an irrelevant specificity that matched with NBspecific antibodies. Immunoperoxidase

(IP) Procedure

The peroxidase-antiperoxidase method described by Steinberg (28), as modified by Gresik et al. (29), was applied for light microscopic immunocytochemistry. Antigen was identified on cytocentrifuged target cells or tissue sections by the sequential application of (i) specific rabbit serum (or hybridoma antibody followed by rabbit anti-mouse Ig antibody), (ii) sheep anti-rabbit IgG antibody; (iii) specifically purified, soluble horseradish peroxidase-antiperoxidase complex (PAP) (Dekopatts A/S); (iv) 3-amino-9-ethylcarbazol (Aldrich Chemicals Co, Inc., Milwakee, Wise.); and H202. Cl300 NB tumors were isolated from mice and kept frozen at -80°C. Cryostat sections ( 10 pm thick) of NB tumors and whole body emryo or newborn BALB/ c mice were attached to glass slides, washed in PBS before being overlayed with antibody for 60 min. Control sections were incubated with PBS, control culture supernatant, or serum. All sections were subsequently incubated with rabbit antimouse Ig serum, PAP reagent for 20 min and finally processed for staining as described above. Rabbit anti-mouse Ig antisera were fractioned with saturated ammonium sulfate (50% final concentration of the salt). The recovered immunoglobulins were then extensively dialyzed, absorbed on a mixture of NB cells, liver, brain, bone marrow, and kidney powders from normal adult BALB/c mice, centrifuged ( 15,OOOgfor 10 min) to remove aggregates, and stored in aliquots at -20°C until use. Immunoglobulins Fab were prepared according to Porter (30).

Cl300 NEUROBLASTOMA-HYBRIDOMA

AUTOANTIBODIES

85

Immunoglobulins and protein A were labeled with ‘*‘I by the chloramine-T catalyzed reaction (3 1). RESULTS Binding of Isoantisera The immune sera (from mice immunized against NB membranes and from tumor-bearing mice) were analyzed by RIA against six lines or clones of NB cells. The sera were also matched against various nonneuroblastoma cell lines. The results of these experiments are summarized in Table 1. As compared to the binding reactivity of the control serum, nonabsorbed sera contained antibodies which reacted with antigens present on some of the NB cells tested, although with significant variations among the different clones (data not reported in the table). In addition we found that the sera had antibodies reacting with antigens present also on syngeneic normal fetal brain and bone marrow cells. There were also antibodies reacting with DBA/2-derived FLC, with a testicular tumor cell line and with a polyoma-transformed fibroblast cell line but no antibodies reacting with normal liver, kidney, or spleen cells of syngeneic mice nor with sarcoma cells, uninfected BALB/c fibroblasts, melanoma, renal adenocarcinoma, or fibrosarcoma cells. The isoantisera were then extensively absorbed with FLC powders and with a mixture of normal brain, kidney, liver, spleen, and bone marrow cell powders of 3-monthold BALB/c mice. After absorption, the two sera still retained strong antibody binding against NB cells but not against the other cell lines. Conversely absorption of the sera with powders of NB6R cells completely removed this reactivity against NB cells. By IP staining smears of NB cells were incubated with PBS or normal BALB/ c mouse serum in place of specific antibodies and then exposed to anti-mouse antibody. An appropriate dilution of the normal mouse serum and of the antimouse immunoglobulin rabbit antiserum were used that did not give nonspecific binding on the target cells. The staining pattern of the same dilution of the sera after absorption was distinctive and localized on the plasma membrane, in the cytoplasm and, perhaps, the nucleus of NB cells (Fig. 1). We found that these immune sera contained antibodies capable of reacting against undifferentiated as well as differentiated NB cells. When the sera were overlayed on sections of neonatal (l- to 3-day-old) BALB/c mice, the staining patterns of the absorbed isoantisera were very distinct, as compared to control sections that were preincubated with PBS or a negative mouse serum control. Control sections showed only a very pale, diffuse, nonspecific staining with cartilagine, lens, uvea, mesenchimal tissues in the skin, and endothelial cells. In contrast the immune sera stained very strongly and specifically to the cellular compartment of the paravertebral nerve sympathetic ganglia (Fig. 2), in addition to the weak nonspecific staining seen in the controls. Absorption of the immune sera on NB6R cell powders completely removed their neuronal specificity, on cryostat sections of neonatal mice. Binding of Monoclonal Antibodies Although NB reactivity was demonstrated by the polyclonal immune isoantisera, extensive serum absorption was required before a restricted neural specificity could be demonstrated. These isoantisera contained, in general, relatively low titers of

86

REVOLTELLA

ET AL.

FIG. 8. Immunoperoxidase staining of NB6R cells by clone 9 hybridoma antibodies: antibodies detected an antigenic determinant present in the cytoplasm.

antibodies and repetitive absorptions markedly limited any further possible analysis. To approach the problem in a different way, we attempted to isolate hybridomas producing monoclonal antibodies with restricted NB specificity. Splenocytes from BALB/c mice with a growing NB tumor were fused with BALB/c myeloma cells, as described under Materials and Methods. Preliminary cloning was done by limiting dilution and hybrids were selected in two stages. Supernatants demonstrating binding by RIA to NB target cells were tested for CF activity against NB membranes, as compared to membrane preparations from allogeneic or syngeneic unrelated cells. A representative example of specific CF or RIA positivity by positive anti-NB monoclonal antibodies is reported in Figs. 3 and 4. Hybridomas scoring positive only against NB cells were recloned at least five times and stable clones were maintained in culture or passed as ascites in adult BALB/c mice. From three experiments of fusion we obtained 23 stable hybrids which produced antibodies reacting with NB6R cells, but most of these antibodies did not react with all NB clones tested, either in suspension or after induced differentiation. Only nine monoclonal antibodies consistently reacted with all NB lines and clones tested, although with significant variations among the clones (Table 2). After screening, six of these latter hybridomas were found to produce antibodies directed against antigenic determinants present also in other, unrelated cell populations, including FLC, BALB/ c, fetal or adult mouse brain cells. None of these latter monoclonal antibodies, however, bound to sarcoma, fibrosarcoma, renal adenocarcinoma, testicular tumor, liver and melanoma cell lines, or fibroblasts. Antibodies from hybridomas 1, 3, 4, and 6 strongly reacted with BALB/c fetal or adult brain cells. Antibodies from clone 2 were probably directed to Ia antigens, as suggested by the positive reactivity with cells expressing Ia specificity and by inhibition RIA tests with a monoclonal

Cl 300 NEUROBLASTOMA-HYBRIDOMA

AUTOANTIBODIES

87

FIG. 91.Immun operoxidase staining by clone 9 hybridoma antibodies of paravertebral nc:rve :S ympathetic cells, on a cryost .at whole body section of a neonatal (3-day-old) BALB/c mouse. FIG. 1IO. Immu moperoxidase staining on a cryostat section of an inguinal lymph node of a 30-day-old BALB/t : mouse with metastatic Cl300 neuroblastoma tumors. A mixture of clones 4, ‘, 8, and 9 hybrid01 ma antibl odies detected positive ceils in the paracortical area of the lymph node., ne V<:r found on serial se:ctions of ’ lymph nodes from normal mice of the same age. To reduce nonspc:cifi C peroxidase activity, tissue se:ctions were treated with H202 before exposure to antibody and PAP.

88

REVOLTELLAETAL.

antibody IgG-Fab against BALB/c mouse Ia specificity (data not reported). Only monoclonal antibodies produced by hybridomas 7, 8, and 9 showed restricted reactivity against NB cells and no reactivity against all other unrelated cells tested (Table 3). These three latter monoclonal antibodies reacted against distinct sites on NB cells. In RIA, NB6R membranes were first saturated with clone 7 antibody and then exposed to ‘251-labeled IgG-anti-mouse Ig-Fab antibodies. The resulting binding was compared after adding clone 8 or 9 antibody to see if further binding occurred. Monoclonal immunoglobulins from an unrelated hybridoma were also tested as control. If NB6R membranes were first saturated with clone 7 antibody, additional binding was obtained with clone 8 or 9 antibodies. The same reciprocal results were obtained with the other two antibodies. The results of these experiments are reported in Fig. 5. IP staining on NB cells revealed that hybridoma 7 antibodies reacted against a membrane component of NB cells (Fig. 6). Absorption with unrelated cell powders did not reduce binding to NB cells, whereas absorption of the antibodies on NB cell powders completely removed this NB membrane positivity. Hybridoma 8 antibodies reacted with NB cells resulting in a mainly peripheral pattern of staining, compatible with membrane staining; apparently these antibodies reacted more against undifferentiated rather than against differentiated NB cells with long processes (Fig. 7). The clone 9 antibody-defined antigen was mainly cytoplasmatic and gave a positive pattern of staining either on differentiated or on undifferentiated NB cells (Fig. 8). On cryostat sections of BALB/c mice, clone 9 monoclonal antibodies produced an intensive cytoplasmic staining on sympathetic cells of the paravertebral ganglia. This reactivity was significantly stronger in l- to 3-day-old mice, than in adult (Zmonth-old) mice (Fig. 9). The pattern of TABLE 3 Binding in RIA of Anti-NB6R Hybridoma Monoclonal Antibodies to Murine Cell Suspensions” Ratio of cell cultures Cells Lines Neuroblastoma” Melanoma’ Renal adeno-carcinoma Fibrosarmma Sarcoma s- 180 Testicular tumor Friend leukemia cells’ Fibroblasts Lines L29. LD Clone IDF Clone BH4 (Py) Normal tissue cell suspensions Bone marrow leukocytes‘ Spleen adherent monocytes Brain Adult Fetal

Strain of origin

3

4

616

516

w

111 O/I 012 O/I O/I 11.2

O/I 01’ 012 01’ O/I *I*

11’ O/l 012 01’ O/I 012

012

O/l

012 O/l O/I

O/l II’

BALB/c BALB/c

013 o/3

313 313

BALB/c BALB/c

o/3 113

o/3 o/3

A, BALB/c (C X DBNFI BALB/c SW CFW BALB/c DBA/Z

GH BALB/c BALB/c

Clone:

I

516

O/I 011 012 011 WI *I*

o/2 Ill

2

absorbing

hybridoma 5

anti-N8

6

antibod& 7

8

9

616 011 O/I 012 O/I O/I

616 WI

O/l O/I

616

616

O/I

O/l

WI

O/I

012

012 O/I O/l

o/2 WI

212

012

012

01.2 012

012

012

O/I

O/I O/l

011 O/I

012 O/l

012

O/l Ill

l/3 013

o/3 o/3

o/3 o/3

013 013

013 313

o/3

o/3 013

113 ‘I3

213

316

O/I O/I 012

O/l

O/I

012 O/I O/I 012

012

O/I

O/I O/I

o/3 013

o/3 013

o/3 o/3

o/3 o/3

o/3 013

o/3 013

‘RIA = 2.0 fig protein of NB6R membranes/well were incubated with hybridoma antibody (1:lO) followed by ‘z’I-labeled rabbit antimouse Ig antibody. Inhibition of specific antibody binding was attained after preabsorption of the antibody with equal volume of cells (at least 10’ cells) of different origin. b Number of positive cell cultures/total number of cultures. ‘Cells expressing Ia specificity, as determined by RIA using hybridoma monoclonal antibodies.

Cl300 NEUROBLASTOMA-HYBRIDOMA

89

AUTOANTIBODIES

TABLE 4 Immunodiagnosis of Neuroblastoma by Hybridoma Monoclonal Antibodies

Site of t”“lOr growth’

Cl300 NB clone Injected

Subcuta

NB-IR

Antigens*

Expression of ncuronal characteristics’ Th

14-3-2

H-2

TWI0r

+

+

+

+

+

+

+

Tissue culture

+

+

+

-

+

+

+

+ +

+ +

‘+

+ + + +

Subculls

NBdR

Tumor T,ssue culture

+ +

+ +

+ +

+ 5

+ +

Suhcut,s

NA

Tumor Tissue culture

+ +

+ +

+ +

t +

+

Clone

I

2

+

3

monoclonal reactivitp

la

Material

NF

Hybridoma

4

5

t

4

-

*

t

6

antlbcdy

7

8

9

PO01 4. 7. x. 4

+

+

-

+

i

+

+

+

+

I

+ +

+ +

+ +

+ r

+ +

i +

+ +

-

+ +

+ +

+ +

+ +

= BALB/c mxe were injected SC with IO6 cloned NB cells (three mice for each clone). After I5 days, the solid tumor masses were removed. The cells from half of each tumor were mechanically dispersed and adapted to grow III culture for 4 to IO passages. The second half of the tumor, one piece was processed for electron microscopy, the second was frozen. ‘Cryostat sections of the tumor or smeared culture cells were analyzed by the PAP test for detecting la antigens and H-2 haplotype using monospccific alloanttbodies. ‘The neuconal speafic 14-3-2 protein was detected on cryostat sections of the tumor or on smeared NB cells by PAP onmunoatd~~~g. using a rabbit anti-14-3-Z protein antiserum. a gift of Dr. A. Grasso, Rome. The presence of catecholamines (tyrosine hydroxylase. Th) was determined m cell extracts as in (34). Neurotilaments (IO nm diameters) were detected by electron microscopy “As determined by PAP immunostaining. An arbitrary scale of comparative reactivity was adopted ranging from negatwe to pos~twe

staining was very similar to the one produced by the isoantisera from the immune mice. Thus, although clone 9 antibody-reacting antigen is located in the cytoplasm of the tumor cells, the specific antibody appeared to react selectively in or on normal sympathetic cells. The epitopes detected by hybridomas 7, 8, and 9 antibodies are not alloantigens exclusively associated with the BALB/c genome. Sympathetic cells of inbred C57B1/6, A, AKR, and outbred Swiss Webster mice were as susceptible to stain with the three monoclonal hybridoma antibodies as the sympathetic cells of BALB/c origin (data not reported). Immunodiagnosis

of Neuroblastoma

Using Monoclonal

Antibodies

A solid tumor mass arising in the left side of a BALB/c mouse injected 15 days before with lo6 NB6R cells was removed. The same was done from tumor masses arising in mice preinoculated with two different NB clones (NBlR, NA). Histologically, all these tumors were composed of small, round, undifferentiated cells and a specific histologic diagnosis was made of “undifferentiated neural tissue.” Electron microscopic evidence of neurofilaments, detection of catecholamines and 14-3-2 protein in the tissue extracts defined these tumor cells as “neuroblastoma.” Isolated cells from each tumor were placed in culture and tested for binding of hybridoma antibodies. The results of IP staining on isolated NB cells or on cryostat sections of the tumors, summarized in Table 4, indicated that cells from all these NB tumors reacted with clones 7, 8, and 9 antibodies. Some tumors also reacted with clones 2, 3, 4, 5, and 6 antibodies but only a few of these tumors reacted with clone 1 antibody. Clone 8 antibodies specifically reacted more with NB cells in culture for 4-10 passagesrather than with NB cells tested directly from the tumor mass. It was of interest to find that a mixture of clones 4, 7, 8, and 9 hybridoma antibodies appeared to be a better probe for detecting NB tumor cells in frozen tissue sections of a neuroblastoma mass rather than the administration of each

90

REVOLTELLA

ET AL.

single antibody. Thus not only the four monoclonal antibodies together did not interfere with their individual binding abilities but also, on the contrary, they appeared to be more effective for NB cell detection than each antibody given alone. The same clones 4,7,8, and 9 antibody mixture was also used as a probe to monitor the presence of NB cells metastases in mice. Sheffler et al. (32) demonstrated that Cl 300 NB cells can originate tumors that give origin to metastaseswhen inoculated id into syngeneic neonatal mice. We exposed 3-day-old BALB/c mice to a light whole body X-ray irradiation (100 rad). The mice were then injected id with lo5 viable NB6R cells (0.02 ml per mouse). All mice developed within 30 days disseminated neuroblastoma tumors, mostly localized in the lungs, liver, bones, and adrenals. IP staining using the pool of hybridomas 4, 7, 8, and 9 showed a definite positivity in the tumors and the intense NB cell staining was clearly distinct from background levels seen in the normal tissues not invaded by the tumor masses.As an example, Fig. 10 shows these antibodies detecting positive NB cells on serial sections of a tumor-draining inguinal lymph node of one of these mice. No positivity was detected in serial sections of lymph nodes of control mice pre-X-treated and then inoculated with CFA instead of NB6R cells. DISCUSSION Our aim has been to use antibodies to detect tumor-associated antigens on mouse Cl 300 neuroblastoma cells. We produced isoantisera from BALB/c mice immunized against membrane preparations of syngeneic cloned NB cells (clone NB6R) and also from mice with a growing NB tumor. We have also shown that hybridomas producing anti-NB autoantibodies can be established using spleen cells, from these mice developing neuroblastomas. Isoantisera demonstrated a wide range of antibody specificities, including reaction in vitro and in vivo with alloantigens expressed on different normal tissues, such as bone marrow, fetal and adult brain, and sympathetic nerve cells. There were also antibodies defining Ia membrane markers and antigens associated to virus-transformed unrelated target cells. Thus conventional anti-NB isoantibodies proved to be highly heterogeneous. However, small quantities of relatively low-titer antibodies were obtained with this approach: purification of antibody specific for NB-common antigens by repeated absorption with cells of different normal tissues, or unrelated allogeneic or syngeneic tumor cell lines, markedly reduced the antibody titer of the isoantisera, thus preventing any further possible approach toward identifying and isolating more restricted NB tumor-associated antigenic components. The development of hybridomas that secrete monoclonal autoantibodies reacting with NB cells has circumvented this main difficulty. Most of the monoclonal antibodies tested reacted with antigenic epitopes also present on unrelated cell lines or normal tissues of the mouse, including bone marrow, and fetal or adult brain. Clone 2 antibodies detected an epitope of the Ia complex system. Only monoclonal 7, 8, and 9 antibodies reacted in vitro with most NB cell lines and clones tested, but not with other unrelated cells. The three antigens detected by these latter antibodies are distinct immunologically and are located in different compartments of the NB cell. The antigen detected by hybridoma 7 antibody is a membrane component present on differentiated as well as undifferentiated NB cells whereas hybridoma 8 antibody detects a membrane component expressed on undifferentiated more than on differentiated NB cells. Clone 9 antibody detects a cytoplasmic epitope. On cryostat sections of NB tumor masses,

Cl300 NEUROBLASTOMA-HYBRIDOMA

AUTOANTIBODIES

91

clones 7 and 9 antibodies reacted strongly and specifically with NB cells. Clone 8 antibody reacted with NB cells adapted in culture for 4 to 10 passagesmore than on primary tumor cell suspensions, suggesting that the corresponding antigen detected by these antibodies is specific for a primitive stage of differentiation of neuroblastoma cells. A combination of clones 4, 7, 8, and 9 hybridoma antibodies proved to be a powerful and specific probe for detecting NB cells in ectopic body areas of the mice such as metastases and tumor-draining lymph nodes. Further characterization of these three monoclonal antibodies on cryostat whole body sections of neonatal and adult BALB/c mice revealed a very important finding, that these three epitopes were also expressed on normal nerve sympathetic cells. In particular, these three antigens are apparently more expressed on nerve sympathetic cells of neonatal (l- and 3-day-old) mice than on those of adult (2-month-old) mice. These antigens appear to be commonly expressed in different mouse strains. These results thus indicate that NB cells express a variety of alloantigens that are capable of eliciting a syngeneic antibody response. They also demonstrate that the use of hybridomas for the production of monoclonal antibodies against differentiation and tissue-specific antigens on tumor cells is an important and useful approach for diagnosis of the disease. In humans, hybridoma monoclonal antibodies have been recently obtained which react with established human neuroblastoma cell lines and tumors (33). These positive hybridoma antibodies anti-human NB also reacted with alloantigens detected on normal spleen, bone marrow cells, and/ or fetal brain cells, and also on different tumors all derived from neuroectoderm. Although these epitopes did not appear to be tumor specific, they proved to be extremely useful for diagnosis of neuroblastoma. It should be stressed again that even if none of our positive antibodies are by themselves tumor specific, the use of a mixture of these antibodies to detect positive cells, in ectopic areas in a host with a growing NB tumor, gives sufficient information about the nature of the positive cells. The question of whether tumor-specific antigens actually exist on NB cells, or whether some antigens that are expressed on normal nerve sympathetic cells at some stage of development and appear in large quantities on tumor cells can be called tumor antigens remains unanswered. Perhaps, our screening and detection methods select for these antigens. Further production and characterization of other anti-NB monoclonal antibodies that react with NB tumors but not with normal adult cells will provide more insights on this problem. Moreover, these antigens should be useful for defining the lineage and, perhaps, for identifying subpopulations of nerve sympathetic cells and isolating cell components to be used as markers of specific stages of their differentiation. ACKNOWLEDGMENTS We wish to thank Mrs. Eliana Vigneti for expert technical assistance,Dr. Roberto Accolla for providing us with anti-mouse Ia hybridoma monoclonal antibody, Dr. E. Gresik for useful advice in performing the immunocytochemical assays.This work was supported in part by the Progetto Finalizzato “Controllo della Crescita Neoplastica” by the Italian National Council of Research.

REFERENCES 1. Dunham, L. J., and Stewart, H. L., J. Not. Cancer Inst. 13, 1299, 1953. 2. Augusti-Tocco, G., and Sato, G., Proc. Nat. Acad. Sci. USA 64, 311, 1969. 3. Schubert, D., Humphreys, S., Baroni, C., and Cohn, M., Proc. Nat. Acad. Sci. USA 64, 3 16, 1969.

92 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14.

REVOLTELLA

ET AL.

Nelson, P., Ruffner, W., and Nirenberg, M., Proc. Nat. Acad. Sri. USA 64, 1004, 1969. Minna, S., Glazer, D., and Nirenberg, M., Nature New Biol. 235, 225, 1972. McMorris, F. A., and Ruddle, F. H., Develop. Biol. 39, 226, 1974. Revoltella, R., Bertolini, L., and Diamond, L., Inr. J. Cancer 18, 521, 1976. Philips, M. E., Cipazak, P. M., Della Penta, D., Schakner, M., J. Nat. Cancer Inst. 59, 1215, 1977. Fiorani, M., Butler, R., Bertolini, L., and Revoltella, R., Eur. J. Cancer 14, 217, 1978. Procicchiani, G., Fiorani, M., and Revoltella, R., Cell. fmmunol. 44, 39, 1979. Revoltella, R., Brahe, C., and Procicchiani, G., Cell. Immunol. 47, 115, 1979. Terman, D. S., Stewart, I., Tavel, A., and Kirch, D., Cancer Res. 35, 1761, 1975. Oldstone, M. B. A., J. Nat. Cancer Inst. 54, 223, 1975. Faraggiana, T., Biasio, L., Marinozzi, V., Butler, R., Bertolini, L., and Revoltella, R., Virchows Arch. B. Cell Pathol. 27, 23, 1978. 15. Bertolini, L., Diamond, L., and Revoltella, R., Cancer Res. 36, 2111, 1976. 16. Lauro, G., Businaro, R., Butler, R., and Revoltella, R., J. Nat. Cancer Inst. 40, 1169, 1980. 17. Esber, E., and Martin, J. W., Gaslini 11, 157, 1979. 18. Revoltella, R., Butler, R., Bertolini, L., and Businaro, R. In “Pontificia Academiae Scientiarum Scripta Varia,” Vol. 45, p. 87, 1980. 19. Della Valle, G., La Bella, F., and Basilica, C., unpublished. 20. Revoltella, R., Bertolini, L., and Friend, C., Proc. Nat. Acad. Sci. USA 76, 1464, 1979. 21. Lowry, 0. H., Rosebrough, N. J., Farr, A. L., Randall, R. J., J. Biol. Chem. 193, 265, 1951. 22. Strober, S., Appella, E., and Law, L. W., Proc. Nat. Acad. Sci. USA 67, 765, 1970. 23. Gefter, M. L., Margulies, D. H., and Scharff, M. D., Somat. Cell Gene?.3, 231, 1977. 24. Littlefield, J. W., Science 145, 709, 1964 25. Wassermann, E., and Levine, L., J. Immunol. 87, 290, 1961. 26. Businaro, R., Butler, R., Rubenstein, A. E., and Revoltella, R., J. Neurosci. Res. 6, 89, 1981. 27. Langone, J. J., J. Immunol. Methods 24, 269, 1978. 28. Steinberg, L., “ImmunoCytochemistry,” Wiley, New York, 1979. 29. Gresik, E., and Barka, T., J. Histochem. Cytochem. 25, 1627, 1977. 30. Porter, R. R., Biochem. J. 73, 119, 1959. 31. Greenwood, F. C., Hunter, W. M., Glower, J. S., Biochem. J. 89, 114, 1963. 32. Sheffer, B. A., Repman, M. A., and Schengrund, C. L., Cancer Res. 39, 711, 1979. 33. Kennett, R. H., and Gilbert, F., Science 203, 1120, 1979. 34. Philipson, D. T., and Sandier, M., Brain Res. 90, 283, 1975.