Acta path. rnicrobiol. scand. Sect. C, 84: 105-111, 1976

TUMOR Fc RECEPTORS AND TUMOR-ASSOCIATED IMMUNOGLOBULINS OLAV TONDER*, ENGIKOLAI C. KRISHNAN, WILLIAM R. JEWELI., PAULA. MORSE,JR. and LOREN J. HUMPHREY Department of Surgery, University of Kansas Medical Center, Kansas City, Kansas, USA

Tonder, O., Krishnan, E. C., Jewel], W. R., Morse, P. A., Jr. & Humphrey, L. J. Tumor Fc receptors and tumor-associated immunoglobulins. Acta path. rnicrobiol. scand. Sect. C, 84: 105-111, 1976. Tissue sections and cell suspensions from ten malignant tumors were tested for Fc receptors using sheep erythrocytes sensitized by rabbit IgG antibodies. Surface bound IgG on cells from the same tumors were estimated using an antiglobulin consumption test with 1 2 5 1 labelled human IgG as reference. The amount of IgG present per 106 cells varied from < l o 0 ng to approximately 600 ng. When these amounts of IgG were plotted against the Fc receptor activity of corresponding tumors, seven of the tumors were distributed along a line showing a n inverse linear relationship; i.e. tumors with large amounts of IgG on their cell surfaces had the lowest Fc reactivity and vice versa. Cells from three of the tumors had lower amounts of IgG on their surface than expected from this relationship. However, the lack of correlation could be explained by the focal distribution of the Fc positive tissue within non-reactive tissue. The cells from these areas presumably carry less IgG on their surface and thereby reduce the quantity of IgG calculated per 106 cells. Prolonged washing of tumor sections resulted in stronger Fc receptor activity, and correspondingly washed cells had lower amount of IgG on their surface. Presumably the Fc receptor can bind IgG in vivo. Key words: Fc receptor; tumor-associated Ig.

0. Tonder, M.D., Broegelmann Research Laboratory for Microbiology, MFH-bygget, 5016 Haukeland sykehus, Bergen, Norway.

Received 16.ix.75

Accepted 22.x.75

Evidence has been presented that many human malignant tissues possess a receptor for the Fc region of IgG, particularly recognizing Fc in antigen-antibody complexes in preference to IgG alone (15, 16). It is also well established that many malignant tumors contain significant amounts of immunoglobulins (8, 14, 18), and the evidence

*

Visiting Professor from Broegelmann Research Laboratory for Microbiology and Department of Microbiology, School of Medicine, University of Bergen, Bergen, Norway.

8 Acta path. microbiol. scand. Sect. C, 84, 2

for actual coating of tumor cells in viuo by immunoglobulins is convincing (reviewed in 18). The mechanism of binding of immunoglobulin may be a specific antigen-antibody reaction, i.e. the Fab portion of antibody reacts with antigens on the tumor cell surface. Several investigators have speculated that this reaction is responsible for the enhancement of tumor growth (6, 12, 18). However, our data (15) suggest that immunoglobulin may also be bound in viuo by the Fc portion in a non-specific manner, 105

mediated by the Fc receptor present in the tissue. Fc receptors coated in vivo would not be available for immediate in nitro detection. Experiments were designed to test this hypothesis. Data presented compare the amounts of immunoglobulin bound in uiuo with the in vitro Fc reactivity of selected human malignant tumors, using tissue sections and cells in suspension. Since the tumor Fc receptor has affinity mainly for IgG (16), these investigations deal with tumor associated IgG only. MATERIALS AND METHODS

Tissues Specimens from malignant tumors were obtained from tissue removed during surgery. Specimens from normal liver and spleen were kindly provided by the Department of Pathology. Tissue not used immediately was frozen in phosphate buffered saline, p H 7.2 (PBS) and wrapped in plastic bags to be stored at -70" C. Cryostat sections of tissue were prepared and handled as described earlier

serum, and this line gave identity with the line produced by a commercial antiserum to IgG. Antiserum to human IgG was produced by injecting rabbits subcutaneously with 200 pg in complete Freund's adjuvant at multiple sites every other week for 10 weeks. The rabbits were bled 10 days after the last injection. The antiserum gave only a n IgG line against human serum in agar diffusion and irnmunoelectrophoresis. Antiserum specific to human IgG was also purchased from Hyland, Costa Mesa, California, USA. Before use the antisera were ultracentrifuged in order to remove complexes caused by storage.

Preparation of Labelled IgG T h e proteins were labelled with 1 2 5 1 according to the method described by Hunter & Greenwood ( 7 ) . The average labelling was 0.5 to 1 atom of Iodine per molecule of protein. T h e labelled preparations were dialysed for 72 hours against PBS and then passed through 1 cm x 10 cm G-50 Sephadex columns.

Radioimmunoassay for Quantitation o f IgG

T h e method has been described in detail elsewhere (8). The optimum ratio of antiserum to labelled and unlabelled human IgG was determined by testing a series of increasing dilutions of anti(15, 1 6 ) . serum against 400 ng of labelled IgG. T h e dilution Histological investigations and classification of of antiserum required to precipitate the maximum tumors were routinely performed in the Departamount of labelled IgG at 35 per cent saturation ment of Pathology. Consecutive sections of frozen of ammonium sulfate was determined. The proper tissue, in addition to those used in the tests, were saturation of ammonium sulfate needed to separate stained with hematoxylin and eosin in order to bound and unbound IgG was determined as deprovide histopathological confirmation of the tissue scribed by Farr ( 2 ) . Diluent was normal rabbit under study. serum 1 in 50 in 0.05 borate buffer p H 8.4. TO obtain the standard curve, 0 to 800 ng of unCell Suspensions labelled human IgG were added to rabbit antiFresh tissue was thoroughly minced with iris human IgG of appropriate dilution as determined scissors and suspended in PBS. The suspension above. The mixtures were incubated at 4" C for was centrifuged at 100 x g for 1 min to sediment 24 h ; 400 ng of labelled IgG was added and inthe coarser particles, and the supernatant was cubation was continued for another 24 h at 4" C. saved. T h e cells were sedimented by centrifugaAmmonium sulfate was added to make up a final tion a t 650 x g for 5 min, washed 3 times in volume of 5 ml with saturation 35 per cent. After PBS and resuspended to approximately 2 X loG centrifugation at 1,100 x g for 30 min, 1 ml of the cells/ml. For some experiments, aliquots of cells supernatant, and the rest including the precipitate, were saved before each washing, and washing was were counted separately. T h e percent precipitated continued up to five times. 1251 IgG - was calculated from the formula: R-4S lZ5I IgG precipitated = where R = Serum and Serum Fractions ( R t S) cpm in the rest, and S = cpm in 1 ml supernatant. Sera from 25 healthy blood donors were pooled. Percent of 1251 IgG precipitated was determined Immunoglobulin G was isolated by exclusion chrofor different amounts of unlabelled IgG and a matography according to the method described by standard curve was obtained (Fig. 1). Flodin & Kitlander (4). I t was further purified T o determine the quantity of IgG on the cells, by treating twice with DEAE Sephadex a t p H 6.5 106 washed cells were incubated in triplicate for (1). The purified preparation produced a single 24 h with antiserum of the same dilution as used IgG line in agar diffusion and immunoelectrophoto obtain the standard curve. After incubation, the resis when tested against antisera to whole human

106

For morphological studies, smears were stained with methylgreen and pyronine.

E X P E R I M E N T S AND R E S U L T S

Fc R e c e p t o r s in Tissue and on Cells For complete experiments only those tumors could be used which gave good single cell suspensions after mechanical disintegration. Ten tumors were selected according to this criterion. Trypsin treatment was avoided " 200 400 600 800 because it would destroy bound IgG. The Ng UNLABELED IgG results obtained with these tumors in tests with sections and EA are shown in Table 1. F i g . 1 . Standard curve for IgG quantitation. Percent labelled IgG precipitated after preincubation The patterns of reactions are also given, as with various amounts of unlabelled IgG. diffuse or focal types, i.e. most of the entire section is covered in contrast to larger or smaller areas. cells were removed by centrifugation, 400 ng 1251 The ten tumors showed very different acIgG was added to the supernatant and the mixture tivity, without any relation to type of tuwas incubated for another 24 h. T h e percent precimors. Cells in suspension from the tumors pitation was determined as described above. The amount of bound IgG was expressed in terms of which adsorbed EA in diffuse patterns ng/108 cells by referring to the standard curve. formed rosettes with EA in percentages which corresponded well with the results obtained Detection of Fc Receptors with sections, while the percent rosettes with Sheep erythrocytes ( E ) sensitized with various cells from the tumors showing focal reactions amounts of corresponding rabbit IgG antibodies was very low ( < 5 per cent). The results (A) expressed as agglutinating units were prepared obtained with sections are therefore used. as described earlier (15, 16). One agglutinating unit is defined as the amount of the highest dilution of antiserum which agglutinates an equal amount of a 1 per cent suspension of E. The techniques for detection of Fc receptors in tissue sections and on single cells in suspensions have also been described (15, 16). I n short, cryostat sections of frozen tissue on large cover glasses were incubated at 20" C (room temperature) in wet chambers with the EA applied using inverted microculture slides with a single concavity allowing the EA (suspended in PBS containing 2 per cent NaCl) to settle on the cover glass. After 30 min at 20" C, the slides were turned coverglass up, and left for detachment of EA from the glass and nonreactive tissue. T h e degree of hemadsorption ( 3 , 2 +, 1 + ) was recorded microscopically when the glass around the tissue was free of erythrocytes. One drop of cells in suspension (approximately 2 x 106 cells/ml) was mixed in glass tubes with 1 drop of the proper EA suspension containing approx. 2 x 108 erythrocytes/ml ( 1 per cent suspension). The mixtures were left at 20" C for 2 h. The sedimented cells were then gently resuspended and the percentage of rosette forming cells determined using phase contrast microscopy at 430 x .

+

IgC on the Celts Cells in suspension from each of the tumors were tested for presence of IgG bound to the cell surface. The mean values of duplicate determinations are given in Table 1, and in Fig. 2 they are plotted against Fc reactivity in sections of corresponding tumors. An inverse relationship was observed, i.e. with high amounts of IgG on the surface the Fc reactivity was low, while with low amounts of IgG bound to the cells, the Fc activity was pronounced. The tumors which fall outside this relationship were tumors showing focal reactions with few and/or small foci mixed with histologically non-malignant tissue. These results indicate that a significant relationship exists between bound immunoglobulins and availability of Fc receptor sites in/on the malignant cells. 107

T 15 T40 T53 T41 T37 TI0 T8

T4 T7 T18

4

2+ 3+ 3+ 3+ 2+ 3+ 3+

3+ 3+ 2+

~

-

2+ 2+ 2+ 3+ 3+

3+

2+

3+

3+

3+

2

-~

8 Numbers in italics considered endpoints.

Controls with unsensitized E.

Thyroid Colon Bladder Leiomyosarcoma

Carcinoma of Breast

Melanoma

T y p e of tissue

.~

_

2+ 2+

2+§ 3+ 3+

1+

2+ 3+

I+ 3+ I+ I+ 2+ 3+

1/2

2+ 3+ I+ I+

_

_

~

_

_

_

Test with tissue sections

2+ 3+

t 2+ -

t

1+

1/16

-

1/8

+-

-

114

~.

EA and no. of agglutinating units of A

~

1

~

_ -

1/32

~- ~-

-

of

E*

-

TABLE 1. Fc Receptor Activity in Tumor Tissue Sections Measured by Adsorption of EA, and Presence

Focal Diffuse Diffuse + Focal Focal Focal Diffuse Diffuse

Focal Diffuse Diffuse

Patterns of ads. EA

400 285 300 350