Cancer Immunol Immunother (1992) 34:294 - 298

ancer mmunology mmunoth Papy © Springer-Verlag 1992

The human repertoire of antibody specificities against Thomsen-Friedenreich and Tn-carcinoma-associated antigens as defined by human monoclonal antibodies 1BioInventAB, S-22370 Lund,and 2Departmentof Immunotechnology,LundUniversity,E O. Box 7031, S-22007 Lund, Sweden

Summary. Human monoclonal antibodies specific for tumour-associated Thomsen-Friedenreich (TF) [Gal(~I3)GalNAc(~)-O-] and Tn [GalNAc(cz)-O-] glycoproteins were prepared using peripheral blood lymphocytes from healthy blood donors. The B lymphocytes were either directly transformed with Epstein-Barr virus (EBV) or transformed after an in vitro stimulation period with synthetic glycoproteins. The EBV-transformed lymphocytes were subsequently fused with a mouse-human heteromyeloma to secure antibody production and stability. IgM antibodies exhibiting different patterns of specificity for synthetic TF and Tn antigens were obtained, including antibodies specific for the ~ and ~ forms of different Gal(~31- 3)GalNAc-O- and GalNAc-O- conjugates and antibodies agglutinating neuraminidase-treated erythrocytes. Several of the human monoclonal antibodies showed an increased binding to cultured carcinoma cells as compared to melanoma cells. This straightforward approach for the production of human monoclonal antibodies demonstrates the possibility of investigating the reactivity pattern of tumour-binding antibodies from peripheral blood lymphocytes. The binding patterns of these monoclonal antibodies show that healthy donors carry different fine specificities against synthetic TF/Tn antigens and that these antibodies react with different tumour cells. Key words: In vitro immunization - Human monoclonal antibody - Thomsen-Friedenrich antigen - Tn antigen T antigen

The Thomsen-Friedenreich (TF) antigen and Tn antigen have been described as tumour-associated antigens present on various forms of carcinomas [6-8, 15, 18]. TF and Tn

antigens are the precursors of the MN blood group antigens and are normally not present on the surface of eukaryotic cells, but appear both during normal maturation and on tumour cells [ 12-14]. Neuraminidase-treated, normal cells do also express the TF antigen in much the same way as the tumour cells, although the latter have lost the capacity to add terminal sialic acid residues. The Tn antigen was first described by Dausset et al. [5] on erythrocytes of patients suffering from the rare Tn syndrome. The presence of the Tn antigen is explained by an inherent genetic block of a single biosynthetic step, i.e. the transfer of D-galactose to N-acetylgalactosamine (GalNAc-O-). The inability of tumour cells to complete the normal carbohydrate synthesis results in the same way in an enrichment of the Tn antigen. Normal human serum agglutinates neuraminidasetreated human erythrocytes and this anti-TF activity is explained by the presence of antibodies with a specificity similar to the Arachis hypogaea lectin [14]. Springer et al. reported a decrease in this titre of natural anti-TF antibodies in sera from carcinoma patients [12, 13] and this decrease was suggested to predict recurrence of tumour. More rarely a rise in the anti-TF antibody titre during tumour growth was observed, which would indicate an immune response against the tumour [12]. Furthermore, it has also been claimed that normal human sera contain some antibodies against the Tn antigen. The specificities of natural anti-TF antibodies have been analysed using pooled sera from healthy donors. Wolf et al. found that not all polyclonal anti-Thomsen-Friedenreich antibodies prepared from a pool of sera from normal humans did bind carcinoma cell lines [17]. However, since these pools are polyclonal by nature the observed specificity is the result of many antigen-antibody interactions and can thus not be used to evaluate the humoral immune response and the diagnostic usefulness of these tumour-associated antigens. This paper describes an approach that allows the analysis of the humoral immune repertoire against TF and Tn antigens, by preparing a panel of human monoclonal antibodies specific for a number of different carbohydrates related to those present on turnout cells.

295 Table 1. Peripheral blood lymphocytes from five healthy donors were transformed with Epstein-Barr virus after treatment with Leu-LeuOMe and then screened after 2 - 3 weeks for the presence of antibodies specific for TFc~ or Tnc~ synthetic glycoproteins Donor no.

Growth positive wells/total no.

TF(z a

Tn~ a

Established clones

1 2 3 4 5

1734/1825 1220/1526 3321/3571 1080/1440 1625/1766

20 39 55 10 16

10 5 23 ND b ND

4B1 8C6 CMG8/CMG9 14B8 7Cll/9D10

a Number of wells with antibodies specific for the TF or Tn antigen that did not bind to a control antigen (human serum albumin) b ND, not determined

Materials and methods In vitro stimulation. The in vitro stimulation was performed as previously described [1], using peripheral blood lymphocytes (PBL) treated with c-leucyl-L-leucine methyl ester (Leu-LeuOMe) (Bachem Feinchemikalien AG, Budendorf, Switzerland), in RPMI-1640 medium containing 10% human serum, 4 mM L-glutamine, 1% (v/v) 100 x uou-essential amino acids and 50 gg gentamicin/ml (Flow Laboratories Inc., Ayrshire, UK). The Leu-LeuOMe treatment was performed as described [11]. Briefly, the isolated PBL were treated with 0.25 mM Leu-LeuOMe in RPMI- 1640 medium, supplemented with 2% pooled normal human ABO sera for 15 min. The cells were washed in the same medium and were either directly transformed with Epstein-Barr virus (EBV) or first stimulated in vitro [10] before the EBV infection. During in vitro stimulation, the medium was further supplemented with 50 gM 2-mercaptoethanol, 5 IU recombinant IL-2, antigen (100 ng/ml) and 25% (v/v) superuatant, isolated from a culture of irradiated (20 Gy) human T cells stimulated ovemight with 10 g pokeweed mitogen/ml [4]. The immunization period was 6 days at 37 ° C in a humidified incubator with an 8% CO2/92% air gas phase, and fresh medium was added to the cultures on day 3.

Infection of lymphocytes with Epstein-Barr virus [10]. The EBV-containing superuatant isolated from the virus-producing marmoset cell line (B95-8) was filtered through a 0.454tm sterile filter, stored at 4 ° C and was used without further treatment. Prior to infection with EBV, lymphocytes stimulated in vitro were washed twice with serum-free RPMI-1640 medium. Cells were infected for 2 h at 37°C with occasional stirring using 1 ml EBV-containing supematant/107 lymphocytes. After infection the cells were washed twice with RPMI-1640 medium containing 10% fetal calf serum. Finally, infected cells were seeded at 5 x 104 cells/well in 96-well microtitre plates with feeder cells (5 x 104 irradiated, 30 Gy, PBL/well) in supplemented RPMI-1640 medium, containing 10% fetal calf serum (Gibco Ltd., Paisley UK).

Establishing EBV-transformed lymphoblastoid cell lines. EBV-transformed cells were cloned after 14 days by limiting dilution using irradiated PBL as feeder cells (5 x 104 irradiated, 30 gy, PBL/weI1) or expanded to (3 - 10) x 106 cells before being fused with K6H6/B5 [2].

ELISA. Antigen-specific enzyme-linked immunosorbent assay (ELISA) for screening was performed by coating antigen (5 gg/tal) in a 0.1 M sodium carbonate buffer (pH 9.6) in 96-well ELISA microtitre plates (M129B, Dynatec Labs Ltd., Billingshurst, UK). The synthetic glycoproteins used were TF~3 = Gal (~I-3)GalNAc-(13)-O-CETE-BSA, Tn~ = GalNAc-(~)-O-CETE-BSA and GG = Gal(~l -3)GlcNAc-([3)-OCETE-BSA (B-1011, B-1018 and B-1012; Carbohydrates International, Arlöv, Sweden); TFc~= Gal([31 - 3)GalNAc-((z)-O-PAP-HSA and Tn(z = GalNAc-(o~)-O-APE-HSA (60/76 and 60/85, BioCarb AB, Lund, Sweden). CETE is the abbreviation for the linker molecule [2-(2-carbomethoxyethylthio)ethyl], while PAP and APE stand for paminophenyl and aminophenylethyl; BSA and HSA are bovine and human serum albumin respectively. The plates were washed with ELISA buffer (0.01 M sodium phosphate buffer pH 7.3, containing 0.15 M sodium chloride and 0.05% Tween 20). Supernatants from hybridoma and EBV clones were diluted 1/10 prior to testing and then incubated for 1 h at room temperature. The plates were washed and incubated with horseradish-peroxidase-conjugated to a goat anti-(humanIg) antibody specific for a, "f and g heavy chains (Zymed Laboratories Inc., San Francisco, Calif.) for 1 h, before 200 gl developing solution (3.7 mM o-phenylenediamine in 0.05 M citric acid) was added to each well. The reaction was stopped after 10 min by adding 50 gl 1 M H2SO4 to each well. The plates were read at 492 um. Biotinylated lectins, with known carbohydrate specificity, were also used for testing the synthetic glycoproteins: 5 gg/tal biotinylated A. hypogaea lectin (L-6135, Sigma Chemical Co., St Louis, Mo.) and Vicia villosa isolectin B4 (L-7638, Sigma) weree incubated for 1 h with the synthetic glycoproteins, as described above. HH8 [3] and 22.19 [14] mouse monoclonal antibodies were a kind gift from Dr. J. Zeuthen (The Fibiger Institute, Copenhagen); they were specific for asialoglycophorin and synthetic TF antigen and used as described in the references.

Cell ELISA. Monolayers of Mel-28 (melanoma), HeLA F (cervical carcinoma), T293 H (lung carcinoma), Kato III (stomach carcinoma) cells were fixed with 0.15% glutaraldehyde in phosphate-buffered saline (PBS) for 10 min before washing with PBS and blocking with PBS containing 10% (v/v) fetal calf serum and 0.1 M glycine, pH 7.3. The dilution of supernatants from EBV clones or hybridomas and other details in the ELISA were the same as described above, except for the enzyme conjugate, where instead horseradish proxidase conjugated to goat anti-(human IgM) (Zymed Laboratories Inc. San Francisco, Calif.) was used. The enzyme substrate was the same as described above. Inhibition ofbinding. The human monoclonal antibodies 5C7 and 6D4 were titrated against TF[3 and Tn~ antigens. An appropriate dilution of the antibody was incubated with 2 mg/ml different carbohydrates or 0.250 mg/tal glycoproteins. The antibody binding was then detected as described above for the ELISA.

Haemagglutination assay. Agglutination was conducted in 96-well microtitre plates with V-shaped bottoms. Samples of 50 gl hybridoma supernatant were mixed with 50 gl 1% v/v suspension of neuraminidasetreated human erythrocytes (O, Rh) in PBS. The erythrocytes (0.5 ml 10% v/v suspension) were treated with 3 mU neuraminidase from Vibrio cholerae. Plates were recorded after 1 h incubation at room temperature. Untreated erythrocytes from the same donor were used as control.

Fusion and cloning ofcells. EBV clones and the human x mouse heterohybridoma K6H6/B5 [2], at a ratio of 1:1, were fused in medium containing 45% (v/v) polyethyleneglycol (Mr = 1540, PEG-1540) and 7.5% dimethylsulphoxide, as described [1]. Cells were seeded at ( 1 3) x 10» cells/well in 96-well plates in supplemented RPMI-1640 medium, containiug 10% fetal calf serum, 100 ~tM hypoxanthine, 0.4 gM aminopterin, 16 gM thymidine, and 1 gM ouabain. Clones were tested with ELISA for antigen-specific antibodies after 2 - 4 weeks. Antigen-positive wells were expanded to 24-well plates and retested after a few days for sustained antibody production. Cloning by limiting dilution was performed at least twice, using 5 x 04 irradiated (30 Gy) PBL/well as feeder cells [10].

Results Peripheral blood lymphocytes from five healthy donors were directly transformed, without any in vitro stimulation, a n d t h e r e s u l t s a r e s h o w n i n T a b l e 1. S u p e m a t a n t f r o m E B V c l o n e s f r o m t h r e e o f t h e s e e x p e r i m e n t s w e r e all s c r e e n d e d a g a i n s t T F a , Tnc~ a n d h u m a n s e r u m a l b u m i n using the enzyme immunoassay. In the other two experiments the transformed cells were initially screened against

296

TF

22.19

I

HH8

=,

Vica V

Tn I

Arachis h. AntiDil 4BI

I

6D4

I

]

5C7

L I

Fig. 1. Enzyme-linked immunosorbent assay

CMG!

I I

CMGI

5

8C6

,.J

14B8 [_

9Dl0

7Cll

3.0

2~0

1.'0

0

0

1.C)

2.b

3.0

Absorbance

reactivity profiles of human monoclonal IgM antibodies. The cell culture supernatant was diluted ten times. Vicia villosa isolectin B4 (Vicia v.), Arachis hypogaea (Arachis h.), and mouse monoclonal anti-TF antibodies (HH8 and 22.19) were used as positive controls. A human anfi-digoxin IgM antibody (Anti-Dig) was used as negative control. Closed and open bars represents the o~and [~forms of the antigens. The concentrations of IgM in the supernatants were in the range of 5 - 37 gg/ml

Table 2. Human monoclonal antibodies tested for their ability to aggluti-

nate neuraminidase-treated erythrocytes mAb

Anti-digoxin 4B 1 6D4 5C7 CMG9 CMG8 8C6 14B8 9Dl0 7C 11

Immortalization methoda EBV EBV EBV EBV EBV EBV EBV EBV EBV EBV

+ fusion + fusion + fusion + fusion + fusion

VV Agglutination; neuraminidasePNA

treated

untreated

+

-

+ + -

-

22.19

7Cll 9Dl0

14B8 ll HeLa F [~T293H

CMG8

Lectins and control murine monoclonal antibody PNA Lectin + VV Lectin -

-

CMG9

22.19 HH8

-

5C7

mAb mAb

+ +

a EBV, Epstein-Barr virus

r.drH~

I

[ ] Kato

DMrL-28 a.__.__

6D4

4B1

o n e o f T F ~ , Th(x, T F ~ a n d Tn[3, u s i n g H S A as a n e g a t i v e control. After the EBV clones had been expanded they were r e t e s t e d a n d at t h i s s e c o n d E L I S A s c r e e n i n g w e a l s o included the control proteins: keyhole limpet haemocyanin, human thyroglobulin, dsDNA, conalbumin, concanaval i n A, d i n i t r o p h e n o l c o n j u g a t e d w i t h H S A ( D N P - H S A ) , phosphorylaseb, ovalbumin, carbonic anhydrase, soyabean trypsin inhibitor, o~-lactalbumin, lactate dehydrog e n a s e , c a t a l a s e a n d f e r r i t i n . A n t i b o d i e s t h a t b o u n d to a n y of these irrelevant proteins were discarded. Antibodies f r o m a n u m b e r o f s t a b l e l y m p h o b l a s t o i d c e l l l i n e s (Fig. 1) were tested. Antibodies 5C7 and 6D4 were producedfrom

-0,2

0

0,5

1,0

1,5

2,0

Absorbance Fig. 2. Analysis of the binding of human monoclonal antibody to glu-

taraldehyde-fixed cell lines. HeLa F, cervical carcinoma; T293H, lung carcinoma; Kato, stomach carcinoma; MEL-28, melanoma. A human monoclonal IgM anti-digoxin antibody was used as negative control, and this background has been deducted from the experimental values

297 Galactose Sucrose Methyl Glycoside Mannose Acetyl-Glucose Glucose Fucose Melobiose TF bens GG

•

ing fixed tumour cells. Supernatant from EBV cell lines and hybrodimas was diluted five times and incubated with monolayers of fixed carcinoma or melanoma cells (Fig. 2). The antibodies showed various degrees of binding to the three carcinoma lines (HeLA, T293H, Kato III) and 2/9 also bound to the melanoma cell line (SK-Mel-28). All human monoclonal antibodies showed a reactivity against the cervical carcinoma (HeLa) cell line. The specificity of two antibodies (5C7 and 6D4) were further characterized using a competitive inhibition test (Fig. 3). The results show that 5C7 was inhibited by both TF~ and T~ antigens and also by a free non-protein-conjugated benzoyl-TFot, while the 6D4 antibody was only inhibited by the Tn[3.

6D4

[] 5C7

TF~x Tn I] Tn BSA 0

50

100

Discussion I n h i b i t i o n (%) Fig. 3. Inhibition of the binding of 5C7 and 6D4 monoclonal antibodies. The supernatants from 5C7 and 6D4 were diluted and mixed with different carbohydrates (2 mg/ml) and glycoproteins (250 ~tg/ml) [TF bens, benzyl-2-acetamido-2-deoxy-3- O-[3-O-galactopyranosyl-«-D-galactopy ranoside). 5C7 was tested against TF~ and 6D4 against Tn[~

two different donors using in vitro stimulation with TF[3 and Tn~, respectively. All other clones were the products of direct EBV transformation and all monoclonal antibodies were of the bt isotype. On the basis of the pattern of specificity when binding to synthetic antigens, the monoclonal antibodies presented in Fig. 1 could be divided into six groups. The first group includes the 4B 1 antibody that bound to all four synthetic glycoproteins tested but still did not bind GG [Gal([313)GlcNAc-([3)-O-CETE-BSA], DNP-HSA or unmodified BSA and HSA. The second group is represented by 6D4 antibody, which reacted predominantly with Tn[3, although a slight reactivity with TF[3 could be detected. In the third group, 5C7, CMG9 and CMG8 antibodies specifically reacted with TF[3, although CMG9 also showed a weak reactivity against the TFo~ antigen. 8C6 represents the fourth group, with an antibody that specifically reacts with Tna, a binding profile similar to that of the lectin from V. villosa. The fifth group contains antibodies 14B8 and 9Dl0, both binding to the o~forms, i.e. TFo~ and Tna. 14B8 and 9Dl0 were both also able to agglutinate neuraminidase-tr¢,kted erythrocytes. The last group is represented by the 7C11 antibody with specificity for Tn antigens, independent of o~ or [3 forms, i. e. exhibiting a crossreaction between T n a and Tn[3. None of the antibodies tested had the same reactivity pattern as A. hypogaea lectin or the mouse monoclonal antibody HH8. A. hypogaea lectin and HH8 showed specificity for the To~ antigen while the binding of V. villosa isolectin B4 is restricted to the Tn« antigen. Despite the fact that 22.19 [16] and HH8 [3] were very effective in agglutinating neuraminidase-treated human erythrocytes only HH8 bound the T F a antigen. The antibodies were also tested in an agglutination assay. Only 4B1, 14B8 and 9Dl0 were able to give weak agglutination reactions (Table 2). Finally the reactivity of the different human monoclonal antobodies was tested us-

Longenecker et al. [9] showed that an antibody produced after immunizing mice with a synthetic TF antigen coupled in the [3position to a carrier protein (TF~) induced antibodies with higher reactivity for carcinomas than antibodies prepared after immunization with the ofform. Our results show that after an in vitro stimulation with the synthetic [3 form of the TF antigen (TF[3) the resulting antibody (5C7) crossreacted between the c, and [3 forms. The antibody also bound to one of the carcinoma cell lines. In vitro stimulation with the synthetic form of the 13 Tn antigen (Tn~) resulted in an antibody (6D4) that did not crossreact between the 13 and o~ forms. Some antibody specificites (CMG9, CMG8), produced without in vitro manipulation, were very similar to antibody specificities that were obtained by in vitro stimulation (5C7). Despite a very similar specificity pattern of 5C7 and CM9 against synthetic antigens, the CM9 but not 5C7 antibody bound a melanom cell line. In much the same way, some antibodies exhibiting the same specificity as the 6D4 antibody were observed during screening of normal blood donors without prior in vitro stimulation (data not shown). The 6D4 and 5C7 inhibition assay revealed that the 5C7 antbody was easily inhibited by the benzoyl-TFo~ antigen and by both the TFo~ and TF[3 antigens. 6D4 was only inhibited by the synthetic glycoprotein containing the Tn~ antigen. This shows that 5C7 is crossreactive between TFo~ and TF[3 antigens, in a similar way to that described for mouse monoclonal antibodies. The different specificity patterns of the human monoclonal antibodies in Fig. 1 demonstrate the nature of a normal repertoire of antibody specificities (anti-TF and anti-Tn) isolated from healthy individuals. The antibody specificities found are some but not necessarily all the types of anti-TF and anti-Tn reactivity that can be found in the B cell repertoire of a normal person. We have also shown that these antibodies can bind to tumour cell lines, indicating the possibility that this type of antibody participates in an immune reaction with carbohydrate moities present in circulation and on the tumour cell surface. Further analysis of the monoclonal antibodies will show the exact nature of the recognized epitopes and the titres of similar antibodies during tumour growth. Knowledge of

298 a n t i b o d y r e a c t i v i t y patterns, as e x e m p l i f i e d b y t h e s e h u m a n m o n o c l o n a l a n t i b o d i e s , w i l l p r o v i d e the basis for s e l e c t i n g the p r o p e r a n t i b o d i e s to b e u s e d in a n a l y s i s and therapy of carcinomas.

Acknowledgements. This investigation was supported by a grant from The Swedish Cancer Society.

References 1. Borrebaeck CAK, Danielsson L, Möller SA (1988) Human monoclonal antibodies produced by primary in vitro immunization of peripheral blood lymphocytes. Proc Natl Acad Sci USA 85:3995 2. Caroll WL, Thielemans K, Dilley J, Levy R (1986) Mouse x human heterhybridomas as fusion partners with human B cell tumors. J Immunol Methods 89:61 3. Clausen H, Stroud M, Parker J, Springer G, Hakomori S-I (1988) Monoclonal antibodies directed to the blood group A associated structure, galactosyl-A: specificity and relation to the ThomsenFriedenreich antigen. Mol Immuno125:199 4. Danielsson L, Möller SA, Borrebaeck CAK (1987) Effect of cytokines on specific in vitro immunization of human peripheral B lymphocytes against T-cell dependent antigens. Immunology 61:51 5. Dausset J, Moullec J, Bernard J (1959) Acquired hemolytic anemia with polyagglutinability of red blood cells due to a new factor present in normal serum (anti-Tn). Blood 14:1079 6. Hull SR, Carraway KL (1988) Mechanism of expression of Thomsen-Friedenreicch (T) antigen at the cell surface of a mammary adenocarcinoma FASEB J 2:2380 7. Itzkowitz SH, Yuan M, Montgomery CK, Kjeldsen T, Takahashi HK, Bigbee WL, Kim YS (1989) Expression of Tn, sialosyl-Tn, and T antigens in human colon cancer. Cancer Res 49:197 8. Lange PH, Winfield HN (1987) Biological markers in urological cancer. Cancer 60:464

9. Longenecker BM, Willans DJ, Maclean GD, Selvaraj S, Suresh MR, Noujaim AA (1987) Monoclonal antibodies and synthetic tumor associated glycoconjugates in the study of the expression of Thomsen-Friedenriech like and Tn-like antigens on human cancers. J Natl Cancer Inst 78:489 10. Ohlin M, Broliden PA, Danielsson L, Wahren B, Rosen J, Jondal M, Borrebaeck CAK (1989) Human monoclonal antibodies against a recombinant HIV envelope antigen produced by primary in vitro immunization. Characterization and epitope mapping. Immunology 68:325 11. Ohlin M, Danielsson L, Carlsson R, Borrebaeck CAK (1989) The effect of leucyl-leucine methyl ester on proliferation and Ig secreüon of EBV-transformed human B lymphocytes. Immunology 66:485 12. Springer GF (1984) T and Tn, general carcinoma autoantigens. Science 224:1198 13. Springer GF, Desai PR, Scanlon EF (1976) Blood group MN precursors as human brest carcinoma-associated antigens and "naturally" occurring human cytotoxins against them. Cancer 37:169 14. Springer GF, Desai PR, Murthy MS, Tegtmeyer H, Scanlon EF (1979) Human carcinoma associated precursor antigens of the blood group MN system and the host immune responses to them. Prog Allergy 26:42 15. Stein R, Chen S, Grossman W, Goldenberg DM (1989) Human lung carcinoma monoclonal antibody specific for the Thomsen-Friedenreich antigen, cancer Res 49:32 16. Steuden I, Duk M, Czerwinski M, Radzikowski C, Lisowska E (1985) The monoclonalantibody, antiasialoglycophorin from human erythrocytes, specific for beta-D-Gal-l-3-alfa-D-GalNAc chains (Thomsen-Friedenreich receptors). Glycoconjugate 2:303 17. Wolf MF, Koerner U, Klump B, Schumacher K (1987) Characterization of Thomsen-Friedenreich antibody subpopulations from normal human serum. Tumor Biol 8:264 18. Yamada T, Fukui I, Yokokawa M, Oshima H (1988) Changing expression of ABH blood group and cryptic T antigens of noninvasive and superficially invasive papillary transitional cell carcinoma of the bladder from initial occurance to malignant progression. Cancer 61:721