Journal oflmmunologicalMethods, 133 (1990) 245-251 Elsevier
245
JIM05720
A method for blocking antigen-independent binding of human IgM to frozen tissue sections when screening human hybridoma antibodies Henrik Ditzel 1, Karin Erb 1, Bjarne Nielsen 2, Per Borup-Christensen 3 and Jens Chr. Jensenius 4 l Department of Medical Microbiology, and 2 Department of Pathology, University of Odense, Odense, Denmark, 3 Department of Surgery, Svendborg Hospital, Svendborg Denmark, and 4 Institute of Medical Microbiology, University of Aarhus, Aarhus, Denmark (Received 4 May 1990, revised received 14 June 1990, accepted 29 June 1990)
Using an indirect immunoperoxidase technique antigen independent binding of both human monoclonal and polyclonal IgM was found to a wide range of frozen sections of normal and malignant human glandular epithelia. Identical binding was found using dimeric human IgA (dIgA), whereas no binding was found with monomeric human IgA or human IgG. Secretory component (SC) was found to be the component in these tissues mediating antigen-independent binding of human IgM and dIgA antibodies. A method for the blocking of this antigen-independent binding of human IgM and dIgA was evaluated. Frozen sections of tissues containing SC were blocked with antibody to endogenous immunoglobulin and preincubated with rabbit anti-secretory component antibody (anti-SC) before applying the human monoclonal antibody. This treatment blocked the binding of control polyclonal and monoclonal human IgM to sections of SC-containing tissues such as respiratory and colonic epithelia. The influence of anti-SC on the binding of dIgA could not be established due to interactions between the anti-SC antibody and the IgA preparation. Using this method human hybridoma supernatants containing IgM could be readily screened for reactivity with frozen tissue sections from patients with colo-rectal cancer. This approach is recommended for the screening of human IgM monoclonal antibodies on frozen human tissue sections. Key words." Monoclonal antibody, human; Immunohistochemistry; Secretory component
Introduction
Interest in the use of monoclonal antibodies for diagnosis and treatment of cancer is growing.
Correspondence to: H. Ditzel, Department of Medical Microbiology, Institute of Medical Biology, University of Odense, J.B. Winslows Vej 19, DK-5000 Odense C, Denmark. Abbreviations: SC, secretory component; anti-SC, antisecretory component antibody; dlgA, dimeric IgA; mIgA, monomeric IgA; PBS, phosphate-buffered saline.
However difficulties have been encountered in the use of murine monoclonal antibodies for therapy due to the induction in the patient of human anti-murine antibodies which may inhibit tumor targeting and give rise to allergic reactions. Human monoclonal antibodies may, therefore, be more useful for in vivo applications. Furthermore human antibodies may react with clinically relevant antigens that are immunogenic in humans but not in mice. Most of the human hybridoma antibodies reported to date are of the IgM class (Boyd and James, 1989).
0022-1759/90/$03.50 © 1990 Elsevier Science Publishers B.V. (Biomedical Division)
246 The screening of hybridoma supernatants for cancer-specific antibody activity is usually carried out by immunohistochemical techniques (Naiem et al., 1982), immunocytochemical techniques (Erber et al., 1983) or cell binding assays on cultured cell fines (Surer et al., 1980). Screening by immunohistochemical techniques is particularly attractive due to the immediate information of the antigen distribution in different tissues as well as its cellular location. However, there have been technical problems with such screening such as background staining due to endogenous immunoglobulin and antigen-independent binding of immunoglobulin to different tissues. Technical difficulties due to endogenous immunoglobulin can be overcome by preincubating the tissue section with Fab' rabbit anti-human IgM or anti-IgG as previously reported (Nielsen et al., 1987). Antigen-independent binding could take place when testing antibodies on epithelial tissue since such tissue may contain secretory component (SC). It is known that secretory component, an 80 kDa glycoprotein present in most glandular epithelial cells, functions as the receptor for human IgM and dIgA. The J chains which are common for IgM and dIgA have been shown to determine the formation of immunoglobulin complexes with SC (Eskeland and Brandtzaeg, 1974). The present report describes the correlation between distribution of secretory component (SC) and antigen-independent binding of human Ig in frozen tissue sections and a method by which this antigen-independent binding can be blocked.
Materials and methods
Human-human hybridomas Human monoclonal IgM antibodies (D4213 and F10279) were secreted by human-human hybridomas derived by fusion between the human B lymphoblastoid cell line, WI-L2-729-HF2 (Abrams et al., 1983) and lymphocytes obtained from mesenteric lymph nodes from patients with colorectal cancer (Borup-Christensen et al., 1990). D4213 was selected due to its reactivity with neoplastic colonic tissue and limited or no binding to other normal and malignant tissues. F10279
was selected to serve as a control, since it showed no specific binding to frozen sections.
Immunoglobulin preparations Three different batches of polyclonal human IgM and one batch of polyclonal human IgG purified from normal human serum were obtained from Cappel Laboratories (Cochranville, PA, U.S.A.). Two different IgM preparations were purified from the sera of patients with WaldenstrOm macroglobulinaemia by affinity chromatography on monoclonal anti-/~ antibody (HB57, ATCC, Rockville, MO, U.S.A.) coupled to Sepharose 4B. Polyclonal human IgA was purified from normal human serum as described by Heide and Schwick (1978). In brief, euglobulins were precipitated by dialysis against water and IgG was removed by precipitation with ZnSO 4 before precipitation of IgA with (NH4)2SO 4. Dimeric and monomeric IgA were further purified by gel chromatography on a Sephadex G-200 column. Dimeric IgA was eluted in a distinct peak between V0 and V~ for IgG. Monomeric IgA was eluted just ahead of the IgG position. The identity of the proteins was verified by SDS-PAGE and crossed immunoelectrophoresis.
Anti-secretory component antibodies The anti-secretory component antibody (antiSC) used for staining of SC and for the blocking of IgM binding was a rabbit polyclonal anti-SC obtained from Dakopatts, Copenhagen, Denmark (code A187, lot 032B). The antibody had been solid phase adsorbed using plasma fractions and lactoferrin and the specificity was tested by crossed immunoelectrophoresis. No precipitate was obtained using human plasma, and only one precipitation line was observed with human colostrum. A mouse monoclonal anti-SC antibody was obtained from Sigma, St. Louis, MO, U.S.A. (ascites, code 1-6635, lot 38F 4802). The reactivity of the two anti-SC antibodies was compared by immunohistochemical staining on frozen tissue sections and on Western blots of non-reduced saliva. Similar staining patterns were observed with both of the anti-SC antibodies using a panel of normal and malignant frozen sections of colonic tissue. On Western blots of non-reduced saliva identical
247 TABLE I B I N D I N G O F H U M A N IgM, d l g A A N D R A B B I T A N T I - S C T O F R O Z E N S E C T I O N S O F N O R M A L BY I N D I R E C T I M M U N O P E R O X I D A S E S T A I N I N G
O e s o p h a g u s s q u a m o u s e p i t h e l i u m (n = 2) G a s t r i c e p i t h e l i u m (n = 5) Small intestine e p i t h e l i u m (n = 3) C o l o n e p i t h e l i u m ( n = 11 ) A n a l tract s q u a m o u s e p i t h e l i u m (n = 1) Liver (hepatocytes) (n = 1) Biliary ductal e p i t h e l i u m (n = 1) Breast g l a n d u l a r e p i t h e l i u m (n = 10) Bronchial respiratory e p i t h e l i u m (n = 2) Bronchial g l a n d u l a r e p i t h e l i u m (n = 2) L u n g aveolar e p i t h e l i u m (n = 2) R e n a l tubuli (n = 1) R e n a l glomeruli (n = 1) B l a d d e r transitional e p i t h e l i u m (n = 2) Skin k e r a t o s q u a m o u s e p i t h e h u m (n = 1) T o n g u e s q u a m o u s e p i t h e l i u m (n = 1) Cervical s q u a m o u s e p i t h e l i u m (n = 1) Cervical g l a n d u l a r e p i t h e l i u m (n = 3) E n d o m e t r i a l g l a n d u l a r e p i t h e l i u m (n = 1) T u b a e e p i t h e l i u m (n = 1) O v a r i a n follicular e p i t h e l i u m (n = 1) S m o o t h muscle (n = 2) T h y r o i d follicular e p i t h e l i u m (n = 2) Tonsil ( l y m p h a t i c c o m p o n e n t ) (n = 2) C o n n e c t i v e tissue (n = 10)
TISSUES ANALYSED
IgM binding
dlgA binding
Presence of SC
+ + +
+ + +
+ + +
+ +
+ +
+ +
___
+
+ +
+
a
+ +
+ +
±
~---
±
+ + + +
+ + + +
+ + + +
a ± indicates that Ig b i n d i n g or the presence of SC o n l y are f o u n d in some areas of the i n v e s t i g a t e d tissue.
bands were visualized with the two anti-SC antibodies. Tissues
Material was obtained from patients undergoing surgical resection and from diagnostic biopsies. The tissues and numbers of samples are listed in Tables I and II. Tissue judged to be normal by conventional microscopy was obtained from nondiseased areas of tissue specimens from cancer patients and from patients with benign disease. Tissues were snap-frozen in Isopentane, cooled in dry ice, or formalin fixed (4% formaldehyde, 0.075 M sodium phosphate, pH 7.0) for 24 h and paraffin-embedded. Sections of 5 /am were then cut, air-dried overnight at 25 ° C and fixed in ice-cold acetone for 10 rain. The formalin-fixed, paraffinembedded sections were deparaffinized and rehydrated through graded alcohols. The formalin-
T A B L E II B I N D I N G O F H U M A N IgM, d l g A A N D R A B B I T A N T I - S C TO FROZEN SECTIONS OF MALIGNANT TISSUES ANALYSED BY INDIRECT IMMUNOPEROXIDASE STAINING IgM IgA Presb i n d i n g b i n d i n g ence of SC C o l o n a d e n o c a r c i n o m a (n = 8) _ Breast a d e n o c a r c i n o m a (n = 5) G a s t r i c a d e n o c a r c i n o m a (n = 1) + Cervical c a r c i n o m a (n = 2) + E n d o m e t r i a l a d e n o c a r c i n o m a (n = 3) + O v a r i a n a d e n o c a r c i n o m a (n = 1) + R e n a l a d e n o c a r c i n o m a (n = 1) B l a d d e r a d e n o c a r c i n o m a (n = 1) D e r m a t o f i b r o s a r c o m a (n = 1) M e l a n o m a ( n = 1) -
+ + ± + + -
+ + + _+ + + -
+ indicates that Ig b i n d i n g or the presence of SC only are found in some areas of the i n v e s t i g a t e d tissue.
248
fixed sections were, before incubation with anti-SC antibody, treated for 20 rain at 37°C with 0.05% trypsin (cat. no. T-8128, Sigma, St. Louis, MO, U.S.A.) in distilled water containing 0.05% CaC12 adjusted to pH 7.8 with NaOH.
Immunohistochemical techniques (a) Binding of immunoglobulin and treatment with anti-SC. Frozen sections were preincubated with the Fab' fragment of rabbit anti-human IgM or Fab' rabbit anti-human IgA, diluted in phosphate-buffered saline pH 7.4 (PBS) containing 1% rabbit serum for 4 h at 4°C in a humidified chamber. This was done to block endogenous IgM or IgA determinants and thereby inhibit the binding of the subsequently applied enzyme-labelled anti-IgM or anti-IgA (Nielsen et al., 1987). To block the antigen-independent immunoglobulin binding, the sections were incubated with rabbit anti-SC (10 mg IgG/ml, Dakopatts, Copenhagen, Denmark), diluted 1/20 in PBS containing 1% rabbit serum for 2 h. Sections (with or without prior anti-SC treatment) were then incubated overnight at 4°C with human monoclonal IgM antibodies or other IgM (5 /xg/ml), IgG or IgA preparations (100 /~g/ml and 400/~g/ml) in PBS containing 10% rabbit serum and 1 mM EDTA. Sections were next incubated for 30 rain at room temperature with horseradish-peroxidase-conjugated rabbit anti-human lgM (HRP-anti-IgM) (1/400), HPR-anti-IgG (1/200) or HRP-anti-IgA (1/40) (Dakopatts, Copenhagen, Denmark) in PBS with 10% rabbit serum. Before the first incubation and between each additional step, sections were washed three times with PBS containing 0.05% Tween 20 (Merck, Darmstadt, F.R.G. (PBSTween)). Staining was developed by incubation in PBS containing diaminobenzidine (0.6/~g/ml) and hydrogen peroxide (0.01%). Formalin-fixed, paraffinembedded tissue sections were treated similarly. (b) Binding of anti-SC. Frozen and formalinfixed sections were incubated with rabbit anti-SC (Dakopatts, Copenhagen, Denmark) diluted 1/1800 (frozen tissue) and 1/600 (formalin-fixed tissue) in 20% pig serum. The incubation was performed for 30 min at room temperature, followed by incubation with horseradish-peroxidaseconjugated pig anti-rabbit lg (Dakopatts) (1/50)
in PBS. Washing and visualisation with chromogenic substrate were as described above. Sections were couterstained with haematoxylin and mounted in Aquamount (Gurr, BDH, Poole, England). Adjacent sections were stained with haematoxylin and eosin for the examination of tissue morphology.
Results
Normal and malignant frozen, acetone-fixed tissue was examined by an indirect immunoperoxidase technique for the presence of secretory component (SC) and binding of human IgM and dIgA. The results are listed in Tables I and II. The binding pattern of three different polyclonal IgM preparations and two different IgM preparations from patients with Waldenstr~Sm macroglobulinaemia were investigated on four sections of neoplastic colonic tissue and adjacent normal epithelia. All preparations used at a concentration of 5 ~g/ml showed the same binding pattern with staining of all normal colonic tissue and focal staining of the neoplastic colonic tissue. The same sections were also investigated using mlgA, dlgA and IgG (all at 100 /~g/ml and 400 /ag/ml). Staining was only found with dIgA at a concentration of 400 tag/ml, dIgA and IgM showed a similar binding pattern, whereas no binding was found to normal or malignant colonic tissue using mIgA or IgG at similar or higher concentrations. Binding of human IgM and dIgA was not observed when the tissue was formalinfixed and paraffin-embedded. As shown in Tables I and II similar binding was found with IgM, dIgA and rabbit anti-SC to different tissues. Controls with normal rabbit IgG were negative. The influence of preincubation with rabbit anti-SC (Dakopatts, Denmark) on the antigen-independent binding of IgM and dIgA was evaluated on frozen, acetone-fixed tissue expressing binding of IgM and dIgA, such as colonic epithelium and respiratory epithelium. Prior to these investigations endogenous IgM or IgA in the tissue was blocked by incubation with Fab' anti-IgM or Fab' anti-IgA. Antigen-independent binding of IgM was
249
Fig. 1. Immunoperoxidase analysis on frozen sections of bronchial tissue for the binding of polyclonal human IgM (Cappel). A: section blocked with Fab' anti-~ only before applying the IgM preparation. Antigen-independent IgM binding is seen in the respiratory epithelium and glandular cells. B: section blocked with both Fab' anti-~ and anti-SC before applying the IgM preparation. Antigen-independent IgM binding is abolished. effectively blocked when the section was preincubated with polyclonal anti-SC at a concentration of 0.5 m g / m l for 2 h at 25 ° C (Fig. 1). The staining with dIgA did not disappear when sections were preincubated with anti-SC, but in most cases was actually enhanced. The combined blocking with Fab' anti-IgM and anti-SC permitted the screening of human hybridoma supernatants containing IgM (F10279 and D4213) on SC-containing frozen tissue sections such as respiratory epithelium and colonic epithelium (Fig. 2).
Discussion The production of human monoclonal antibodies has become a widely used technique and the
development of reliable screening methods such as immunohistochemistry is therefore increasingly important. It has been observed that m a n y antibodies react with determinants which are better preserved in sections of fresh-frozen tissue than in formalin-fixed tissue (Mason et al., 1982). However, most immunohistochemical investigations are carried out on formalin-fixed paraffin-embedded tissues due to problems of routine freezing, storage and preparation of tissue sections. In this study we have shown that antigen-independent binding of h u m a n IgM is a problem when screening human monoclonal antibody on sections of frozen tissue. Antigen-independent binding was also found with dIgA, however, only when dIgA was applied at concentrations higher than normally used in immunohistochemical screening. It could be that only a small fraction of the dIgA
250
Fig. 2. Immunoperoxidase analysis on frozen sections of colonic epithelium for the binding of the human monoclonal IgM antibodies, F10279 and D4213. A: section blocked with Fab' anti-/~ only before applying the human antibody F10279 or D4213. Due to antigen-independent IgM binding the antibody specificity of the monoclonal antibody cannot be determined. B: section blocked with both Fab' anti-t~ and anti-SC before applying the human antibody F10279. There is no specific binding to the colonic epithelium. C: section blocked with both Fab' anti-# and anti-SC before applying the human antibody D4213. In this case there is specific binding to the colonic epithelium.
purified from serum was binding in the latter case. I g G and mIgA on the other hand did not bind nonspecifically to frozen tissue. Table I shows the tissue distribution of secretory component. As noted in the introduction section, secretory component is present in most glandular epithelial cells and functions as the receptor for human IgM and dlgA. Anti-SC, human IgM and human dlgA showed a similar staining pattern. The results show the tissues in which antigen-independent binding of IgM and IgA are to be found and where problems in the interpretation of antibody specificity are to be expected. We examined the possibility of blocking the antigen-independent IgM and dIgA binding with rabbit anti-SC antibody in order to improve the screening of human hybridoma antibodies on frozen tissue. Non-specific staining with IgM was completely prevented when the frozen tissue sec-
tion was blocked for endogenous immunoglobulin by Fab' anti-IgM followed by anti-SC, before incubation with non-specific IgM. The staining was not prevented, but was enhanced, when the frozen tissue sections were incubated with anti-SC before incubation with dlgA. This staining was due to interactions between the dlgA preparation and the rabbit anti-SC antibody. Since dlgA binding only takes place when it is present at very high concentrations (400 /zg/ml) the blocking of dlgA binding in the screening of human monoclonal antibodies may be of minor importance. However, we have not been able to test this with human monoclonal IgA antibody. It is of importance when working with techniques involving anti-Ig antibodies to ensure that problems do not arise due to cross species reactivities. This may often occur when using a primary antibody raised in one species and a secondary
251
antibody raised in another species, However, this problem does not occur when, as in this work, only rabbit antibodies are used. Furthermore, the protein source used for nonspecific blocking must be from the same species as the antibodies in order to avoid problems with cross reactions. The method of blocking antigen-independent IgM binding with anti-SC described here is simple and should be of general use for the screening of human monoclonal antibodies of IgM class on human tissue.
Acknowledgements This investigation was supported by grants from the Danish Cancer Society and the Danish Medical Research Council. Mrs. A. Klim, Mrs. J. Bohrk, Mrs. K. Hartmann and Miss K. Kristensen provided skilled technical assistance.
References Abrams, P.G., Knost, J.A., Clarke, G., Wilburn, S., Oldham, R.K. and Foon, K.A. (1983) Determination of the optimal human cell lines for development of human hybridomas. J. Immunol. 131, 1201. Borup-Christensen, P., Erb, K., Ditzel, H., Nielsen, B., Larsen, J., Svehag, S.-E. and Jensenius, J..C. (1990) Human-human hybridoma producing monoclonal antibodies against colorectal cancer-associated antigens. APMIS, in press.
Boyd, J.E. and James, K. (1989) Human monoclonal antibodies: Their potential, problems, and prospects. In: A. Mizrahi (Eds.), Monoclonal Antibodies: Production and Application. Alan R. Liss, New York, p. 1. Erber, W.N., Falini, B., Chosh, A.K., Moir, D.J. and Mason, D.Y. (1983) lmmuno-alkaline phosphatase labelling of haematological samples with monoclonal antibodies. In: S. Avrameas et al. (Eds.), Immunoenzymatic Techniques. Elsevier, Amsterdam, p. 29. Eskeland, T. and Brandtzaeg, P. (1974) Does J chain mediate the combination of 19 S IgM and dimeric IgA with the secretory component rather than being necessary for their polymerization? Immunochemistry 11, 161. Heide, K. and Schwick, H.G. (1978) Salt fractionation of immunoglobulins. In: D.M. Weir (Ed.), Handbook of Experimental Immunology, vol. 1. Blackwell Scientific Publishers, Ch. 7. Mason, D.Y., Naim, M., Abdulaziz, Z., Gatter~ K., Nash, J.R.G. and Stein, H. (1982) In: J.W. Fabre and A.J. Mc Michael (Eds.), Monoclonal Antibodies in Clinical Medicine. Academic Press, London, p. 585. Naiem, M., Gerdes, A.Z., Sunderland, C.A., Allington, M.J., Stein, H. and Mason, D.Y. (1982) The value of immunohistological screening in the production of monoclonal antibodies. J. Immnunol. Methods 50, 145. Nielsen, B., Borup-Christensen, P., Erb, K., Jensenius, J.C. and Husby, S. (1987) A method for the blocking of endogenous immunoglobulin on frozen tissue sections in the screening of human hybridoma antibody culture supernatants. Hybridoma 6, 103. Surer, L., Bruggen, J. and Sorg, C. (1980). Use of an enzymelinked immunosorbent assay (ELISA) for screening of hybridoma antibodies against cell surface antigens. J. Immunol. Methods 39, 407.
245
JIM05720
A method for blocking antigen-independent binding of human IgM to frozen tissue sections when screening human hybridoma antibodies Henrik Ditzel 1, Karin Erb 1, Bjarne Nielsen 2, Per Borup-Christensen 3 and Jens Chr. Jensenius 4 l Department of Medical Microbiology, and 2 Department of Pathology, University of Odense, Odense, Denmark, 3 Department of Surgery, Svendborg Hospital, Svendborg Denmark, and 4 Institute of Medical Microbiology, University of Aarhus, Aarhus, Denmark (Received 4 May 1990, revised received 14 June 1990, accepted 29 June 1990)
Using an indirect immunoperoxidase technique antigen independent binding of both human monoclonal and polyclonal IgM was found to a wide range of frozen sections of normal and malignant human glandular epithelia. Identical binding was found using dimeric human IgA (dIgA), whereas no binding was found with monomeric human IgA or human IgG. Secretory component (SC) was found to be the component in these tissues mediating antigen-independent binding of human IgM and dIgA antibodies. A method for the blocking of this antigen-independent binding of human IgM and dIgA was evaluated. Frozen sections of tissues containing SC were blocked with antibody to endogenous immunoglobulin and preincubated with rabbit anti-secretory component antibody (anti-SC) before applying the human monoclonal antibody. This treatment blocked the binding of control polyclonal and monoclonal human IgM to sections of SC-containing tissues such as respiratory and colonic epithelia. The influence of anti-SC on the binding of dIgA could not be established due to interactions between the anti-SC antibody and the IgA preparation. Using this method human hybridoma supernatants containing IgM could be readily screened for reactivity with frozen tissue sections from patients with colo-rectal cancer. This approach is recommended for the screening of human IgM monoclonal antibodies on frozen human tissue sections. Key words." Monoclonal antibody, human; Immunohistochemistry; Secretory component
Introduction
Interest in the use of monoclonal antibodies for diagnosis and treatment of cancer is growing.
Correspondence to: H. Ditzel, Department of Medical Microbiology, Institute of Medical Biology, University of Odense, J.B. Winslows Vej 19, DK-5000 Odense C, Denmark. Abbreviations: SC, secretory component; anti-SC, antisecretory component antibody; dlgA, dimeric IgA; mIgA, monomeric IgA; PBS, phosphate-buffered saline.
However difficulties have been encountered in the use of murine monoclonal antibodies for therapy due to the induction in the patient of human anti-murine antibodies which may inhibit tumor targeting and give rise to allergic reactions. Human monoclonal antibodies may, therefore, be more useful for in vivo applications. Furthermore human antibodies may react with clinically relevant antigens that are immunogenic in humans but not in mice. Most of the human hybridoma antibodies reported to date are of the IgM class (Boyd and James, 1989).
0022-1759/90/$03.50 © 1990 Elsevier Science Publishers B.V. (Biomedical Division)
246 The screening of hybridoma supernatants for cancer-specific antibody activity is usually carried out by immunohistochemical techniques (Naiem et al., 1982), immunocytochemical techniques (Erber et al., 1983) or cell binding assays on cultured cell fines (Surer et al., 1980). Screening by immunohistochemical techniques is particularly attractive due to the immediate information of the antigen distribution in different tissues as well as its cellular location. However, there have been technical problems with such screening such as background staining due to endogenous immunoglobulin and antigen-independent binding of immunoglobulin to different tissues. Technical difficulties due to endogenous immunoglobulin can be overcome by preincubating the tissue section with Fab' rabbit anti-human IgM or anti-IgG as previously reported (Nielsen et al., 1987). Antigen-independent binding could take place when testing antibodies on epithelial tissue since such tissue may contain secretory component (SC). It is known that secretory component, an 80 kDa glycoprotein present in most glandular epithelial cells, functions as the receptor for human IgM and dIgA. The J chains which are common for IgM and dIgA have been shown to determine the formation of immunoglobulin complexes with SC (Eskeland and Brandtzaeg, 1974). The present report describes the correlation between distribution of secretory component (SC) and antigen-independent binding of human Ig in frozen tissue sections and a method by which this antigen-independent binding can be blocked.
Materials and methods
Human-human hybridomas Human monoclonal IgM antibodies (D4213 and F10279) were secreted by human-human hybridomas derived by fusion between the human B lymphoblastoid cell line, WI-L2-729-HF2 (Abrams et al., 1983) and lymphocytes obtained from mesenteric lymph nodes from patients with colorectal cancer (Borup-Christensen et al., 1990). D4213 was selected due to its reactivity with neoplastic colonic tissue and limited or no binding to other normal and malignant tissues. F10279
was selected to serve as a control, since it showed no specific binding to frozen sections.
Immunoglobulin preparations Three different batches of polyclonal human IgM and one batch of polyclonal human IgG purified from normal human serum were obtained from Cappel Laboratories (Cochranville, PA, U.S.A.). Two different IgM preparations were purified from the sera of patients with WaldenstrOm macroglobulinaemia by affinity chromatography on monoclonal anti-/~ antibody (HB57, ATCC, Rockville, MO, U.S.A.) coupled to Sepharose 4B. Polyclonal human IgA was purified from normal human serum as described by Heide and Schwick (1978). In brief, euglobulins were precipitated by dialysis against water and IgG was removed by precipitation with ZnSO 4 before precipitation of IgA with (NH4)2SO 4. Dimeric and monomeric IgA were further purified by gel chromatography on a Sephadex G-200 column. Dimeric IgA was eluted in a distinct peak between V0 and V~ for IgG. Monomeric IgA was eluted just ahead of the IgG position. The identity of the proteins was verified by SDS-PAGE and crossed immunoelectrophoresis.
Anti-secretory component antibodies The anti-secretory component antibody (antiSC) used for staining of SC and for the blocking of IgM binding was a rabbit polyclonal anti-SC obtained from Dakopatts, Copenhagen, Denmark (code A187, lot 032B). The antibody had been solid phase adsorbed using plasma fractions and lactoferrin and the specificity was tested by crossed immunoelectrophoresis. No precipitate was obtained using human plasma, and only one precipitation line was observed with human colostrum. A mouse monoclonal anti-SC antibody was obtained from Sigma, St. Louis, MO, U.S.A. (ascites, code 1-6635, lot 38F 4802). The reactivity of the two anti-SC antibodies was compared by immunohistochemical staining on frozen tissue sections and on Western blots of non-reduced saliva. Similar staining patterns were observed with both of the anti-SC antibodies using a panel of normal and malignant frozen sections of colonic tissue. On Western blots of non-reduced saliva identical
247 TABLE I B I N D I N G O F H U M A N IgM, d l g A A N D R A B B I T A N T I - S C T O F R O Z E N S E C T I O N S O F N O R M A L BY I N D I R E C T I M M U N O P E R O X I D A S E S T A I N I N G
O e s o p h a g u s s q u a m o u s e p i t h e l i u m (n = 2) G a s t r i c e p i t h e l i u m (n = 5) Small intestine e p i t h e l i u m (n = 3) C o l o n e p i t h e l i u m ( n = 11 ) A n a l tract s q u a m o u s e p i t h e l i u m (n = 1) Liver (hepatocytes) (n = 1) Biliary ductal e p i t h e l i u m (n = 1) Breast g l a n d u l a r e p i t h e l i u m (n = 10) Bronchial respiratory e p i t h e l i u m (n = 2) Bronchial g l a n d u l a r e p i t h e l i u m (n = 2) L u n g aveolar e p i t h e l i u m (n = 2) R e n a l tubuli (n = 1) R e n a l glomeruli (n = 1) B l a d d e r transitional e p i t h e l i u m (n = 2) Skin k e r a t o s q u a m o u s e p i t h e h u m (n = 1) T o n g u e s q u a m o u s e p i t h e l i u m (n = 1) Cervical s q u a m o u s e p i t h e l i u m (n = 1) Cervical g l a n d u l a r e p i t h e l i u m (n = 3) E n d o m e t r i a l g l a n d u l a r e p i t h e l i u m (n = 1) T u b a e e p i t h e l i u m (n = 1) O v a r i a n follicular e p i t h e l i u m (n = 1) S m o o t h muscle (n = 2) T h y r o i d follicular e p i t h e l i u m (n = 2) Tonsil ( l y m p h a t i c c o m p o n e n t ) (n = 2) C o n n e c t i v e tissue (n = 10)
TISSUES ANALYSED
IgM binding
dlgA binding
Presence of SC
+ + +
+ + +
+ + +
+ +
+ +
+ +
___
+
+ +
+
a
+ +
+ +
±
~---
±
+ + + +
+ + + +
+ + + +
a ± indicates that Ig b i n d i n g or the presence of SC o n l y are f o u n d in some areas of the i n v e s t i g a t e d tissue.
bands were visualized with the two anti-SC antibodies. Tissues
Material was obtained from patients undergoing surgical resection and from diagnostic biopsies. The tissues and numbers of samples are listed in Tables I and II. Tissue judged to be normal by conventional microscopy was obtained from nondiseased areas of tissue specimens from cancer patients and from patients with benign disease. Tissues were snap-frozen in Isopentane, cooled in dry ice, or formalin fixed (4% formaldehyde, 0.075 M sodium phosphate, pH 7.0) for 24 h and paraffin-embedded. Sections of 5 /am were then cut, air-dried overnight at 25 ° C and fixed in ice-cold acetone for 10 rain. The formalin-fixed, paraffinembedded sections were deparaffinized and rehydrated through graded alcohols. The formalin-
T A B L E II B I N D I N G O F H U M A N IgM, d l g A A N D R A B B I T A N T I - S C TO FROZEN SECTIONS OF MALIGNANT TISSUES ANALYSED BY INDIRECT IMMUNOPEROXIDASE STAINING IgM IgA Presb i n d i n g b i n d i n g ence of SC C o l o n a d e n o c a r c i n o m a (n = 8) _ Breast a d e n o c a r c i n o m a (n = 5) G a s t r i c a d e n o c a r c i n o m a (n = 1) + Cervical c a r c i n o m a (n = 2) + E n d o m e t r i a l a d e n o c a r c i n o m a (n = 3) + O v a r i a n a d e n o c a r c i n o m a (n = 1) + R e n a l a d e n o c a r c i n o m a (n = 1) B l a d d e r a d e n o c a r c i n o m a (n = 1) D e r m a t o f i b r o s a r c o m a (n = 1) M e l a n o m a ( n = 1) -
+ + ± + + -
+ + + _+ + + -
+ indicates that Ig b i n d i n g or the presence of SC only are found in some areas of the i n v e s t i g a t e d tissue.
248
fixed sections were, before incubation with anti-SC antibody, treated for 20 rain at 37°C with 0.05% trypsin (cat. no. T-8128, Sigma, St. Louis, MO, U.S.A.) in distilled water containing 0.05% CaC12 adjusted to pH 7.8 with NaOH.
Immunohistochemical techniques (a) Binding of immunoglobulin and treatment with anti-SC. Frozen sections were preincubated with the Fab' fragment of rabbit anti-human IgM or Fab' rabbit anti-human IgA, diluted in phosphate-buffered saline pH 7.4 (PBS) containing 1% rabbit serum for 4 h at 4°C in a humidified chamber. This was done to block endogenous IgM or IgA determinants and thereby inhibit the binding of the subsequently applied enzyme-labelled anti-IgM or anti-IgA (Nielsen et al., 1987). To block the antigen-independent immunoglobulin binding, the sections were incubated with rabbit anti-SC (10 mg IgG/ml, Dakopatts, Copenhagen, Denmark), diluted 1/20 in PBS containing 1% rabbit serum for 2 h. Sections (with or without prior anti-SC treatment) were then incubated overnight at 4°C with human monoclonal IgM antibodies or other IgM (5 /xg/ml), IgG or IgA preparations (100 /~g/ml and 400/~g/ml) in PBS containing 10% rabbit serum and 1 mM EDTA. Sections were next incubated for 30 rain at room temperature with horseradish-peroxidase-conjugated rabbit anti-human lgM (HRP-anti-IgM) (1/400), HPR-anti-IgG (1/200) or HRP-anti-IgA (1/40) (Dakopatts, Copenhagen, Denmark) in PBS with 10% rabbit serum. Before the first incubation and between each additional step, sections were washed three times with PBS containing 0.05% Tween 20 (Merck, Darmstadt, F.R.G. (PBSTween)). Staining was developed by incubation in PBS containing diaminobenzidine (0.6/~g/ml) and hydrogen peroxide (0.01%). Formalin-fixed, paraffinembedded tissue sections were treated similarly. (b) Binding of anti-SC. Frozen and formalinfixed sections were incubated with rabbit anti-SC (Dakopatts, Copenhagen, Denmark) diluted 1/1800 (frozen tissue) and 1/600 (formalin-fixed tissue) in 20% pig serum. The incubation was performed for 30 min at room temperature, followed by incubation with horseradish-peroxidaseconjugated pig anti-rabbit lg (Dakopatts) (1/50)
in PBS. Washing and visualisation with chromogenic substrate were as described above. Sections were couterstained with haematoxylin and mounted in Aquamount (Gurr, BDH, Poole, England). Adjacent sections were stained with haematoxylin and eosin for the examination of tissue morphology.
Results
Normal and malignant frozen, acetone-fixed tissue was examined by an indirect immunoperoxidase technique for the presence of secretory component (SC) and binding of human IgM and dIgA. The results are listed in Tables I and II. The binding pattern of three different polyclonal IgM preparations and two different IgM preparations from patients with Waldenstr~Sm macroglobulinaemia were investigated on four sections of neoplastic colonic tissue and adjacent normal epithelia. All preparations used at a concentration of 5 ~g/ml showed the same binding pattern with staining of all normal colonic tissue and focal staining of the neoplastic colonic tissue. The same sections were also investigated using mlgA, dlgA and IgG (all at 100 /~g/ml and 400 /ag/ml). Staining was only found with dIgA at a concentration of 400 tag/ml, dIgA and IgM showed a similar binding pattern, whereas no binding was found to normal or malignant colonic tissue using mIgA or IgG at similar or higher concentrations. Binding of human IgM and dIgA was not observed when the tissue was formalinfixed and paraffin-embedded. As shown in Tables I and II similar binding was found with IgM, dIgA and rabbit anti-SC to different tissues. Controls with normal rabbit IgG were negative. The influence of preincubation with rabbit anti-SC (Dakopatts, Denmark) on the antigen-independent binding of IgM and dIgA was evaluated on frozen, acetone-fixed tissue expressing binding of IgM and dIgA, such as colonic epithelium and respiratory epithelium. Prior to these investigations endogenous IgM or IgA in the tissue was blocked by incubation with Fab' anti-IgM or Fab' anti-IgA. Antigen-independent binding of IgM was
249
Fig. 1. Immunoperoxidase analysis on frozen sections of bronchial tissue for the binding of polyclonal human IgM (Cappel). A: section blocked with Fab' anti-~ only before applying the IgM preparation. Antigen-independent IgM binding is seen in the respiratory epithelium and glandular cells. B: section blocked with both Fab' anti-~ and anti-SC before applying the IgM preparation. Antigen-independent IgM binding is abolished. effectively blocked when the section was preincubated with polyclonal anti-SC at a concentration of 0.5 m g / m l for 2 h at 25 ° C (Fig. 1). The staining with dIgA did not disappear when sections were preincubated with anti-SC, but in most cases was actually enhanced. The combined blocking with Fab' anti-IgM and anti-SC permitted the screening of human hybridoma supernatants containing IgM (F10279 and D4213) on SC-containing frozen tissue sections such as respiratory epithelium and colonic epithelium (Fig. 2).
Discussion The production of human monoclonal antibodies has become a widely used technique and the
development of reliable screening methods such as immunohistochemistry is therefore increasingly important. It has been observed that m a n y antibodies react with determinants which are better preserved in sections of fresh-frozen tissue than in formalin-fixed tissue (Mason et al., 1982). However, most immunohistochemical investigations are carried out on formalin-fixed paraffin-embedded tissues due to problems of routine freezing, storage and preparation of tissue sections. In this study we have shown that antigen-independent binding of h u m a n IgM is a problem when screening human monoclonal antibody on sections of frozen tissue. Antigen-independent binding was also found with dIgA, however, only when dIgA was applied at concentrations higher than normally used in immunohistochemical screening. It could be that only a small fraction of the dIgA
250
Fig. 2. Immunoperoxidase analysis on frozen sections of colonic epithelium for the binding of the human monoclonal IgM antibodies, F10279 and D4213. A: section blocked with Fab' anti-/~ only before applying the human antibody F10279 or D4213. Due to antigen-independent IgM binding the antibody specificity of the monoclonal antibody cannot be determined. B: section blocked with both Fab' anti-t~ and anti-SC before applying the human antibody F10279. There is no specific binding to the colonic epithelium. C: section blocked with both Fab' anti-# and anti-SC before applying the human antibody D4213. In this case there is specific binding to the colonic epithelium.
purified from serum was binding in the latter case. I g G and mIgA on the other hand did not bind nonspecifically to frozen tissue. Table I shows the tissue distribution of secretory component. As noted in the introduction section, secretory component is present in most glandular epithelial cells and functions as the receptor for human IgM and dlgA. Anti-SC, human IgM and human dlgA showed a similar staining pattern. The results show the tissues in which antigen-independent binding of IgM and IgA are to be found and where problems in the interpretation of antibody specificity are to be expected. We examined the possibility of blocking the antigen-independent IgM and dIgA binding with rabbit anti-SC antibody in order to improve the screening of human hybridoma antibodies on frozen tissue. Non-specific staining with IgM was completely prevented when the frozen tissue sec-
tion was blocked for endogenous immunoglobulin by Fab' anti-IgM followed by anti-SC, before incubation with non-specific IgM. The staining was not prevented, but was enhanced, when the frozen tissue sections were incubated with anti-SC before incubation with dlgA. This staining was due to interactions between the dlgA preparation and the rabbit anti-SC antibody. Since dlgA binding only takes place when it is present at very high concentrations (400 /zg/ml) the blocking of dlgA binding in the screening of human monoclonal antibodies may be of minor importance. However, we have not been able to test this with human monoclonal IgA antibody. It is of importance when working with techniques involving anti-Ig antibodies to ensure that problems do not arise due to cross species reactivities. This may often occur when using a primary antibody raised in one species and a secondary
251
antibody raised in another species, However, this problem does not occur when, as in this work, only rabbit antibodies are used. Furthermore, the protein source used for nonspecific blocking must be from the same species as the antibodies in order to avoid problems with cross reactions. The method of blocking antigen-independent IgM binding with anti-SC described here is simple and should be of general use for the screening of human monoclonal antibodies of IgM class on human tissue.
Acknowledgements This investigation was supported by grants from the Danish Cancer Society and the Danish Medical Research Council. Mrs. A. Klim, Mrs. J. Bohrk, Mrs. K. Hartmann and Miss K. Kristensen provided skilled technical assistance.
References Abrams, P.G., Knost, J.A., Clarke, G., Wilburn, S., Oldham, R.K. and Foon, K.A. (1983) Determination of the optimal human cell lines for development of human hybridomas. J. Immunol. 131, 1201. Borup-Christensen, P., Erb, K., Ditzel, H., Nielsen, B., Larsen, J., Svehag, S.-E. and Jensenius, J..C. (1990) Human-human hybridoma producing monoclonal antibodies against colorectal cancer-associated antigens. APMIS, in press.
Boyd, J.E. and James, K. (1989) Human monoclonal antibodies: Their potential, problems, and prospects. In: A. Mizrahi (Eds.), Monoclonal Antibodies: Production and Application. Alan R. Liss, New York, p. 1. Erber, W.N., Falini, B., Chosh, A.K., Moir, D.J. and Mason, D.Y. (1983) lmmuno-alkaline phosphatase labelling of haematological samples with monoclonal antibodies. In: S. Avrameas et al. (Eds.), Immunoenzymatic Techniques. Elsevier, Amsterdam, p. 29. Eskeland, T. and Brandtzaeg, P. (1974) Does J chain mediate the combination of 19 S IgM and dimeric IgA with the secretory component rather than being necessary for their polymerization? Immunochemistry 11, 161. Heide, K. and Schwick, H.G. (1978) Salt fractionation of immunoglobulins. In: D.M. Weir (Ed.), Handbook of Experimental Immunology, vol. 1. Blackwell Scientific Publishers, Ch. 7. Mason, D.Y., Naim, M., Abdulaziz, Z., Gatter~ K., Nash, J.R.G. and Stein, H. (1982) In: J.W. Fabre and A.J. Mc Michael (Eds.), Monoclonal Antibodies in Clinical Medicine. Academic Press, London, p. 585. Naiem, M., Gerdes, A.Z., Sunderland, C.A., Allington, M.J., Stein, H. and Mason, D.Y. (1982) The value of immunohistological screening in the production of monoclonal antibodies. J. Immnunol. Methods 50, 145. Nielsen, B., Borup-Christensen, P., Erb, K., Jensenius, J.C. and Husby, S. (1987) A method for the blocking of endogenous immunoglobulin on frozen tissue sections in the screening of human hybridoma antibody culture supernatants. Hybridoma 6, 103. Surer, L., Bruggen, J. and Sorg, C. (1980). Use of an enzymelinked immunosorbent assay (ELISA) for screening of hybridoma antibodies against cell surface antigens. J. Immunol. Methods 39, 407.
