J. Cancer
(1991), 63, Suppl. XIV,
Br. Br. J. Cancer (1991), 63, Suppi.
XIV,
78-81 81 78
NI, ©
Macmillan Press Ltd., 1991 Macmillan
Immunotargeting of human small cell lung cancer xenografts in athymic mice using a monoclonal antibody (RNL-1) against a neuroendocrine-related antigen E.P. Mijnheere', O.C. Boerman', J.L.V. Broers2, M. Klein Rot', G.P. Vooijs' & F.C.S. Ramaekers2 'Department of Pathology, University Hospital Nijmegen, The Netherlands and 2Department of Molecular Cell Biology, University of Limburg, PO Box 616, 6200 MD Maastricht, The Netherlands. Summary A mouse monoclonal antibody (RNL-1) was raised against the variant small cell lung cancer (SCLC) cell line NCI-H82. Immunohistochemical studies on frozen sections showed that the antibody was reactive with most SCLC (15 out of 16) and lung carcinoids (six out of seven), while in general adenocarcinomas and squamous cell carcinomas of the lung were negative. Immunocytochemical studies on 29 different cell lines derived from human lung tumours confirmed the neuroendocrine-related expression of the RNL-1 defined antigenic determinant. Immunoelectron microscopy showed that RNL-1 recognises an extracellular membrane domain, concentrated at adhesion sites between adjacent cells. The tissue distribution of the RNL-1 defined antigen was mainly restricted to neural and neuroendocrine tissues. These immunohistochemical data suggest that RNL-1 is directed against a neuroendocrine-related cell adhesion molecule. Being reactive with an epitope expressed on the surface of most neuroendocrine malignant cells, RNL-1 (IgGI isotype) is a potential vehicle for targeting SCLC in vivo. We evaluated the ability of radiolabelled RNL-1 to localise human SCLC xenografts in nude mice as a first step in determining the in vivo value for radioimmunodetection. RNL-1 was radioiodinated using the Bolton-Hunter labelling technique. Nude mice bearing NCI-H82 xenografts were injected intravenously with the radiolabelled RNL-1 preparations, and animals were dissected 4, 24, 48, 72 and 120 h post injection (p.i.) to determine the biodistribution of the radiolabel. The iodine-125 label accumulated in the tumour up to 48 h p.i. (6.5% injected dose per gram of tissue [ID g-'I]), while the label content of the normal tissues decreased with time. Tumour/non-tumour ratios 72 h p.i. ranged from 47 (tumour/brain) to 3.1 (tumour/lung). These data suggest that RNL-1 is a promising candidate for in vivo applications.
Lung cancer is still one of the main causes for cancer deaths, especially amongst men in Western countries. Small cell lung cancer (SCLC) accounts for 20-25% of all new cases of primary lung cancer. Unlike the other major histological types of lung cancer (collectively referred to as non-SCLC), SCLC is highly sensitive to both chemotherapeutic agents and radiation therapy, so that 90% of patients with SCLC initially will achieve a clinical remission after cytotoxic therapy. However, the options for the treatment of recurrent disease occurring in most cases, are extremely limited. Given the difficulties in treating SCLC by means of conventional therapeutic prototcols, there is a continued interest in the application of immunological techniques for the management of this disease. A series of monoclonal antibodies (MAbs) against antigens associated with SCLC has been developed in various laboratories (see for review: Souhami et al., 1988). Labelled with appropriate radionuclides these antibodies may be used for radioimmunodetection of (un)suspected tumour lesions. Linked to drugs, toxins or radionuclides anti-SCLC MAbs may also have therapeutic potential. In this report we describe the development and characterisation of a new MAb, RNL-1, reactive with SCLCs and lung carcinoids. As a first step to evaluate the potential application of the antibody for in vivo diagnostic and therapeutic use, we investigated its ability to localise human SCLC in a nude mouse model.
Materials and methods Tissues, cell lines and immunocytochemical procedures Human tissue samples were obtained freshly at surgery or autopsy, snap-frozen in liquid nitrogen and stored either in liquid nitrogen or at - 80C. Cryostat sections (4-7 ,sm thick) were air dried and fixed Correspondence: F.C.S. Ramaekers.
in acetone for 5 min. Indirect immunofluorescence and immunoperoxidase assays were performed as decribed previously (Broers et al., 1986). Twenty-one SCLC cell lines and eight non-SCLC cell lines were used in this study (Carney et al., 1985a,b; Gazdar et al., 1985; de Leij et al., 1985a; Bepler et al., 1987a,b, 1988). All cell lines were grown in RPMI 1640-based culture medium (Gibco, Paisly, UK) supplemented with 13% foetal calf serum (Flow, Herts, UK).
Production of the monoclonal antibody RNL-1 Balb/c mice were immunised with cell suspensions of NCIH82 cells. For this purpose cells were harvested and washed once with 50 mM phosphate-buffered saline, pH 7.4 (PBS). On day 0 and day 14, mice received i.p. injections of 107 cells mixed with equal volumes of Freunds complete and incomplete adjuvant, respectively. Thereafter, mice were boosted four times at weekly intervals, without addition of adjuvant to the cell suspension. Three days after the last booster mice were sacrificed and spleen cells were fused with Sp2/0-Agl4 myeloma cells at a ratio of 2:1, essentially as described by Kohler and Milstein (1975), using PEG 4000 (E. Merck, Darmstadt, FRG). Hybridoma clones producing relevant monoclonal antibodies (MAbs) were selected using an enzyme-linked immunosorbent assay (ELISA) as described (Suster et al., 1980). In the ELISA, NCI-H82 cells coated onto the wells of 96-well plates were used as the solid phase. In addition, positive wells were immediately screened using the indirect immunofluorescence technique on cytospin preparations of NCI-H82 and NCI-H69, as well as on frozen sections of a SCLC, an adenocarcinoma, and a squamous cell carcinoma of the lung. Clones reacting only with SCLC were selected, subcloned three times and used for further studies.
Immunoelectron microscopy A suspension of 2.5 x 106 NCI-H82 cells was subjected to a standard indirect immunoperoxidase procedure. After washing
IMMUNOTARGETING USING A MONOCLONAL ANTIBODY, RNL-1
in PBS the cells were fixed in 1.5% glutaraldehyde in PBS (10 min, 20°C). After subsequent washing the peroxidase activity was detected with 3,3'-diaminobenzidine-HCI (Sigma Chem. Co., St. Louis, MO) in the presence of 0.01% H202. Thereafter, cells were fixed in 3% glutaraldehyde (1 h), washed in PBS and postfixed in 1% osmium-tetroxide (1 h). Cells were dehydrated in an ascending series of ethanol, embedded in Epon 812, and viewed with a Philips EM 300 electron microscope (Philips, Eindhoven, The Netherlands).
79
specifically reacted with SCLC. Therefore RNL-1 was further tested in an extended series of lung cancer cell lines and solid lung tumours, but also in a panel of human non-lung cancers
and normal human tissues. These results have been extensively described by Broers et al., 1991. In brief, all SCLC cell lines were clearly stained at the cell surface. More specifically, strong reactivity was seen at adherent junctions between adjacent cells, while cell surface regions not in contact with each other showed a less intense staining reaction (Figure
la).
Radioiodination of antibody RNL-J The antibody RNL-1 (IgGI subclass) was purified from ascitic fluid generated by inoculation of hybridoma cells into pristane-primed Balb/c mice, using a Protein A-Sepharose 4B column (Pharmacia, Uppsala, Sweden) (Ey et al., 1978). Antibody concentrations were quantitated by absorbance at 280 nm, assuming that 1 mg ml-' corresponds with an extinction of 1.38. The preparations were stored at - 20°C until use. The purified RNL-1 was radioiodinated according to the lodogen method (Fraker & Speck, 1975) as well as according to the Bolton-Hunter method (Bolton & Hunter, 1973). For the iodogen labelling procedure 100lig of RNL-1 in 50mM phosphate buffer (pH 7.2) and 0.3 mCi 1251 (Amersham Int., Amersham, UK) were added to glass tubes that had been precoated with 1,3,4,6-tetrachloro-3a, 6a-di-phenylglycoluril (Pierce Chemical Co., Rockford, IL) (15 jig 100 l1-'), and the reaction was allowed to proceed for 5 min. For the Bolton-Hunter labelling method, 400 fig of RNL-l in 0.1 M borate buffer, pH 8.5, were added to 1 mCi of dried Bolton and Hunter reagent (iodinated hydroxysuccinimide ester of 3-(4-hydroxyphenyl)propionic acid; Amersham Int., Amersham, UK) and incubated for 15 min at 0°C. Following both labelling reactions, the reaction mixtures were applied to a Sephadex G-25 column (Pharmacia, Sweden) and eluted with PBS. The fractions containing the labelled antibody were pooled and used for in vivo studies within 24 h. Immunoreactivity of the labelled antibody preparations was checked in a competitive binding assay as described earlier (Boerman et al., 1990), using NCI-H82 coated 96-well plates as the solid phase. Biolocalisation studies Male nude Balb/c mice bearing a subcutaneously (s.c.) growing human SCLC xenograft (NCI-H82) were used. The tumour model was established by s.c. injection of 3 x 107 NCI-H82 cells. Tumours were resected aseptically, and small pieces of 2-3 mm diameter were serially transplanted s.c. under ether anaesthesia. Four weeks after transplantation tumours measured 0.7 to 1.5 cm in diameter (0.3-2.0 g). At that time mice were used for biodistribution studies. NCI-H82 xenograft-bearing mice received 10 jig of 1251_ labelled RNL-1 in the tail vein. At five different time points after injection (4 h, 24 h, 48 h, 72 h, 120 h) groups of three mice were bled and sacrificed under ether anaesthesia. Tumours and selected tissues were removed, washed in PBS, blotted dry, weighed and analysed in a gamma counter. The results were expressed as the percentage of the injected dose per gram of tissue (% ID g-'), and as the radioacivity in tumour as compared to that in normal tissues (tumour/nontumour ratios). All experiments were done on two different occasions to check the reproducibility of the method.
Immunoelectron microscopy showed that the RNL- 1 defined antigen was indeed localised on the cell surface and especially at the cell adhesion sites (Figure 1 b). Nearly all SCLC tumour samples examined (13 out of 14, Figure lc), and lung carcinoids (six out of seven) were positive. Some squamous cell carcinomas of the lung (11 out of 38) did also react with RNL-1, while all adenocarcinomas of the lung (n = 15) were negative. With respect to expression of the RNL- 1 epitope in normal human tissues it can be stated that the antibody appears to be reactive in all neuronal tissues. Strong staining was seen in all nerves and nerve fibers. In addition, RNL-1 was reactive with neuroendocrine tissues, such as pancreatic islet cells, some cells in the pituitary gland and in the adrenal medulla. Expression was also observed in Leydig cells of the testis, in the thyroid and in smooth muscle cells of the small intestine, colon and bladder. All other normal human tissues examined (brain Purkinje cells, epidermis, tongue, oesophagus, stomach,
spleen, liver, mammary gland, cervix, ovary, prostate, kidney, lymph nodes, blood cells and bone marrow) were negative.
Biolocalisation studies Radioiodination of the purified RNL-1 antibody appeared to be critical and influenced its immunoreactivity. According to the competitive binding assay on 0.025% glutaraldehydefixed NCI-H82 cells, the antibody lost its immunoreactivity upon oxidative radioiodination (iodogen method), while ester linkage of the 251I-labelled Bolton-Hunter reagent preserved the immunoreactivity of RNL-1. With the latter preparation 50% displacement of the 1251-RNL-1 was obtained at an unlabelled antibody concentration of 6 jig ml 1, indicating an affinity constant of 2.7 x I07 M-. This preparation was used in the biolocalisation studies. Following i.v. injection of the '251-RNL-1 preparation into xenograft bearing mice, the '25I-label accumulated in the tumour, reaching a maximum level of 6.5% ID/g 48 h p.i. (Figure 2). The label content of the normal mouse tissues decreased during the experiment. Throughout the experiment the 125I content of the different solid tissues varied considerably, being the highest for kidney and the lowest for brain tissue (Figure 3). Blood clearance of the radiolabel was slow (7.4% IDg-' at 24h, 3.4% IDg-' at 120h), and
Results Selection and cross-reactivity of RNL-J The RNL-1 antibody was selected on basis of its reactivity with two human SCLC cell lines and three lung cancer specimens of different histology. From tests on this limited panel of lung cancer cell lines and tumour specimens we obtained the impression that amongst lung tumours RNL-1
Figure I Reaction pattern of RNL-l with a classic SCLC cell line (a, NCI-H69, using immunofluorescence) and a variant SCLC cell line (b, NCI-H82, immunoelectron microscopy), clearly indicating that the RNL-1 defined epitope is localised on the cell surface. The membranous staining pattern of RNL-1 is also observed in cryostat sections of SCLCs (c, indirect immunoperoxidase technique). Magnification: a, 275 x, b, lOOOx, c, 225 x.
80
E.P. MIJNHEERE et al.
8~~~~~~~~~~~~~~~
2i--
4
24
c}
- -,t
a
,g-^
48
72
120 hr
-Tumor --+-Kidney -*-Liver - Muscle Figure 2 Label uptake in tumour xenografts and several normal tissues of athymic Balb/c mice during 5 days after injection of '25l-RNL-1. The mean value of three experiments is presented.
7
6 5-
4
2-
0
Tumor
Muscle Spleen Intest Kidney Liver Heart Lung Brain Blood
Figure 3 Biodistribution of the radiolabel (expressed as percentage of the injected dose per gram of tissue) 48 h after injection of 10 fig '25I-Iabelled RNL-1. The mean value (± s.d.) of three experiments is presented.
therefore tumour/blood ratio is smaller than 1.0 during the first 3 days of the experiment. For all solid tissues 72 h p.i. tumour/non-tumour ratios exceeded four, except the tumour/ lung ratio being 3.1 at that time. Discussion In recent years several newly developed monoclonal antibodies (MAbs) have been described which are to some extent specific for certain subtypes of lung cancer. While some of these MAbs ae thought to recognize antigens characteristic of non-SCLC (Sobol et al., 1986) others are characteristic for (a subset of) SCLC (De Leij et al., 1985b; Takahashi et al., 1986; Tong et al., 1986; Ball et al., 1987; Waibel et al., 1987). RNL-1, the MAb described in this paper, shows clear similarities in its reactivity pattern with SCLC-cluster 1 MAbs as summarised at the First International Workshop on Small Cell Lung Cancer Antigens (Souhami et al., 1988)
and was grouped in cluster 1 at the Second Workshop. The expression of the RNL-1 antigen in normal human tissues, as well as our immunoelectron micr6scopical observations are in agreement with the findings of Kibbelaar et al. (1989), suggesting a close relationship between cluster 1 antigens and the neural cell adhesion molecule (NCAM). In addition, cluster 1 antibodies have been shown to react with NCAM using tranfectants (Patel et al., 1989). Initially our biolocalisation studies with RNL-1 were hampered because of the loss of immunoreactivity upon oxidative radioiodination. This problem could be overcome by using the Bolton-Hunter labelling technique. We suspect that the iodogen labelling technique may cause the modification of particular tyrosine residues of the MAb that are essential for antigen binding. In contrast, the BoltonHunter reagent is linked to free amino-groups of the antibody molecule. The immunoreactive fraction of the radioiodinated antibody could not be deduced from the binding assay because the assay was not performed at antigen excess conditions. Our immunotargeting studies showed that radioactivity levels in all normal tissues of the mice decreased during the biolocalisation experiments, while the uptake into the xenografted human SCLC increased up to 6.5% ID g- during the first 48 h after injection. This indicates a specific tumour accumulation of the RNL-1 antibody. Although RNL-1 is reactive with a wide range of neuroendocrine tissues, including human brain, no accumulation is seen in murine brain during our biolocalisation experiments. Whether this is due to lack of cross-reactivity of the MAb with murine neuroendocrine tissues or the inability of the immunoconjugate to penetrate the blood-brain barrier, is unclear. Furthermore, it remains to be defined whether the expression of the cluster 1 antigen on nerve fibers and other neural tissues will give rise to severe side-effects when these antibodies are applied in man. After 48 h the 1251 content of the xenografts decreased gradually with time. Probably this decrease is due to dehalogenation of the radioiodinated antibody at the tumour site, a generally observed phenomenon when radioiodinated antibodies are applied in in vivo studies (Halpern & Dillman, 1987). Indeed, preliminary results indicate that label uptake into the xenografts is higher and more persistent when In-l 11 labelled RNL-1 is used in this in vivo model. It has been suggested that the mode of presentation of a target antigen in the tumour may be essential for tumour uptake of antibody (Matzku et al., 1988). Theoretically, the RNL-1 antigen, predominantly expressed at adherent junctions between adjacent cells, could be considered a suboptimal target for tumour targeting. Our results, however, indicate that even such 'hidden' antigens can be reached by antibodies. It would therefore be worthwhile to compare the biolocalisation characteristics of different cluster 1 MAbs in a nude mouse model. Such a comparative study could demonstrate whether the relatively limited accessibility of the cluster 1 antigen actually limits its potential as a target for radioimmunoimaging and possibly radio-immunotherapy of SCLC. We wish to thank Drs G. Bepler, L. de Leij, D.N. Carney and Sj.Sc. Wagenaar for providing cell lines and tissues, and G. Schaart, 0. Moesker, W. van den Broek, J. Koedam and G. Grutters for their skilled technical assistance.
References
BALL, E.D., KEEFE, K.A..& COLBY, E. (1987). Expression of antigens associated with small cell carcinoma of the lung on hematopoietic progenitor cells. Cancer Res., 47, 6556. BEPLER, G., JAQUES, G., KOEHLER, A., GROPP, C. & HAVEMANN, K. (1987a) Markers and characteristics of human SCLC cell lines.
Neuroendocrine markers, classical tumor markers, and chromosomal characteristics of permanent human small cell lung cancer cell lines. J. Cancer Res. Clin. Oncol., 113, 253.
BEPLER, G., JAQUES, J., NEUMAN, K., AUMUELLER, G., GROPP, C.
& HAVEMANN, K. (1987b). Establishment, growth properties,
and morphological characteristics of permanent human small cell lung cancer cell lines. J. Cancer Res. Clin. Oncol., 113, 31. BEPLER, G., KOEHLER, A., KIEFER, P. & 5 others (1988). Characterization of the state of differentiation of six newly established human non-small-cell lung cancer cell lines. Differentiation, 37, 158.
IMMUNOTARGETING USING A MONOCLONAL ANTIBODY, RNL-1 BOERMAN, O.C., MAKKINK, W.K., THOMAS, C.M.G. & 4 others
(1990). Monoclonal antibodies that discriminate between human ovarian carcinomas and benign ovarian tumours. Eur. J. Cancer Clin. Oncol., 26, 117. BOLTON, A.E. & HUNTER, W.M. (1973). The labelling of proteins to high specific radioactivities by conjugation to a '25I-containing acylating agent. Application to the radioimmunoassay. Biochem. J., 133, 529. BROERS, J.L.V., CARNEY, D.N., KLEIN ROT, M. & 4 others (1986). Intermediate filament proteins in classic and variant types of small cell lung carcinoma cell lines: a biochemical and immunochemical analysis using a panel of monoclonal and polyclonal antibodies. J. Cell Sci., 83, 37. BROERS. J.L.V., MIJNHEERE, E.P., KLEIN ROT, M. & 4 others (1991). Novel antigens characteristic of neuroendocrine malignancies. Cancer, 67 (in press). CARNEY, D.N., GAZDAR, A.F., BEPLER, G. & 5 others (1985a). Establishment and identification of small cell lung cancer cell lines having classic and variant features. Cancer Res., 45, 2913. CARNEY, D.N., GAZDAR, A.F., NAU, M., MINNA, J.D. (1985b). Biological heterogeneity of small cell lung cancer. Sem. Oncol., 12, 289. DE LEIJ, L., POSTMUS, P.E., BUYS, C.H.C.M. & 7 others (1985a). Characterization of three new, 'variant-type' cell lines derived from small cell carcinoma of the lung. Cancer Res., 45, 6024. DE LEIJ, L., POPPEMA, S., KLEIN NULEND, J. & 5 others (1985b). Neurendocrine differentiation antigen on human lung carcinoma and Kulchitsky cells. Cancer Res., 45, 2192. EY, P.L., PROWSE, S.J. & JENKINS, C.R. (1978). Isolation of pure IgG1, IgG2. and IgG2b immunoglobulins from mouse serum using protein A-sepharose. Immunochemistry, 15, 429. FRAKER, P.J. & SPECK, J.C. (1978). Protein and cell membrane iodination with a sparingly soluble chloramide 1,3,4,6-tetrachloro3a, 6c-diphenylglucoril. Biochem. Biophys. Res. Commun., 80, 849. GAZDAR, A.F., CARNEY, D.N., NAU, M.M. & MINNA, J.D. (1985).
Characterization of variant subclasses of cell lines derived from small cell lung cancer having distinctive biochemical, morphological, and growth properties. Cancer Res., 45, 2924.
81
HALPERN, S.E. & DILLMAN, R.O. (1987). Problems associated with radioimmunodetection and possibilities for future solutions. J. Biol. Resp. Modif., 6, 235. KIBBELAAR, R.E., MOOLENAAR, C., MICHALIDES, R., VAN BODEGOM, P., WAGENAAR, S. & MOOI, W. (1989). N-CAM, NE markers and prognosis in lung carcinoma. In: Proceedings of the 30th Dutch Federation Meeting. 211. KOEHLER, G. & MILSTEIN, C. (1975). Continuous cultures of fused cells secreting antibody of predefined specificity. Nature, 256, 495. MATZKU, S., TILGEN, W., KALTHOFF, H., SCHMIEGEL, W.H. &
BROEKER, E.-A. (1988). Dynamics of antibody transport and internalization. Int. J. Cancer, 2 (supplement), 11. PATEL, K., MOORE, S.E., DICKSON, G. & 4 others (1989). Neural cell adhesion molecule (NCAM) is the antigen recognized by monoclonal antibodies of similar specificity in small-cell lung carcinoma and neuroblastoma. Int. J. Cancer, 44, 573. SOBOL, R.E., PETERS, R.E., ASTARITA, R.W. & 8 others (1986). A novel monoclonal antibody-defined antigen which distinguishes human non-small cell from small cell lung carcinomas. Cancer Res., 46, 4746. SOUHAMI, R.L., BEVERLEY, P.C.L. & BOBROW, L. (1988). Proceedings of the First International Workshop on Small Cell Lung Cancer Antigens. Lung Cancer, 4, 1. SUSTER, L., BRUEGGEN, J. & SORG, C. (1980). Use of enzyme-linked immunosorbent assay (ELISA) for screening of hybridoma antibodies against cell surface antigens. J. Immunol. Methods, 39, 407. TAKAHASHI, T., UEDA, R., SONG, X. & 10 others (1986). Two novel cell surface antigens on small cell lung carcinoma defined by mouse monoclonal antibodies NE-25 and PE-35. Cancer Res., 46, 4770. TONG, A.W., LEE, J.C. & STONE, M.J. (1986). Discrimination of human small cell and non-small cell lung tumors by a panel of monoclonal antibodies. J. Natl Cancer Inst., 77, 1023. WAIBEL, R., O'HARA, C.J. & STAHEL, R.A. (1987). Characterization of an epithelial and tumor-associated human small cell lung carcinoma glycoprotein antigen. Cancer Res., 47, 3766.
(1991), 63, Suppl. XIV,
Br. Br. J. Cancer (1991), 63, Suppi.
XIV,
78-81 81 78
NI, ©
Macmillan Press Ltd., 1991 Macmillan
Immunotargeting of human small cell lung cancer xenografts in athymic mice using a monoclonal antibody (RNL-1) against a neuroendocrine-related antigen E.P. Mijnheere', O.C. Boerman', J.L.V. Broers2, M. Klein Rot', G.P. Vooijs' & F.C.S. Ramaekers2 'Department of Pathology, University Hospital Nijmegen, The Netherlands and 2Department of Molecular Cell Biology, University of Limburg, PO Box 616, 6200 MD Maastricht, The Netherlands. Summary A mouse monoclonal antibody (RNL-1) was raised against the variant small cell lung cancer (SCLC) cell line NCI-H82. Immunohistochemical studies on frozen sections showed that the antibody was reactive with most SCLC (15 out of 16) and lung carcinoids (six out of seven), while in general adenocarcinomas and squamous cell carcinomas of the lung were negative. Immunocytochemical studies on 29 different cell lines derived from human lung tumours confirmed the neuroendocrine-related expression of the RNL-1 defined antigenic determinant. Immunoelectron microscopy showed that RNL-1 recognises an extracellular membrane domain, concentrated at adhesion sites between adjacent cells. The tissue distribution of the RNL-1 defined antigen was mainly restricted to neural and neuroendocrine tissues. These immunohistochemical data suggest that RNL-1 is directed against a neuroendocrine-related cell adhesion molecule. Being reactive with an epitope expressed on the surface of most neuroendocrine malignant cells, RNL-1 (IgGI isotype) is a potential vehicle for targeting SCLC in vivo. We evaluated the ability of radiolabelled RNL-1 to localise human SCLC xenografts in nude mice as a first step in determining the in vivo value for radioimmunodetection. RNL-1 was radioiodinated using the Bolton-Hunter labelling technique. Nude mice bearing NCI-H82 xenografts were injected intravenously with the radiolabelled RNL-1 preparations, and animals were dissected 4, 24, 48, 72 and 120 h post injection (p.i.) to determine the biodistribution of the radiolabel. The iodine-125 label accumulated in the tumour up to 48 h p.i. (6.5% injected dose per gram of tissue [ID g-'I]), while the label content of the normal tissues decreased with time. Tumour/non-tumour ratios 72 h p.i. ranged from 47 (tumour/brain) to 3.1 (tumour/lung). These data suggest that RNL-1 is a promising candidate for in vivo applications.
Lung cancer is still one of the main causes for cancer deaths, especially amongst men in Western countries. Small cell lung cancer (SCLC) accounts for 20-25% of all new cases of primary lung cancer. Unlike the other major histological types of lung cancer (collectively referred to as non-SCLC), SCLC is highly sensitive to both chemotherapeutic agents and radiation therapy, so that 90% of patients with SCLC initially will achieve a clinical remission after cytotoxic therapy. However, the options for the treatment of recurrent disease occurring in most cases, are extremely limited. Given the difficulties in treating SCLC by means of conventional therapeutic prototcols, there is a continued interest in the application of immunological techniques for the management of this disease. A series of monoclonal antibodies (MAbs) against antigens associated with SCLC has been developed in various laboratories (see for review: Souhami et al., 1988). Labelled with appropriate radionuclides these antibodies may be used for radioimmunodetection of (un)suspected tumour lesions. Linked to drugs, toxins or radionuclides anti-SCLC MAbs may also have therapeutic potential. In this report we describe the development and characterisation of a new MAb, RNL-1, reactive with SCLCs and lung carcinoids. As a first step to evaluate the potential application of the antibody for in vivo diagnostic and therapeutic use, we investigated its ability to localise human SCLC in a nude mouse model.
Materials and methods Tissues, cell lines and immunocytochemical procedures Human tissue samples were obtained freshly at surgery or autopsy, snap-frozen in liquid nitrogen and stored either in liquid nitrogen or at - 80C. Cryostat sections (4-7 ,sm thick) were air dried and fixed Correspondence: F.C.S. Ramaekers.
in acetone for 5 min. Indirect immunofluorescence and immunoperoxidase assays were performed as decribed previously (Broers et al., 1986). Twenty-one SCLC cell lines and eight non-SCLC cell lines were used in this study (Carney et al., 1985a,b; Gazdar et al., 1985; de Leij et al., 1985a; Bepler et al., 1987a,b, 1988). All cell lines were grown in RPMI 1640-based culture medium (Gibco, Paisly, UK) supplemented with 13% foetal calf serum (Flow, Herts, UK).
Production of the monoclonal antibody RNL-1 Balb/c mice were immunised with cell suspensions of NCIH82 cells. For this purpose cells were harvested and washed once with 50 mM phosphate-buffered saline, pH 7.4 (PBS). On day 0 and day 14, mice received i.p. injections of 107 cells mixed with equal volumes of Freunds complete and incomplete adjuvant, respectively. Thereafter, mice were boosted four times at weekly intervals, without addition of adjuvant to the cell suspension. Three days after the last booster mice were sacrificed and spleen cells were fused with Sp2/0-Agl4 myeloma cells at a ratio of 2:1, essentially as described by Kohler and Milstein (1975), using PEG 4000 (E. Merck, Darmstadt, FRG). Hybridoma clones producing relevant monoclonal antibodies (MAbs) were selected using an enzyme-linked immunosorbent assay (ELISA) as described (Suster et al., 1980). In the ELISA, NCI-H82 cells coated onto the wells of 96-well plates were used as the solid phase. In addition, positive wells were immediately screened using the indirect immunofluorescence technique on cytospin preparations of NCI-H82 and NCI-H69, as well as on frozen sections of a SCLC, an adenocarcinoma, and a squamous cell carcinoma of the lung. Clones reacting only with SCLC were selected, subcloned three times and used for further studies.
Immunoelectron microscopy A suspension of 2.5 x 106 NCI-H82 cells was subjected to a standard indirect immunoperoxidase procedure. After washing
IMMUNOTARGETING USING A MONOCLONAL ANTIBODY, RNL-1
in PBS the cells were fixed in 1.5% glutaraldehyde in PBS (10 min, 20°C). After subsequent washing the peroxidase activity was detected with 3,3'-diaminobenzidine-HCI (Sigma Chem. Co., St. Louis, MO) in the presence of 0.01% H202. Thereafter, cells were fixed in 3% glutaraldehyde (1 h), washed in PBS and postfixed in 1% osmium-tetroxide (1 h). Cells were dehydrated in an ascending series of ethanol, embedded in Epon 812, and viewed with a Philips EM 300 electron microscope (Philips, Eindhoven, The Netherlands).
79
specifically reacted with SCLC. Therefore RNL-1 was further tested in an extended series of lung cancer cell lines and solid lung tumours, but also in a panel of human non-lung cancers
and normal human tissues. These results have been extensively described by Broers et al., 1991. In brief, all SCLC cell lines were clearly stained at the cell surface. More specifically, strong reactivity was seen at adherent junctions between adjacent cells, while cell surface regions not in contact with each other showed a less intense staining reaction (Figure
la).
Radioiodination of antibody RNL-J The antibody RNL-1 (IgGI subclass) was purified from ascitic fluid generated by inoculation of hybridoma cells into pristane-primed Balb/c mice, using a Protein A-Sepharose 4B column (Pharmacia, Uppsala, Sweden) (Ey et al., 1978). Antibody concentrations were quantitated by absorbance at 280 nm, assuming that 1 mg ml-' corresponds with an extinction of 1.38. The preparations were stored at - 20°C until use. The purified RNL-1 was radioiodinated according to the lodogen method (Fraker & Speck, 1975) as well as according to the Bolton-Hunter method (Bolton & Hunter, 1973). For the iodogen labelling procedure 100lig of RNL-1 in 50mM phosphate buffer (pH 7.2) and 0.3 mCi 1251 (Amersham Int., Amersham, UK) were added to glass tubes that had been precoated with 1,3,4,6-tetrachloro-3a, 6a-di-phenylglycoluril (Pierce Chemical Co., Rockford, IL) (15 jig 100 l1-'), and the reaction was allowed to proceed for 5 min. For the Bolton-Hunter labelling method, 400 fig of RNL-l in 0.1 M borate buffer, pH 8.5, were added to 1 mCi of dried Bolton and Hunter reagent (iodinated hydroxysuccinimide ester of 3-(4-hydroxyphenyl)propionic acid; Amersham Int., Amersham, UK) and incubated for 15 min at 0°C. Following both labelling reactions, the reaction mixtures were applied to a Sephadex G-25 column (Pharmacia, Sweden) and eluted with PBS. The fractions containing the labelled antibody were pooled and used for in vivo studies within 24 h. Immunoreactivity of the labelled antibody preparations was checked in a competitive binding assay as described earlier (Boerman et al., 1990), using NCI-H82 coated 96-well plates as the solid phase. Biolocalisation studies Male nude Balb/c mice bearing a subcutaneously (s.c.) growing human SCLC xenograft (NCI-H82) were used. The tumour model was established by s.c. injection of 3 x 107 NCI-H82 cells. Tumours were resected aseptically, and small pieces of 2-3 mm diameter were serially transplanted s.c. under ether anaesthesia. Four weeks after transplantation tumours measured 0.7 to 1.5 cm in diameter (0.3-2.0 g). At that time mice were used for biodistribution studies. NCI-H82 xenograft-bearing mice received 10 jig of 1251_ labelled RNL-1 in the tail vein. At five different time points after injection (4 h, 24 h, 48 h, 72 h, 120 h) groups of three mice were bled and sacrificed under ether anaesthesia. Tumours and selected tissues were removed, washed in PBS, blotted dry, weighed and analysed in a gamma counter. The results were expressed as the percentage of the injected dose per gram of tissue (% ID g-'), and as the radioacivity in tumour as compared to that in normal tissues (tumour/nontumour ratios). All experiments were done on two different occasions to check the reproducibility of the method.
Immunoelectron microscopy showed that the RNL- 1 defined antigen was indeed localised on the cell surface and especially at the cell adhesion sites (Figure 1 b). Nearly all SCLC tumour samples examined (13 out of 14, Figure lc), and lung carcinoids (six out of seven) were positive. Some squamous cell carcinomas of the lung (11 out of 38) did also react with RNL-1, while all adenocarcinomas of the lung (n = 15) were negative. With respect to expression of the RNL- 1 epitope in normal human tissues it can be stated that the antibody appears to be reactive in all neuronal tissues. Strong staining was seen in all nerves and nerve fibers. In addition, RNL-1 was reactive with neuroendocrine tissues, such as pancreatic islet cells, some cells in the pituitary gland and in the adrenal medulla. Expression was also observed in Leydig cells of the testis, in the thyroid and in smooth muscle cells of the small intestine, colon and bladder. All other normal human tissues examined (brain Purkinje cells, epidermis, tongue, oesophagus, stomach,
spleen, liver, mammary gland, cervix, ovary, prostate, kidney, lymph nodes, blood cells and bone marrow) were negative.
Biolocalisation studies Radioiodination of the purified RNL-1 antibody appeared to be critical and influenced its immunoreactivity. According to the competitive binding assay on 0.025% glutaraldehydefixed NCI-H82 cells, the antibody lost its immunoreactivity upon oxidative radioiodination (iodogen method), while ester linkage of the 251I-labelled Bolton-Hunter reagent preserved the immunoreactivity of RNL-1. With the latter preparation 50% displacement of the 1251-RNL-1 was obtained at an unlabelled antibody concentration of 6 jig ml 1, indicating an affinity constant of 2.7 x I07 M-. This preparation was used in the biolocalisation studies. Following i.v. injection of the '251-RNL-1 preparation into xenograft bearing mice, the '25I-label accumulated in the tumour, reaching a maximum level of 6.5% ID/g 48 h p.i. (Figure 2). The label content of the normal mouse tissues decreased during the experiment. Throughout the experiment the 125I content of the different solid tissues varied considerably, being the highest for kidney and the lowest for brain tissue (Figure 3). Blood clearance of the radiolabel was slow (7.4% IDg-' at 24h, 3.4% IDg-' at 120h), and
Results Selection and cross-reactivity of RNL-J The RNL-1 antibody was selected on basis of its reactivity with two human SCLC cell lines and three lung cancer specimens of different histology. From tests on this limited panel of lung cancer cell lines and tumour specimens we obtained the impression that amongst lung tumours RNL-1
Figure I Reaction pattern of RNL-l with a classic SCLC cell line (a, NCI-H69, using immunofluorescence) and a variant SCLC cell line (b, NCI-H82, immunoelectron microscopy), clearly indicating that the RNL-1 defined epitope is localised on the cell surface. The membranous staining pattern of RNL-1 is also observed in cryostat sections of SCLCs (c, indirect immunoperoxidase technique). Magnification: a, 275 x, b, lOOOx, c, 225 x.
80
E.P. MIJNHEERE et al.
8~~~~~~~~~~~~~~~
2i--
4
24
c}
- -,t
a
,g-^
48
72
120 hr
-Tumor --+-Kidney -*-Liver - Muscle Figure 2 Label uptake in tumour xenografts and several normal tissues of athymic Balb/c mice during 5 days after injection of '25l-RNL-1. The mean value of three experiments is presented.
7
6 5-
4
2-
0
Tumor
Muscle Spleen Intest Kidney Liver Heart Lung Brain Blood
Figure 3 Biodistribution of the radiolabel (expressed as percentage of the injected dose per gram of tissue) 48 h after injection of 10 fig '25I-Iabelled RNL-1. The mean value (± s.d.) of three experiments is presented.
therefore tumour/blood ratio is smaller than 1.0 during the first 3 days of the experiment. For all solid tissues 72 h p.i. tumour/non-tumour ratios exceeded four, except the tumour/ lung ratio being 3.1 at that time. Discussion In recent years several newly developed monoclonal antibodies (MAbs) have been described which are to some extent specific for certain subtypes of lung cancer. While some of these MAbs ae thought to recognize antigens characteristic of non-SCLC (Sobol et al., 1986) others are characteristic for (a subset of) SCLC (De Leij et al., 1985b; Takahashi et al., 1986; Tong et al., 1986; Ball et al., 1987; Waibel et al., 1987). RNL-1, the MAb described in this paper, shows clear similarities in its reactivity pattern with SCLC-cluster 1 MAbs as summarised at the First International Workshop on Small Cell Lung Cancer Antigens (Souhami et al., 1988)
and was grouped in cluster 1 at the Second Workshop. The expression of the RNL-1 antigen in normal human tissues, as well as our immunoelectron micr6scopical observations are in agreement with the findings of Kibbelaar et al. (1989), suggesting a close relationship between cluster 1 antigens and the neural cell adhesion molecule (NCAM). In addition, cluster 1 antibodies have been shown to react with NCAM using tranfectants (Patel et al., 1989). Initially our biolocalisation studies with RNL-1 were hampered because of the loss of immunoreactivity upon oxidative radioiodination. This problem could be overcome by using the Bolton-Hunter labelling technique. We suspect that the iodogen labelling technique may cause the modification of particular tyrosine residues of the MAb that are essential for antigen binding. In contrast, the BoltonHunter reagent is linked to free amino-groups of the antibody molecule. The immunoreactive fraction of the radioiodinated antibody could not be deduced from the binding assay because the assay was not performed at antigen excess conditions. Our immunotargeting studies showed that radioactivity levels in all normal tissues of the mice decreased during the biolocalisation experiments, while the uptake into the xenografted human SCLC increased up to 6.5% ID g- during the first 48 h after injection. This indicates a specific tumour accumulation of the RNL-1 antibody. Although RNL-1 is reactive with a wide range of neuroendocrine tissues, including human brain, no accumulation is seen in murine brain during our biolocalisation experiments. Whether this is due to lack of cross-reactivity of the MAb with murine neuroendocrine tissues or the inability of the immunoconjugate to penetrate the blood-brain barrier, is unclear. Furthermore, it remains to be defined whether the expression of the cluster 1 antigen on nerve fibers and other neural tissues will give rise to severe side-effects when these antibodies are applied in man. After 48 h the 1251 content of the xenografts decreased gradually with time. Probably this decrease is due to dehalogenation of the radioiodinated antibody at the tumour site, a generally observed phenomenon when radioiodinated antibodies are applied in in vivo studies (Halpern & Dillman, 1987). Indeed, preliminary results indicate that label uptake into the xenografts is higher and more persistent when In-l 11 labelled RNL-1 is used in this in vivo model. It has been suggested that the mode of presentation of a target antigen in the tumour may be essential for tumour uptake of antibody (Matzku et al., 1988). Theoretically, the RNL-1 antigen, predominantly expressed at adherent junctions between adjacent cells, could be considered a suboptimal target for tumour targeting. Our results, however, indicate that even such 'hidden' antigens can be reached by antibodies. It would therefore be worthwhile to compare the biolocalisation characteristics of different cluster 1 MAbs in a nude mouse model. Such a comparative study could demonstrate whether the relatively limited accessibility of the cluster 1 antigen actually limits its potential as a target for radioimmunoimaging and possibly radio-immunotherapy of SCLC. We wish to thank Drs G. Bepler, L. de Leij, D.N. Carney and Sj.Sc. Wagenaar for providing cell lines and tissues, and G. Schaart, 0. Moesker, W. van den Broek, J. Koedam and G. Grutters for their skilled technical assistance.
References
BALL, E.D., KEEFE, K.A..& COLBY, E. (1987). Expression of antigens associated with small cell carcinoma of the lung on hematopoietic progenitor cells. Cancer Res., 47, 6556. BEPLER, G., JAQUES, G., KOEHLER, A., GROPP, C. & HAVEMANN, K. (1987a) Markers and characteristics of human SCLC cell lines.
Neuroendocrine markers, classical tumor markers, and chromosomal characteristics of permanent human small cell lung cancer cell lines. J. Cancer Res. Clin. Oncol., 113, 253.
BEPLER, G., JAQUES, J., NEUMAN, K., AUMUELLER, G., GROPP, C.
& HAVEMANN, K. (1987b). Establishment, growth properties,
and morphological characteristics of permanent human small cell lung cancer cell lines. J. Cancer Res. Clin. Oncol., 113, 31. BEPLER, G., KOEHLER, A., KIEFER, P. & 5 others (1988). Characterization of the state of differentiation of six newly established human non-small-cell lung cancer cell lines. Differentiation, 37, 158.
IMMUNOTARGETING USING A MONOCLONAL ANTIBODY, RNL-1 BOERMAN, O.C., MAKKINK, W.K., THOMAS, C.M.G. & 4 others
(1990). Monoclonal antibodies that discriminate between human ovarian carcinomas and benign ovarian tumours. Eur. J. Cancer Clin. Oncol., 26, 117. BOLTON, A.E. & HUNTER, W.M. (1973). The labelling of proteins to high specific radioactivities by conjugation to a '25I-containing acylating agent. Application to the radioimmunoassay. Biochem. J., 133, 529. BROERS, J.L.V., CARNEY, D.N., KLEIN ROT, M. & 4 others (1986). Intermediate filament proteins in classic and variant types of small cell lung carcinoma cell lines: a biochemical and immunochemical analysis using a panel of monoclonal and polyclonal antibodies. J. Cell Sci., 83, 37. BROERS. J.L.V., MIJNHEERE, E.P., KLEIN ROT, M. & 4 others (1991). Novel antigens characteristic of neuroendocrine malignancies. Cancer, 67 (in press). CARNEY, D.N., GAZDAR, A.F., BEPLER, G. & 5 others (1985a). Establishment and identification of small cell lung cancer cell lines having classic and variant features. Cancer Res., 45, 2913. CARNEY, D.N., GAZDAR, A.F., NAU, M., MINNA, J.D. (1985b). Biological heterogeneity of small cell lung cancer. Sem. Oncol., 12, 289. DE LEIJ, L., POSTMUS, P.E., BUYS, C.H.C.M. & 7 others (1985a). Characterization of three new, 'variant-type' cell lines derived from small cell carcinoma of the lung. Cancer Res., 45, 6024. DE LEIJ, L., POPPEMA, S., KLEIN NULEND, J. & 5 others (1985b). Neurendocrine differentiation antigen on human lung carcinoma and Kulchitsky cells. Cancer Res., 45, 2192. EY, P.L., PROWSE, S.J. & JENKINS, C.R. (1978). Isolation of pure IgG1, IgG2. and IgG2b immunoglobulins from mouse serum using protein A-sepharose. Immunochemistry, 15, 429. FRAKER, P.J. & SPECK, J.C. (1978). Protein and cell membrane iodination with a sparingly soluble chloramide 1,3,4,6-tetrachloro3a, 6c-diphenylglucoril. Biochem. Biophys. Res. Commun., 80, 849. GAZDAR, A.F., CARNEY, D.N., NAU, M.M. & MINNA, J.D. (1985).
Characterization of variant subclasses of cell lines derived from small cell lung cancer having distinctive biochemical, morphological, and growth properties. Cancer Res., 45, 2924.
81
HALPERN, S.E. & DILLMAN, R.O. (1987). Problems associated with radioimmunodetection and possibilities for future solutions. J. Biol. Resp. Modif., 6, 235. KIBBELAAR, R.E., MOOLENAAR, C., MICHALIDES, R., VAN BODEGOM, P., WAGENAAR, S. & MOOI, W. (1989). N-CAM, NE markers and prognosis in lung carcinoma. In: Proceedings of the 30th Dutch Federation Meeting. 211. KOEHLER, G. & MILSTEIN, C. (1975). Continuous cultures of fused cells secreting antibody of predefined specificity. Nature, 256, 495. MATZKU, S., TILGEN, W., KALTHOFF, H., SCHMIEGEL, W.H. &
BROEKER, E.-A. (1988). Dynamics of antibody transport and internalization. Int. J. Cancer, 2 (supplement), 11. PATEL, K., MOORE, S.E., DICKSON, G. & 4 others (1989). Neural cell adhesion molecule (NCAM) is the antigen recognized by monoclonal antibodies of similar specificity in small-cell lung carcinoma and neuroblastoma. Int. J. Cancer, 44, 573. SOBOL, R.E., PETERS, R.E., ASTARITA, R.W. & 8 others (1986). A novel monoclonal antibody-defined antigen which distinguishes human non-small cell from small cell lung carcinomas. Cancer Res., 46, 4746. SOUHAMI, R.L., BEVERLEY, P.C.L. & BOBROW, L. (1988). Proceedings of the First International Workshop on Small Cell Lung Cancer Antigens. Lung Cancer, 4, 1. SUSTER, L., BRUEGGEN, J. & SORG, C. (1980). Use of enzyme-linked immunosorbent assay (ELISA) for screening of hybridoma antibodies against cell surface antigens. J. Immunol. Methods, 39, 407. TAKAHASHI, T., UEDA, R., SONG, X. & 10 others (1986). Two novel cell surface antigens on small cell lung carcinoma defined by mouse monoclonal antibodies NE-25 and PE-35. Cancer Res., 46, 4770. TONG, A.W., LEE, J.C. & STONE, M.J. (1986). Discrimination of human small cell and non-small cell lung tumors by a panel of monoclonal antibodies. J. Natl Cancer Inst., 77, 1023. WAIBEL, R., O'HARA, C.J. & STAHEL, R.A. (1987). Characterization of an epithelial and tumor-associated human small cell lung carcinoma glycoprotein antigen. Cancer Res., 47, 3766.
