JLP5B9: new monoclonal antibody against polysialylated neural cell adhesion molecule is of value in phenotyping lung cancer

Journal of Immunological Methods 233 Ž2000. 21–31 www.elsevier.nlrlocaterjim

JLP5B9: new monoclonal antibody against polysialylated neural cell adhesion molecule is of value in phenotyping lung cancer Maguy Del Rio a,1, Pascal Demoly b,1, Aurelia M.C. Koros c , Jean-Claude Laurent d , Jean-Claude Mani a , Bernard Pau a , Jean-Louis Pujol b,) b

a CNRS-UMR 9921, Centre de Recherche en Cancerologie, CRLC Val d’Aurelle, Montpellier, France ´ Maladies Respiratoires, Hopital Arnaud de VilleneuÕe, Centre Hospitalier UniÕersitaire (Prof. F.B. Michel), 34295 Montpellier Cedex 5, ˆ France c Department of Infectious Diseases and Microbiology, UniÕersity of Pittsburgh, Graduate School of Public Health, Pittsburgh, PA, USA d Biocytex, Departement BioserÕices, Marseille, France

Received 4 September 1998; received in revised form 24 June 1999; accepted 13 September 1999

Abstract Non-small-cell lung cancer ŽNSCLC. is currently one of the most prevalent malignant tumors. It displays a wide variety of phenotypes which includes neuroendocrine markers commonly found on small-cell lung cancers ŽSCLC. such as the neural cell adhesion molecule ŽNCAM. and in particular its highly polysialylated isoform, embryonic NCAM ŽeNCAM.. NSCLC with neuroendocrine differentiation may represent a subset of tumors whose cells have a more aggressive biological behavior. A tumor marker that distinguishes this latter sub-type could be of clinical relevance. Accordingly, we have raised a monoclonal antibody of the IgM type ŽJLP5B9. directed against capsular polysaccharides of N. meningitidis B which bears polysialic acid groups. We have demonstrated that JLP5B9 recognizes eNCAM with high affinity and that it is specifically directed against the polysialic acid moieties of NCAM. JLP5B9 was also found to react with human SCLC, NSCLC and neuroblastoma cell lines. We then used JLP5B9 as a specific probe for the detection of tissue eNCAM and found that it was expressed on up to 20% of tumor cells obtained from 5 out of 13 patients with NSCLC. This mAb deserves further investigation to evaluate its potential as a tool for serodiagnosis of lung cancer. q 2000 Elsevier Science B.V. All rights reserved. Keywords: Polysialic acid; NCAM; IgM; Lung cancer

1. Introduction The neural cell adhesion molecule ŽNCAM. is a membrane glycoprotein that can promote cell–cell adhesion through a homophilic binding mechanism ) Corresponding author. Tel.: q33-4-67-33-61-36; fax: q33-467-33-61-41; e-mail: [email protected] 1 MDR and PD are equal co-authors.

ŽRutishauser et al., 1982.. A unique structural feature of NCAM is the presence of polysialic acid, a developmentally regulated carbohydrate composed of a linear homopolymer of a-Ž2,8.-linked sialic acid residues ŽFinne et al., 1983a.. Unlike most carbohydrates found on the cell surface, polysialic acid in vertebrates appears to be exclusively associated with. NCAM. Indeed, in NCAM-deficient mice, lack of NCAM correlates with the almost entire loss of

0022-1759r00r$ - see front matter q 2000 Elsevier Science B.V. All rights reserved. PII: S 0 0 2 2 - 1 7 5 9 Ž 9 9 . 0 0 1 7 9 - 9

22

M. Del Rio et al.r Journal of Immunological Methods 233 (2000) 21–31

polysialic acid ŽCremer et al., 1994.. Although polysialylated NCAM ŽeNCAM. is abundant in embryonic tissues, the majority of NCAM in adult tissues lacks this glycan. The adhesive properties of NCAM are hampered by this large, negatively charged carbohydrate ŽRutishauser et al., 1988. and it is possible that this property contributes to the plasticity of neurons from the embryonic brain. The expression of eNCAM has also been detected in some carcinoids, neuroblastomas, renal carcinomas, small-cell lung cancers ŽSCLC. and non-small-cell lung cancers ŽNSCLC. ŽRoth et al., 1988; Moolenaar et al., 1990; Rygaard et al., 1992.. Several clues suggest that the same mechanism may contribute to the lack of cell adhesion in human lung cancers and neuroblastomas. This phenomenon is possibly a first step in the progession towards a metastatic phenotype. We and others have previously demonstrated that patients with NCAM-positive NSCLC have a shorter survival than those with NCAM-negative NSCLC and that NCAM expression is correlated with nodal status and indicates a poor prognosis ŽPujol et al., 1993.. This observation is in accordance with the hypothesis that the phenotypic diversification Ži.e., expression of NCAM by nonendocrine lung tumors. results in tumor progression. However, the antibodies used in these studies recognized mainly noncarbohydrate epitopes of NCAM, and it has never been possible to determine whether polysialic acid is involved in the negative effect of NCAM on cell adhesion. Thus, the polysialic acid moiety of NCAM may be related to the invasive and metastatic growth potential of human lung tumor cells. A significant association between the presence of polysialic acid in lung tumor cells and an aggressive and immature subtype of the tumor was demonstrated in vitro by Michalides et al. Ž1994.. Therefore, NCAM and possibly eNCAM might be expressed by a subset of lung tumor cells with a more aggressive biological behavior. In this study, we have characterized a monoclonal antibody ŽJLP5B9. raised against capsular polysaccharides of N. meningitidis B which have similarities with brain glycopeptides bearing polysialosyl units ŽFinne et al., 1983b.. We demonstrated that JLP5B9 is specifically directed against the polysialic acid moieties of NCAM. We used JLP5B9 as a

specific probe for the detection of eNCAM and found that it was expressed on up to 20% of tumor cells obtained from 5 out of 13 patients with NSCLC.

2. Materials and methods 2.1. Tissues and cells Tissues from the nervous system of 17-day rat embryos were used for the preparation of eNCAMenriched preparations. Tissues were homogenized at a ratio 1r10 Žwtrvol. in 20 mM Tris–HCl buffer ŽpH 8.0. containing 1% Nonidet P40, 5 mM EDTA and 1 mM phenylmethylsulfonyl fluoride, 10 mgrml leupeptin, 10 mgrml apoprotinin as protease inhibitors. The homogenate was centrifuged for 30 min at 48C and 40 000 = g; then the supernatant was removed and stored at y808C. The protein concentration was determined by the BCA protein assay ŽPierce, Rockford, IL, USA.. Well-characterized prototype cell lines representing SCLC ‘‘classic’’ ŽNCI-H69. SCLC ‘‘ variant’’ ŽNCI-N417. and other cell lines are shown in Table 1 and were kindly provided by Drs. A.F. Gazdar and H. Oie ŽCarney et al., 1988; Deftos et al., 1989; Lu et al., 1989.. Cultured cells were grown in RPMI 1640 medium with 10% heat-inactived fetal bovine serum without antibiotics in an atmosphere of 5% CO 2 at 378C and sub-cultured 24–96 h before being dispersed mechanically. 2.2. Monoclonal antibodies mAb 123C3 ŽIgG1. which recognizes an epitope within the extracellular part of human NCAM was purchased from Interchim ŽMontlucon, France.. mAb 0B11 ŽIgG1. localizes the transmembrane isoforms of NCAM Ž180 and 140 kDa. and was purchased from Sigma ŽSt. Louis, MO, USA.. mAb 735 ŽIgG2a., mAb 5A5 ŽIgM. and mAb anti-MenB ŽIgM. were generous gifts from Dr. Bitter-Suermann ŽFinne et al., 1987., Dr. Rutishauser ŽYang et al., 1994. and Dr. Rougon ŽRougon et al., 1986., respectively; these three mAbs are directed against polysialic acid. AntiNKH1 ŽN901; gift from Coulter Immunology, Hialeah, FL. and mAb MOC-1 ŽDako, Glostrup, Denmark. recognize the CD56 antigen ŽNCAM..

M. Del Rio et al.r Journal of Immunological Methods 233 (2000) 21–31

23

Table 1 Reactivity determined by flow cytometric analysis of monoclonal antibodies JLP5B9 and anti-NKH1 Žinternal control. against various cell lines Cell lines

Source

Classification

Immunoreactivity a of JLP5B9

NKH1

NCI-H69 NCI-H146 NCI-H345 NCI-H510 NCI-H82 NCI-H417 SHP-77 NCI-H460 NCI-H520 NCI-H676 SMS KCN SMS KCNR SMS KAN Pitt 1 IMR 32

Dr. Adi Gazdar Dr. Adi Gazdar Dr. Adi Gazdar Dr. Adi Gazdar Dr. Adi Gazdar Dr. Adi Gazdar Dr. Adi Gazdar and J. Paulson Dr. Adi Gazdar Dr. Adi Gazdar Dr. Adi Gazdar Dr. C.P. Reynolds Dr. C.P. Reynolds Dr. C.P. Reynolds Dr. Aurelia Koros American Type Culture Collection

Small-cell carcinoma-lung ‘‘classic’’ Small-cell carcinoma-lung ‘‘classic’’ Small-cell carcinoma-lung ‘‘classic’’ Small-cell carcinoma-extra pulmonary Small-cell carcinoma-lung ‘‘ variant’’ Small-cell carcinoma-lung ‘‘ variant’’ Small-cell carcinoma-lung unclassified Large cell carcinoma Squamous cell carcinoma Adenocarcinoma Neuroblastoma Neuroblastoma Neuroblastoma Neuroepithelioma Neuroblastoma

64 64 81 72 83 84 78 62 57 30 82 91 60 72 62

77 84 80 46 75 50 84 77 22 36 78 97 68 30 ND b

a b

Immunoreactivity is expressed in percentage of stained cells. Not determined.

mAb JLP5B9 ŽIgM. was raised against the polysaccharide of N. meningitidis B, a highly polymeric form of a-Ž2,8.-sialic acid. It was obtained by fusion of spleen cells from a 6-week-old BALBrc mouse ŽByJico, Iffa Credo, Lyon, France. immunized four times by i.p. injection with an inactivated bacterial suspension and SP2r0-Ag14 myeloma cells. Hybridoma JLP5B9 was selected by an indirect ELISA using the bacteria as antigen and immunoblotting using extracts of embryonic rat brain as antigen. After cloning by limiting dilution, JLP5B9 was produced in ascitic fluid from pristane-pretreated BALBrc mice and purified by gel filtration chromatography ŽSepharose CL-6B, Pharmacia, Uppsala, Sweden..

in Trisrglycine buffer Ž2.5 mM Tris, 192 mM glycine, and 20% Žvolrvol. methanol. using a semidry transfer method ŽPharmacia.. Non-specific binding sites were blocked for 1 h at room temperature with a solution of 10 mM Tris–HCl, pH 7.5, 150 mM NaCl ŽTBS., 0.05% Žvolrvol. Tween 20 ŽTBST. containing 5% Žwtrvol. skim dry milk. Immunostaining was then performed with a primary mAb anti-eNCAM in ascites fluid diluted 1r500 in TBS-T. After 2 h at room temperature, peroxidase conjugated goat anti-mouse isotype antibodies ŽSigma. was added for 1 h at room temperature. A tablet of 4-chloro-1-naphthol ŽSigma. in 50 ml TBS, 10 ml methanol and 0.1% Žvolrvol. H 2 O 2 was used as substrate for detection.

2.3. SDS-PAGE and immunoblotting 2.4. Immunopurification of eNCAM Samples were diluted in the sample buffer of Laemmli Ž1970. containing 5% Žwtrvol. 2-mercaptoethanol and heated at 808C for 3 min. The proteins were separated by electrophoresis on 8% SDS-polyacrylamide gels. Resolved proteins were then electrophoretically transferred to a nitrocellulose membrane ŽSchleicher & Schuell, Keene, NH.

One-hundred microliters of embryonic rat brain extracts Ž100 mgrml. were incubated for 4 h at 378C with anti-MenB immobilized microtiter plates. Antibody-bound proteins were eluted with 20 ml of Laemmli’s sample buffer containing 5% Žwtrvol. 2-mercaptoethanol. These eluates were pooled, then

24

M. Del Rio et al.r Journal of Immunological Methods 233 (2000) 21–31

analyzed on 8% SDS-polyacrylamide gels and tranferred to a nitrocellulose membrane. The reactivity of JLP5B9 was tested by immunoblotting as described above. 2.5. Neuraminidase treatment Embryonic brain extracts or cells were incubated for 2 h at 378C with an equal volume of either neuraminidase from Clostridium perfringens Žtype X, 1 Urml, Sigma. in 25 mM sodium acetate, pH 5, or endoneuraminidase from purified K1 bacteriophage Ža generous gift from Dr. Rougon. ŽWang et al., 1994. diluted 1r1000 in phosphate-buffered saline ŽPBS.. 2.6. Inhibition by colominic acid mAb JLP5B9 at 10 mgrml was incubated with increasing concentrations Ž1 to 50 mgrml. of colominic acid ŽSigma. for 2 h at 378C. Blot experiments were then performed as described above. 2.7. Affinity constant determination with BIACORE We determined the affinity constants for the binding of JLP5B9 to eNCAM by SPR Žsurface plasmon resonance. analysis using BIACORE ŽBiacore, Uppsala, Sweden. ŽFagerstam and Karlsson, 1993; ¨ Malmqvist, 1993.. BIACORE permits real-time analysis of biospecific interactions without the need for labelling either isotopic or enzymatic. BIACORE technology is based on optical surface plasmon resonance for detecting small changes in the refractive index on the surface of a thin gold film coated with a dextran matrix to which one of the two reactants is covalently linked, while the other one is introduced in a flow passing over the surface. To analyse JLP5B9reNCAM interactions, JL5B9 was immobilized onto a polyclonal rabbit anti-mouse IgM antibody ŽSigma. covalently linked to the dextran layer of a sensor chip CM5 ŽBiacore., by the well-known EDC-NHS protocol ŽJohnson et al., 1991.; then eNCAM was injected followed by mAb 123C3. All antibodies were diluted at 50 mgrml in 10 mM HEPES-buffered saline pH 7.4 containing 3.4 mM EDTA and 0.005% P20 surfactant Žrunning buffer..

The injection was performed at a flow rate of 10 mlrmin, and regeneration of the sensor chip was performed using 30 ml of 2 M formic acid. Rat embryonic brain extract at 6 mgrml of proteins was used as the source of antigen. To determine kinetic and equilibrium constants, we used two tumor extracts containing eNCAM Ža generous gift from Dr. Figarella-Branger, Hopital de la Timone, Marseille. whose concentration was determined by an immunoassay described by Dubois et al. Ž1995.. Tumor extracts no. 2226 and no. 2912 contained respectively 4.32 and 0.31 nM eNCAM. The kinetic parameters of the binding reactions were calculated from the sensorgrams corresponding to four concentrations of eNCAM Ž72, 108, 216 and 432 pM. using BIAevaluation 3.0 software ŽBiacore. and the global determination protocol ŽKarlsson and Falt, 1997.. Three sensorgrams were generated for each sample. 2.8. Flow cytometric analysis Cells were washed once with 48C Hank’s balanced salt solution ŽHBSS. without Ca2q or Mg 2q containing 0.1% sodium azide and 2% fetal bovine serum Žmonoclonal wash solution.. Approximately 5 = 10 5 cells were resuspended in 200 ml of monoclonal wash solution and 50 ml of the JLP5B9 mAb diluted 1r30 in monoclonal wash solution were added. The control mAb, anti-NKH1 was in ascites fluid diluted 1r30. Cells were incubated at 48C overnight, washed twice in monoclonal wash solution, resuspended in 200 ml monoclonal wash and stained at 48C with 50 ml of a 1r60 dilution of fluorescein-conjugated goat FŽabX . 2 anti-mouse IgG q IgM, human absorbed fragments Ž Tago, Burlingame, CA.. Cells were washed twice in monoclonal wash, resuspended in HBSS without Ca2q or Mg 2q containing 1% paraformaldehyde and analysed by flow cytometry using a single argon laser ŽProfile II, Coulter Electronics.. Five-thousand cells were analysed for each antibody. Bit maps were drawn around cell populations falling in regions that were not clumped and which were free of debris. The autofluorescence of each cell type was used to set the flow cytometer cursors at a baseline. The reactivity of each cell type with the goat FŽabX . 2 anti-mouse IgG q IgM was also determined.

M. Del Rio et al.r Journal of Immunological Methods 233 (2000) 21–31 Table 2 Patient characteristics and results of immunohistochemical analyses n

Sex

Age Žyear.

Histology a

Immunoreactivity b of JLP5B9 MOC-1

1 2 3 4 5 6 7 8 9 10 11 12 13

F M F F M F M M F F M F F 5 Mr8 F

42 62 54 41 67 31 55 68 66 77 39 38 36 52"14

LCC SQC SQC SQC SQC SQC Ad LCC SQC SQC Ad SQC SQC

1 1 1 0 2 0 2 0 0 0 0 0 0 5r13

3 1 1 2 1 1 1 1 1 2 2 0 1 12r13

a

Ad, adenocarcinoma; SQC, squamous cell carcinoma and LCC, large cell carcinoma. b Staining score: 0, no positive cells; 1, 1%–5% positive cells; 2, 5%–20% positive cells and 3, over 20% positive cells.

25

ŽDako.; finally, the alkaline phosphatase substrate, mixed with the Fast Red dye indicator ŽDako. was added, the slides washed, and counterstained with haematoxilin and mounted in glycergel w ŽDako.. Control slides were prepared by using an irrelevant mouse antibody of the same isotype at the same concentration as the specific primary mAb ŽDako.. Two different investigators, both unaware of the clinical findings, examined the slides. Positive cells were counted at a high power field according to the following scale: 0, no positive cells; 1, 1%–5% positive cells; 2, 6%–20% positive cells and 3, over 20% positive cells.

3. Results JLP5B9, a monoclonal antibody raised against capsular polysaccharides of N. meningitidis B was first selected by an indirect ELISA using bacteria as antigen and then on embryonic rat brain extracts by immunoblotting experiments. As shown in Fig. 1,

2.9. Immunohistochemistry Surgically resected specimens of lung cancer were tested immunohistochemically for reactivity with JLP5B9. They were obtained during surgical resection of histologically proven lung cancers from 13 patients ŽTable 2.. During the surgical procedure, a specimen was sampled from a non-necrotic area of each primary tumor and frozen in liquid nitrogen. In addition, slides prepared by cytocentrifugation of the IMR32 cell line were used as a positive control. Four-micrometer tumor sections were prepared and fixed in 3.7% Žwtrvol. formaldehyde. A first section was stained with hematoxylin and eosin for routine histologic evaluation. The percentage of positive cells was assessed using immune complexes of alkaline phosphatase and monoclonal anti-alkaline phosphatase ŽCordell et al., 1984.. Briefly, the primary mAb ŽJLP5B9 or MOC-1. at the final concentration of 10 mgrml was incubated for 30 min at 378C in humid chambers. Slides were washed three times in TBS Ž10 min each. and then incubated for 30 min with a rabbit anti-mouse IgG Žfinal dilution 1r100. ŽDako.. Slides were again washed three times and then incubated with the alkaline phosphatase-anti-alkaline phosphatase antibody Žfinal dilution 1r50.

Fig. 1. Immunoblot analysis of polysialylated NCAM from embryonic rat brain extracts Ž30 mg of protein per lane.. After separation by 8% SDS-PAGE under reducing conditions, transferred proteins were allowed to react with mAb JLP5B9 Žlane 1., mAb 5A5 Žlane 2., mAb anti-MenB Žlane 3., mAb 735 Žlane 4. and mAb OB11 Žlane 5.. Lane 6 represents the negative control with only the peroxidase-conjugated goat anti-mouse antibody. The molecular weights of the marker proteins ŽkDa. are indicated on the left.

26

M. Del Rio et al.r Journal of Immunological Methods 233 (2000) 21–31

JLP5B9 recognized a broad band of 200 to 250 kDa Žlane 1., compatible with the highly polysialylated isoform of NCAM. In the same experiment, we used as controls three anti-polysialic acid mAbs whose specificity is well documented, namely, 5A5 ŽYang et al., 1994., anti-MenB ŽRougon et al., 1986. and 735 ŽFinne et al., 1987. and a commercial antiNCAM antibody, OB11 which localizes the high molecular weight isoform of NCAM in mammalian tissues. The same staning was observed with all the mAbs tested: JLP5B9 Žlane 1., 5A5 Žlane 2., antiMenB Žlane 3. and 735 Žlane 4., and OB11 Žlane 5., indicating that JLP5B9 is probably directed against eNCAM contained in the rat embryonic brain. To ascertain whether the molecule recognized by JLP5B9 was eNCAM, we performed an immunopurification of eNCAM from extracts of embryonic rat brain using mAb anti-MenB Žsee Section 2.. The immunopurified eNCAM tranferred to nitrocellulose was strongly recognized by JLP5B9 Ždata not shown. demonstrating that the polysialylated protein recognized by JLP5B9 was indeed eNCAM. To further characterize JLP5B9, two inhibition experiments were carried out. In the first experiment, colominic acid, which is composed of oligomers of a-Ž2,8.-linked sialic acid, was used to inhibit JLP5B9 binding to eNCAM by saturation of the antibody binding sites. Increasing amounts of colominic acid were preincubated with JLP5B9 Ž10 mgrml. before immunoblotting on eNCAM. Fig. 2A shows that inhibition was concentration-dependent and reached 100% at 50 mgrml of colominic acid Žlane 1.. In the second experiment, rat embryonic brain extracts were pretreated with neuraminidase from C. perfringens Žknown to cleave NeuAc a-Ž2,8. linkages. for 2 h at 378C, before immunoblotting. The reactivity of JLP5B9 was completely abolished on pretreated extracts, providing evidence that desialylated NCAM was no longer recognized by JLP5B9 ŽFig. 2B, lane 1 vs. lane 2.. The efficiency of the enzyme was demonstrated in a control experiment where mAb OB11 recognized bands at 140 and 180 kDa Žlane 4., the two high molecular weight desialylated NCAM isoforms. Taken together, these data demonstrate that JLP5B9 is specifically directed against polysialic acid. Using BIACORE technology, we analyzed the molecular interaction of eNCAM with JLP5B9 cap-

Fig. 2. Specificity of JLP5B9. ŽA. Inhibition of JLP5B9 binding by preincubating colominic acid at 50 mgrml Žlane 1., 10 mgrml Žlane 2., 5 mgrml Žlane 3. and 1 mgrml Žlane 4. for 2 h at 378C with JLP5B9 at 10 mgrml. The reactivity of JLP5B9 with eNCAM without inhibitor is shown in lane 5. Molecular weights of marker proteins ŽkDa. shown on left. ŽB. The reactivity of JLP5B9 Žlanes 1 and 2. and OB11 Žlanes 3 and 4. was tested against embryonic rat brain extracts pretreated Žq. or not Žy. with neuraminidase for 2 h at 378C. Molecular weights of marker proteins ŽkDa. shown on right.

tured by a rabbit anti-mouse IgM previously covalently linked to the dextran layer of the sensor chip. We verified that the antigen captured by JLP5B9 from an embryonic rat brain extract was eNCAM by showing the sequential binding of the anti-NCAM mAB 123C3 to this complex ŽFig. 3A.. To demonstrate the specificity of the interaction, we showed that this capture was not possible when eNCAM was desialylated Ždata not shown.. JLP5B9 was also able to detect eNCAM in two tumor extracts Ždata not shown.. The association and dissociation rate constants of this interaction were determined by injec-

M. Del Rio et al.r Journal of Immunological Methods 233 (2000) 21–31

27

Fig. 3. Analysis of the binding of eNCAM to mAb JLP5B9 by BIACORE technology. mAb JLP5B9 was immobilized onto a polyclonal rabbit anti-mouse IgM antibody covalently bound to a CM5 sensor chip. ŽA. The capture of eNCAM from embryonic rat brain extract followed by injection of mAb 123C3. ŽB. Capture of the polysialylated neural cell adhesion molecule from four dilutions of a tumor extract containing 4.32 nM eNCAM. The kinetic parameters of the binding reaction were calculated from these sensorgrams using BIAevaluation 3.0 software ŽBiacore. and the global determination protocol ŽKarlsson and Falt, 1997..

tion of various dilutions of the the tumor extract containing 4.32 nM eNCAM. The kinetic analysis showed that binding of eNCAM to immobilized JLP5B9 comprised a rapid association rate and a slow dissociation rate ŽFig. 3B.. This behavior is characteristic of a stable complex and a high affinity constant. Indeed, the kinetic rate constants for mAb

JLP5B9 were found to be an association rate constant of 7.6 = 10 6 1rMs, a dissociation rate constant of 5.2 = 10y4 1rs and an affinity for eNCAM of 6.9 = 10y1 1 M. Futhermore, to check the possible involvement of JLP5B9 multivalence in the high affinity found for the binding of eNCAM to JLP5B9, we ran two sets of kinetic determinations using

28

M. Del Rio et al.r Journal of Immunological Methods 233 (2000) 21–31

BIAevaluation 3.0 software. In the first set, we used a 1:1 model, the fitting between the experimental and calculated curves was almost perfect, with only a very slight deviation Žless than 0.1%. in the dissociation phase. In order to evaluate whether this small effect was significant or not, we analyzed the same curve with a kinetic model called ‘‘Heterogeneous Ligand-Parallel Reactions’’ used to determine affinity rate constants when the ligand Žhere JLP5B9. is bi- or multivalent. Under these conditions, we calculated two dissociation rate constants Žone identical to the k d found with the 1:1 model and one higher.; but the involvement of the second parameter was negligible as shown by the ratio of R eq Ž R value at equilibrium., this lower affinity binding representing less than 0.1% of total binding.

We next studied eNCAM expression on human tumors by immunocytochemistry and immunohistochemistry. By flow cytometric analysis, we determined the reactivity of JLP5B9 against various cell lines including ‘‘classic’’ or ‘‘ variant’’ SCLC, NSCLC and neuroblastomas ŽTable 1. and used antiNKH1 Žanti-NCAM. as an internal positive standard. As shown in Table 1, JLP5B9 reacted with all cell lines tested as expected. Overall, 30% to 91% of the cells were positively stained with JLP5B9. In accordance with previous data showing evidence of eNCAM expression by both neuroblastoma and SCLC ŽPatel et al., 1989., 64% to 83% of SCLC cells and 60% to 91% of neuroblastoma cells were stained. By comparison, less NSCLC cells expressed polysialic acid Ž30% to 62%. indicating that JLP5B9 could be

Fig. 4. FACS analysis using ŽA. mAb 123C3 Žan NCAM-specific antibody. and ŽB. JLP5B9 Ža polysialic acid-specific antibody. of IMR32 . with endoneuraminidase for 2 h at 378C. cells treated Ž – – – . or not Ž

M. Del Rio et al.r Journal of Immunological Methods 233 (2000) 21–31

use to phenotype a sub-group of NSCLC cells. In addition, endoneuraminidase pretreatment of the IMR32 cell line led to an 86% inhibition of JLP5B9 reactivity, demonstrating that this antibody specifically binds polysialic acid, whereas immunoreactivity with mAb 123C3 is not altered by endoneuraminidase treatment of the cells ŽFig. 4.. Futhermore, 5 out of 13 NSCLC tumor biopsies showed positive JLP5B9 immunoreactivity ŽTable 2.. JLP5B9-positive cancer cells were found on some islets distributed throughout the tumor specimen. None of the non-malignant cells Žmostly inflammatory cells. surrounding or infiltrating the tumor showed positive JLP5B9 staining. No staining was observed in the control sections in which an unrelated antibody was used as the primary antibody. Since NCAM is the major carrier of polysialic acid in vertebrates ŽCremer et al., 1994., our immunocytochemistry and immunohistochemistry results demonstrate that JLP5B9 is a specific probe for the detection of eNCAM in human tissues.

4. Discussion In this study we have characterized a new mAb, JLP5B9, raised against capsular polysaccharides from N. meningitidis B, a highly polymeric form of a-Ž2,8. sialic acid. We have shown that JLP5B9 cross-reacts with eNCAM Žthe embryonic form of NCAM. and in particular with the polysialic acid moieties of eNCAM. These data confirm the similarity between capsular polysaccharides from N. meningitidis B and polysialic acid from human and rat brain, first discribed by Finne et al. Ž1983b.. The specificity against polysialic acid was demonstrated by inhibiting JLP5B9 binding to brain extracts both by competition with colominic acid composed of random-sized polymers of a-Ž2,8. sialic acid and by neuraminidase treatment. In both cases, the reactivity of JLP5B9 on eNCAM was abolished. Experiments are in progress to define precisely the structure of the epitope recognized by JLP5B9. BIACORE analysis has confirmed the specificity of JLP5B9 for eNCAM in a soluble form and determined its high affinity. To our knowledge, this is the first mAb directed against polysialic acid-dependent determinants on NCAM whose affin-

29

ity has been established using surface plasmon resonance technology. Futhermore, JLP5B9 reactivity on various human cell lines has demonstrated its capacity to recognize human eNCAM. In addition, JLP5B9 may be a useful tool for phenotyping and subclassifying the classic NSCLC histologic group as indicated by immunohistochemistry experiments. The possibility that JLP5B9 recognizes polysialyl residues expressed in other proteins in tissue specimens seems unlikely. In fact, the almost total absence of polysialic acid in mice with a genetic deletion of NCAM obtained by homologous recombination ŽCremer et al., 1994. provides definitive evidence that NCAM is the major carrier of this carbohydrate. Lung cancer presents with a wide variety of phenotypes. According to histological criteria, this malignant disease is classified into two majors groups: SCLC and NSCLC. SCLC, characterized by neuroendocrine differentiation markers, has a poor prognosis due to an early and widespread metastatic process, but it is sensitive to initial chemotherapy and radiation. NSCLC, in contrast, can be surgically resected when localized. Recently, several studies have suggested that the convenient separation between SCLC and NSCLC might be too theoretical inasmuch as these groups frequently overlap resulting in a composite SCLC-NSCLC. In this particular setting, up to 20% of NSCLC express a neuroendocrine phenotype, particularly NCAM. Those NSCLC with neuroendocrine differentiation may represent a subset of patients whose tumor cells present a more aggressive biological behavior. A tumor marker which can distinguish this latter subtype could be of clinical relevance. In this regard, the prognostic significance of the expression of NCAM in both types of lung cancer has been analysed in several studies. We have previously shown that patients with NCAM-positive NSCLC have a shorter survival time than those with NCAM-negative NSCLC and that NCAM expression is correlated with nodal status and indicates a poor prognosis ŽPujol et al., 1993.. These results confirm those of Kibbelaar et al. Ž1991. who showed that patients with NCSLC positive for mAb 123C3 had overall post-operative and disease-free survival times which were significantly shorter than patients with 123C3negative NSCLC. In vitro studies showed that there was a correlation between the presence of polysialic

30

M. Del Rio et al.r Journal of Immunological Methods 233 (2000) 21–31

acid in lung tumor cells and an aggressive and immature sub-type of the tumor ŽMichalides et al., 1994.. The NCAM molecule belongs to the immunoglobulin superfamily and consists of five extracellular immunoglobulin-like homology domains and two fibronectin type III units ŽRutishauser et al., 1982.. Alternative splicing of a single gene gives rise to several mRNAs encoding three major isoforms of 180, 140 and 120 kDa, the latter being linked to membranes via a glycan-phosphatidylinositol ŽGPI. anchor. In addition, the three isoforms can undergo polysialylation. The highly polysialylated isoform, eNCAM, was the first cell adhesion molecule shown to be an oncodevelopmental antigen ŽRoth et al., 1988.. It has clearly been demonstrated that eNCAM can be regarded as a specific serum marker for SCLC ŽJaques et al., 1993; Takamatsu et al., 1994.. Soluble eNCAM could arise from cancer cell necrosis and hydrolysis of the GPI anchor. Non-sialylated soluble NCAM can be found in normal subjects but soluble eNCAM is only found in cancer patients ŽSCLC. and appears to be tumor-associated ŽTakamatsu et al., 1994.. Our results indicate that mAb JLP5B9 can be used as a specific probe for the detection of eNCAM in lung tumor tissue. Studies are in progress to ascertain the clinical applicability of this mAb for the detection of eNCAM in sera from lung cancer patients.

Acknowledgements The authors thank Dr. J. Simony-Lafontaine, Mrs. N. Lequeux and Mrs. M. Radal ŽLaboratoire d’Anatomopathologie, Centre Val d’Aurelle, Montpellier. for the immunohistochemistry experiments, ŽCentre de Recherche en CanceroloDr. A. Pelegrin ` ´ gie, Centre Val d’Aurelle, Montpellier. and Dr. G. Rougon ŽLaboratoire de Genetique et Physiologie du ´ ´ Developpement, Marseille. for critical comments and ´ Dr. S.L. Salhi ŽLaboratoire d’Immunologie et Biotechnologie, Faculte´ de Pharmacie, Montpellier. for help in preparing the manuscript. This study was jointly supported by a grant from the G.E.F.L.U.C, from Elf Aquitaine and from the Scientific Commi-

tee of the Centre Val d’Aurelle, Montpellier and a grant to Aurelia Koros from the Cancer Federation of Banning, CA, USA .

References Carney, D.N., Gazdar, A.F., Bepler, G., Guccion, J.G., Marangos, P.J., Moody, T.W., Zweig, M.H., Minna, J.D., 1988. Establishment and identification of small-cell cancer cell lines having classic and variant features. Cancer Res. 45, 2913. Cordell, J.L., Falini, B., Erber, W.N., Ghosh, H.K., Zainalabideen, A., McDonald, S., Pulford, K., Stein, H., Mason, D.Y., 1984. Immunoenzymatic labelling of monoclonal antibodies using immune complexes of alkaline phosphatase and monoclonal anti-alkaline phosphatase ŽAPAAP. complexes. J. Histochem. Cytochem. 32, 219. Cremer, H., Lange, R., Christoph, A., Plomann, M., Vopper, G., Roes, J., Brown, R., Baldwin, S., Kraemer, P., Scheff, S., Barthels, P., Rajewsky, K., Wille, W., 1994. Inactivation of the N-CAM gene in mice results in size reduction of the olfactory bulb and deficits in spatial learning. Nature 367, 455. Deftos, L.J., Gazdar, A.F., Ikeda, K., Broadus, A.E., 1989. The parathyroid-hormone-related protein asociated with malignancy is secreted by neuro-endocrine tumors. Mol. Endocrinol. 3, 503. Dubois, C., Okandze, A., Figarella-Bringer, D., Rampini, C., Rougon, G., 1995. A monoclonal antibody against Meningococcus group B polysaccharides used to capture and quantify polysialylated NCAM in tissue and biological fluids. J. Immunol. Methods 181, 125. Fagerstam, L.G., Karlsson, R., 1993. Biosensor techniques. In: ¨ Van Regenmortel, M.H.V., Van Oss, C.J. ŽEds.., Immunochemisty. Marcel Dekker, p. 949. Finne, J., Finne, U., Deagostini-Bazin, M., Goridis, C., 1983a. Occurence of alpha 2.8 linked polysialosyl units in a neural cell adhesion molecule. Biochem. Biophys. Res. Commun. 112, 482. Finne, J., Leinonen, M., Makela, ¨ ¨ H., 1983b. Antigenic similarities between brain components and bacteria causing meningitis. Lancet II, 355. Finne, J., Bitter-Suermann, D., Goridis, C., Finne, U., 1987. An IgG monoclonal antibody to group B meningococci cross-reacts with developmentally regulated polysialic acid units of glycoproteins in neural and extraneural tissues. J. Immunol. 138, 4402. Jaques, G., Auerbach, B., Pritsch, M., Wolf, M., Madry, N., Havemann, K., 1993. Evaluation of serum neural cell adhesion molecule as a new tumor marker in small-cell lung cancer. Cancer 72, 418. Johnson, B., Lofas, S., Lindquist, G., 1991. Immobilization of ¨ proteins to a carboxymethyldextran modified gold surface for biospecific interaction analysis in surface plasmon resonance. Anal. Biochem. 198, 268. Karlsson, R., Falt, A., 1997. Experimental design for kinetic

M. Del Rio et al.r Journal of Immunological Methods 233 (2000) 21–31 analysis of proteinrprotein interactions with surface plasmon resonance. J. Immunol. Methods 200, 121. Kibbelaar, R.E., Moolenaar, K.E.C., Michalides, R.J.A.M., Van Bodegom, P.C., Vanderschueren, R.G.J.R.A., Wagenaar, S.S., Dingemans, K.P., Bitter-Suermann, D., Dalesio, O., Van Zandwijk, N., Mooi, W.J., 1991. Neural cell adhesion molecule expression, neuroendocrine differentiation and prognosis in lung carcinoma. Eur. J. Cancer 27, 431. Laemmli, U.K., 1970. Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227, 680. Lu, C., Ikeda, K., Deftos, L.J., Gazdar, A.F., Mangin, M., Broadus, A.E., 1989. Glucocorticoid regulation of parathyroid hormone-related peptide gene transcription in human neuroendocrine cell line. Mol. Endocrinol. 3, 2034. Malmqvist, M., 1993. Biospecific interaction analysis using biosensor technology. Nature 361, 186. Michalides, R., Kwa, B., Springall, D., Van Zandwijk, P., Koopman, J., Hilkens, J., Mooi, W., 1994. NCAM and lung cancer. Int. J. Cancer 8, S37. Moolenaar, C.E.C., Muller, E.J., Schol, D.J., Figdor, C.G., Bock, E., Bitter-Suermann, D., Michalides, R.J.A.M., 1990. Expression of neural cell adhesion molecule-related sialoglyco-protein in small-cell lung cancer cell lines and neuroblastoma cell lines H69 and CHP 212. Cancer Res. 50, 1102. Patel, K., Moore, S.E., Dickson, G., Rossell, R.J., Beverly, P.C., Kemshead, J.T., Walsh, F.S., 1989. Neural cell adhesion molecule ŽNCAM. is the antigen recognized by monoclonal antibodies of similar specificity in small-cell lung carcinoma and neuroblastoma. Int. J. Cancer 44, 573. Pujol, J.L., Simony, J., Demoly, P., Charpentier, R., Laurent, J.C., Daures, ` J.P., Lehmann, M., Guyot, V., Godard, P., Michel,

31

F.B., 1993. Neural cell adhesion molecule and prognosis of surgically resected lung cancer. Am. Rev. Respir. Dis. 148, 1071. Roth, J., Zuber, C., Wagner, P., Taatjes, D.J., Weisgerber, C., Heitz, P.U., Goridis, C., Bitter-Suermann, D., 1988. Reexpression of polyŽsialic acid. units of the neural cell adhesion molecule in Wilms tumor. Proc. Natl. Acad. Sci. U. S. A. 85, 2999. Rougon, G., Dubois, C., Buckley, N., Magnani, J.L., Zollinger, W.A., 1986. Monoclonal antibody against Meningococcus group B polysaccharides distinguishes embryonic from adultNCAM. J. Cell Biol. 103, 2429. Rutishauser, U., Hoffman, S., Edelman, G.M., 1982. Binding properties of a cell adhesion molecule from neural tissue. Proc. Natl. Acad. Sci. U. S. A. 79, 685. Rutishauser, U., Acheson, A., Hall, A.K., Mann, D.M., Sunshine, J., 1988. The neural cell adhesion molecule ŽNCAM. as a regulator of cell–cell interactions. Science 240, 53. Rygaard, K., Moller, C., Bock, E., Spang-Thomsen, M., 1992. Expression of cadherin and NCAM in human small-cell lung cancer cell lines and xenografts. Br. J. Cancer 65, 573. Takamatsu, K., Auerbach, B., Gerardy-Schahn, R., Eckhardt, M., Jaques, G., Madry, N., 1994. Characterization of tumor-associated neural cell adhesion molecule in human serum. Cancer Res. 54, 2598. Wang, C., Rougon, G., Kiss, J.Z., 1994. Requirement of polysialic acid for the migration of the O-2A glial progenitor cell from neurohypophyseal explants. J. Neurosci. 14, 4446. Yang, P., Major, D., Rutishauser, U., 1994. Role of charge and hydration in effects of polysialic acid on molecular interactions and between cell membranes. J. Biol. Chem. 16, 23039.