Isolation and preliminary characterization of a human ‘phage display’-derived antibody against neural adhesion molecule-1 antigen interfering with fibroblast growth factor receptor-1 binding

Abstract

BACKGROUND:

The NCAM or CD56 antigen is a cell surface glycoprotein belonging to the immunoglobulin super-family involved in cell-cell and cell-matrix adhesion. NCAM is also over-expressed in many tumour types and is considered a tumour associated antigen, even if its role and biological mechanisms implicated in tumour progression and metastasis have not yet to be elucidated. In particular, it is quite well documented the role of the interaction between the NCAM protein and the fibroblast growth factor receptor-1 in metastasis and invasion, especially in the ovarian cancer progression.

OBJECTIVE:

Here we describe the isolation and preliminary characterization of a novel human anti-NCAM single chain Fragment variable antibody able to specifically bind NCAM-expressing cells, including epithelial ovarian cancer cells.

METHODS:

The antibody was isolate by phage display selection and was characterized by ELISA, FACS analysis and SPR experiments. Interference in EOC migration was analyzed by scratch test.

RESULTS:

It binds a partially linear epitope lying in the membrane proximal region of two fibronectin-like domains with a dissociation constant of 3.43 $\times$ 10 ${}^{-8}$ M. Interestingly, it was shown to interfere with the NCAM-FGFR1 binding and to partially decrease migration of EOC cells.

CONCLUSIONS:

According to our knowledge, this is the first completely human antibody able to interfere with this newly individuated cancer mechanism.

Keywords

NCAM single-chain variable fragment phage display FGFR1 epithelial ovarian cancer

1. Introduction

The neural cell adhesion molecule-1 antigen (NCAM) or CD56 antigen is a membrane glycoprotein, a cell adhesion molecule structurally belonging to the immunoglobulin super-family with a molecular mass of 175 to 220 kDa under non-reducing conditions. NCAM is encoded by a single gene located on 11q23-24 that generates a wide variety of glycoprotein isoforms by alternative splicing. More than 24 distinct mRNAs of NCAM may be produced that are expressed preferentially in different tissues and stages of embryonic development.

The three major isoforms of NCAM differ mainly in overall size and in the type of membrane attachment. Two transmembrane-anchored glycoproteins of 140 and 180 kDa, respectively and a 120 kDa glycosyl phosphatidylinositol-anchored glycoprotein result from the use of different exons for the carboxyl terminus of the molecule. 120 kDa NCAM isoform is predominantly expressed in normal and well differentiated tissues, while both 140 and 180 kDa isoforms are found predominantly in less differentiated or malignant cell types [1, 2].

The extracellular domain has five IgG-like domains at the N-terminus and two fibronectin type III-like domains on the membrane proximal region and it is responsible for NCAM mediated cell-cell and cell-matrix interactions, by means of homophilic binding and various heterophilic interactions with other molecules.

NCAM is modified post-translationally by the addition of O-linked and N-linked oligosaccharides as well as $\alpha$ -2,8-linked polysialic acid and by sulfation. All of these modifications play a critical role in the regulation of NCAM-mediated cell adhesion and are essential for the correct differentiation of various tissues [3].

NCAM is expressed in several normal tissues, including natural killer (NK) cells, activated T cells, neurons, glial cells, neuroendocrine glands, skeletal muscle cells and cardiomyocytes. It contributes to a number of physiological processes such as embryogenesis and development, neuronal differentiation and synaptic plasticity [4].

The function of NCAM on human NK cells and T cells remains unclear, although its expression seems not to be essential for NK and T-cell development and function. Although a correlation exists between MHC-unrestricted cytotoxic function and expression of NCAM on lymphocytes, the NCAM antigen itself appears not to participate directly in NK killing of most target cells; notwithstanding, a recent report suggests that NCAM can mediate homotypic adhesion of NK cells to NCAM positive tumour cell lines and pathogen and tumour-cell recognition [5, 6, 7, 8].

NCAM is also considered a tumour-associated antigen, since it is overexpressed in almost all neuroblastoma [9], rhabdomyosarcoma [10], 98% of small cell lung cancer [11], 78% of multiple myeloma [12], and 53% of acute myeloid leukemia patients [13], some ovarian cancers [14], neuroendocrine cancers [15], pancreatic and gastric cancers [16], Wilms tumour [17] and Ewing sarcoma [18] patients also have elevated NCAM.

The functional role of NCAM in these tumours has to be elucidated, but actually a plethora of scientific data has emerged on the ability of NCAM to promote tumour progression and metastasis [14]. For example, NCAM upregulation subsequent to the loss of E-cadherin expression is implicated in the epithelial-mesenchymal transition (EMT) [19], a process that occurs when cancer cells disseminate from a primary tumour and infiltrate the surrounding tissue, leaving the tumour mass not only by gaining migratory and invasive capabilities but also by dissolving cell-cell contacts, known as tight and adherent junctions [20].

Moreover, recently, NCAM has been proposed to regulate the function of members of the fibroblast growth factor receptor (FGFR) family. These growth factor receptors are tyrosine kinases that trigger signaling cascades when bound by growth factors and other molecules, and aberrant FGFR signaling has been implicated in the progression of some cancers, including epithelial ovarian cancer, where its expression is associated with high tumour grade.

Published data have demonstrated that NCAM stimulates EOC cell migration and invasion in vitro and promotes metastatic dissemination in mice [14]. Several lines of evidence taken from in vitro and in vivo studies underline how this activity implicate the specific interaction between the adhesion molecule NCAM and members of the fibroblast growth factor receptor family, in particular FGFR1.

FGFR1 signaling induced by NCAM stimulation differs from the pathway initiated by other ligands such as fibroblast growth factor, FGF. In the absence of FGF, NCAM-mediated activation of FGFR specifically promotes characteristic cellular trafficking and receptor recycling that results in sustained FGFR signaling, leading to an enhanced cell migration with invasive and aggressive biological behavior [21].

NCAM-FGFR interactions are mediated by the second and the third Ig module of FGFR and the two Fibronectin-like type 3 (FN3) modules of NCAM. Even if the affinity of the NCAM-FGFR interaction (10 $\mu$ M K ${}_{\text{D}}$ ) may seem biologically insignificant when compared with the affinity of FGFs for FGFR (K ${}_{\text{D}}$ about 10 pM), the abundant expression of NCAM on cells surface can take account of the FGFR activation by NCAM. The regions directly of NCAM involved in this interaction were identified in FRM peptide from A-B loop of the first FN3 module (aa 504-516 of NCAM) and in FGL peptide from F-G loop of the second FN3 module (aa 670-684 of NCAM) [22].

It was also presumed that association of NCAM to FGFR occurs when NCAM is polysialylated and has not a tight packing of homophilic interactions [23]. Moreover, the change in NCAM glycosylation in some tumours, in particular an increase of polysialic acids (PSA) levels within the fifth immunoglobulin domain of NCAM by two polysialyltransferases [24], appears to play a predominant role in the metastasis formation and tumour progression, by preventing NCAM interactions and downstream signaling [25, 26, 27].

NCAM over expression in different tumour types together with its role in promoting migration, invasion and metastasis, especially in EOC, make this antigen a potential novel drug development target. In this context, the development of an antibody-based target therapy could be a welcome addition to the available treatment choices.

In this paper, we describe the isolation and characterization of a novel, fully human antibody, in a ‘single-chain’ format specific for NCAM and also able to interfere with its interaction with FGFR1.

2. Materials and methods

2.1 Antibody phage library, NCAM antigen, human PBMC from healthy donors and cell lines

The IORISS naive human antibody library consists of a large array (more than 10 ${}^{9}$ antibody combination) of scFv polypeptides displayed on the surface of M13 phage. It was constructed by recovering the light and heavy antibody chains genes from a group of healthy donors, then randomly joining them by means of an appropriate peptide linker sequence, and cloning the resulting scFvs genes into a phagemid vector as described elsewhere [28]. It is noteworthy that the scFv inserts have been cloned upstream of two tag sequences, Flag tag and 6-His tag, respectively suitable to detect scFv in immunochemical assays and to purify them by nickel chromatography.

Peripheral blood mononuclear cells (PBMC) were isolated by Ficoll-Hypaque (Cedarlane, CL5120) density gradient centrifugation from heparinised blood samples of healthy donors (see Ethics approval and consent to participate section). Natural Killer cells were isolated via a magnetic cell sorter by using NK cell isolation kit II (Miltenyi Biotec, 130-092-657) according to the manufacturer’s protocol. The purity of isolated CD3-CD56 $+$ NK cells was $>$ 95%.

NK-92 is an interleukin-2 (IL-2, CellGS, GFH12) dependent Natural Killer cell line, and was a kind gift of Professor L. Moretta, (Director of the Istituto Gaslini, Genova, Italy) who obtained the cells from ATCC repository (ATCC; CRL-2407). The following human melanoma cells have different source: Me20842, Me30966, MelC and Mel501 are metastatic melanoma cell lines, LM2, LM3, LM4, LM5, LM7, INT10, INT18, INT30 and ATT are primary and metastatic tumour cells derived from lesions of melanoma patients and adapted to standard cell culture conditions. All these cells were kindly supplied by Dr Licia Rivoltini (Fondazione IRCCS Istituto Nazionale dei Tumori, Milan, Italy). Lymphoblastoid cells which include CCRF-CEM, HL60, K562, THP1 and MOLT-4 cell lines all obtained from the ATCC. A panel of carcinoma cell lines (breast, colon, and lung carcinoma) MCF7, HT29, LoVo, H69, H1299, H460, H292, H1650 and Calu were all purchased from ATCC. Cells were grown under standard conditions for mammalian cells cultures. The basic medium (BM) for cell culturing consisted of RPMI-1640 (Aurogene, AU-L0501) supplemented with 10% heat-inactivated foetal bovine serum (FBS, Lonza, DE14-801F), L-glutamine (L-Glu, Euroclone, EC13000D)), penicillin (100 U/mL)/streptomycin (100 U/mL) (P/S, Euroclone, 8ECB3055D).

Human osteosarcoma and Ewing’s Sarcoma cell lines, U-2OS, Saos-2 and SK-ES-1 were obtained from the American Type Collection. The TC-71 and 6647 cell lines were kindly provided by T. J. Triche (Children’s Hospital, Los Angeles, CA). IOR/BRZ was obtained from the Laboratorio di Oncologia Sperimentale, Istituto Ortopedico Rizzoli, Bologna, Italy. These cell lines are cultured in Iscove’s modified Dulbecco’s medium (IMDM, EuroClone, ECM0192L), supplemented with L-Glu (2 mM), P/S (100 U/ml) and 10% FBS. All cells in this study were maintained at 37 ${}^{\circ}$ C in a humidified 5% CO2 atmosphere.

Human epithelial ovarian cancer cell lines, OVCAR3, Hey, A2780 WT, gently provided from Oncotech Tustin, CA, were cultured at 37 ${}^{\circ}$ C in an atmosphere of 5% CO2 in RPMI 1640 Medium (Sigma-Aldrich, R0883) and Skov3 in Dulbecco’s Modified Eagle’s Medium (DMEM, Sigma-Aldrich, D6546) with 10% fetal bovine serum (FBS, Euroclone, Milan, Italy), 50 IU/ml penicillin, 50 mg/ml streptomycin, 50 mg/ml gentamicin and 0,3 mg/ml glutamine.

NCAM antigen used for biopanning strategies was a purified recombinant protein corresponding to the extracellular domain of human NCAM (neural cell adhesion molecule 1 isoform 1 precursor [Homo sapiens] NCBI Reference Sequence: NP_000606.3, aa 20-718 of the whole sequence) with 6xHis-tag (NCAMecd), produced in E. Coli by BiologicsCorp protein manufacturer (USA) on request.

2.2 Biopanning strategy for selection NCAM extracellular domain recombinant protein specific scFvB1

Specific antibodies in scFv format were isolated from IORISS phage library as previously described [28]. In brief, an aliquot of phages containing about 10 ${}^{13}$ colony forming units (cfu), was adsorbed to the immunotubes (Nunc Maxisorp, T468-3) coated overnight (ON) at room temperature (RT) with 5 $\mu$ g/ml of recombinant purified NCAMecd in 4 ml phosphate buffered saline (PBS). After biopanning, phages were eluted with 1 ml of 100 mM triethylamine and the solution was immediately neutralised by adding 0.5 ml of 1 M Tris-HCl pH 7.4. Eluted phages were used to infect log phase TG1 E. coli bacteria [supE hsd15 thi 1(lacproAB) F’(traD36 proAB $+$ lacIqlacZ1M15)] and amplified for the next round of selection as described by Ascione et al. [29]. Briefly, 50 ml of 2 $\times$ YT medium (Sigma-Aldrich, Y2627) with 100 mg/ml ampicillin (Sigma-Aldrich, A9518) (2 $\times$ TYA) and glucose 2% were inoculated with enough bacterial suspension to yield an optical density at 600 nm (OD600 nm) of 0.05–0.1. The culture was grown up to OD600 nm of 0.4–0.5 and infected with M13 K07 helper phage at a ratio of around 20:1 phage/bacteria. The rescued phages were concentrated by precipitation with polyethylene glycol (PEG) 6000 and used for subsequent rounds of panning. Plating on agar of TG1 cells infected with a pool of phage antibodies from last cycle of selection allowed individual clones harbouring phagemid to grow. Several bacterial clones were tested for their ability to secrete functional anti-NCAM scFvs. For soluble scFv preparation, the individual colonies were grown in 96 flat bottomed wells (Corning, CLS3997-50EA) for 2 h at 37 ${}^{\circ}$ C in 180 ml 2 $\times$ YTA medium and glucose 0.1% in 96 well plates and induced with 50 ml 2 $\times$ TYA medium and 6 mM isopropyl $\beta$ -dithiogalactopyranoside (IPTG, Thermo Fischer Scientific, 15529019). The following day the plates were spun down at 1800 g for 10 min and the supernatants containing soluble scFvs were recovered and tested for specificity by ELISA.

2.3 DNA characterization and sequences

Plasmid DNA from scFvB1 bacterial clone was digested with specific endonucleases; complementarity determining regions (CDR) were sequenced with an automated DNA sequencer (Eurofins Italia/, Italy) using scFvFor (5’TTGTCACGAATTCGGCCATGGCC 3’) and scFvRev (5’TTTGGGCTCGCTAGCTGCGG CCGC 3’) primers.

2.4 Soluble ScFv purification

Soluble scFv purification was performed as described by Ascione et al. [29]. For production of soluble scFvs, specific phage-TG1 E. coli infected cells were cultured at 37 ${}^{\circ}$ C in 2 $\times$ TY containing 100 $\mu$ g/mL ampicillin and 0.1% glucose up to OD ${}_{\text{600nm}}$ 0.5. After induction of antibody expression by adding IPTG 1 mM to culture, cells were incubated ON at 30 ${}^{\circ}$ C. Then, the bacterial culture was centrifuged and scFv containing supernatant collected. ScFv antibodies were precipitated with ammonium sulfate and dialyzed in PBS. His-tagged scFv antibodies were purified by immobilised metal affinity chromatography using Ni2 $+$ -nitriloacetic acid agarose (Ni-NTA resin, Qiagen, 1018244). ScFv fragments were eluted with 250 mM imidazole in PBS, dialysed, quantified by UV absorbance at 280 nm, ELISA tested for specific antigen recognition, and stored at $-$ 80 ${}^{\circ}$ C. One and five microliters of purified scFv were analysed in SDS-PAGE.

2.5 NCAM extracellular fragments cloning and production

PBMC from a healthy donor were separated by Ficoll-Hypaque gradient centrifugation and then cultured in complete of RPMI 1640 medium with 20% FBS, 2 mM L-Glu, 100 $\mu$ g/ml P/S with 500 U/mL IL-2 (CellGS, GFH12).

After three and five days, the level of NCAM expression was evaluated by flow cytometry: an aliquote of cells was stained with the anti-CD56 PE (Becton-Dickinson, 555516) and anti-CD3 FITC (Becton-Dickinson, 349201) antibodies according the manufacturer’s instructions.

After five days, NK were isolated by magnetic separation, and mRNA was extracted using QuickPrep Micro mRNA Purification Kit (Analytik Jena, 845-KS-20300) according to the manufacturer’s instructions. mRNA (1 $\mu$ g) was reverse-transcribed for double-strand cDNA synthesis using a smart PCR Synthesis kit (BD Bioscience Clontech, 634902). NK cells derived cDNA was the template for the amplification of NCAM extracellular domain (NCAMecd) fragments coding sequence by PCR with the NCAM 20-307For, 5’-AT CTATGGATCCCTGCAGGTGGATATTGTT-3’, and NCAM 20-307Rev, 5’-ACGTTGAAGCTTTTTTGCA AAGACTTTGAGGTG-3’, for the sequence encoding for the amino acids from 20 to 307 of NCAMecd; NCAM 302-505For, 5’-ATCTATGGATCCCTCAAAG TCTTTGCAAAAC-3’ and NCAM 302-505Rev, 5’-ACGTTGAAGCTTAACAAGGATGAATTCCAAG-3’, for the sequence encoding for amino acids from 302 to 505; NCAM 492-699For, 5’-AGACTAGGATCCTTCA TCCTTGTTCAAGCAG-3’ and NCAM 492-699Rev, 5’-ACTTGAAAGCTTGATGGCTGTGGGCTG-3’ for the sequence encoding for amino acids from 492 to 699. All the sense primers contain the BamHI restriction site and the sequence coding for the first six amino acids of the respective NCAM fragment, while all the antisense primers contains the sequences encoding for the end part of the respective NCAM fragment and HindIII restriction site. The PCR fragments were cloned in pQE30Xa vector between BamHI and HindIII restriction sites. This expression vector inserts a 6 $\times$ His tag at the N-terminal of the NCAMecd fragment (NCAM aa/His) for purification. Primers were designed on the basis of the Neuronal adhesion molecule 1 mRNA nucleotidic sequence number NM_000615.6 (GI:336285433) in NCBI database. PCRs were performed using Pwo enzyme (Roche Diagnostics, 11644947001) and the resulting PCR fragments were agarose-purified using the High Pure PCR Product Purification Kit (Roche Diagnostics, 11 732 668 001).

TG1 E.coli (supE hsd $\Delta$ 5 thi $\Delta$ (lac-proAB) F’ [traD36 proAB $+$ lacIqlacZ $\Delta$ M15]) cells were electroporated with the recombinant plasmids pQE30Xa-NCAM(20-307)/His, pQE30Xa-NCAM(302-505)/His, pQE30Xa-NCAM(492-699)/His. Transformant clones were analysed by PCR on colony. DNAs from positive clones were sequenced by Biofab Research SRL (Italy, EU).

Clones that showed the best induction were used for purification of proteins. Bacteria were grown in 100 mL 2 $\times$ TY broth supplemented with 100 $\mu$ g/mL ampicillin and 0.1% glucose in a 37 ${}^{\circ}$ C shaker until OD600 nm $=$ 0.6. IPTG was added to a final concentration of 1.5 mM. Cells were harvested 3 hours (hrs) later, centrifuged at 10,800 g for 20 min at 4 ${}^{\circ}$ C and lysed by sonication in lysis buffer (50 mM NaH2PO4, 300 mM NaCl, 10 mM imidazole, pH 8). Proteins that showed the expected bio-molecular specificities, were purified by affinity chromatography on Ni-NTA resin, using native protocol according to the manufacturer’s instructions. Purified proteins were eluted with 250 mM imidazole elution buffer, dialyzed in PBS, aliquoted and stored at $-$ 80 ${}^{\circ}$ C. Purified proteins were analysed by SDS-PAGE.

2.6 Elisa on recombinant NCAM proteins

96-well ELISA plates (ThermoFischer Scientific, 44-2404) were coated ON with 50 $\mu$ L/well of 10 $\mu$ g/mL NCAMecd or NCAMecd fragments in PBS at 4 ${}^{\circ}$ C. Next day a blocking solution consisting of 2% non-fat milk in PBS (MPBS) was added and after 2 hrs the plates were washed with PBS containing 0.1% Tween 20 (TPBS). Plates were incubated for 2 hrs at room temperature (RT) with 50 $\mu$ L of supernatants containing soluble scFv antibodies or 10 ug/ml of purified scFvs, anti-Flag M2 antibody (Sigma, F3165) and HRP-conjugated anti-mouse antibody (Dako, P0447). All antibodies were resuspended in 2% MPBS.

The reaction was visualized with 3,3’-5,5’- tetramethylbenzidine (BM blue, POD substrate, Roche, 10203700103) and stopped by adding 50 $\mu$ L of 1 M hydrogen sulfide. The reaction was detected with ELISA reader (Model 680 microplate reader, Biorad, CA, USA). Results were expressed as absorbance A $=$ A (450 nm) $-$ A (620 nm).

For the ELISA conducted in heat-denatured conditions, the plate was coated in parallel with 0.5 $\mu$ g/well of intact recombinant NCAMecd protein and NCAMecd protein boiled for 5 min (or glucose oxidase (GO) antigen, Sigma-Aldrich, G2133-10KU) in the same conditions. Then the procedures were the same described above.

2.7 Determination of binding affinity

The determination of antibody affinity was performed by equilibrium saturation analysis using competition ELISA assay according to Rath et al. [30], with minor modifications as described by Gellini et al. [31].

A first ELISA step was performed to determinate the antibody concentration exhibiting 50% binding (EC50). Microtiter plates were coated with 50 $\mu$ L of the antigen (10 $\mu$ g/mL) in PBS buffer and incubated ON at 4 ${}^{\circ}$ C. Microtiter plates were washed three times with TPBS, and blocked with MPBS for 2 hrs RT. Microtiter plates were then washed three times with TPBS and incubated for 2 hrs at RT with 50 $\mu$ L of serial diluitions of scFv antibodies in MPBS, together with 10 $\mu$ L of a freshly prepared mixture consisting of an anti-FLAG M2, (Sigma, F3165) (12.5 $\mu$ g/mL) and an antimouse HRP-conjugated antibody, (Dako, P0447) (22 $\mu$ g/mL). After the incubation period, the plates were washed three times with TPBS and three times with PBS. The colour development system was added and the colour intensities were measured as above indicated. The EC50 was calculated based on obtained absorbance values.

A second ELISA step was performed to determinate the molar concentration of the free antibody at the equilibrium. Serial dilutions of the target antigen in solution were incubated in microtiter plates with the specific scFv at the determined EC50 (calculated from the first ELISA) and incubated for 1 hr to reach the equilibrium.

After the blocking step, the solution with the antigen at the varying concentrations and with the respective antibody at the determined EC50 (calculated from the direct ELISA) and allowed to incubate for about 60 min, was added to the antigen-coated wells together with a freshly prepared mixture consisting of an anti-FLAG M2, (Sigma, F3165) (12.5 $\mu$ g/mL) and an anti-mouse HRPconjugated antibody (Dako, P0447) (22 $\mu$ g/mL) which will reveal free scFv antibodies at the equilibrium. The microtiter plates were rinsed thoroughly, the substrate solution applied, and then the enzyme reaction developed and measured as already described. The molar concentration of solution antigen required to obtain 50% inhibition (IC50) represents an estimation of average antibody affinity.

2.8 Epitope mapping

Epitope mapping procedure was performed by Takis srl (Castel Romano, Italy).

A human NCAM peptide library was designed based on the sequence of NCAMecd 492-699 region and a total of 40 peptides were purchased from Genscript. The library was composed of peptides of 15 aa in length with 10 overlapping residues covering aa 492 to 699 of human NCAM1 precursor protein (NP_000606). Peptides were purchased in lyophilised form to avoid degradation.

Peptides were resuspended in 100% DMSO at 1 mg/ml concentration. Maxisorp 96 well ELISA plates were coated with the peptides diluted at 10 ug/ml concentration in 10 mM Carbonate buffer pH:9.6, in 100 ul/well volume, and incubated at 4 ${}^{\circ}$ C. After ON incubation, plates were washed three times in 1 $\times$ PBS, 0.05% Tween and blocked with 200 ul of 5% BSA in TPBS. ScFv, diluted in 1% BSA/TPBS at 10 ug/ml, was added to the plate and incubated ON, at $+$ 4 ${}^{\circ}$ C. Following incubation, the plates were washed 3 times and secondary Ab was added (anti-6His or anti-Flag diluted 1:1000) (Anti-Flag AP conjugated: Sigma, A9469, anti-6His HRP conjugated: Bethyl, A190-113P) and incubated 1 h at RT. Finally, plates were washed 3 times with PBST and 3 times with PBS and substrate was added. Depending on the secondary conjugated Ab used, absorbance was measured at 405 nm for AP conjugate antibody or at 450 nm for HRP conjugated antibody.

2.9 SDS-PAGE and western blotting

For analysis of purified scFvB1 one and five microliters were diluted in reducing SDS-PAGE loading buffer (NuPAGE ${}^{\text{TM}}$ LDS Sample Buffer (4X) Invitrogen ${}^{\text{TM}}$ ) containing 10 mM DTT and heated at 90 ${}^{\circ}$ C for 10 minutes using a Termomixer block for protein denaturation. Samples were then separeted by high-performance precast polyacrylamide gels (4–12%) (NuPAGE Bis-Tris Gels, Invitrogen ${}^{\text{TM}}$ ), and visualized by colloidal Coomassie staining (Simply blue safe staining, Invitrogen ${}^{\text{TM}}$ ) according to manifacture product.

Purified NCAM fragments eluitions and corresponding bacterial lysates were fractionated by SDS PAGE. At the time of the choice of the best bacterial clone for protein production, bacterial lysates from not induced and 3 hours induced bacteria from different clones were fractionated by SDS/PAGE and transferred on nitrocellulose membrane. The membrane was blocked in 2% non-fat dry milk/PBS. After washing, the membrane was incubated 1 hr at RT with peroxidase conjugated anti-HIS antibody (Sigma). The reaction was detected with the DAB (3,3-Diaminobenzidine) reagent (Sigma).

For analysis of scFvB1 binding to recombinant NCAMecd protein in western blot assay, 0.5 $\mu$ g of both recombinant NCAMecd and, as control, GO, were loaded in multiple wells and run on 12% SDS-PAGE and transferred to a nitrocellulose membrane using standard procedures. Filter strips were blocked in 2% MPBS ON at RT. The membranes were washed and incubated for 2 hours with either anti-6 $\times$ His antibody AD1.1.10 (BioRad, MCA1396GA ) 1 $\mu$ g/mL diluted in 2%MPBS (taking advantage of the presence of 6 $\times$ His-tag at the C-terminus of the recombinant NCAMecd protein), or 10 $\mu$ g/ml of purified scFvB1 or scFvGO (i.e., a human single-chain Ab against the yeast GO enzyme) [29] diluited in 2% MPBS, washed with 0.05% TPBS and incubated again with 5 $\mu$ g/ml anti-Flag M2 mouse antibody (Sigma, F3165) in 2% MPBS (only for purified scFvs). After an additional incubation for 1 hour at RT in presence of 5 $\mu$ g/ml of a HRP-conjugated goat anti-mouse antibody (Dako, P0447) both for anti-His and scFvs primary antibodies, the reaction was developed using DAB (3,3’-diaminobenzidine) Colour Development Solution (Sigma-Aldrich, D5905).

2.10 NCAM detection on cell surface

The ability of scFvB1 to bind NCAM on the surface of cells was evaluated by flow cytometry investigations on intact/living PBMC and NK cells and a consistent panel of human cell lines which include CCRF-CEM and Molt4 (T lymphoblast), K-562 (chronic myelogenous leukemia), THP-1 (monocytic leukemia), HL-60 (promyelocytic leukemia), NK92 (Natural Killer cell line), H69 (small cell lung carcinoma), Calu (lung adenocarcinoma), H292 (lung epithelial), H1299 (nonsmall cell lung carcinoma), H460 (large cell lung cancer), H1650 (bronchoalveolar carcinoma), MCF7 (breast adenocarcinoma), HT29 and LoVo (colorectal carcinoma), U-2OS and Saos-2 (osteosarcoma), TC-71, 6647, SK-ES-1, IOR/BRZ (Ewing’s Sarcoma), Me20842, Me30966, MelC and Mel50 (metastatic melanoma cell lines), LM2, LM3, LM4, LM5, LM7, INT10, INT18, INT30 and ATT (are primary and metastatic tumour cells derived from lesions of melanoma patients), OVCAR3, SKOV 3, Hey, A2780 WT (ovarian cancer cell lines), Hela (cervix epitheloid carcinoma cell line).

Cells in exponential phase of growth were collected, carefully washed in PBS and pelleted.

For screening of scFv clones that were reactive on cells, about 2.5 $\times$ 10 ${}^{5}$ NK92 cells were resuspended in 200 $\mu$ l of bacterial supernatant, while for characterisation of binding profiles of scFvB1, cells of different cell lines were resuspended in 200 $\mu$ l of PBS containing 25 $\mu$ g/mL of purified scFv antibody and incubated for 1 h at RT. In both cases cells were incubated for 1 h with scFv primary antibody. After several washings, cells were resuspended for 30 min at 4 ${}^{\circ}$ C in PBS containing 10 ug/ml anti-6 $\times$ His antibody AD1.1.10 (BioRad, MCA1396GA) and after washings, cells were incubated with 6 $\mu$ g/mL of an FITC-goat anti-mouse IgG (ThermoFischer scientific, 31569) for 30 min at 4 ${}^{\circ}$ C. The irrelevant antibody scFvGO [32] was used as negative control. As positive control, cells were marked with commercial anti-CD56 PE conjugated for 30 min at 4 ${}^{\circ}$ C. After staining, the cell samples were washed, maintained at 4 ${}^{\circ}$ C and immediately analysed by FACScan (Becton-Dickinson, NJ, USA) equipped with 15 mW, 488-nm, argon laser.

NK cells were purified from PBMC from healthy donors and activated (where indicated) with 100 ng/mL IL-2 (CellGS, GFH12) for three days. Cells were then washed and marked with LIVE/DEAD ${}^{\text{TM}}$ Fixable Aqua Dead Cell Stain reagent (Invitrogen, L34957) in order to exclude dead cells during FACS analysis. Then NK cells were washed and marked with antibodies as described for cell lines.

For PBMC, 1 $\times$ 10 ${}^{6}$ cells/sample were collected, washed with PBS and marked with LIVE/DEAD ${}^{\text{TM}}$ Fixable Aqua Dead Cell Stain reagent (Invitrogen, L34957) in order to exclude dead cells during FACS analysis. Then they were washed and marked as described above for cell lines. As control, PBMC were marked with CD3-FITC (Becton-Dickinson, 349201), CD16-PE (Beckton-Dickinson, 561313) and CD56-Alexafluor647 (Beckton-Dickinson, 563443). Samples were acquired using Gallios cytometer and analysed with KALUZA software (Beckman Coulter, CA, US).

2.11 Competition assay with FGFR1-Fc

OVCAR3 or A2780 cells in exponential phase of growth were collected, carefully washed in PBS and pelleted. About 2.5 $\times$ 105 cells were resuspended in 100 ul PBS containing scFvB1 (50 ug/ml) or scFvGO (50 ug/ml) and incubated for 30 min at RT. Then, 100 ul of PBS (200 ul in the sample without scFvs) containing 10 $\mu$ g/mL of Recombinant Human FGFR1 alpha (IIIc) Fc Chimera (FGFR1-Fc, R&D system, 661-FR) and incubated for 1 h at RT. After washings, cells were resuspended for 30 min at 4 ${}^{\circ}$ C in PBS and incubated with FITC-conjugated goat anti-human IgG (Invitrogen, 31529) for 30 min at 4 ${}^{\circ}$ C. After staining, the cell samples were washed, maintained at 4 ${}^{\circ}$ C and immediately analyzed by FACScan (Becton-Dickinson, NJ, USA).

2.12 Surface Plasmon Resonance assay

Surface Plasmon Resonance (SPR) experiments were carried out using a SensiQ Pioneer system [33]. The sensor chip (COOH5) was activated chemically by a 35 $\mu$ l injection of a 1:1 mixture of N-ethyl-N’-(3-(diethylaminopropyl) carbodiimide (200 mM) and N-hydroxysuccinimide (50 mM) at a flow rate of 5 $\mu$ l/min. The ligand, i.e. FGFR1-Fc, was immobilised on activated sensor chips via amine coupling. The immobilisations were carried out in 20 mM sodium acetate at pH 4.0; the remaining unreacted groups were blocked by injecting 1 M ethanolamine hydrochloride (35 $\mu$ l). The immobilisation level of FGFR1-Fc was 110 RU.

The analytes (NCAMecd and scFvB1) were dissolved in PBS buffer $+$ 0.005% surfactant P20 (PBSP) to a concentration of 20 $\mu$ M, automatically further diluted in buffer PBSP (running buffer) and were injected on the sensor chip at the following concentrations: 625 nM, 1.25 $\mu$ M, 2.5 $\mu$ M, 5 $\mu$ M, 10 $\mu$ M, 20 $\mu$ M (NCAMecd) and 39 nM, 78 nM, 156 nM, 312 nM, 625 nM, 1.25 $\mu$ M, 2.5 $\mu$ M, 5 $\mu$ M, 10 $\mu$ M (scFvB1) at a constant flow (30 $\mu$ l/min). The increase in RU relative to baseline indicates complex formation; the plateau region represents the steady-state phase of the interaction (RUeq), whereas the decrease in RU after 240 s represents dissociation of analytes from immobilised ligand after injection of buffer. Competition experiments were carried out in the same conditions, by injecting 5 $\mu$ M NCAMecd alone or premixed with 1 $\mu$ M, 2 $\mu$ M, 5 $\mu$ M, 10 $\mu$ M scFvB1. Regeneration procedures are based on two long (2000 s and 500 s) injections of buffer, separated by a brief (5 s) injection of 10 mM NaOH. Sensorgrams analysis was carried out using Qdat (SensiQ, ForteBio), and performing Scatchard analysis.

2.13 Modeling and docking

Modeling of scFvB1 three-dimensional structure was carried out using the server I-TASSER [34, 35, 36]. I-TASSER modelling starts from the structure templates identified by LOMETS from the PDB library. LOMETS is a meta-server threading approach containing multiple threading programs, where each threading program can generate tens of thousands of template alignments. I-TASSER only uses the templates of the highest significance in the threading alignments, the significance of which are measured by the Z-score, i.e. the difference between the raw and average scores in the unit of standard deviation; among the top-threading templates, I-TASSER used 5YD5A, 5WN9A, 5WN9H, 2GKI, 6UL6 and 4BUHA. For each target, I-TASSER simulations generate a large ensemble of structural conformations, called decoys. To select the final models, I-TASSER uses the SPICKER program to cluster all the decoys based on the pair-wise structure similarity, and reports up to five models, which corresponds to the five largest structure clusters. The confidence of each model is quantitatively measured by C-score that is calculated based on the significance of threading template alignments and the convergence parameters of the structure assembly simulations. C-score is typically in the range of [ $-$ 5, 2], where a C-score of a higher value signifies a model with a higher confidence and vice-versa. TM-score and RMSD are estimated based on C-score and protein length following the correlation observed between these qualities.

The best scFvB1 model was docked to the structure of the two FN3 domains of NCAM (Protein Data Bank 2VKW, chain A) using ClusPro [37] and HADDOCK [38], yielding similar top clusters of docked models.

2.14 Wound healing assay

Cells (3 $\times$ 105 cells/well) were seeded in a 24-well plate, and the tip of a 200- $\mu$ l micropipette was used to make a straight scratch on a confluent monolayer of cells to create a wound. The detached cells were rinsed with PBS twice, and 1 ml RPMI plus 10% FBS, with scFvB1 and scFvG0 (40 $\mu$ g/ml) were then added. The wound closure in the area was imaged in brightfield using a Juli Smart microscope (Digital Bio). Microphotographs were taken after incubation at 0 and 48 h, with a 4 $\times$ magnification. The wound closure areas were selected randomly and the width of the wound at each time point was quantified in ImageJ (v1.8, National Institutes of Health). The results of three independent experiments were imaged and the percentage average and standard errors have been calculated. Subsequently $T$ -test student for statistical analysis has been performed.

2.15 Statistical analysis

For all ELISA experiments the cut-off value separating positive from negative samples was calculated as 3 standard deviation values (SD) above the mean, of the values obtained using the negative controls.

For Wound healing assay and flow cytometry competition assay data are expressed as means $\pm$ standard error of the mean (SEM). The student’s $t$ test was performed to determine significance. A $p$ value of $<$ 0.05 was considered statistically significant.

Figure 1.

Isolation of scFvB1 antibody. A. Schematic representation of panning cycle for the isolation of anti-NCAM scFvs from IORISS phage library. B. Screening procedures to identify monoclonal antibodies positive for NCAM antigen and able to recognise the antigen on cells. On the top, anti-NCAM ELISA carried out with supernatants of 90 bacterial colonies isolated from a polyclonal population of phage from three rounds of selection. The plate was coated with 0.5 $\mu$ g of antigen per microwell. The cut-off (O.D. $=$ 0.12) value (red line) separating positive from negative samples was calculated as 3 standard deviation values (SD) above the mean of the values obtained using the irrelevant anti-GO scFv (OD $=$ 0.046). O.D. at nm 450 ref 620: optical density. The red arrow indicates the B1 clone. Down, flow cytometric analysis on NK92 cells of bacterial supernatants from a non-reactive clone (blue), polyclonal population after three round of selections (pink) and scFvB1 clone (green). In black the control of anti-His antibody used as secondary antibody.

Figure 2.

scFvB1 nucleotidic and amino acid sequence. In bold CDR1, CDR2 and CDR3 of VH (light blue) and VL fragments (light yellow). The sequence of the linker between the VH and VL fragments is in red.

Figure 3.

scFvB1 binding properties. A. Affinity binding (expressed as binding dissociation constant) calculated by a two steps ELISA, as described in methods. A first ELISA step (on the top) was performed to determinate the antibody concentration exhibiting 50% binding (EC50). Wells were coated with 0.5 ug of NCAMecd protein and then serial diluitions of scFvB1 (X axis, ng/ml) were tested. A second ELISA (down) was performed to determinate the molar concentration of the free antibody at the equilibrium. Wells were coated with 0.5 ug of NCAMecd protein and then serial dilutions of the same antigen (X axis, ng/mL) were incubated in solution with the specific scFv at the determined EC50 (calculated from the first ELISA). EC50 value from this second ELISA was used for the dissociation constant. B. On the top, schematic representation of the three cloned NCAM extracellular fragments (aa 20-307, aa 302-505, aa 492-699) and, down, results from ELISA of scFvB1 on the two available purified fragments (aa 20-307, aa 492-699). The plate was coated with 0.5 $\mu$ g of antigens per microwell. In orange, the anti-His antibody recognizing the His-tag cloned at N-terminal of the three fragments and in blue, the scFB1 reactivity on the three fragments. scFvB1 was detected with an anti-FLAG secondary antibody and then with anti-mouse-HRP conjugated antibody. O.D. nm 450 ref 620: optical density. Represented results were the average of three independent experiments; Bars represent standard deviation.

Figure 4.

scFvB1 epitope analysis. A. WB analysis of ScFvB1 binding to NCAMecd antigen or irrelevant GO antigen in denatured conditions. NCAM antigen was revealed with anti-His antibody (lane 1) and scFvB1 (lane 4). GO antigen was only recognised by its specific scFvGO (lane 7). scFvB1 and scFvGO were detected with an anti-FLAG secondary antibody and then with anti-mouse-HRP conjugated antibody. Kilodaltons (kDa) of molecular markers (lane 3 and 8) are reported on the right. Nitrocellulose membrane was cut at level of the center of the two markers lane in order to perform different antibodies blotting and the reconstituted in the original manner for the figure. B. OD results from ELISA performed on 15 aa peptides (10 aa overlapping) mapping the NCAMecd 492-699 protein region coated at 10 ug/ml (NCAMecd 492-699 protein was coated as positive control). ScfvB1 was used at 10 ug/ml. scFvB1 was detected with anti-Flag AP conjugated (orange) or anti-6His HRP conjugated (blue) as secondary antibodies. w/o peptide: no coating. Control $+$ : irrelevant protein with 6His or Flag. C. OD results from ELISA performed on NCAMecd protein (or irrelevant GO protein) in native and denatured (5 min boiling) conditions with scFvB1 or anti-His HRP conjugated antibodies. scFvB1 was detected with an anti-FLAG secondary antibody and then with anti-mouse-HRP conjugated antibody. O.D. nm 450 ref 620: optical density. Represented results were the average of three independent experiments; Bars represent standard deviation. Star represents statistically significant difference, $p$ value of $<$ 0.05.

Figure 5.

Flow cytometry reactivity. A. Dose-response binding of purified scFvB1 on NK92 cells. As control, NK92 cells were labeled with the secondary Abs alone (black). Green 0.25 ug/ml scFvB1, pink 2.5 ug/ml, light blue 25 ug/ml, blue 250 ug/ml, orange 1250 ug/ml. B. Flow cytometric analysis of purified scFvB1 (25 ug/mL) on purified native and IL-2 activated NK cells from healthy donor (right) in comparison with commercial anti-CD56 PE antibody (left). As control, cells were labeled with the secondary Abs alone or isotype control. C. Flow cytometric analysis of purified scFvB1 on PBMC isolated from healthy donor (left). The same sample was marked with commercial CD56 AlexaFluor647 and CD16PE in order to individuate the CD56 $+$ NK cells percentage in this PBMC sample as control of scFvB1 reactivity (right).

3. Results

3.1 Isolation of human scFvB1 single chain antibody specific for recombinant and cell surface-exposed human NCAM by phage display technology

Human recombinant antibodies against NCAM were obtained by an in vitro affinity enrichment process commonly known as biopanning; in essence, an aliquot of phage-displayed scFvs representing the complexity of the naive IORISS library [28], has been subjected to cycles of absorption and elution to recombinant NCAMecd protein – used as a bait immobilised on immunotubes (Fig. 1A).

After three rounds of selections, we were able to isolate a phage-antibody population specifically recognizing NCAMecd. Figure 1B (top) reports results of monoclonal ELISA screening of soluble scFvs from single colonies obtained after plating bacteria TG1 infected with the enriched phages. Several of them proved to be specific for NCAMecd protein, while no signal was detected against recombinant Glucose Oxidase (GO) protein used as irrelevant antigen (data not shown).

Then, we analysed positive supernatants containing soluble scFvs for their reactivity on physiologically expressed NCAM, by flow cytometry on NK92 cells. Only one of tested scFvs was able to recognise NCAM on cell surface (Fig. 1B), therefore, hereafter, we will refer to this single NCAM-specific scFv, as scFvB1.

The DNA encoding for scFvB1 was sequenced; Fig. 2 shows scFvB1 DNA and amino acid sequence.

ScFvB1 was produced in bacteria and purified by metal chromatography using the six-histidine tag positioned at the C-terminal end of the antibody. After purification, scFvB1 was checked by SDS-PAGE (Supplementary Fig. 1) and the protein presents the expected molecular weight of about 27–30 KDa. An additional higher band at about 70 KDa is present, probably corresponding to scFv aggregates.

3.2 Evaluation of scFvB1 binding affinity and recognized epitope

First, the purified scFv was tested against the recombinant antigen used for selection panning in ELISA, in order to determine its binding affinity, that was calculated as Kd: 3.43 $\times$ 10 ${}^{-8}$ M (Fig. 3A). Then, in order to evaluate the region of the NCAM extracellular domain that is recognized by our scFvB1, we cloned the sequence of three different portions of the NCAM extracellular domain: one includes the first 3 Ig-like domains (NCAM aa 20-307), the second includes the last 2 Ig-like domains (NCAM aa 302-505) and the last includes the 2 fibronectin-like domains (NCAM aa 492-699).

Unfortunately, we were able to obtain producing bacterial clones only for the first NCAM aa 20-307 and the third NCAM aa 492-699 fragments. These two NCAM fragments were produced and purified (Supplementary Fig. 2).

ELISA tests on these two recombinant regions showed that scFvB1 binds the region 492-699, that is close to the cellular surface (Fig. 3B) and involved in the interaction with the second and third Ig-like domain of FGFR1.

Since scFvB1 was able to recognize NCAMecd antigen also in denatured conditions in western blot analysis (Fig. 4A), we supposed that it binds to a linear epitope.

For this reason, we performed an epitope mapping with linear peptides of 15 amino acids with 10 amino acids overlapping regions covering the NCAMecd 492-699 region, recognised by the scFvB1.

Surprisingly, no signal was detected on any peptides in ELISA assay performed with 10 ug/ml of scFvB1 (Fig. 4B).

So we decided to conduct an ELISA on the heat-denatured NCAM protein (by boiling). In this case, the OD signal with scFvB1 was significantly lower but not abolished in respect to native NCAM protein, whereas the signal with anti-HIS antibody was almost the same on the two antigen (Fig. 4C).

Based on these results, we can state that the epitope recognized by scFvB1 is not completely linear, it being at least partially dependent on structure that SDS Page and boiling do not completely disassemble.

3.3 Evaluation of scFvB1 binding on tumor cells expressing NCAM antigen

First of all, a pilot flow cytometric analysis on NK92 cells with different concentrations of purified scFvB1 was performed to choose the optimal concentration for binding studies on different tumour cell lines. This experiment confirmed the ability of scFvB1 to bind naturally expressed NCAM on the surface of cells expressing the antigen and let us to identify the most balanced concentration of scFvB1 (25 ug/ml) for the successive wider cytofluorimetric studies (Fig. 5A).

scFvB1 was also tested on NK cells purified from a healthy donor before and after IL2-driven activation in order to promote a higher expression of NCAM, thus demonstrating the ability of scFvB1 to effectively detect target antigen level variations on cells surface (Fig. 5B). scFvB1 was also able to bind NCAM expressing cells in a more variegated population, that is PBMC purified from a healthy donor (Fig. 5C); in this case the percentage of NK population we found in PBMC was similar to the percentage detected with commercial fluorochrome-conjugated antibodies, thus confirming the good affinity and specificity of scFvB1.

Then, scFvB1 was tested on a large panel of different tumour cell lines available in our cellular bank. As shown in Table 1, scFvB1 was able to recognize NCAM as the commercial antibody control or better as for osteosarcoma cell lines and colon carcinoma. In particular, scFvB1 recognized 5/13 melanoma cell lines, 2/2 osteosarcoma cell lines, 3/4 Ewing sarcoma cell lines, 2/6 lung carcinoma cell lines (in particular small cell lung carcinoma), 1/2 colon carcinoma. Noteworthy, scFvB1 was positive on 1/5 EOC cell lines tested, OVCAR3 cells, according to literature data which report a percentage of NCAM-positive EOC cells of about 23% [14]. Supplementary Fig. 3 shows scFvB1 binding profiles to NCAM positive cell lines.

Table 1
Flow cytometry determination of the binding level of scFvB1 on different tumor cell lines

Tumor type	Cell lines	Mab 56	ScFv B1	ScFv GO
Melanoma	LM2	$-$	$-$	$-$
	LM3	$-$	$-$	$-$
	LM4	$++$	$++$	$-$
	LM5	$-$	$-$	$-$
	LM7	$-$	$-$	$-$
	INT10	$-$	$-$	$-$
	INT18	$-$	$-$	$-$
	INT30	$+++$	$+++$	$-$
	Me20842	$-$	$-$	$-$
	MelC	$+++$	$+++$	$-$
	Me501	$++$	$++$	$-$
	Me30966	$+/-$	$+/-$	$-$
	ATT	$-$	$-$	$-$
Epithelial ovarian	A2780	$-$	$-$	$-$
cancer	OVCAR3	$+++$	$+++$	$-$
	Hey	$-$	$-$	$-$
	Skov	$-$	$-$	$-$
	HeLa	$-$	$-$	$-$
Osteosarcoma	Saos-2	$+$	$+++$	$-$
	U2os	$+$	$++$	$-$
Ewing sarcoma	TC-71	$++$	$++$	$-$
	6647	$-$	$-$	$-$
	SK-ES1	$++$	$++$	$-$
	IOR/BRZ	$+++$	$+++$	$-$
Lung carcinoma	Calu	$-$	$-$	$-$
	H1650	$-$	$-$	$-$
	H292	$-$	$-$	$-$
	H69	$+++$	$+++$	$-$
	H1299	$-$	$-$	$-$
	H460	$+$	$+$	$-$
Breast cancer	MCF7	$-$	$-$	$-$
Colon cancer	Lovo	$-$	$+/-$	$-$
	HT29	$-$	$-$	$-$
Multiple myeloma	HL60	$-$	$-$	$-$
	K562	$-$	$-$	$-$
	THP1	$-$	$-$	$-$
ALL	Molt 4	$-$	$-$	$-$
AML	CEM	$-$	$-$	$-$
NK	NK92	$+++$	$+++$	$-$

A commercial anti-CD56PE was used as positive control, while an scFv against glucose oxidase antigen (GO) was used as negative control. Stained cells: $-$ , (0–5%); $+/-$ , (1–10%); $+$ , (11–30%); $++$ , (31–60%); $+++$ , (61–100%). ScFv B1 strongly reacts with NK92, small cell lung carcinoma (H69), melanoma metastatic cells, osteosarcoma and Ewing’s sarcoma cells, epithelial ovarian cancer (OVCAR3) and to a lesser extent, with colon carcinomas (HT29, LoVo), primary melanomas, in contrast no binding with breast carcinoma (MCF7), and lymphoblastoid cells.

3.4 scFvB1 interferes with NCAM-FGFR1 interaction

Since scFvB1 was able to recognize the region of NCAM involved in the interaction with FGFR1, we tried to verify whether it was also able to interfere with this binding.

Since NCAM-FGFR1 interaction, intracellular signaling and subsequent effects on migration are well documented for epithelial ovarian cancer (EOC), we focus our experimental evaluation on the NCAM-positive OVCAR3 cell line.

First, we tested the ability of scFvB1 to compete with exogenous FGFR1-Fc recombinant protein for the binding to OVCAR3 cells. As reported in Fig. 6 and Supplementary Fig. 4, the incubation of OVCAR3 cells with scFvB1 reduces the binding of FGFR1-Fc to the cell’s surface, thus suggesting a competition between scFvB1 specific for NCAM1, and FGFR1-Fc. The competition effect seems to be antigen-specific since we didn’t observe the same effect with the irrelevant antibody scFvGO [32]; on the contrary, we observed the opposite effect, such as an increase of FGFR1-Fc signal. This is a counterintuitive effect we observed in other occasions (data not shown), probably caused by a binding stabilization in turn due by interactions between the proteins present in the sample. Moreover, On the basis of these results, we evaluated the ability of scFvB1 to interfere between NCAM and FGFR1 by studying the protein-protein interaction by surface plasmon resonance experiments. First we confirmed literature data of about 10 $\mu$ M binding affinity between NCAM and FGFR1 (Supplementary Fig. 5A and B). Then we observed a dose-response decrease of NCAM1 recombinant extracellular protein binding to immobilised FGFR-Fc protein in presence of scFvB1 (Fig. 7), thus confirming data obtained on cells with flow cytometry competition assay. No interaction was observed between scFvB1 and immobilised FGFR1-Fc (Supplementary Fig. 6).

Figure 6.

FGFR1-Fc competition assay by flow cytometry on OVCAR3 cells. OVCAR3 cells were marked with FGFR1-FC protein (10 ug/ml, 0.125 uM) and anti-human IgG FITC as secondary antibody, alone or in presence of scFvB1 or scFvGO (50 ug/ml, 1.66 uM). Represented histograms were the average of four independent experiments; Bars represent standard error. Star represents statistically significant difference, $p$ value of $<$ 0.05.

Figure 7.

SPR competition experiment. Interaction between NCAMecd and immobilised FGFR1-Fc in the presence of scFvB1. Sensorgrams measured upon injection of 5 $\mu$ M NCAMecd, alone and in the presence of 1 $\mu$ M, 2 $\mu$ M, 5 $\mu$ M, 10 $\mu$ M scFvB1. The increase in RU relative to baseline indicates complex formation; the plateau region represents the steady-state phase of the interaction (RUeq), whereas the decrease in RU after 240 s represents dissociation of analytes from immobilised ligand after injection of buffer. The presence of scFvB1 decreases binding of NCAMecd to FGFR1.

Figure 8.

Docking analysis. A. Docking of NCAM-FN3 structure (blue) and scFv model (green), performed with HADDOCK 2.0. B. Close-up of the interface of the NCAM-FN3 – scFv docking, with the residues involved in the interaction between the proteins: residues Q507, E509, Y511 and Q516 belonging to the FRM region of 1FN3, and Y672, Q679, K681and K683, belonging to the FGL region of 2FN3, are among the residues establishing interactions with scFv. Figure designed with Pymol.

Figure 9.

Wound healing assay. A. Graphic representation of percentage of scratch area repopulation at 48 hrs of OVCAR3 cells treated with scFvB1 and scFvG0 (40 $\mu$ g/ml). No treated cells were used as negative control. Represented results were the average of three independent experiments; Bars represent standard error. Star represents statistically significant difference, $p$ value of $<$ 0.05. B. Microphotographs of wound closure at T0 and 48 h, with a 4x magnification. The wound closure areas were selected randomly and the width of the wound at each time point was quantified in ImageJ (v1.8, National Institutes of Health).

3.5 Modeling and docking

In order to further investigate the interesting ability of scFvB1 to interfere with NCAM-FGFR1 binding, we performed a docking computational study. The top model (model 1) of scFvB1 three-dimensional structure predicted by I-TASSER had a C-score of 0.94, estimated TM-score of 0.84 $\pm$ 0.08 and RMSD of 3.9 $\pm$ 2.7 Å, and was structurally similar (TM-score $>$ 0.85) to 5YD5A, 5LX9H, 5NPIA, 4BUHA, 5WN9H, 6EHWB, 3UMTA and 2GHWB. Predicted binding regions included residues 34–36, 55–62, 102–105.

The best scFvB1 model was docked to the structure of the two FN3 domains of NCAM; the best cluster of structures (cluster 7), has a HADDOCK score of $-$ 85.4 $\pm$ 6.3, with Van der Waals energy $-$ 83.8 $\pm$ 5.1, Electrostatic energy $-$ 277.9 $\pm$ 17.8, Desolvation energy $-$ 4.3 $\pm$ 0.7, Restraints violation energy 582.3 $\pm$ 81.1, Buried Surface Area, 2211.6 $\pm$ 88.1, Z-Score $-$ 1.8 (Supplementary Fig. 7). The model of the interaction between scFvB1 and the structure of the two FN3 domains of NCAM is shown in Fig. 8. The interface displays a series of intermolecular contacts (distance $<$ 5.0 Å), involving S34, W36, Y55, D60, R62, Y104, R106, H196 and D236 of scFvB1 (all belonging to CDRs of VH chain except for D236 residue belonging to VL CDR3), and Q507, E509, Y511, Y511, Q516, R599, E600, Y672, K681 and K683 of NCAM FN3 domains (Fig. 8B). In particular, residues Q507, E509, Y511, Y511 and Q516 belong to the FRM sequence of FN3 module 1, and Y672, K681 and K683 are part of the FGL sequence of FN3 module 2. The FRM and FGL sequences, responsible for the interaction with FGFR1 (22), therefore appear to establish contact with residues of scFvB1: this may represent the structural basis of the competition between scFvB1 and FGFR1 for NCAM binding.

3.6 scFvB1 interferes with migration of EOC cells in vitro

Finally, we decided to evaluate whether the molecular interference in the interaction between NCAM and FGFR1 mediated by scFvB1 translates in an inhibition of EOC cell migration. OVCAR3 cells were scratch wounded and treated with scFvB1 or scFvGO as described in materials and methods. First significant results were obtained after 48 hrs; as reported in Fig. 9, repopulation of the scratched area was significantly lower when cells were treated with scFvB1 compared to the other controls, suggesting a partial inhibition of cells migration process. Experiments were conducted in triplicate; Fig. 9 reports images from a representative experiment.

4. Discussion

NCAM has been shown to have a high level of expression in a variety of malignancies and has a role in tumour biology, in particular migration and metastasis [39]. This aspect makes this cell surface molecule an attractive antigen target for immunotherapy and diagnostic approaches in spite of its expression also on some normal tissue and cells including NK, cardiomyocytes and peripheral neural tissues. Indeed, investigators were not impeded to pursue an anti-NCAM approaches for cancer therapy by this potential drawback, as evidenced by a steadily rising number of indications and strategies are being published.

To date, only three non-mouse anti-NCAM antibodies are actively described, all of them developed for an antibody-drug conjugated (ADC) approach:

1.
Lorvotuzumab (huN901, humanized N901) linked to the cytotoxic maytansinoid effector molecule DM1 via a disulfide linkage (mertansine) (LM, huN901-SPP-DM1, BB-10901, IMGN901) that is being evaluated in clinical trial for different solid tumours [40] and for multiple myeloma [41] with good results [42].
2.
m900 and m906 conjugated with pyrrolobenzodiazepine dimer (PBD) (m900PBD and m906PBD), both human scFvs tested on neuroblastoma cell lines, which exhibited different killing activities, correlated both to their different ability to be internalised and to the heterogeneity of dynamics of NCAM on the different cell lines [43].
3.
Promiximab, a chimeric anti-NCAM antibody linked to two different drugs, which proved to be a good efficacy in small cell lung cancer (SCLC) cell line xenograft models, even if not all SCLC cell lines were bound by the antibody [44, 45].

However, recently a great interest is arising on the interaction between the NCAM1 and FGFR1 and the consequent induction of cell migration and metastatic dissemination, with particular attention to the EOC. Clearly, the interplay between an adhesion molecule and a growth factor may represent a new target for novel drug development for EOC, or other tumors that share this mechanism at the basis of their genesis and development, including during EMT [46, 47, 48].

These observations open the possibility to evaluate another mechanism of action for anti-NCAM antibody-based therapeutic strategies. In this context, the generation of a number of different highly specific antibodies against human NCAM is a prerequisite for the development of appropriated therapeutic strategies.

Here, we describe a new NCAM-specific fully human single chain antibody named scFvB1, endowed with good binding affinity and capable to selectively recognize various NCAM expressing tumor cell lines.

Our data indicate that the scFvB1 binds to a complex epitope that is not completely linear; in fact, even if scFvB1 was able to recognize the antigen in both WB and ELISA under reduction conditions, no reactivity was observed on overlapping linear peptides derived from the primary NCAM sequence. To this regard, it is important to point out that some antibodies recognising linear epitopes will bind to a specific spatial conformation of the linear epitope and thus it still recognises a conformational-specific epitope in the case of native proteins [49].

At the same time, there are different examples of antibodies toward conformational epitopes that showed ability, even if limited, to bind the target protein in Western blot [50].

Interestingly, we demonstrated that scFvB1 recognised the region of the two fibronectin like domains on extracellular domain of NCAM which is involved in the interaction with FGFR1 [21, 22].

This binding was also proved to translate in an interference into the molecular interaction between NCAM and FGFR with two different competition assays. The fact that scFvB1 doesn’t completely abolish, but only reduces the fluorescent signal of FGFR1-Fc, in spite of the higher binding affinity of scFvB1 for NCAM compared to FGFR and NCAM, can have different explanations. It could be the effect of a partial competition, which means that the epitope of scFvB1 is involved in the interaction but not limited to it. However, the phenomen may also be determined by the fact that FGFR1-Fc binding to OVCAR3 is not strictly limited to the NCAM molecule; this is a reasonable explanation since it is known that FGFR forms homophilic and heterophilic bindings with other molecules on cells surface.

Importantly, treatment of OVCAR3 cells with scFvB1 caused a partial inhibition of cells migration.

Docking analysis provided us the possible structural explanation for the competition ability of scFvB1, since the best obtained model of the interaction between scFvB1 and the structure of the two FN3 domains of NCAM displays a series of intermolecular contacts that involve amino acids of FRM and FGL regions of NCAM that are responsible for FGFR1 interactions.

Interestingly, scFvB1 amino acids involved in antibody-NCAM antigen interactions are located almost all in the VH chain; this finding is in line with several previous studies pointing out the central role of the heavy chain in antigen binding and specificity [51, 52, 53].

According to our knowledge, scFvB1 is the first completely human antibody described so far, with the potential to specifically target and interfere with the NCAM/FGFR interplay, known as involved in multiple progression cancer mechanisms. In the context of innovative therapeutic strategies, this context, a blocking antibody specific for NCAM could probably be more safe in respect to an inhibitor of FGFR1 receptor, since NCAM expression occurs specifically in transformed cancer cells while no antigen was observed in corresponding normal tissues.

Clearly, a deeper study of scFvB1 is necessary, but, it is important to underline that the structural flexibility of the scFvB1 format allows to further optimize the properties of scFvB1, for example by improving its affinity or steric hindrance and its competitive function; furthermore, scFvB1 can be considered a starting block to develop antibody constructs with a more balanced molecular weight in order to satisfy both the requirements of adequate pharmacokinetics and effective tumor penetration.

These properties make it a starting molecule for the development of potential therapeutic applications.
5. Conclusions

Recent studies are underlying the role that NCAM can have in cell migration and metastatic dissemination, via FGFR1 receptor. Different papers also examined in depth the structural basis for this interplay. We isolated a completely human scFv, named scFvB1, that was able to specifically bind NCAM molecule and interfere with the FGFR1 binding as demonstrated by functional and computational results. The structural flexibility of the scFv would allow to optimize its functional and kinetics properties, while its completely human origin is already an essential starting point for the development of therapeutic agents.

Availability of data and materials

The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.

Authors’ contributions

MF isolated scFvB1; GC performed SPR experiments and docking analysis; AA performed affinity experiments and ELISA for region mapping; MLD carried out characterization of scFvB1 on cell lines; EP and RR carried out OVCAR3 scratched tests; MA was on charge with FACS analyses; CR performed SDS PAGE and western blot analysis; MG was in charge to purify scFvB1; AB analyzed scratched tests images; SB performed recombinant protein purification; SV supervised the study design on the basis of the scFv isolation; AM designed experiments, wrote the paper and coordinated the study. All authors have seen and agreed to the final version of the paper.

Ethics approval and consent to participate

Human specimens used in some experiments described in our manuscript are PBMCs from healthy donors. Multiple stocks of human peripheral blood mononuclear cells (PBMCs) isolated from a large number of healthy donors ( $n>$ 20), have been stored in liquid nitrogen in our laboratory. The PBMCs have been purified from buffy coats obtained from the Transfusion Center of the University of Rome “La Sapienza” as waste material resulting from the procedures of plasma/platelet/red blood isolation from whole blood of healthy donors, who provided a written standard informed consent, which include the consent for the use of biologic material for research purposes, established with the Legislative Decree of the Italian Ministry of Health of November 2, 2015 “Dispositions relative to requirements of quality and safety of blood and blood derivatives” (Official Gazette of Italian Republic n. 300, published on December 28, 2015).

They were volunteers; accordingly to the Italian law on this matter (Legislative Decree of the Italian Ministry of Health of January 25, 2001, published on the Official Gazette of April 3, 2001) donor ages ranged between 21–60 years for women and 21–65 years for men. There have been no restrictions for blood donors as to sex and ethnic background, except for children, mentally disabled individuals and pregnant women. All donors were healthy, non-medicated consulting adults, and were prescreened for exposure to infectious agents. Blood was obtained by venipuncture following standard medical procedures using sterile techniques and a compression bandage in the area of blood removal. A record of the donor and hematocrit result is stored on file at the Transfusion Center. Research investigators do not have any access to personal data or other information concerning the donors in agreement with the Italian law Number 675 of December 31, 1996.

Our study did not require an ethical approval for the use of described cell lines for the evaluation of the ability of our scfv to recognize CD56 on their surface, since: the first legislation that is applicable to our study was emanated by the Data Protection authority in 2012 (General authorization for the treatment of personal data for research uses. Official Gazette of Italian Republic n. 72 published on March 26, 2012) and cell lines described in our study were obtained in a previous time. Moreover, there is not possibility to find the identity of the patients from which cell lines originated.

Funding

This work was supported by research funds of Italian Ministry of Health with the Grant number: GR-2010-2306432.

Supplementary data

The supplementary files are available to download from https://dx-doi-org.web.bisu.edu.cn/10.3233/HAB-200431.

Footnotes

Acknowledgments

We thank Dr. Massimo Sanchez and Dr. Valentina Tirelli of the ISS Service of instruments and core facilities for their technical support.

Conflict of interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

References

Crnic

et al., Loss of neural cell adhesion molecule induces tumor metastasis by up-regulating lymphangiogenesis, Cancer Res 64(23) (2004), 8630–8638. doi: 10.1158/0008-5472.CAN-04-2523.

Cavallaro

and Christofori

, Multitasking in tumor progression: signaling functions of cell adhesion molecules, Ann N Y Acad Sci 1014 (2004), 58–66. doi: 10.1196/annals.1294.006.

Jensen

and Berthold

, Targeting the neural cell adhesion molecule in cancer, Cancer Lett 258(1) (2007), 9–21. doi: 10.1016/j.canlet.2007.09.004.

Cunningham

B.A.

et al., Neural cell adhesion molecule: structure, immunoglobulin-like domains, cell surface modulation, and alternative RNA splicing, Science 236(4803) (1987), 799–806. doi: 10.1126/science.3576199.

Robertson

M.J.

and Ritz

, Biology and clinical relevance of human natural killer cells, Blood 76(12) (1990), 2421–2438.

Nitta

et al., Involvement of CD56 (NKH-1/Leu-19 antigen) as an adhesion molecule in natural killer-target cell interaction, J Exp Med 170(5) (1989), 1757–1761. doi: 10.1084/jem.170.5.1757.

Ziegler

et al., CD56 is pathogen recognition receptor on human natural killer cells, Sci Rep 7(1) (2017), 6138. doi: 10.1038/s41598-017-06238-4

Valgardsdottir

et al., Direct involvement of CD56 in cytokine-induced killer-mediated lysis of CD56+ hematopoietic target cells, Exp Hematol 42(12) (2014), 1013–1021. doi: 10.1016/j.exphem.2014.08.005.

Wachowiak

et al., Universal expression of cell adhesion molecule NCAM in neuroblastoma in contrast to L1: implications for different roles in tumor biology of neuroblastoma? Pediatr Surg Int 24(12) (2008), 1361–1364. doi: 10.1007/s00383-008-2264-z.

10.

Mechtersheimer

et al., Expression of the natural killer cell-associated antigens CD56 and CD57 in human neural and striated muscle cells and in their tumors, Cancer Research 51 (1991), 1300–1307. PMID: 1705171.

11.

Zheng

et al., Utility of the quantitative Ki-67 proliferation index and CD56 together in the cytologic diagnosis of small cell lung carcinoma and other lung neuroendocrine tumors, Acta Cytol 57(3) (2013), 281–290. doi: 10.1159/000346394.

12.

Kaiser

et al., The neural cell adhesion molecule NCAM in multiple myeloma, Leuk Lymphoma 20(5–6) (1996), 389–395. doi: 10.3109/10428199609052420.

13.

Ikushima

et al., Expression of CD56/NCAM on hematopoietic malignant cells. A useful marker for acute monocytic and megakaryocytic leukemias, Int J Hematol 54(5) (1991), 395–403. PMID: 1721853

14.

Zecchini

et al., The adhesion molecule NCAM promotes ovarian cancer progression via FGFGR signaling, EMBO Mol Med 3 (2010), 480–494. doi: 10.1002/emmm.201100152.

15.

Carbone

D.P.

et al., Neural cell adhesion molecule expression and messenger RNA splicing patterns in lung cancer cell lines are correlated with neuroendocrine phenotype and growth morphology, Cancer Res 51(22) (1991), 6142–6149.

16.

Sakamoto

et al., Expression of neural cell adhesion molecule in normal gastric mucosa and in gastric carcinoid tumors, Eur Surg Res 26(4) (1994), 230–239. doi: 10.1159/000129340.

17.

Vasei

et al., Immunohistochemical profiling of Wilms tumor: a tissue microarray study, Appl Immunohistochem Mol Morphol 16(2) (2008), 128–134. doi: 10.1097/PAI.0b013e31804d6825.

18.

Ash

et al., Excellent prognosis in a subset of patients with Ewing sarcoma identified at diagnosis by CD56 using flow cytometry, Clin Cancer Res 17(9) (2011), 2900–2907. doi: 10.1158/1078-0432.CCR-10-3069.

19.

Loret

et al., The role of epithelial-to-mesenchymal plasticity in ovarian cancer progression and therapy resistance. Cancers (Basel) 11(6). doi: 10.3390/cancers11060838.

20.

Friedl

and Wolf

, Tumor-cell invasion and migration: diversity and escape mechanisms, Nat Rev Cancer 3(5) (2003 May), 362–374.

21.

Francavilla

et al., The binding of NCAM to FGFR1 induces a specific cellular response mediated by receptor trafficking, J Cell Biol 187(7) (2009), 1101–1116. doi: 10.1083/jcb.200903030. Erratum in: J Cell Biol. 2019 Apr 1; 218(4): 1422. PubMed PMID: 20038681; PubMed Central PMCID: PMC2806277.

22.

Carafoli

et al., Structure of the tandem fibronectin type 3 domains of neural cell adhesion molecule, J Mol Biol 377(2) (2008), 524–534. doi: 10.1016/j.jmb.2008.01.030.

23.

Kiselyov

V.V.

et al, Structural biology of NCAM homophilic binding and activation of FGFR, J Neurochem 94(5) (2005), 1169–1179. Review. doi: 10.1111/j.1471-4159.2005.03284.x.

24.

Eggers

et al., Polysialic acid controls NCAM signals at cell-cell contacts to regulate focal adhesion independent from FGF receptor activity, J Cell Sci 124(Pt 19) (2011), 3279–3291. doi: 10.1242/jcs.084863.

25.

Suzuki

et al., Polysialic acid facilitates tumor invasion by glioma cells, Glycobiology 15(9) (2005), 887–894. doi: 10.1093/glycob/cwi071.

26.

Tanaka

et al., Expression of polysialic acid and STX, a human polysialyltransferase, is correlated with tumor progression in non-small cell lung cancer, Cancer Res (11) (2000), 3072–3080. PMID: 10850459.

27.

Seifert

et al., Polysialylation of the neural cell adhesion molecule: interfering with polysialylation and migration in neuroblastoma cells, Arch Biochem Biophys 524(1) (2012), 56–63. doi: 10.1016/j.abb.2012.04.011.

28.

Pasello

et al., Design and construction of a new human naive single-chain fragment variable antibody library, IORISS1, Journal of Biotechnology 224 (2016), 1–11. doi: 10.1016/j.jbiotec.2016.02.034.

29.

Ascione

et al., Development of a novel human phage display-derived anti-LAG3 scFv antibody targeting CD8+ T lymphocyte exhaustion, BMC Biotechnol 19(1) (2019), 67. doi: 10.1186/s12896-019-0559-x.

30.

Rath

et al., An inhibition enzyme immunoassay for estimating relative antibody affinity and affinity heterogeneity, J Immunol Methods 106(2) (1988), 245–249. doi: 10.1016/0022-1759(88)90204-9.

31.

Gellini

et al., Generation of human single-chain antibody to the CD99 cell surface determinant specifically recognizing Ewing’s sarcoma tumor cells, Curr Pharm Biotechnol 14(4) (2013), 449–463. doi: 10.2174/1389201011314040011.

32.

Ascione

et al., Isolation and characterization of the human monoclonal antibodies C10 in single-chain fragment variable (scFv) format to glucose oxidase from aspergillus niger, Hybrid Hybridomics 23(6) (2004), 380–384. doi: 10.1089/hyb.2004.23.380.

33.

Ilari

et al., Structural basis of sorcin-mediated calcium-dependent signal transduction, Sci Rep 5 (2015), 16828. doi: 10.1038/srep16828.

34.

Yang

et al., The I-TASSER suite: protein structure and function prediction, Nature Methods 12 (2015), 7–8. doi: 10.1038/nmeth.3213.

35.

Ambrish

et al., I-TASSER: a unified platform for automated protein structure and function prediction, Nature Protocols 5 (2010), 725–738. doi: 10.1038/nprot.2010.5.

36.

Zhang

, I-TASSER server for protein 3D structure prediction, BMC Bioinformatics 9 (2008), 40. doi: 10.1186/1471-2105-9-40.

37.

Kozakov

et al., The ClusPro web server for protein-protein docking, Nature Protocols 12(2) (2017), 255–278. doi: 10.1038/nprot.2016.169.

38.

Van Zundert

G.C.P.

et al., The HADDOCK2.2 web server: ser-friendly integrative modeling of biomolecular complexes, J Mol Biol 428(4) (2016), 720–725. doi: 10.1016/j.jmb.2015.09.014.

39.

Zeromski

et al., Significance of cell adhesion molecules, CD56/NCAM in particular, in human tumor growth and spreading, Folia Histochem Cytobiol 39(Suppl 2) (2001), 36–37.

40.

Shah

M.H.

et al., Phase i study of IMGN901, a CD56-targeting antibody-drug conjugate, in patients with CD56-positive solid tumors, Invest New Drugs 34(3) (2016), 290–299. doi: 10.1007/s10637-016-0336-9.

41.

Ailawadhi

et al., A phase i study to assess the safety and pharmacokinetics of single-agent lorvotuzumab mertansine (IMGN901) in patients with relapsed and/or refractory CD-56-positive multiple myeloma, Clin Lymphoma Myeloma Leuk 19(1) (2019), 29–34. doi: 10.1016/j.clml.2018.08.018.

42.

Socinski

M.A.

et al., Phase 1/2 study of the CD56-targeting antibody-drug conjugate lorvotuzumab mertansine (IMGN901) in combination with carboplatin/etoposide in small-cell lung cancer patients with extensive-stage disease, Clin Lung Cancer 18(1) (2017), 68–76. doi: 10.1016/j.cllc.2016.09.002.

43.

Feng

et al., Differential killing of CD56-expressing cells by drug-conjugated human antibodies targeting membrane-distal and membrane-proximal non-overlapping epitopes, MAbs 8(4) (2016), 799–810. doi: 10.1080/19420862.2016.1155014.

44.

et al., Promiximab-duocarmycin, a new CD56 antibody-drug conjugates, is highly efficacious in small cell lung cancer xenograft models, Oncotarget 9 (2018), 5197–5207. doi: 10.18632/oncotarget.23708.

45.

et al., Preparation and anti-cancer evaluation of promiximab-MMAE, an anti-CD56 antibody drug conjugate, in small cell lung cancer cell line xenograft models, J Drug Target 10 (2018), 905–912. doi: 10.1080/1061186X.2018.1450413.

46.

Colombo

and Cavallaro

, The interplay between NCAM and FGFR signaling underlies ovarian cancer progression, Ecancer Medical Science 5 (2011), 226. doi: 10.3332/ecancer.2011.226.

47.

Ćirović

et al., NCAM and FGFR1 coexpression and colocalization in renal tumors, Int J Clin Exp Pathol 7(4) (2014), 1402–1414. Published 2014 Mar 15.

48.

Lehembre

et al., NCAM-induced focal adhesion assembly: a functional switch upon loss of E-cadherin, EMBO J 27(19) (2008), 2603–2615. doi: 10.1038/emboj.2008.178.

49.

Forsström

et al., Proteome-wide epitope mapping of antibodies using ultra-dense peptide arrays, Molecular & Cellular Proteomics 13(6) (2014), 1585–1597. doi: 10.1074/mcp.M113.033308.

50.

Forsström

et al., Dissecting antibodies with regards to linear and conformational epitopes, PLoSONE 10(3) (2015), e0121673. doi: 10.1371/journal.pone.0121673.

51.

Ohno

et al., Antigen-binding specificities of antibodies are primarily determined by seven residues of VH, Proc Natl Acad Sci U S A 82(9) (1985), 2945–2949. doi: 10.1073/pnas.82.9.2945.

52.

Ren

et al., The role of the light chain in the structure and binding activity of two cattle antibodies that neutralize bovine respiratory syncytial virus, Mol Immunol 112 (2019), 123–130. doi: 10.1016/j.molimm.2019.04.026.

53.

Clarke

S.C.

et al., Multispecific antibody development platform based on human heavy chain antibodies, Front Immunol 9 (2019), 3037. Published 2019 Jan 7. doi: 10.3389/fimmu.2018.03037.

Isolation and preliminary characterization of a human ‘phage display’-derived antibody against neural adhesion molecule-1 antigen interfering with fibroblast growth factor receptor-1 binding

Abstract

BACKGROUND:

OBJECTIVE:

METHODS:

RESULTS:

CONCLUSIONS:

Keywords

1. Introduction

2. Materials and methods

2.1 Antibody phage library, NCAM antigen, human PBMC from healthy donors and cell lines

2.2 Biopanning strategy for selection NCAM extracellular domain recombinant protein specific scFvB1

2.3 DNA characterization and sequences

2.4 Soluble ScFv purification

2.5 NCAM extracellular fragments cloning and production

2.6 Elisa on recombinant NCAM proteins

2.7 Determination of binding affinity

2.8 Epitope mapping

2.9 SDS-PAGE and western blotting

2.10 NCAM detection on cell surface

2.11 Competition assay with FGFR1-Fc

2.12 Surface Plasmon Resonance assay

2.13 Modeling and docking

2.14 Wound healing assay

2.15 Statistical analysis

3.1 Isolation of human scFvB1 single chain antibody specific for recombinant and cell surface-exposed human NCAM by phage display technology

3.2 Evaluation of scFvB1 binding affinity and recognized epitope

3.3 Evaluation of scFvB1 binding on tumor cells expressing NCAM antigen

Table 1 Flow cytometry determination of the binding level of scFvB1 on different tumor cell lines

3.6 scFvB1 interferes with migration of EOC cells in vitro

4. Discussion

Availability of data and materials

Authors’ contributions

Ethics approval and consent to participate

Funding

Supplementary data

Footnotes

Acknowledgments

Conflict of interest

References

Table 1
Flow cytometry determination of the binding level of scFvB1 on different tumor cell lines