HK1183492B - Recombinant antibodies ana fragments recognising ganglioside n-glycolyl-gm3 and use thereof in the diagnosis and treatment of tumours - Google Patents

Description

Recombinant antibodies and fragments recognizing N-glycolyl GM3 ganglioside and their use in diagnosis and treatment of tumors

Technical Field

The present invention relates to the field of immunology and in particular to the use of genetic engineering techniques to obtain less immunogenic immunoglobulins for use in cancer therapy. The invention relates in particular to peptide sequences encoding monoclonal antibodies or fragments derived thereof which recognize antigens comprising N-glycolyl GM3 ganglioside, but excluding other related gangliosides such as N-glycolyl or N-acetyl, and sulfated glycolipids.

Background

Gangliosides are sialic acid-containing glycosphingolipids and are present in the plasma membrane of vertebrates (Stults CLM et al, Methods Enzymology 179: 167-214, 1989). Some of these molecules have been reported in the literature as tumor-associated antigens or tumor markers (Hakamori SH: Curr Opin Immunol 3: 646-653, 1991), and then the use of anti-ganglioside antibodies in the diagnosis and treatment of Cancer has been described (Hougton AN et al, NAS USA 82: 1242-1246, 1985; Zhang S et al, Int J Cancer 73: 42-49, 1997). The sialic acids that are more prevalent in mammals are N-acetyl (NAc) and N-glycolyl (NGc) (Corfield AP et al, Cell Biol Monogr 10: 5-50, 1982). NGc is not normally expressed in normal human and chicken tissues but is widely found in other vertebrates (Leeden RW et al, Biological Role of clinical acid. Rosemberg A and Shengtrun CL (Eds) Plenum Press, New York, 1-48, 1976, KawaiT et al, Cancer Research 51: 1242-1246, 2001). However, reports from the literature show that anti-NGc antibodies recognize some human tumors and tumor cell lines (Higashi H et al, Jpn J. cancer Res.79: 952-956, 1988, Fukui Y et al, biochem Biophys Res Commun 160: 1149-1154, 1989). Increased levels of GM3(NGc) gangliosides have been found in human breast Cancer (Marquina G et al, Cancer Research 56: 5165-.

The filed patent EP0972782 a1 describes murine monoclonal antibodies produced by the hybridoma deposited under the Budapest treaty with the accession number ECACC 98101901. The monoclonal antibody has the IgG1 isotype and is produced by immunizing Balb/c mice immunized with NGcGM3 ganglioside hydrophobically conjugated to low density lipoprotein (VLDL) and in the presence of freund's adjuvant, which preferably binds to NGcGM3 ganglioside and recognizes antigens expressed in human breast and melanoma tumors. (Carr A et al, Hybridoma 19: 3: 241-247, 2000). The antibody exhibits strong cytolytic activity against cells containing the sequence, which makes it a therapeutic tool.

It is also known that tumor-induced neovascularization constitutes one of the main parameters in the control of tumor growth. This parameter has directed research towards finding new therapeutic weapons involving inhibition of antigenic processes. There is experimental evidence for the efficacy of the 14F7 antibody involved in this process. The evidence has been found in matrix gel assays (Angigenesis. Vol.4, pags.113-121, 2001; Anti-Cancer Res. Vol.18, pags.783-790, 1998). The matrix gel is a mixture of basement membrane proteins and extracellular matrix surrounding endothelial cells. Its effectiveness in the neovascularization process lies in the physiological importance of endothelial-matrix interactions during the development of new blood vessels. Also, components of the gel such as laminin are considered to be important markers for this process.

This model type has also been used in studies involving the mechanisms of tumor-induced angiogenesis, allowing the evaluation of drugs capable of modulating events directly involving tumors, thus affecting their growth and metastasis.

Treatment of humans with non-human monoclonal antibodies such as murine 14F7 has certain disadvantages, particularly in repeated treatment regimens. For example, murine monoclonal antibodies have a relatively short blood half-life. Furthermore, when used in humans, they lose important immunological properties such as effector function.

And more importantly, murine monoclonal antibodies contain a significant number of amino acid sequences that are immunogenic when injected into humans. Numerous studies have shown that the immune response generated in a patient following administration of exogenous antibodies can be quite strong and substantially eliminates the therapeutic efficacy of the antibodies after initial treatment. Furthermore, even after treatment of patients with murine monoclonal antibodies, subsequent treatment with non-related murine antibodies may be ineffective and even dangerous due to cross-reactivity known as HAMA response (Khazaeli, M.B., et al, Journal of immunotherapy 15: 42-52, 1994).

Mateo et al (U.S. Pat. No. US 5712120) describe methods for reducing the immunogenicity of murine antibodies. The antibodies modified by the method described by Mateo et al (Mateo C et al, Hybridoma 19: 6: 463-471, 2000) retain the recognition and binding ability of the original antibody to the antigen and result in less immunogenicity, the reduction of which increases their therapeutic effectiveness. Modified antibodies with low amounts of mutations are obtained by this method, which antibodies exhibit reduced immunogenicity when compared to chimeric antibodies.

From the above it is concluded that there is a need to obtain less immunogenic versions of therapeutic antibodies in a simple and economical way and thus to produce sufficient therapeutic preparations for human use.

The use of antibody fragments is also known to be very useful in the immunodiagnosis of diseases. Ira Pastan (filed European patent EP 0796334A 1) describes, among others, the construction of Fv-type single-chain fragments using the variable region of an antibody specifically recognizing a carbohydrate involved in a Lewis Y antigen. Based on the fragments, the authors developed a method for detecting cells containing the antigen and also provided evidence of the inhibitory effect of these fragments on the cells containing the antigen.

Due to the potential of this antigen in immunotherapy of various diseases, knowledge of the 14F7 monoclonal antibody binding site is of practical and theoretical interest. Souriiu, C et al in expetpain. biol. ther.1, 845-855, 2001 and Chester, k.a., et al in dis. markers.16, 53-62, 2000 show that the use thereof can lead to very good therapeutic uses by exploiting the pharmacological properties of scFv-type antibody fragments, such as better penetration in tumor tissue.

Construction of a mini-library of antibody fragments from hybridomas exposed on filamentous phage enables selection of molecules in the library that have specific recognition capabilities for the antigen. The aim is to rapidly obtain antibody fragments that bind the original antibody in a smaller molecular size and that can be produced by a bacterial host. This technique allows the selection of variants of multiple antibody fragments which are not functional or impossible to produce in bacteria, obtained during the isolation and manipulation of the variable regions of hybridoma genes using conventional cloning techniques (Rovers, R.C. et al Br.J.cancer.78, 1407-1416, 1998).

Antibody fragment presentation techniques in filamentous phage provide a unique opportunity to access the antibody binding site and to introduce modifications into the antibody to increase its affinity (Chames, P. et al, J.Immunol.161, 5421-5429, 1998, Lammmaki, U et al, J.mol.biol.291, 589-602, 1999, Parhami-Seren et al, J.Immunol.methods.259, 43-53, 2002) or to modulate its specificity (Iba, Y. et al, prot.Engn., 11, 361-370, 1998, Miyazaki, C. et al, prot.Engn.12, 407-415, 1999, Darveau, R.P. et al, J.Clin.Immunolay.15, 25-29, 1992). Expression of antibody fragments in filamentous phages allows exchange of chains (Lantto, J. et al, Methods mol. biol.178, 303. sup. 316, 2002) which are combinations of various VLs and original antibody VH, or vice versa, thus allowing investigation of the role of each chain in antigen recognition and its effect on antibody affinity (Kabat, E.A., Wu, T.T., J.Immunol.147, 1709. sup. 1719, 1991, Barbas III, C.F., Lerner, R.A. Methods: A composition Methods in Enzymology 2, 119. sup. 124, 1991). Strand exchanges make it possible to modify the properties of the binding sites and also to combine immunoglobulin sequences from different species and to select variants which retain the recognition specificity (Klimka, A. et al, Br. J. cancer.83, 252. 260, 2000, Steinberger, P. et al, J. biol. chem.275, 36073. 36078, 2000, Rader, C. et al, Proc. Natl. Acad. Sci.USA.95, 8910. 8915, 1998, Beiboer, S.W.H. et al, J.mol. biol.296, 833. 849, 2000).

The present invention relates to modified antibodies derived from the murine 14F7 monoclonal antibody that recognize the original antigen with the same affinity and that are less immunogenic when administered to a patient. Another finding of the present invention is that fragments derived from said antibodies have been obtained, which fragments comprise a light chain variable region that is different from the original antibody but retains its properties related to specificity, affinity and recognition properties, and can be expressed and presented as a soluble molecule on the surface of filamentous bacteriophage by a bacterial host. Fragments derived from the 14F7 antibody are useful as therapeutic weapons. Furthermore, the expression of scFv-type fragments of the 14F7 antibody on the surface of M13 filamentous phages makes it possible to manipulate the binding site and to obtain variants with higher affinity for therapeutic purposes.

It is an object of the present invention to characterize antibody fragments that only partially preserve the sequence of 14F7 but still retain their specificity, affinity and recognition properties.

Both the modified antibody and the obtained fragment specifically recognize tumor cells expressing the NGcGM3 antigen, thus making it useful for the diagnosis or treatment of said tumors, and having a lower immunogenicity than the murine antibody from which it is derived.

Novel therapeutic compositions for localizing or treating tumors expressing said antigen comprising a modified 14F7 antibody or Fv single chain fragment thereof, wherein said antibody or fragment can bind a radioisotope, are objects of the present invention.

Also an object of the present invention is a method for radioimmunodesis or radioimmunotherapy of ngcm 3 ganglioside expressing tumors using a pharmaceutical composition comprising a modified 14F7 antibody or fragment thereof conjugated with a radioisotope.

Detailed Description

Described in the present invention is a chimeric antibody derived from the murine 14F7 monoclonal antibody produced by the hybridoma deposited with ECACC98101901, characterized by the hypervariable region (CDRs) sequences of the following heavy and light chains.

Heavy chain

CDR1：SYWIH

CDR2：YIDPATAYTESNQKFKD

CDR3：ESPRLRRGIYYYAMDY

Light chain

CDR1：RASQSISNNLH

CDR2：YASQSIS

CDR3：QQSNRWPLT

Preferably the chimeric antibody is characterized by the following sequences of Framework Regions (FRs) of the heavy and light chains:

heavy chain

FR1：QVQLQQSGNELAKPGASMKMSCRASGYSFT

FR2：WLKQRPDQGLEWIG

FR3：KAILTADRSSNTAFMYLNSLTSEDSAVYYCAR

FR4：WGQGTTVTVSS

Light chain

FR1：DLVLTQSPATLSVTPGDSVSFSC

FR2：WYQQRTHESPRLLIK

FR3：GIPSRFSGSGSGTDFTLSIISVETEDFGMYFC

FR4：FGAGTKLELKRA

Furthermore, the chimeric antibody is characterized by comprising the heavy chain constant region of human IgG1 and the light chain constant region of human Ck. In a preferred embodiment, the present invention relates to a modified antibody derived from the murine 14F7 monoclonal antibody produced by the hybridoma deposited with the accession number ECACC98101901, characterized in that the framework regions of the heavy and light chain comprise any of the following mutations:

heavy chain：

Position 5: v replaces Q

Position 9: a replaces N

Position 11: v replaces L

Position 12: v replaces A

Position 18: v replaces M

Position 19: r replaces K

Position 20: v replaces M

Position 38: r replaces K

Position 40: a replaces R

Position 42: g for D

Position 48: v is substituted for I

Light chain：

Position 39: k replaces R

Position 40: p instead of T

Position 41: g replaces H

Position 42: q instead of E

Position 58: v is substituted for I

Preferably, the modified antibody is characterized by comprising the heavy chain constant region of human IgG1 and the light chain constant region of human Ck.

Another aspect of the invention relates to an Fv single chain type fragment derived from the 14F7 antibody, said fragment comprising the sequence of the heavy chain variable region obtained from the hybridoma (accession number ECACC 98101901) and preserving the recognition specificity of 14F 7.

Preferably the Fv fragment comprises a light chain variable region derived from a non-immunized mouse or human that is different from the light chain variable region produced by the hybridoma, which Fv fragment allows its production by a bacterial host as an antibody fragment or soluble molecule presented on a filamentous bacteriophage.

The invention also relates to cell lines expressing said chimeric and humanized antibodies and bacterial clones producing fragments derived from 14F 7.

Furthermore, the present invention relates to pharmaceutical compositions for the treatment and/or localization and determination of malignant breast and melanoma tumors and their metastases and relapses, characterized in that they comprise one of the recombinant monoclonal antibodies or fragments thereof of the invention and excipients suitable for their use. The antibodies or fragments may bind to a radioisotope for use in locating and/or treating a malignant tumor.

Finally, the invention relates to the use of the described recombinant antibodies for the production of pharmaceutical compositions for the treatment and/or localization and identification of malignant tumors.

The following detailed description is directed to the implementation and use of the invention.

PCR (polymerase chain reaction) synthesis and amplification of cDNA for the variable region of 14F7 murine antibodies.

Murine 14F7 antibody was obtained by immunizing Balb/c mice with GM3(NeuGc) ganglioside, which is hydrophobically conjugated to low density lipoprotein, in the presence of Freund's adjuvant (EP 0972782A 1, Carr A et al, Hybridoma 19: 3: 241-Asonic 247, 2000). From 10 using TRIZOL extraction method (GIBCO BRL, NY) according to the manufacturer's instructions⁶RNA was obtained from hybridoma cells producing a 14F7 monoclonal antibody (murine IgG1 Mab) that recognizes NGcGM3 ganglioside.

By using the kit according to the manufacturer's instructions: access RT-PCR (Promega, USA) performs the synthesis of complementary DNA (cDNA) and the amplification of VH and VL variable regions. Briefly, at 25 picomoles of dT oligonucleotide and corresponding to each variable region: the reaction was performed from 5. mu.g RNA in the presence of the end oligonucleotides of VH and VL, designed to hybridize to the poly-A tail of the RNA. Incubate at 60 ℃ for 10 minutes, then add a mixture containing 0.2mM each deoxyribonucleotide (dNTPs), 1mM MgSO 2 in reaction buffer₄5 μ AMV-reverse transcriptase and 5 μ DNA polymerase in a total volume of 50 μ l. The samples were incubated at 48 ℃ for 45 minutes, 94 ℃ for 2 minutes, and then subjected to 40 PCR cycles: 94 ℃ (30 seconds), 60 ℃ (1 minute), 68 ℃ (2 minutes), and finally incubation at 68 ℃ for 7 minutes.

2. Construction of chimeric genes and sequencing of amplified cDNA.

The PCR products of VH and VK were degraded with the restriction enzymes Eco RV-Nhe I (for VH) and Eco RV-Sal I (for VK) and cloned into the respective expression vectors (Coloma MJ et al, J Immunol Methods 152: 89-104, 1992). The VH regions were cloned into a PAH4604 vector already containing the human IgG1 constant region and the anti-L-histidinol selectable marker gene. The VK region was cloned into the PAG4622 vector carrying the microphenolic acid resistant gene and the human kappa constant region. The resulting gene constructs were called 14F7VH-PAH4604 and 14F7VK-PAG4622, respectively. Both constructs were sequenced by the dideoxynucleotide method (Sanger F et al, PNASA 74: 5463-.

3. Expression of chimeric antibodies

10 μ g of 14F7VH-PAH4604 linearized with the enzyme Pvu I and 10 μ g of 14F7VK-PAG4622 were electroporated into NSO cells. DNAs were precipitated with ethanol, mixed and dissolved in 50. mu.L PBS (phosphate buffered saline). When about 10⁷Cells were harvested by centrifugation as they grew to half confluence. The cells were then resuspended with DNA in 0.5mL PBS in an electroporation cuvette. After 10 minutes in ice bath, the cells were pulsed at 200V and 960F and kept on ice for 10 minutes. The cells were cultured in Modified Dulbecco (DMEM-F12) selection medium containing 10% Fetal Calf Serum (FCS) and 10mM L-histidinol in 96-well plates. Transfected clones were observed 10 days after addition of selection medium.

Production of chimeric immunoglobulins was determined by ELISA of the clone supernatants. For this purpose, polystyrene plates (High binding, Costar) were coated with goat anti-human IgG serum in 100mM bicarbonate buffer (pH9.8) overnight at 4 ℃. The plate was then washed with PBS-Tween (phosphate buffered saline, 0.5% Tween-20, pH7.5), a sample of culture supernatant diluted in PBS-Tween-FCS was added and incubated at 37 ℃ for 1 hour. The plate was washed again with PBS-Tween and then incubated with peroxidase (Jackson) conjugated goat anti-human kappa chain serum for 1 hour at 37 ℃. The plate was then washed in the same manner and with pH 4.2 containing substrate: o-fennelenediamine in citrate-phosphate buffer. The absorbance was measured at 492nm after 15 minutes.

The ability of the chimeric antibody to recognize antigen was determined by performing a competition ELISA using NGcGM3 and NacGM3 ganglioside. Briefly, polystyrene plates (Polysorp, Nunc) were coated with a solution of 4 μ g/mL NGcGM3 or NAcGM3 in 50 μ L of methanol and incubated at 37 ℃ for 1 hour. The plate was then blocked with 200. mu.L of a solution of Bovine Serum Albumin (BSA) in 1% Tris-HCl (pH7.8-8.0) buffer for 1 hour at 37 ℃. After washing the plate 3 times with PBS, the sample diluted in 1% Tris-HCl-BSA was added at a concentration ranging between 2. mu.g/mL and 0.01. mu.g/mL in the presence of 1mg/mL biotinylated murine 14F7 antibody and incubated at 37 ℃ for 2 hours. After washing the plate with PBS, the reaction was developed with streptavidin-peroxidase conjugate (Jackson) for 1 hour at 37 ℃. The absorbance was measured at 492 nm.

4. Construction of a humanized Ab 147FhT by humanization of the T epitope

Prediction of T epitopes

The sequence of the 14F7 variable domain was analyzed using the AMPHI program (Margalit H et al, J Immunol 138: 2213-2229, 1987) which allows the identification of fragments of 7 or 11 amino acids with an amphipathic helical structure associated with T cell immunogenicity. In this case these algorithms predict fragments involved in T epitope presentation in the heavy and light chain variable regions of the murine 14F7 monoclonal antibody.

Analysis of homology with human immunoglobulins

The amino acid sequence of the variable domain of 14F7 was compared to the reported variable region sequences of human immunoglobulins to identify the human immunoglobulin with the highest homology to the murine molecule being analyzed. This can be achieved by the BLAST program (Altschul S F et al, Nucleic Acids Res 25: 3389-.

Analysis of reduced immunogenicity

The essence of the method is that reduced immunogenicity is obtained by fragmentation or humanization of possible T epitopes, with minimal mutations in FRs (excluding those sites that participate in the three-dimensional structure of the antigen recognition site), particularly those fragments with amphipathic helical structures.

According to said method, the sequences of the VH and VL variable regions of the murine immunoglobulin are compared with those of the most homologous human immunoglobulin and the residues that differ between the murine and human sequences are identified only in the amphiphatic regions within the FRs region (Kabat E, sequence of proteins of immunological interest, Fifth Edition, national institute of Health, 1991). These "murine" residues may be interchanged with residues found at the same positions in the human sequence. Residues present at FRs positions that are responsible for the canonical structure, residues present in the Vernier zone or residues involved in the VH-VL interface interaction cannot be mutated as they can influence the three-dimensional structure of the antibody variable domain and thus the binding to antigen. Additional information concerning the effect of substitutions on tertiary structure can be obtained by variable region molecular modeling.

Due to the fact that e.g. proline residues are present in the amphipathic helix or the fact that: some murine residues do not appear at the same positions in the highest homology human sequences but often appear in other human immunoglobulins, so it is possible to obtain versions containing the most mutations, i.e. all murine residues different from the human sequence are mutated, but other versions with different combinations of mutations can also be obtained.

Following identification of potential T epitopes in the murine 14F7 sequence and identification of residues within these fragments other than human, substitutions can be made by conventional directed mutagenesis techniques.

Cloning and expression of humanized 14F7hT antibody in NSO cells.

After obtaining the gene constructs corresponding to the VH and VL regions of the 14F7hT antibody by the aforementioned method, they were cloned into respective expression vectors similar to those described above in the case of the chimeric antibody construction, to obtain the following gene constructs: 14F7hTVK-PAG4622 and 14F7hTVH-PAH 4604. These genes were transfected into NSO cells under exactly the same conditions as described for the chimeric antibodies. Clones producing the humanized antibody were also tested by ELISA.

The ability of the humanized antibody to specifically recognize the antigen was determined by competitive ELISA using NGcGM3 and NAcGM3 ganglioside. The methods are consistent with those described for chimeric antibodies.

5. Construction of a library of antibody fragments from 14F7 hybridoma based on filamentous phage.

Messenger RNA was isolated from 14F7 hybridoma cells, complementary DNA was synthesized, and sequences corresponding to the variable regions of the heavy and light chains were amplified separately using two sets of oligonucleotides designed to hybridize to a broad spectrum of murine variable regions. Each amplified heavy chain variable region was purified, degraded with an appropriate enzyme and ligated to a phage pHG-1m vector (designed to present single chain antibodies on the surface of filamentous phage) previously degraded with the same enzyme. The product of the binding reaction was purified and introduced into the bacteria of TG1 strain escherichia coli (e. The transformed cells constitute a heavy chain variable region half-library of limited diversity, all of which are derived from the 14F7 hybridoma. The DNA corresponding to this half-library was purified and ligated into pools of light chain variable regions (previously purified and degraded with an appropriate enzyme), respectively. Finally, the gene construct is introduced into TG1 strain bacteria by electroporation to form independent libraries of different types. One of these was a mini library of limited diversity constructed with the light chain variable region also obtained from the 14F7 hybridoma. A strand-exchange library was also obtained in which heavy chain variable regions from hybridomas are combined with various pools of light chain variable regions obtained from murine and human lymphocytes.

6. Isolation and characterization of clones producing antibody fragments against N-glycolyl GM 3.

Clones of each library were randomly picked from which phage were generated by using the helper phage M13K 07. The recognition of N-glycolyl GM3 by the phage carrying the antibody fragment was analyzed directly by the solid phase immunoenzyme assay (ELISA). Thus clones producing functional fragments were located. Phage carrying antibody fragments (from the total mixture of transformed bacteria forming each library) were generated to obtain preparations enriched for functional antibody fragments and to investigate the diversity of the libraries. The resulting phage mixture was contacted with the antigen N-glycolyl GM3 bound to a solid surface, and the phage carrying the antibody fragment with binding ability was retained while the rest was washed well. Eluting the bound phage by changing the pH and propagating the bound phage to obtain a new mixture of phage for use as a starting material for a new round of selection against an antigen. After each round of selection, the recognition function of the antibody fragments present on the phage from the colonies was analyzed. Additional clones producing functional fragments diluted in the starting library were thus identified.

Characterization of the antibody fragments produced by the clones included analysis of specificity against groups of related gangliosides by ELISA, tumor recognition studies by immunohistochemistry, estimation of their ability to be produced as soluble functional antibody fragments outside phage, and complete sequencing of their variable regions.

Examples

In the examples that follow, all restriction enzymes or modifications and reagents and materials used were obtained from commercial sources unless otherwise indicated.

EXAMPLE 1 obtaining of 14F7 chimeric monoclonal antibody

Murine VH and VK variable region cDNAs were synthesized and amplified by PCR with Vent polymerase using specific oligonucleotides with restriction sites Eco RV-Nhe I (for VH) and Eco RV-SalI (for VK). The oligonucleotide primers were as follows:

for VH:

oligonucleotide 1 (hybridized at signal peptide):

5’ggg gatatc cacc atg gaa agg cac tgg atc ttt ctc ttcctg 3’

oligonucleotide 2 (hybridized at JH 1):

5’ggg gctagc tga gga gac ggt gac cgt ggt 3’

for VK:

oligonucleotide 1 (hybridized at signal peptide):

5’ggg gatatc cacc atg gt(at)t(tc)c(ta)ca cct cag(at)t(ac)ctt gga ctt 3’

oligonucleotide 2 (hybridized at VLJ 5):

5’agc gtcgac tta cgt ttc agc tcc agc ttg gtc cc 3’

the PCR products of VH and VK were degraded with the restriction enzymes Eco RV-Nhe I (for VH) and Eco RV-Sal I (for VK) and cloned into their respective expression vectors PAH4604 and PAG4622 corresponding to VH and VK, respectively. These vectors are used for expressing immunoglobulins in mammalian cells and are supplied by Sherie Morrison (UCLA, California, USA). Vector PAH4604 contains the human IgG1 constant region PAG4622 contains the human kappa constant region (Coloma J et al, J Immunol Methods 152: 89-104, 1992). After cloning of the VH and VK regions of 14F7 into the previous vectors, constructs 14F7VH-PAH4604 and 14F7VK-PAG4622 were generated.

12 independent clones were sequenced by the dideoxynucleotide method (Sanger F et al, DNAsequencing with chain-sequencing inhibitors. PNAS USA 1979; 74: 5463-. According to the Kabat classification (Kabat E, Sequences of proteins of immunological interest, Fifth Edition, National Institute of Health, 1991), VH and VK Sequences are highly related to subtypes IIB and V, respectively (FIGS. 1A and B).

NSO cells were electroporated with 10. mu.g of 14F7VH-PAH4604 linearized with Pvu I and 10. mu.g of 14F7VK-PAG 4622. DNAs were precipitated with ethanol, mixed and dissolved in 50. mu.L PBS. The culture was collected by centrifugation to approximately 10 at half confluence⁷Cells were resuspended in 0.5mL PBS along with DNA in an electroporation cuvette. After 10 minutes of ice-bath, the cells were pulsed at 200V and 960F, followed by 10 minutes of ice-bath. The cells were cultured in Dulbecco Modified (DMEM-F12) selection medium containing 10% FCS and 10mM L-histidinol in 96-well plates. Transfected clones were observed 10 days after addition of selection medium. Use gramThe supernatant of the clones was assayed for production of chimeric immunoglobulins by ELISA. For this purpose, polystyrene plates (High binding, Costar) were coated with goat anti-human IgG serum (Sigma) in 100mM bicarbonate buffer (pH9.8) and left overnight at 4 ℃. The plates were then washed with PBS-Tween (phosphate buffered saline, 0.5% Tween-20, pH7.5), samples of culture supernatant diluted in PBS-Tween-FCS were added and incubated for 1 hour at 37 ℃. The plate was then washed with PBS-Tween and incubated with peroxidase-conjugated goat anti-human kappa chain serum (Jackson) for 1 hour at 37 ℃. The plate was then washed in the same way and with a solution containing the substrate: o-fennelenediamine in citrate phosphate buffered saline (pH 4.2). The absorbance was measured at 492nm after 15 minutes.

Example 2.14 reactivity of the F7Q antibody to NGcGM3 and NAcGM 3.

The reactivity of the 14F7 chimeric antibody (Ab) was determined by competition ELISA. FIG. 2 shows the recognition specificity of chimeric 14F7 for polystyrene plates coated with GM 3N-glycolylganglioside (Polysorp, Nunc) compared to murine 14F 7. Both Abs showed 50% inhibition of antigen recognition by murine biotinylated 14F7 Ab at similar concentrations. However, when plates were coated with N-acetylated GM3 variant, no immunoreactivity was observed for any of the antibodies. Murine C5 Mab was used as a non-relevant Ab.

Example 3. different humanized antibody versions were obtained.

The 14F7VH and VK sequences were compared to the human sequence database to obtain human 14F7 antibody sequences with the highest homology to the VH and VK regions (fig. 1A and B). In addition, the amphiphilic regions or potential T epitopes in both sequences were determined. The mutations required to convert murine VH and VK sequences to human sequences were then determined by humanization of the T epitope. We next refer to the largest possible mutation that can be made.

In the case of the VH region, mutations are introduced at positions 5, 9, 11, 12, 18, 19, 20, 38, 40, 42 and 48, wherein amino acids Q, N, L, a, M, K, R, D and I are substituted with V, a, V, R, a, G and V, respectively. These mutations were carried out by overlapping PCR products (Kamman M et al, Nucleic Acids Research 17: 5404-. The oligonucleotide sequences used are shown below:

oligonucleotides for the 5, 9, 11 and 12 mutations of the heavy chain:

oligonucleotide 1: 5 'gtc cag ctt gtg cag tct ggg gct gaa gtg gtaaaa cct ggg 3'

Oligonucleotide 2: 5 'ggg gctagc tga gga gac ggt gac cgt ggt 3'

Oligonucleotide 3: 5 'ccc agg ttt tac cac ttc agc ccc aga ctg cacaag ctg gac 3'

Oligonucleotide 4: 5 'ggg gat atc cacc atg gaa agg cac tgg atc tttctc ttc ctg 3'

After sequencing and identification of the previous mutations, mutations were introduced at positions 18, 19 and 20 of the DNA carrying them to replace position M, K, M with amino acid V, R, V, respectively.

Oligonucleotides 1 and 2 and 3 and 4 used for the mutations are shown below. The overlap of PCR products was performed as described previously.

Oligonucleotides for the 18, 19 and 20 mutations of the heavy chain:

oligonucleotide 1: 5 'ggg gcc tca gtg agg gtg tcc tgc agg 3'

Oligonucleotide 2: 5 'ggg gctagc tga gga gac ggt gac cgt ggt 3'

Oligonucleotide 3: 5 'cct gca gga cac cct cac tga ggc ccc 3'

Oligonucleotide 4: 5 'ggg gatatc cacc atg gaa agg cac tgg atc tttctc ttc ctg 3'

Also after sequencing and identification of the previous mutations, mutations R, A, G were introduced at positions 38, 40 and 42 of the DNAs carrying them to replace amino acid K, R, D, respectively.

Oligonucleotides 1 and 2 and 3 and 4 for these mutations are shown below. The overlap of PCR products was performed as described previously.

Oligonucleotides for the 38, 40 and 42 mutations of the heavy chain:

oligonucleotide 1: 5 'cac tgg tta aga cag gca cctggc cag ggt ctg 3'

Oligonucleotide 2: 5 'ggg gctagc tga gga gac ggt gac cgt ggt 3'

Oligonucleotide 3: 5 'cag acc ctg gcc agg tgc ctg tct taa cca gtg 3'

Oligonucleotide 4: 5 'ggg gat atc cacc atg gaa agg cac tgg atc tttctc ttc ctg 3'

Finally, after the previous mutations were identified by sequencing, mutations were introduced at position 48 of the DNAs carrying them to replace I with amino acid V. Oligonucleotides 1 and 2 and 3 and 4 used for this mutation are shown below. The overlap of PCR products was performed as described previously.

Oligonucleotide for the 48 mutations of the heavy chain:

oligonucleotide 1: at the time of the start of the 5' ctg gaa tgg gtt gga tac att 3,

oligonucleotide 2: at the time of the start of the 5' ggg gctagc tga gga gac ggt gac cgt ggt 3,

oligonucleotide 3: at the time of the start of the 5' aat gta tcc aac cca ttc cag 3,

oligonucleotide 4: 5 'ggg gat atc cacc atg gaa agg cac tgg atc tttctc ttc ctg 3'

All of these mutations were identified by sequencing. The resulting gene construct was designated 14F7 hTVH.

In the case of the heavy chain, there are no substitutions of amino acids that differ between VH14F7 and the most similar human antibody, since these substitutions involve amino acids in the Vernier zone or are critical positions in antigen recognition.

For the light chain, mutations were made at positions 39, 40, 41, 42 and 58, replacing R, T, H, E, I with K, P, G, Q and V, respectively. Mutations were introduced in the same manner as in the heavy chain. The oligonucleotide sequences used are described below.

Oligonucleotides for the 39, 40, 41 and 42 mutations of the light chain:

oligonucleotide 1: 5 'tat caa caa aaa cca ggt cag tct cca agg 3'

Oligonucleotide 2: 5 'agc gtcgac tta cgt ttc agc tcc agc ttg gtccc 3'

Oligonucleotide 3: 5 'cct tgg aga ctg acc tgg ttt ttg ttg ata 3'

Oligonucleotide 4: 5 'ggg gatatc cacc atg gt (at) t (tc) c (ta) cacct cag (at) t (ac) ctt gga ctt 3'

After the previous mutations were identified by sequencing, mutations were introduced at position 58 of the DNAs carrying them to replace I with amino acid V. Oligonucleotides 1 and 2 and 3 and 4 used for this mutation are shown below. The overlap of PCR products was performed as described previously.

Oligonucleotide for the 58 mutation of the light chain:

oligonucleotide 1: 5 'att tct ggg gtc ccc tcc agg 3'

Oligonucleotide 2: 5 'agc gtcgac tta cgt ttc agc tcc agc ttg gtccc 3'

Oligonucleotide 3: 5 'cct gga ggg gac ccc aga aat 3'

Oligonucleotide 4: 5 'ggg gatatc cacc atg gt (at) t (tc) c (ta) ca cctcag (at) t (ac) ctt gga ctt 3'

This was determined by sequencing after the mutation was performed.

The resulting gene construct was designated 14F7 hTVK.

Humanized VK and VH regions were cloned into vectors PAG4622 and PAH4604, resulting in constructs 14F7hTVH-PAH4604 and 14F7hTVK-PAG4622, respectively. NSO cells were electroporated with 10 μ g of vectors 14F7hTVH-PAH4604 and 14F7hTVK-PAG4622, respectively, previously linearized with Pvu I, each carrying a humanized variable region. The electroporation method and detection method for clones expressing the 14F7hT humanized antibody were the same as described for the chimeric antibody.

Example 4: effect of 14F7 Ab on reduction of tumor growth in an in vivo model of angiogenesis.

Melanoma B16 tumor cell line mixed with matrix gel as an angiogenesis inducer was used as a model for angiogenesis in vivo.

Tumor lines with matrix gel (B16) were implanted subcutaneously at the mid-abdominal line of C57/BL6 male mice and the induction of tumor vascularization process was measured. To achieve this, melanoma B16 cells were mixed with 0.5mL of matrigel and 64 units/mL of heparin and administered by subcutaneous injection into the abdominal region of the animals, either alone as a control or in combination with other antibodies. After 15 days of inoculation, animals were sacrificed and the gel beads were removed along with dermal and epidermal remnants and the tumor size was macroscopically observed. Figure 3 shows that the tumor mass size of the mice tumors treated with 14F7 was significantly reduced.

To differentiate new blood vessels, histopathological analysis was performed on gels stained with hematoxylin/eosin and immunostained with anti-PECAM Mab (endothelial marker). Fig. 4A and B show that a reduction in the number of blood vessels was observed in tissue sections from animals treated with 14F7 compared to tumor sections from untreated animals.

Example 5: libraries of antibody fragments of filamentous phage were constructed from the 14F7 hybridoma.

From 5.0X10 using TriPure Isolation Reagent (Boehringer Mannheim, Germany)⁶Isolation of messenger RNA from 14F7 hybridoma cells and use of Pro-STAR FirstStrand RT-PCR reagentThe cassette (Stratagene, USA) synthesizes complementary DNA. 26 cycles of polymerase chain reaction were performed to amplify the heavy and light chain variable regions from the hybridomas according to the following conditions: at 94 ℃ for 50 seconds, at 55 ℃ for 1 minute and at 72 ℃ for 1 minute. For amplification, all possible sets of degenerate oligonucleotide combinations (designed to hybridize to a wide range of mouse variable regions) were used, which are shown in the following table:

oligonucleotides for amplifying VH and VL. In the oligonucleotide sequences, degenerate positions appear in parentheses. The sequence of the restriction site is underlined.

5’VH Apa LI

5’...ttc tat tct cac agt gca cag(gc)ag gt(gt)cag ct(gt)ac(ag)cag tc(at)gga 3’

5’...ttc tat tct cac agt gca cag gc(ag)gt(ct)ca(ag)ctg cag cag ct(ct)ggg gc 3’

5’...ttc tat tct cac agt gca cag ga(ag)gtg aag ct(gt)(gc)t(gc)gag tct gg(ag)gga 3’

3’VH Sfi I

5’...ga acc agt act cca ggc ctg agg ggc cgc aga gac agt gac cag agt ccc 3’

5’...ga acc agt act cca ggc ctg agg ggc cga gga gac ggt gac tga ggt tcc 3’

5’...ga acc agt act cca ggc ctg agg ggc cga gga gac(gt)gt ga(gc)(ca)gt gga(gc)cc...3’

5’V_L Sal I

5’...gta ctc cag tcg acg aca ttg tg(ca)tg(at)t(ac)c agt ctc c...3’

5’...gta ctc cag tcg acg ata tcc aga tga c(ac)c a(ga)a ct(ac)c...3’

5’...gta ctc cag tcg ac(cg)aaa ttg t(tg)c tca ccc agt ctc c...3’

3’V_L Not I

5’...aag gaa aaa agc ggc cgc ttt(tc)a(tg)(tc)tc cag ctt ggt...3’

5’...aag gaa aaa agc ggc cgc ttt(tg)a(tg)ctc caa ctt gt(gt)...3’

The amplified heavy chain variable region was purified and subsequently degraded with the restriction enzymes Sfi I and ApaL I. The degradation products were purified and ligated with DNA of phagomidium vector pHG-1m (Heber Biotec S.A., Cuba) (FIG. 5). The E.coli strain TG1 was transformed with the resulting gene construct by electroporation. Transformed bacterial population composition 2.3X10⁴A half library of medium-sized heavy chain variable regions of individual members, said heavy chain variable regions being derived from the 14F7 hybridoma. Plasmid DNA from the half-library was purified and subsequently degraded with Not I and Sal I enzymes. This DNA was ligated to 4 light chain variable region aggregates previously degraded with the same enzyme, respectively. The TG1 strain bacteria were transformed with these new gene constructs to obtain 4 independent libraries. One of them was 1.6x10 in size comprising the heavy and light chain variable regions derived from the 14F7 hybridoma⁶A mini library of limited diversity for an individual. The other three are strand exchange libraries that incorporate multiple pools of light chain variable regions previously obtained from mouse (kappa isotype) and human (kappa and lambda isotypes) and are 1.0x10 in size, respectively⁷、1.4x10⁶And 1.2x10⁶And (4) each member.

Example 6: isolation of clones producing antibody fragments against N-glycolyl GM 3.

Phage carrying antibody fragments from 92 clones randomly picked from each library were generated in wells of a microtiter plate according to the method described previously (Marks, j.d. et al, j.mol.biol.222, 581-597.1991). The specificity was estimated by performing ELISA on polyvinyl chloride plates (Costar, USA) coated with 1. mu.g/ml of N-glycolyl GM3 and N-acetyl GM 3. Bound phage were detected with anti-M13 monoclonal antibody conjugated to peroxidase (Amersham, Sweden). In addition to characterization of randomly selected individual clones, selection of phage with the ability to bind to N-glycolyl GM3 was performed from a mixture of phage generated from the population of cells forming each library. For this purpose, the helper phage M13K 07 was used to generate phages, which were purified by precipitation with polyethylene glycol according to the mentioned method and contacted with N-glycolyl GM3 bound to a plastic surface (Inmuno tubes, Nunc, Denmark). The tube was washed thoroughly and the phage eluted in the presence of 100mmol/l triethylamine. The eluate was neutralized and used to infect bacteria of the E.coli strain TG 1. Antibody-bearing phages were generated from the infected bacterial population and the phages were purified again, and the purified phages were used as starting material for a new round of screening under the same conditions as those described. After 4 rounds of screening, 92 phage-infected bacterial clones generated from each library were randomly picked and identified with N-glycolyl GM3 in a manner similar to that described for the evaluation of randomly picked clones from the library.

Based on a direct analysis of the clones that generated the library of clones (isolated from phages carrying antibody fragments that recognize N-glycolyl GM 3), it was found that the clones differed in frequency among the three strand-exchange libraries. There were no positive clones in the mini-library formed exclusively from the variable region derived from the 14F7 hybridoma. After multiple rounds of selection, additional clones were isolated from the strand-exchange library and none of the clones were found in the mini-library of original sequences. The following table shows the number of clones producing antibody fragments recognizing N-glycolyl GM3 and the frequency with which they appear in the total number of clones from each library analyzed. Only one clone produced phage that exhibited cross-reactivity with N-acetyl GM 3.

One clone had cross-reactivity with N-acetyl GM 3.

Example 7 characterization of an antibody fragment recognizing N-glycolyl GM3 derived from 14F 7.

The variable regions of the 11 antibody fragments were completely sequenced using an automated sequencer ALFExpress II (Pharmacia, Sweden) by using oligonucleotides that hybridized to regions of the pHG-1m vector flanking the 5 'and 3' ends of the sequences encoding the inserted antibody fragments. It was determined that the sequence of all heavy chain variable regions corresponded to one reported heavy chain variable region of 14F7, with changes occurring only in the framework region 1 and framework region 4 regions, presumably introduced by the use of degenerate oligonucleotides in PCR. The sequence comprising the 3 CDRs was not changed. In contrast, the light chain variable region sequences of all 11 clones differed from that of 14F7 and were divided into 5 different sequence groups, 2 murine-derived groups and 1 human-derived (kappa isotype) groups and 2 other lambda isotype-derived human-derived groups. One representative clone was selected for each sequence set for subsequent characterization.

The following table shows the differences between the light chain variable regions of selected clones in terms of their identity of origin, their origin and isotype, and subgroup classification (according to Kabat classification).

	Source of VL	Isoforms	And V of 14F7LUniformity of	Subgroup
					AcM 14F7	Murine animal	κ	-	V
ScFv 2Am	Murine animal	κ	59.46％	III
					ScFv 3Fm	Murine animal	κ	59.81％	V
ScFv 7Bhk	Human being	κ	60.74％	I
					ScFv 7Ahl	Human being	λ	35.45％	I
ScFv 8Bhl	Human being	λ	37.04％	III

The differences between the cloned light chain variable region and the light chain variable region of the 14F7 antibody reported initially were located over the entire sequence including the three complementarity determining regions and are shown below (FIG. 6)

Fr1

14F7Mab D L V L T Q S P A T L S V T P G D S V S F S C

2Am scFv . I . M F . . . . S . A . S L . Q R A T I . .

3Fm scFv . I Q M . . T . S S . . A S L . . R . T I . .

7Bhk scFv . I Q M . . T . S S . . A S V . . R . T I T .

7Ahl scFv Q S V V . . P . S A . G G . . Q .L T I . .

8B hl scFv S S E L . . D . . V . . A L . Q T . R I T .

CDR1

R A S Q S I S N N L H

. . . . S V S S . S Y S Y M .

. . . . D . . . Y . N

. . . . . . . S F . N

T G T S S D V G G Y . H V S

Q G D S L R . Y Y A S

Fr2

W Y Q Q R T H E S P R L L I K

. . . . K P G Q P . K . . . .

. . . . K P D G T V K . . . V

. . . . K P G K A . K . . . Y

. . . . H P G K A . K . M . Y

. . . . K P G Q A . V . V . Y

CDR2

Y A S Q S I S

. . . N L E .

. T . R L H .

A . . N L Q .

D V . K R P .

G K N N R P .

Fr3

G I P S R F S G S G S G T D F T L S I I A V E T E D F G M Y F C

. V . A . . . . . . . . . . . . . N . H P . . E . . A A T . Y .

. V . . . . . . . . . . . . Y S . T . S N L . Q . . I A T . . .

. V . . . . . . R . . . . . . . . T . S S L Q P . . . A A . Y .

. V . H . . . . . K . . N T A S . T V S G L Q A . . E A V . Y .

. . . D . . . . . S . . N T A S . T . T G A Q A . . E A D . Y .

CDR3

Q Q S N R W P L T F

. H . R D V . . . .

. . G . T L . P . .

. . G Y T T . . . .

S S Y A G S N N . V .

N S R D S S G N H V V .

Fr4

G A G T K L E L K

. . . . . . . I .

. Q . . . . . . .

. G . . . V T V L

. G . . . . T V L

The specificity of binding was determined by ELISA performed as described using a panel of related gangliosides comprising various N-acetyl gangliosides and one N-glycolylated ganglioside in addition to N-glycolylated GM3 and N-acetyl GM 3. FIG. 7 shows the recognition of different antigens by 5 clones producing characteristic antibody fragments. Binding to the target antigen, N-glycolyl GM3, was only detected.

The cloning was induced to produce soluble antibody fragments (in vitro in phage) in the presence of 1mmol/l isopropyl-thiogalactopyranoside. The supernatant was collected and the periplasmic fraction was obtained according to the established method (de Haard, H.J.biol.chem.274, 18218-18230, 1999). The recognition activity of N-glycolyl GM3 was measured on culture supernatants and periplasmic fractions by ELISA performed on plates coated with the antigen, in a similar manner to that described, with the following changes. Bound antibody fragments were detected using 10. mu.g/ml of 1E10 monoclonal antibody (directed against the c-myc peptide fused to the antibody fragment on the gene construct) and anti-mouse immunoglobulin conjugate (Sigma, USA). Antibody fragments were purified from the periplasmic fraction in a single Affinity chromatography with metal ions using the matrix HIS-Select HC Nickel Affinity Gel (Sigma, USA) according to the manufacturer's instructions. FIG. 8 shows the activity of purified soluble fragments as determined by ELISA.

Brief Description of Drawings

FIG. 1: amino acid sequences of the vh (a) and vk (b) variable regions of 14F7 Mab and the respective most homologous human sequences. CDRs are underlined and amphipathic regions and potential T epitopes are shaded. The humanized variable region sequences obtained by mutation are also shown.

FIG. 2: the recognition characteristics of the 14F7Q antibody were measured by competition ELISA.

An X axis: indicates the antibody concentration.

Y-axis: absorbance measured at 492 nm.

The ior C5 monoclonal antibody was used as a negative control.

FIG. 3: effect of the 14F7 monoclonal antibody on growth of murine tumors (melanoma B16 tumor cell line).

FIG. 4: A) immunohistochemical study of the effect of the 14F7 monoclonal antibody on angiogenesis. Arrows indicate immunostained vessels.

B) Inhibition of angiogenesis induced by the 14F7 monoclonal antibody.

FIG. 5: a diagram showing a pHG-1m vector is shown.

FIG. 6: the VL region of the 14F7 monoclonal antibody and the protein sequence of the selected Fv single-chain fragment. Dots indicate homology to the VL sequence of 14F 7.

FIG. 7: recognition specificity of phage carrying antibody fragments. It was observed that the exposed fragments in the phage inhibited the binding of 14F7 Ab to gangliosides. Plates were coated with N-glycolyl GM3 and the phage incubated with virus particles and various concentrations of 14F 7.

FIG. 8: identification of purified soluble fragments for N-glycolyl GM 3.

Claims (5)

1. A single chain Fv fragment derived from the murine 14F7 monoclonal antibody, wherein said murine 14F7 monoclonal antibody is produced by the hybridoma deposited with the accession number ECACC98101901, characterized in that said fragment comprises the sequence of the heavy chain variable region of said 14F7 monoclonal antibody and the light chain variable region whose sequence is as shown below:

Fr1

D I V M F Q S P A S L A V S L G Q R A T I S C

CDR1

R A S Q S V S S S S Y S Y M H

Fr2

W Y Q Q K P G Q P P K L L I K

CDR2

Y A S N L E S

Fr3

G V P A R F S G S G S G T D F T L N I H P V E E E D A A T Y Y C

CDR3

Q H S R D V P L T F

Fr4

G A G T K L E I K。

2. a single chain Fv fragment derived from the murine 14F7 monoclonal antibody, wherein said murine 14F7 monoclonal antibody is produced by the hybridoma deposited with the accession number ECACC98101901, characterized in that said fragment comprises the sequence of the heavy chain variable region of said 14F7 monoclonal antibody and the light chain variable region whose sequence is as shown below:

Fr1

D I Q M T Q T P S S L S A S L G D R V T I S C

C DR1

R A S Q D I S N Y L N

Fr2

W Y Q Q K P D G T V K L L I V

CDR2

Y T S R L H S

Fr3

G V P S R F S G S G S G T D Y S L T I S N L E Q E D I A T Y F C

CDR3

Q Q G N T L P P T F

Fr4

G A G T K L E L K。

3. a single chain Fv fragment derived from the murine 14F7 monoclonal antibody, wherein said murine 14F7 monoclonal antibody is produced by the hybridoma deposited with the accession number ECACC98101901, characterized in that said fragment comprises the sequence of the heavy chain variable region of said 14F7 monoclonal antibody and the light chain variable region whose sequence is as shown below:

Fr1

D I Q M T Q T P S S L S A S V G D R V T I T C

C DR1

R A S Q S I S S F L N

Fr2

W Y Q Q K P G K A P K L L I Y

CDR2

A A S N L Q S

Fr3

G V P S R F S G R G S G T D F T L T I S S L Q P E D F A A Y Y C

CDR3

Q Q G Y T T P L T F

Fr4

G Q G T K L E L K。

4. a single chain Fv fragment derived from the murine 14F7 monoclonal antibody, wherein said murine 14F7 monoclonal antibody is produced by the hybridoma deposited with the accession number ECACC98101901, characterized in that said fragment comprises the sequence of the heavy chain variable region of said 14F7 monoclonal antibody and the light chain variable region whose sequence is as shown below:

Fr1

Q S V V T Q P P S A S G G P G Q S L T I S C

C DR1

T G T S S D V G G Y N H V S

Fr2

W Y Q Q H P G K A P K L M I Y

CDR2

D V S K R P S

Fr3

G V P H R F S G S K S G N T A S L T V S G L Q A E D E A V Y Y C

CDR3

S S Y A G S N N L V F

Fr4

G G G T K V T V L。

5. a single chain Fv fragment derived from the murine 14F7 monoclonal antibody, wherein said murine 14F7 monoclonal antibody is produced by the hybridoma deposited with the accession number ECACC98101901, characterized in that said fragment comprises the sequence of the heavy chain variable region of said 14F7 monoclonal antibody and the light chain variable region whose sequence is as shown below:

Fr1

S S E L T Q D P A V S V A L G Q T V R I T C

C DR1

Q G D S L R S Y Y A S

Fr2

W Y Q Q K P G Q A P V L V I Y

CDR2

G K N N R P S

Fr3

G I P D R F S G S S S G N T A S L T I T G A Q A E D E A D Y Y C

CDR3

N S R D S S G N H V V F

Fr4

G G G T K L T V L。