Humanized antibodies against globo H and their use in cancer therapy

Abstract

A humanized anti-globo H antibody, or an scFv or Fab fragment thereof, comprising a heavy chain variable domain having three complementarity regions consisting of HCDR1, HCDR2, and HCDR3, and a light chain variable domain having three complementarity regions consisting of LCDR1, LCDR2, and LCDR3, wherein the HCDR1 sequence is GYISSDQILN (sequence number 4), the HCDR2 sequence is RIYPVTGVTQYXHKFVG (sequence number 5, where X is any amino acid), the HCDR3 sequence is GETFDS (sequence number 6), the LCDR1 sequence is KSNQNLLX'SGNRRYZLV (sequence number 7, where X' is F, Y, or W, and Z is C, G, S, or T), the LCDR2 sequence is WASDRSF (sequence number 8), and the LCDR3 sequence is QQHLDIPYT (sequence number 9).

Description

[Technical Field]

[0001] The present invention relates to humanized antibodies that specifically bind to globo H. The present invention also relates to methods for treating and / or diagnosing cancer using these antibodies. [Background technology]

[0002] Globo H (Fuc-α1,2-Gal-β1,3-GalNAc-β1,3-Gal-α1,4-Gal-β1,4-Glc-β1,1-Cer) is a hexasaccharide that belongs to a number of tumor-associated carbohydrate antigens that are overexpressed on the surface of various epithelial cancer cells, including breast, colon, ovarian, pancreatic, lung, and prostate cancer cells. In addition to globoH, other known carbohydrate antigens include Tn (GalNAc-α-O-Ser / Thr), sialyl Tn (Neu5Ac-α2,6-GalNAc-α-O-Ser / Thr, STn), GD2, GD3, GD3L, fucosyl-GM1, Lewis antigens (Lex, Ley, Lea, sialyl Lex, sialyl Lea), and TF (Gal-β1,3-GalNAc-α-O-Ser / Thr), which are also used as target antigens for cancer immunotherapy (Susan F Slovin et al., Carbohydrate Vaccines as Immunotherapy for Cancer, Immunology and Cell Biology (2005) 83, 418-428; Zhongwu Guo and Qianli Wang, Recent Development in Carbohydrate-Based Cancer Vaccines, Curr. Opin. Chem. Biol. 2009). December;13(5-6):608-617;Therese Buskas et al., Immunotherapy for Cancer:Synthetic Carbohydrate-based Vaccines,Chem.Commun.(Comb)2009 September 28;(36):5335-5349).

[0003] However, most carbohydrate antigens are often tolerated by the immune system, resulting in limited immunogenicity. Furthermore, the production of antibodies against a specific immunogen typically involves coordinated interactions between two types of lymphocytes: B cells and helper T cells. Globo H alone cannot activate helper T cells, which is also attributed to its low immunogenicity. Therefore, immunization with globo H alone often results in low titers of immunoglobulin M (IgM), an inability to class switch to immunoglobulin G (IgG), and ineffective antibody affinity maturation.

[0004] Recently, it has been demonstrated that the addition of an appropriate adjuvant can induce antibody responses to globoH, including class switching from IgM to IgG. Therefore, globoH is a promising therapeutic target for cancer vaccination. This approach is currently undergoing various stages of clinical trials for various cancers, including breast, ovarian, prostate, and lung cancer. Summary of the Invention [Problem to be solved by the invention]

[0005] Although antibodies against globo H have shown promise in cancer diagnosis and treatment, there remains a need for better anti-globo H antibodies. [Means for solving the problem]

[0006] Embodiments of the present invention relate to humanized globo H antibodies that specifically bind to globo H and methods of using such antibodies in the diagnosis and / or treatment of cancer. By specifically binding to globo H, the antibodies of the present invention can be used to diagnose and / or treat cancers that overexpress globo H, including various epithelial cancers.

[0007] One aspect of the present invention relates to a humanized anti-globo H antibody. According to one embodiment of the present invention, the humanized anti-globo H antibody, or its scFv or Fab fragment, comprises a heavy chain variable domain having three complementary regions consisting of HCDR1, HCDR2, and HCDR3, and a light chain variable domain having three complementary regions consisting of LCDR1, LCDR2, and LCDR3, wherein the sequence of HCDR1 is GYISSDQILN (SEQ ID NO: 4), the sequence of HCDR2 is RIYPVTGVTQYXHKFVG (SEQ ID NO: 5, where X is any amino acid), the sequence of HCDR3 is GETFDS (SEQ ID NO: 6), the sequence of LCDR1 is KSNQNLLX'SGNRRYZLV (SEQ ID NO: 7, where X' is F, Y, or W, and Z is C, G, S, or T), the sequence of LCDR2 is WASDRSF (SEQ ID NO: 8), and the sequence of LCDR3 is QQHLDIPYT (SEQ ID NO: 9).

[0008] According to some embodiments of the present invention, X in SEQ ID NO:5 is asparagine or glutamine. According to some embodiments of the present invention, X' in SEQ ID NO:7 is tryptophan. According to some embodiments of the present invention, Z in SEQ ID NO:7 is serine or threonine.

[0009] One aspect of the present invention relates to a method for diagnosing, preventing, and / or treating cancer. The method according to one embodiment of the present invention comprises administering an effective amount of an anti-globo H antibody to a subject in need thereof. An effective amount is an amount that can achieve a desired therapeutic result. Those skilled in the art will recognize that an effective amount may vary based on the patient's age, sex, weight, condition, etc. Such an effective amount is routinely determined based on the patient and condition. Those skilled in the art will be able to determine an effective amount without undue experimentation. The cancer is breast cancer, colon cancer, ovarian cancer, pancreatic cancer, lung cancer, liver cancer, or prostate cancer. [Brief explanation of the drawings]

[0010] [Figure 1] FIG. 1 shows a flow chart illustrating the method for obtaining antibody clones against globo H.

[0011] [Figure 2] FIG. 2 shows a schematic diagram of murine, chimeric, and humanized anti-globo H antibodies.

[0012] [Figure 3] FIG. 3 shows the binding curve of the chimeric anti-globo H antibody, illustrating the high binding affinity with a sub-nanomolar binding constant.

[0013] [Figure 4] FIG. 4 shows a model of the variable domain of an anti-globo H antibody used in computer modeling.

[0014] [Figure 5] FIG. 5 shows the results of various anti-globo H antibodies binding to MCF7 using flow cytometry.

[0015] [Figure 6] Figure 6 shows the results of various amino acid substitutions of R50 in the heavy chain variable domain (VH). The results show that substitution of this R50 with any other amino acid significantly reduces the binding affinity of anti-globo H antibodies, indicating that this R50 is essential for antigen-antibody interaction.

[0016] [Figure 7A] Figure 7A shows the results of various amino acid substitutions at W27 in the light chain variable domain (VL). The results show that substitution of W27 with other aromatic amino acids results in a slight decrease in antibody affinity. However, substitution with non-aromatic amino acids significantly reduces the binding affinity of anti-globo H antibodies.

[0017] [Figure 7B]Figure 7B shows the results of various amino acid substitutions at N55 in the heavy chain variable domain (VH). The results indicate that substitution of N55 with many other amino acids is tolerated at this site. Among the various amino acid substitutions, glutamine (Q) substitution actually produced a more robust antibody.

[0018] [Figure 8A] Figure 8A shows the primary sequence alignment of the framework regions of the VH segments of various anti-globo H antibodies (GBH): murine GBH(M), humanized GBH(H), back-mutated humanized GBH(B1), further refined GBH(Re2), and GBH(B11).

[0019] [Figure 8B] Figure 8B shows the primary sequence alignment of the framework regions of the VL segments of various anti-globo H antibodies (GBH): murine GBH(M), humanized GBH(H), back-mutated humanized GBH(B1), further purified GBH(Re2), and GBH(B13).

[0020] [Figure 9A] Figure 9A shows the sequences of HCDR1, HCDR2, and HCDR3 of the consensus clone GBH(C) based on the results from alanine scanning and various amino acid substitutions in the CDR regions. Also shown is an exemplary clone GBH(B11) containing a glutamine (Q) at the non-critical residue X in HCDR2.

[0021] [Figure 9B] Figure 9B shows the sequences of LCDR1, LCDR2, and LCDR3 of the consensus clone GBH(C) based on the results from alanine scanning and various amino acid substitutions in the CDR regions. Also shown is an exemplary clone GBH(B13), which contains a tryptophan (W) at the aromatic residue X of LCDR1 and a threonine (T) at the flexible residue Z.

[0022] [Figure 10]Figure 10 shows the recognition of globo H-expressing breast cancer cells by anti-globo H antibodies. Panels A-C show immunofluorescence images of antibody recognition of MCF7 cells. Panels A'-C' show corresponding images of cells visualized by light microscopy.

[0023] [Figure 11] Figures 11(A)-11(G) show the FACS results, demonstrating the expression of globo H in various cancer cells as detected by anti-globo H antibody. Figure 11(A) shows that MCF7 breast cancer cells express both Her2 antigen (upper panel) and globo H antigen (lower panel). Figure 11(B) shows that HCC1428 hepatocellular carcinoma cells express both Her2 antigen (upper panel) and globo H antigen (lower panel). Figure 11(C) shows that BT474 breast cancer cells express Her2 (upper panel) but not globo H (lower panel). Similarly, Figure 11(D) shows that Capan-1 pancreatic cancer cells express globo H. Figure 11(E) shows that A-431 squamous cell carcinoma cells express globo H. Figure 11(F) shows that NCI-N87 gastric cancer cells express globo H. FIG. 11(G) shows that HT-29 colorectal cancer cells express low levels of globoH.

[0024] [Figure 12] Figures 12A to 12F show the results of antibody-dependent cellular cytotoxicity (ADCC) mediated by anti-globo H antibodies. Figure 12A shows the results for MCF7 cells. Figure 12B shows the results for HCC1428 cells. Figure 12C shows the results for BT474 cells. Figure 12D shows the results for Capan-1 cells. Figure 12E shows the results for NCI-N87 cells. Figure 12F shows the results for A431 cells. These results indicate that expression of globo H is required for ADCC mediated by anti-globo H antibodies.

[0025] [Figure 13]Figures 13A-13E show the results of complement-dependent cytotoxicity (CDC) mediated by anti-globo H antibodies. Figure 13A shows the results for MCF7 cells. Figure 13B shows the results for HCC1428 cells. Figure 13C shows the results for BT474 cells. Figure 13D shows the results for Capan-1 cells. Figure 13E shows the results for NCI-N87 cells. These results indicate that expression of globo H is required for CDC mediated by anti-globo H antibodies.

[0026] [Figure 14] Figure 14 shows the results of the inhibition of anti-globo H antibody-mediated cytotoxicity by synthetic glycans. The results show that globo H can compete with anti-globo H antibodies and inhibit anti-globo H antibody-mediated cytotoxicity in a dose-dependent manner, whereas Lewis b tetrasaccharide cannot. This specific competition indicates that anti-globo H-mediated cytotoxicity is due to binding to globo H expressed on the cell surface.

[0027] [Figure 15] FIG. 15 shows that humanized anti-globo H antibodies can inhibit tumor growth in a dose-dependent manner in a predictive model.

[0028] [Figure 16] FIG. 16 shows that humanized anti-globo H antibodies can dose-dependently inhibit tumor growth in a therapeutic model.

[0029] [Figure 17] FIG. 17 shows the sequences of the GBH light chain variable domains of various anti-globo H antibodies (SEQ ID NO: 1, SEQ ID NO: 2, and SEQ ID NO: 3). DETAILED DESCRIPTION OF THE INVENTION

[0030] definition Unless otherwise defined, scientific and technical terms used herein shall have the meanings that are commonly understood by those skilled in the art.Furthermore, unless otherwise required by context, singular terms include plurals, and plural terms include singulars.In general, the nomenclature used in connection with the described cell and tissue culture, molecular biology, and protein and oligo- or polynucleotide chemistry and hybridization is well known and commonly used in the art.

[0031] Standard techniques are used for recombinant DNA, oligonucleotide synthesis, tissue culture, and transformation (e.g., electroporation, lipofection). Enzymatic reactions and purification techniques are carried out according to manufacturer's specifications, or as commonly practiced in the art, or as described herein. The above techniques and procedures are generally carried out according to conventional methods well known in the art, and as described in various general and more specific references cited and discussed throughout this specification. See, for example, Sambrook et al., Molecular Cloning: A Laboratory Manual (3rd ed., Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY (2001)), which is incorporated herein by reference.

[0032] As used herein, the term "chimeric antibody" refers to an antibody that contains sequences from multiple sources (eg, species).

[0033] As used herein, the term "humanized antibody" refers to an antibody in which minimal portions of a non-human antibody have been introduced into an otherwise human antibody.

[0034] As used herein, the term "human antibody" refers to an antibody in which substantially all portions of the protein are substantially non-immunogenic in humans and have only minor sequence variations or mutations.

[0035] As used herein, the term "antigen-binding fragment" refers to a fragment of an antibody that retains the ability to bind to an antigen. Such antigen-binding fragments may include scFv, Fab, F(ab')2, etc.

[0036] The term "CDR region" or "CDR" is intended to refer to the hypervariable region of an immunoglobulin heavy or light chain, as defined by Kabat et al., 1991 (Kabat, E. A. et al., (1991) Sequences of Proteins of Immunological Interest, 5th Edition. U.S. Department of Health and Human Services, Public Service, NIH, Washington) and subsequent editions. Antibodies typically contain three heavy chain CDRs and three light chain CDRs. The term CDR(s) is used herein to refer to one or several or even the entirety of these regions, as the case may be, that contain most of the amino acid residues involved in binding to the antigen or epitope it recognizes.

[0037] The term "set of CDRs" referred to herein includes CDR1, CDR2 and CDR3. Thus, a set of HCDRs refers to HCDR1, HCDR2 and HCDR3 (HCDR refers to heavy chain CDR), and a set of LCDRs refers to LCDR1, LCDR2 and LCDR3 (LCDR refers to light chain CDR). Unless otherwise specified, a "set of CDRs" may include HCDRs and / or LCDRs.

[0038] Two amino acid sequences are "homologous" if there is partial or complete identity between their sequences. For example, 85% homology means that when the two sequences are aligned for maximum matching, 85% of the amino acids are identical. Gaps (in either of the two matching sequences) are allowed when maximizing matching, with gap lengths of 5 or less being preferred, and 2 or less being more preferred. It should be recognized that there may be regions of different homology within two orthologous sequences. For example, functional regions of mouse and human orthologs may have a higher degree of homology than non-functional regions.

[0039] An antigen-binding site is generally formed by variable heavy (VH) and variable light (VL) immunoglobulin domains, with the antigen-binding interface formed by six surface polypeptide loops called complementarity-determining regions (CDRs). Each VH (HCDR1, HCDR2, HCDR3) and each VL (LCDR1, LCDR2, LCDR3) contains three CDRs along with framework regions (FRs).

[0040] An antibody antigen-binding site, consisting of a VH domain and a VL domain, is typically formed by six loops of the polypeptide: three from the light chain variable domain (VL) and three from the heavy chain variable domain (VH). Analysis of antibodies of known atomic structure can reveal the relationship between the sequence and three-dimensional structure of the antibody binding site.

[0041] Sequence-structure relationship studies can be used to predict the residues of antibodies with known sequences but unknown three-dimensional structures, and these residues are important for maintaining the three-dimensional structure of the CDR loops, thus maintaining binding specificity. In structural approaches, antibody molecule models can be created using freely available or commercially available packages such as WAM. Protein visualization and analysis software packages such as Insight II (Accelrys, Inc.) or DeepView can then be used to evaluate possible substitutions at each position in the CDR. This information can then be used to make substitutions that are likely to have minimal or beneficial effects on activity.

[0042] Techniques for making amino acid substitutions within the sequences of CDRs and antibody VH or VL domains are available in the art.

[0043] Detailed Description Embodiments of the present invention relate to anti-globo H antibodies and their use in the diagnosis and treatment of cancer. The practice of the present invention employs techniques, including conventional techniques of cell biology, cell culture, antibody technology, and genetic engineering, which are within the ordinary skill of the art.

[0044] The following examples describe the development of humanized anti-globo H antibodies and their use in cancer diagnosis and treatment. Those skilled in the art will recognize that these examples are for illustrative purposes only and are not intended to limit the scope of the invention.

[0045] As described above, globo H is a hexasaccharide that is overexpressed on the surface of epithelial cancer cells, including breast, colon, ovarian, pancreatic, lung, and prostate cancer cells. Antibodies against globo H have been shown to be useful in the treatment and diagnosis of such cancers. For use in patients, the antibody should have minimal or no side effects, such as no undesired immunological responses. Embodiments of the present invention relate to humanized anti-globo H antibodies. These humanized antibodies have good binding efficiency and no or minimal undesired immunological responses.

[0046] According to embodiments of the present invention, a general method for producing a humanized anti-globo H antibody includes obtaining a hybridoma that produces a monoclonal antibody against globo H, obtaining CDR sequences from the hybridoma, and cloning the CDR sequences into human framework sequences to produce a humanized antibody. The humanized antibody may be further optimized, for example, to improve the sequences of the framework regions and / or the CDR sequences. Figure 1 shows a flowchart outlining a method according to one embodiment of the present invention for obtaining a humanized anti-globo H antibody.

[0047] Methods for various procedures are known in the art. The following specific examples describe exemplary embodiments. However, those skilled in the art will recognize that modifications or variations are possible without departing from the scope of the invention.

[0048] Molecular cloning of the V region of anti-globo H antibodies. First, anti-globin H hybridomas (e.g., mouse GBH hybridomas) were generated. Such hybridomas can be generated using standard protocols for producing monoclonal antibodies. Then, total RNA from the hybridomas was isolated using, for example, TRIzol® reagent. Next, a single-stranded cDNA synthesis kit (Superscript III) and oligo(dT 20 ) primer or Ig-3′ constant region primer, cDNA was synthesized from total RNA.

[0049] The heavy and light chain variable regions of immunoglobulin genes were then cloned from the cDNA. For example, the VH and VL variable regions of anti-globulin H mAb were amplified from mouse GBH hybridoma cDNA by PCR using the mouse Ig-5' primer set (Novagen). The PCR product could be directly cloned into an appropriate vector (e.g., pJET1.2 vector) using the CloneJet™ PCR Cloning Kit (Ferments). The pJET1.2 vector contains a lethal insert that survives selection conditions only if the desired gene is cloned into this lethal region. This facilitates the selection of recombinant colonies. Finally, recombinant colonies were screened for the desired clones, and the DNA of those clones was isolated and sequenced. Immunoglobulin (IG) nucleotide sequences could be analyzed on the International ImMunoGeneTics Information System (IGMT) website.

[0050] Antibody expression and purification For antibody production, the isolated clones can be expressed in any suitable cells. As an example, F293 cells (Life Technologies) were transfected with an anti-globo H mAb expression plasmid and cultured for 7 days. Anti-globo H antibodies were purified from the culture medium using a Protein A affinity column (GE). Protein concentration can be measured using a Bio-Rad protein assay kit using procedures known in the art or according to the manufacturer's instructions, and analyzed by 12% SDS-PAGE.

[0051] ELISA assay Antibody affinity can be assessed by any suitable method known in the art, such as ELISA or Biacore. For example, globoH-NH2 (Oligotech) diluted in sodium carbonate buffer (pH 9.5) was coated onto a 96-well plate overnight at 4°C. After blocking (e.g., with BSA), two-fold serial dilutions of anti-globoH antibody were added to the wells and incubated for 2 hours at 37°C. After binding, goat anti-human IgG-HRP (1:15,000) was added and incubated for 1 hour at 37°C. Color development was then performed using 3,3',5,5'-tetramethylbenzidine (TMB) substrate, and the reaction was stopped by adding 1N H2SO4. The extent of antigen-antibody binding was determined by reading the plate, i.e., measuring absorbance at 450-655 nm using an ELISA reader (BioRad Model 680). Data can be analyzed using appropriate software, such as GraphPad Prism 5 software.

[0052] Humanization of anti-globo H antibody An embodiment of the present invention relates to the humanization of anti-globo H antibodies. Figure 2 shows a schematic diagram of chimeric and humanized antibodies. A chimeric antibody has variable and constant regions from different sources (e.g., different species). For example, the variable region of a mouse anti-globo H antibody cloned from the above procedure can be grafted onto the constant region of a human antibody to generate a chimeric antibody.

[0053] Figure 3 shows a binding assay of a chimeric antibody according to one embodiment of the present invention. The binding assay was performed by ELISA as described above. From this binding curve, the binding constant of this chimeric antibody can be estimated to be better than 1.0 nM, indicating that, as expected, grafting the antibody variable domain onto the constant domain of a human antibody does not impair binding.

[0054] Referring to Figure 2, a humanized antibody comprises CDRs from one species and framework and constant regions from a human immunoglobulin. An exemplary process for generating a humanized antibody can be as follows.

[0055] 1) Selection of human V-region frameworks. The frameworks from human immunoglobulins with the highest homology to the framework regions of the cloned variable regions from mice were selected from the IMGT database. Based on the homology comparison, the heavy chain framework of the VH3 subgroup and the light chain framework of the Vk1 subgroup were selected for the humanized anti-globo H mAb (GBH), respectively.

[0056] 2) Construction of CDR-grafted anti-globo H antibodies. A human framework (VL kappa subgroup I and VH subgroup III) with the six complete mouse CDR sequences was constructed by PCR and then subcloned into an antibody expression vector. Any suitable vector known in the art can be used. This generates hybrid variable regions (VH and VL).

[0057] 3) Reversion mutation. The CDR grafting into the framework results in variable domains (VH and VL) from different sources. Such heterologous domains may not have optimal sequences. Therefore, the affinity of the antibody may not be the best. To improve binding affinity, some amino acids can be backmutated to other species.

[0058] As shown in Figure 4, important amino acid residues that may affect antibody binding may be analyzed by computer modeling. Modeling may consider the upper core region and interface region (e.g., within a 5 Å region). Modeling may also apply prior knowledge based on successful cases. Based on modeling, amino acid substitutions may be performed (i.e., back mutations), for example, to replace amino acid residues with corresponding amino acid residues from the original species. The mutant antibodies may then be evaluated for binding to select improved antibodies.

[0059] Therefore, clone GBH(B1) was selected from the first round of backmutation. Clones GBH(Re1) and GBH(Re2) were generated from further mutations based on the B1 clone with the following additional considerations: (i) avoiding most structurally conserved strands of the Fv β-barrel; (ii) ranking resurfacing sites (murine amino acids) by relatively high surface accessibility (e.g., greater than 30%); and (iii) classifying commonly reported risk site frameworks.

[0060] As shown in Figure 5, the chimeric antibody (cGBH) binds well to MCF7 cells. However, humanization significantly reduces this binding (see GBH(HH)). After an initial round of backmutation, GBH(B1) restored most of the antibody's binding activity. Further mutations, GBH(Re1Re2) and GBH(re2Re2), significantly improved binding affinity. The sequences of the light and heavy chain variable domains of these mutants are shown in Figures 8A and 8B.

[0061] 4) CDR affinity optimization and alanine scanning of key amino acid residues. In addition to the above-mentioned backmutations in the framework regions, antibody affinity can be further improved by optimizing the CDR sequences. Based on computer modeling and computer docking of the globo H antigen and the GBH(Re2Re2) antibody, selective CDR mutant clones were generated by site-directed mutagenesis. The binding affinity of the mutant clones can be analyzed by any suitable method, such as ELISA, Biacore, or ForteBio.

[0062] Table 1 shows examples of alanine scanning results using GBH(Re2Re2) as the starting antibody. These results show that alanine substitutions at four sites (I33 of CDRH1, R50 of CDRH2, E96 of CDRH3, and W27 of CDRL1) significantly reduced antibody affinity, indicating that these four residues are important for antibody binding. On the other hand, alanine substitutions at other sites (e.g., N58 of CDRH2, N28 of CDRL1, D93 of CDRL1, and I94 of CDRL3) did not significantly affect antibody binding, indicating that these residues are not important for antigen-antibody interactions. [Table 1]

[0063] The critical residues from the alanine scanning above can be further examined to confirm their importance. As shown in Figure 6, R50 of CDRH2 is a critical residue, and substitution with other amino acids essentially abolished binding.

[0064] On the other hand, some residues clearly identified by alanine scanning may tolerate similar amino acids. For example, as shown in Figure 7A, W27 of LCDR1 (see Figure 9B) is important for binding, but substitution with other aromatic amino acid residues (Y and F) was well tolerated, whereas other amino acid substitutions at this site essentially abolished antibody binding. These results indicate that antibody binding activity is substantially maintained as long as an amino acid residue with an aromatic side chain is present at this position.

[0065] On the other hand, amino acids that are not critical from the alanine scanning results can be further optimized. For example, residue 58 (shown as X' in HCDR2 of GBH(C) in Figure 9A) is not critical, and many amino acids are well tolerated at this position (Figure 7B). Among them, glutamine (Q) confers the highest binding activity. Similarly, residue 32 (shown as Z in LCDR1) (Figure 9B) can tolerate several amino acids with similar polarity. In the exemplary antibody GBH(B13), this Z is threonine (Figure 9B).

[0066] Other residues in the CDR were also examined to see whether substitution with other amino acids (i.e., other than alanine) would improve binding. As shown in Table 2, only substitution of N58 with Q (clone name: VHB11) resulted in a slight improvement in binding. All other amino acid substitutions at key residues resulted in loss or reduction of binding. [Table 2]

[0067] Figure 8A shows a sequence alignment of the framework regions of the heavy chain variable domains of murine clone (M) (SEQ ID NO: 10), humanized clone (H) (SEQ ID NO: 11), backmutation clone (B1) (SEQ ID NO: 12), repeated backmutation clone (Re2) (SEQ ID NO: 13), and VHB11 clone (B11) (SEQ ID NO: 14). Figure 8B shows a sequence alignment of the framework regions of the light chain variable domains of murine clone (M) (SEQ ID NO: 15), humanized clone (H) (SEQ ID NO: 16), backmutation clone (B1) (SEQ ID NO: 17), repeated backmutation clone (Re2) (SEQ ID NO: 18), and VHB11 clone (B11) (SEQ ID NO: 19).

[0068] In CDRL1, residue 32 is not important for binding. Several residues are tolerated: C, S, G, and T. As an example, clone B13, which has a T at this position, is shown in Figure 9B.

[0069] Figure 9A shows the sequences of the heavy chain variable domains of consensus clone GBH(C) (SEQ ID NO:20) and exemplary clone GBH(B13) (SEQ ID NO:21), and Figure 9B shows the sequences of the light chain variable domains of consensus clone GBH(C) (SEQ ID NO:22) and exemplary clone GBH(B13) (SEQ ID NO:23).

[0070] Affinity Measurement Assay - ForteBio To evaluate the affinity of anti-globo H antibodies, globo H-amine was immobilized on an amine-reactive biosensor according to the manufacturer's instructions. All measurements were performed at 30 °C using a ForteBio Octet Red 96. Affinity binding curve fitting was performed using a predefined model (1:1 binding) provided by the Octet Data Analysis software.

[0071] Table 3 shows the results of some example ForteBio assays. [Table 3]

[0072] The antibodies of the present invention can be used to visualize globoH-expressing cells, for example, in the diagnosis of globoH-expressing cancers. For example, to visualize anti-globoH antibodies bound to globoH expressed on the cell surface (e.g., MCF7 cells), MCF7 cells were cultured on glass slides (NUNC) and then fixed with 4% paraformaldehyde. After washing with PBS, the cells were stained with anti-globoH antibodies, followed by goat anti-human FITC antibodies (1:1000; Thermo Scientific). Cells without primary antibodies were included as controls. The slides containing the samples were examined by FITC fluorescence and brightfield imaging using a fluorescence microscope (Olympus).

[0073] Figure 10 shows the results of immunostaining MCF7 cells with anti-globo H antibody. Panels A to C show fluorescent imaging of MCF7 cells using anti-globo H Ab, and panels A' to C' show dark-field (BF) imaging of the same cells. These results demonstrate that the antibodies of the present invention can indeed bind to globo H on the cell surface.

[0074] Affinity Measurement Assay - Biacore™ Antibody affinity can be assessed using ELISA or BiaCore™. For the BiaCore™ assay, globo H-amine was immobilized on a Biacore™ CM5 chip (Biacore, Uppsala, Sweden) using standard amine chemistry according to the manufacturer's instructions. Independent serial dilutions of anti-globo H antibodies were prepared on a microplate. Each sample was injected for 2.5 minutes at a flow rate of 50 μL / min over two flow cells: one for control cells and one for cells immobilized with globo H. Binding rates were measured at the end of the injection. After each sample, the chip was regenerated by injecting 10 mM glycine pH 2.5 / 1.5 (v / v = 1) for 45 seconds at a flow rate of 30 μL / min. All experiments were performed in HBS-EP buffer (Biacore™) at a constant temperature of 25.0°C using a Biacore T100 instrument. Affinity binding curve fitting was performed using a predefined model (1:1 binding) provided by the Biacore T100 evaluation software 2.0.

[0075] Fluorescence-activated cell sorting (FACS) Anti-globo H antibodies can be used to detect or sort cells expressing globo H on the cell surface, for example, using FACS. Globo H-expressing cells, MCF7 or HCC1428, were harvested and resuspended in 5% PBS / FBS buffer. Cells (1x10 5) were incubated with anti-globo H antibody (10 μg / ml) or Herceptin (10 μg / ml, negative control) for 1 hour at 4°C, and then stained with goat anti-human IgG FITC conjugate (1 / 1000) for 1 hour at 4°C. For each assay, two additional controls were prepared: one without primary antibody and one without any antibody. All processed samples were analyzed using FACSVerse (Becton Dickinson), and the results were processed using FACSuite software (Becton Dickinson).

[0076] Using FACS, various cancer cells can be evaluated to determine whether they express globoH. Figures 11(A)-11(G) show the results of globoH detection from various cell lines using anti-globoH antibodies and FACS. Figure 11(A) shows that MCF7 breast cancer cells express both Her2 antigen (upper panel) and globoH antigen (lower panel). Figure 11(B) shows that HCC1428 hepatocellular carcinoma cells express both Her2 antigen (upper panel) and globoH antigen (lower panel). Figure 11(C) shows that BT474 breast cancer cells express Her2 (upper panel) but not globoH (lower panel). Similarly, Figure 11(D) shows that Capan-1 pancreatic cancer cells express globoH. Figure 11(E) shows that A-431 squamous cell carcinoma cells express globoH. Figure 11(F) shows that NCI-N87 gastric cancer cells express globo H. Figure 11(G) shows that HT-29 colorectal cancer cells express low levels of globo H. These results demonstrate the utility of anti-globo H antibodies in the diagnosis of various cancers.

[0077] Antibody-dependent cell-mediated cytotoxicity (ADCC) Antibody-dependent cell-mediated cytotoxicity (ADCC) plays an important role in antibody-based cancer therapy. The humanized anti-globoH antibody of the present invention is a promising therapeutic agent for treating globoH-expressing cancers. To evaluate the ADCC activity of anti-globoH antibodies, purified human NK cells were incubated with antibody-treated human breast cancer cells (e.g., HCC1428, MCF7, Capan-1, NCI-N87, A431, or BT474 cells) at an E / T ratio of 5:1 for 3 hours. For the ADCC assay, an anti-IL20 antibody was used as a negative control, and Herceptin (Roche) was used as a positive control. The rate of cell death was measured using TDA release with the DELFIA EuTDA Cytotoxicity Reagents Kit (PerkinElmer). Fluorescence was measured using a time-resolved fluorometer (CLARIO, BGM).

[0078] Figures 12A-12F show that anti-globoH antibodies can induce ADCC in various cancer cells, including MCF7 (Figure 12A), HCC1428 (Figure 12B), Capan-1 (Figure 12D), NCI-N87 (Figure 12E), and A431 (Figure 12F) cancer cells. However, anti-globoH antibodies failed to induce ADCC in BT474 cells (Figure 12C), consistent with the observation that BT474 cells do not express globoH on their surface. These results confirm that ADCC induced by anti-globoH antibodies is dependent on the expression of globoH on the cell surface and confirm that anti-globoH antibodies are effective in inducing the killing of cancer cells expressing globoH.

[0079] Complement-dependent cytotoxicity (CDC) Similar to ADCC, complement-dependent cytotoxicity (CDC) plays an important role in antibody-based cancer therapy. To evaluate the ability of the anti-globoH antibodies of the present invention to induce CDC, the following experiments were performed. In these studies, 40% normal human serum (NHS) (v:v) was added to antibody-treated human cancer cells, including HCC1428, MCF7, BT474, Capan-1, or NCI-N87 cells, for 3 hours. Both anti-IL20 antibody and Herceptin were used as negative controls for the CDC assay. The rate of cell death was measured using TDA release with the DELFIA EuTDA Cytotoxicity Reagents Kit (PerkinElmer). Fluorescence was measured using a time-resolved fluorometer (CLARIO, BGM).

[0080] Figures 13A-13E show that anti-globoH antibodies can induce CDC in MCF7 cells (Figure 13A), HCC1428 cells (Figure 13B), Capan-1 cells (Figure 13D), and NCI-N87 cells (Figure 13E). However, anti-globoH antibodies were unable to induce CDC in BT474 cells. These results are consistent with the fact that MCF7, Capan-1, NCI-N87, and HCC1428 cells express globoH on their surface, whereas BT4747 cells do not. These results confirm that anti-globoH antibodies can induce CDC and that CDC induced by anti-globoH antibodies depends on the expression of globoH on the cell surface.

[0081] Globo H competition assay To confirm that anti-globo H antibody-induced cytotoxicity is globo H-dependent, globo H competition can be performed. 1 μM anti-globo H antibody was preincubated with various concentrations of either synthetic globo H or Lewis b-tetrasaccharide (Sigma) for 1 h at 37 °C. Lewis b-tetrasaccharide was used as a negative control. 40% NHS (v:v) was added to the anti-globo H antibody-glycan mixture, which was then incubated with the human breast cancer cell line MCF7 for 3 h. The rate of cell death was measured using TDA release with the DELFIA EuTDA Cytotoxicity Reagents Kit (PerkinElmer). Fluorescence was measured with a time-resolved fluorometer (CLARIO, BGM).

[0082] Figure 14 shows that the cytotoxicity induced by anti-globo H can be dose-dependently competed with globo H (fucose-galactose-N-Ac-galactosamine-galactose-galactose-glucose), but not with Lewis b tetrasaccharide (fucose-galactose-N-Ac-glucosamine-fucose). These results further confirm that the cytotoxicity induced by anti-globo H antibodies is globo H-dependent.

[0083] Xenograft animal model The fact that anti-globo H can induce ADCC and CDC of globo H-expressing cells indicates that these antibodies are useful for preventing and / or treating cancers that express globo H on their surface.

[0084] To evaluate the ability of anti-globo H antibodies in the prevention and / or treatment of cancer, use groups of five female (NOD / SCID) mice weighing 20-24 g (6-7 weeks old) and inoculate them with viable human breast cancer HCC1428 cells (provided by the organizers, 1 x 10 cells in 0.2 ml of Matrigel). 7) is injected subcutaneously into the dorsum of nude mice. Estol-Depot (100 μg / mouse) is injected subcutaneously as a supplement twice a week for 7 weeks, starting 1 week before cell transplantation. Test reagents (anti-globo H antibody, taxol, or vehicle) are administered intravenously starting 2 hours (preventive model) or 7 days (therapeutic model) after cell transplantation. Body weight and tumor size are recorded twice a week for up to 60 days. Tumor weight (mm 3 ) is the formula for a prolate ellipsoid: length (mm) x [width (mm)] 2 x 0.5. Tumor growth in animals treated with test compounds is calculated as T / C (treated / control) x 100%; a T / C value of 42% is considered significant in demonstrating antitumor activity.

[0085] Figure 15 shows that in a prophylactic model, anti-globo H antibodies were able to prevent tumor growth in a dose-dependent manner. At a dose of 10 mg / Kg (mpk), the antibodies were as effective as taxol, and at 30 mpk, the antibodies were more effective than taxol in preventing cancer growth.

[0086] Figure 16 shows that in a therapeutic model, anti-globo H antibodies were able to suppress tumor growth in a dose-dependent manner. At a dose of 10 mg / Kg (mpk), the antibodies were as effective as taxol, and at doses of 20 mpk and 30 mpk, the antibodies were more effective than taxol in preventing cancer growth.

[0087] Results from these in vivo models indicate that humanized anti-globo H antibodies can be used to prevent and treat various epithelial cancers, including those that express globo H, such as breast cancer, colon cancer, ovarian cancer, pancreatic cancer, lung cancer, liver cancer, and prostate cancer.

Claims

1. a heavy chain variable domain having three complementarity determining regions consisting of HCDR1, HCDR2, and HCDR3, and a light chain variable domain having three complementarity determining regions consisting of LCDR1, LCDR2, and LCDR3; the sequence of HCDR1 is GYISSDQILN (SEQ ID NO: 4); the sequence of HCDR2 is RIYPVTGVTXYNHKFVG (SEQ ID NO: 5, X is Q); the sequence of HCDR3 is GETFDS (SEQ ID NO: 6); the sequence of LCDR1 is KSNQNLLX'SGNRRYZLV (SEQ ID NO: 7, X' is F, Y or W, and Z is C); the sequence of LCDR2 is WASDRSF (SEQ ID NO: 8); the sequence of LCDR3 is QQHLDIPYT (SEQ ID NO: 9); the heavy chain variable domain has the corresponding framework sequence in SEQ ID NO: 13; The light chain variable domain has the corresponding framework sequence in SEQ ID NO:

18. A humanized anti-globo H antibody, or an antigen-binding fragment thereof.

2. X' in SEQ ID NO: 7 is tryptophan; The humanized anti-globo H antibody, or antigen-binding fragment thereof, of claim 1.

3. the heavy chain variable domain comprises the sequence of SEQ ID NO: 13 and the light chain variable domain comprises the sequence of SEQ ID NO: 18; The humanized anti-globo H antibody, or antigen-binding fragment thereof, of claim 1.

4. A compound comprising the antibody or antigen-binding fragment thereof according to any one of claims 1 to 3. A pharmaceutical composition for use in treating or preventing cancer.

5. The cancer is breast cancer, colon cancer, ovarian cancer, pancreatic cancer, lung cancer, liver cancer, or prostate cancer; The pharmaceutical composition according to claim 4.

Images